Banding with the E-M5 and the 20/1.7: A potential work-around

Started Oct 19, 2012 | Discussions
Anders W
Forum ProPosts: 17,838Gear list
Like?
Banding with the E-M5 and the 20/1.7: A potential work-around
Oct 19, 2012

As most people who own this combo (E-M5 and 20/1.7) are probably aware, it tends to produce banding at really high ISOs. At the beginning of June, Olympus acknowledged the problem and said: "we are continuing to study how we can eliminate this".

http://www.dpreview.com/news/2012/06/12/Olympus-acknowledges-OM-D-E-M5-banding-with-panasonic-20mm-f1-7-lens

When firmware version 1.5 for the E-M5 was made available a couple of weeks ago, there were some tentative reports from other forum members that it might have helped reduce or even eliminate the banding (although there was no mention of anything like that in the firmware announcement). In the process of testing this for myself, I eventually found that the banding was still very much there in at least some cases. However, I also found that I could sometimes produce  shots at very high ISOs that were just as banding-free as those from other lenses.

Whether the latter was possible with earlier firmware as well is something I cannot really tell at this point. But I don't remember seeing it in the tests I did in the past and the fact that I now saw it of course prompted me to pinpoint the conditions under which the banding was there and not there.

After quite a bit of experimentation and testing over the past few days, I am now reasonably (if not perfectly) convinced that I have found at least one such condition. If you leave the camera turned off for at least an hour or two and then turn it on, the first shots are very likely to show strong banding, at least if you start shooting right away. But if you fire a few shots (five or so), the banding will gradually disappear.

The test images posted on top show a prototypical exemple. Both are shot at ISO 6400 and underexposed a couple of stops. In the first image, shot right after turning the camera on, there is very evident banding. However, the second, which was shot less than one minute later after firing four additional shots inbetween and briefly looking at them on the LCD, is about as free from banding as we can hope anything to be at such an ISO and level of exposure. I have repeated this test (turn camera off, let it rest for a while, turn it on, fire five to ten shots and see what happens) quite a few times over the last few days and the result has so far always been the same: evident banding at first but then gradually disappearing.

I should mention here that I am not really done testing what happens when the camera has merely been asleep rather than turned off by means of the switch. However, my preliminary results indicate that the banding you get on start-up is less pronounced and disappears more quickly in that case. I should also mention that, for convenience, I have kept the 20/1.7 more or less permanently mounted while performing these tests, which means that I don't really know if lens swaps at this or that stage of my testing process would have an impact on the results in one way or the other.

Of course, I can provide no guarantees whatsoever, that the work-around I have found will work the same way for other copies of the E-M5 and the 20/1.7 as it does on mine. We are in all likelihood dealing with some form of electro-magnetic interference and the components involved may well behave slightly differently in one copy of the camera or lens than another. But I certainly hope it will help at least some others and I thought, on that basis, that it was my responsibility to report what I had found.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 19, 2012

P.S. I forgot to mention that the test shots were shot RAW and processed via LR with everything at default.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Detail Man
Forum ProPosts: 15,000
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 19, 2012

Anders W wrote:

After quite a bit of experimentation and testing over the past few days, I am now reasonably (if not perfectly) convinced that I have found at least one such condition. If you leave the camera turned off for at least an hour or two and then turn it on, the first shots are very likely to show strong banding, at least if you start shooting right away. But if you fire a few shots (five or so), the banding will gradually disappear.

The test images posted on top show a prototypical exemple. Both are shot at ISO 6400 and underexposed a couple of stops. In the first image, shot right after turning the camera on, there is very evident banding. However, the second, which was shot less than one minute later after firing four additional shots inbetween and briefly looking at them on the LCD, is about as free from banding as we can hope anything to be at such an ISO and level of exposure. I have repeated this test (turn camera off, let it rest for a while, turn it on, fire five to ten shots and see what happens) quite a few times over the last few days and the result has so far always been the same: evident banding at first but then gradually disappearing.

Interesting. From my background in electronics design - and just some guesses - I might wonder if a power supply circuit (possibly using DC to DC conversion using switching techniques to generate certain power-supply DC voltages) may go through a thermal stabilization cycle when started "cold" that places related high-frequency switching-noise on one or more power-supply lines, and then improves once having had a few minutes to thermally stabilize.

While digital switching circuitry has high power-suppy noise-immunity, potential entry-points involving analog circuitry, and grounds/ground-planes/shields, having some finite inductance, allow voltage-drops across themselves.

Such a scenario (involving active semiconductor components) might well be more likely than any such temperature-related changes in power-supply system capacitors/inductors involved in filtering the internally-regulated power-supply voltages. While active semiconductor circuitry in the image-sensor data read-electronics could be involved, the power-supply electronics themselves might be a more likely (active semiconductor component) source ...

Could it be that the 20mm F/1.7 lens draws more (average) power-supply current that the other lenses which do not show such a problem ? The existence of a higher load current on the power-supply system (might) cause the camera body's power-supply to generate more high-frequency noise on one or more internally regulated power-supply lines.

Such a higher (average) power-supply current load on the camera's internal power-supply system (could, possibly) cause such a problem as described above to arise - whether or not the electronic circuitry/inter-connections within/to the 20mm F/1.7 lens are themselves directly involved.

One possible experiment related to the 20mm F/1.7 lens' average power-supply current load might be to set the camera up for continuous auto-focusing upon power-up - as a way of being able to (very likely further) increase the lens' average supply-current load, and observe whether the recorded banding characteristics may change (in their nature as recorded in images, and/or in the length of time following power-up from a cold state during which banding appears to persist).

It may (possibly) be that such high-frequency switching noise might (to some extent, also) appear on the battery supply-line running directly from the camera battery to the camera itself. The placing of a reasonably high bandwidth oscilloscope (with a suitable low-capacitance/inductance probe), AC-coupled, and with a reasonably high channel-gain (Y-axis scale-factor) might make the observation of such switching-noise possible.

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Detail Man, Oct 19, 2012

Detail Man wrote:

Anders W wrote:

After quite a bit of experimentation and testing over the past few days, I am now reasonably (if not perfectly) convinced that I have found at least one such condition. If you leave the camera turned off for at least an hour or two and then turn it on, the first shots are very likely to show strong banding, at least if you start shooting right away. But if you fire a few shots (five or so), the banding will gradually disappear.

The test images posted on top show a prototypical exemple. Both are shot at ISO 6400 and underexposed a couple of stops. In the first image, shot right after turning the camera on, there is very evident banding. However, the second, which was shot less than one minute later after firing four additional shots inbetween and briefly looking at them on the LCD, is about as free from banding as we can hope anything to be at such an ISO and level of exposure. I have repeated this test (turn camera off, let it rest for a while, turn it on, fire five to ten shots and see what happens) quite a few times over the last few days and the result has so far always been the same: evident banding at first but then gradually disappearing.

Interesting. From my background in electronics design - and just some guesses - I might wonder if a power supply circuit (possibly using DC to DC conversion using switching techniques to generate certain power-supply DC voltages) may go through a thermal stabilization cycle when started "cold" that places related high-frequency switching-noise on one or more power-supply lines, and then improves once having had a few minutes to thermally stabilize.

While digital switching circuitry has high power-suppy noise-immunity, potential entry-points involving analog circuitry, and grounds/ground-planes/shields, having some finite inductance, allow voltage-drops across themselves.

Such a scenario (involving active semiconductor components) might well be more likely than any such temperature-related changes in power-supply system capacitors/inductors involved in filtering the internally-regulated power-supply voltages. While active semiconductor circuitry in the image-sensor data read-electronics could be involved, the power-supply electronics themselves might be a more likely (active semiconductor component) source ...

Could it be that the 20mm F/1.7 lens draws more (average) power-supply current that the other lenses which do not show such a problem ? The existence of a higher load current on the power-supply system (might) cause the camera body's power-supply to generate more high-frequency noise on one or more internally regulated power-supply lines.

Such a higher (average) power-supply current load on the camera's internal power-supply system (could, possibly) cause such a problem as described above to arise - whether or not the electronic circuitry/inter-connections within/to the 20mm F/1.7 lens are themselves directly involved.

Interesting ideas DM. What seems pretty clear to me at this point is that we are dealing with electric interference of one kind or another caused by components in the camera and/or lens. The other two types of hypotheses that people tossed around at an early stage (the nature of the light used, the lens acting as some kind of antenna) are, I would say, already refuted.

But I have hitherto assumed that it must be something in the lens that interferes with the body, more specifically the sensor and associated analog read-out circuitry (if the ADC are not on chip). What you point out is that it might also be camera components interfering with one another as a result of unusual external conditions, such as high power drain.

In addition, it is not at all unlikely that the 20/1.7 does indeed consume greater than average power at peak, due to its old-fashioned focusing system (no internal focus; the motor has to move the whole package of lens elements in and out). On the other hand, this is hardly continuous consumption. Since I have the camera set so as to reset the lens on startup, the AF motor is engaged briefly when I switch on the camera, moving things in and out to check that they are working and then setting focus to infinity. But this takes no more than a second or so. And even though I AF before each exposure, I don't AF during the exposure, or, more importantly, during readout. And how could high power drain at one time cause interference at another time, particularly the time during which the sensor is read?

One possible experiment related to the 20mm F/1.7 lens' average power-supply current load might be to set the camera up for continuous auto-focusing upon power-up - as a way of being able to (very likely further) increase the lens' average supply-current load, and observe whether the recorded banding characteristics may change (in their nature as recorded in images, and/or in the length of time following power-up from a cold state during which banding appears to persist).

It may (possibly) be that such high-frequency switching noise might (to some extent, also) appear on the battery supply-line running directly from the camera battery to the camera itself. The placing of a reasonably high bandwidth oscilloscope (with a suitable low-capacitance/inductance probe), AC-coupled, and with a reasonably high channel-gain (Y-axis scale-factor) might make the observation of such switching-noise possible.

No easy access to an oscilloscope I am afraid. But I will try your test with AF-C when I find the time. Ironically, the 20 doesn't really allow AF-C because of its old-fashioned AF system. On the G1 it was simply impossible to turn AF-C on with this lens. The camera would simply tell me to switch to AF-S. But if I remember correctly, the E-M5 is more tolerant in this respect and will allow me to push the lens to its AF limits.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Detail Man
Forum ProPosts: 15,000
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 19, 2012

Anders W wrote:

Detail Man wrote:

Anders W wrote:

After quite a bit of experimentation and testing over the past few days, I am now reasonably (if not perfectly) convinced that I have found at least one such condition. If you leave the camera turned off for at least an hour or two and then turn it on, the first shots are very likely to show strong banding, at least if you start shooting right away. But if you fire a few shots (five or so), the banding will gradually disappear.

The test images posted on top show a prototypical exemple. Both are shot at ISO 6400 and underexposed a couple of stops. In the first image, shot right after turning the camera on, there is very evident banding. However, the second, which was shot less than one minute later after firing four additional shots inbetween and briefly looking at them on the LCD, is about as free from banding as we can hope anything to be at such an ISO and level of exposure. I have repeated this test (turn camera off, let it rest for a while, turn it on, fire five to ten shots and see what happens) quite a few times over the last few days and the result has so far always been the same: evident banding at first but then gradually disappearing.

Interesting. From my background in electronics design - and just some guesses - I might wonder if a power supply circuit (possibly using DC to DC conversion using switching techniques to generate certain power-supply DC voltages) may go through a thermal stabilization cycle when started "cold" that places related high-frequency switching-noise on one or more power-supply lines, and then improves once having had a few minutes to thermally stabilize.

While digital switching circuitry has high power-suppy noise-immunity, potential entry-points involving analog circuitry, and grounds/ground-planes/shields, having some finite inductance, allow voltage-drops across themselves.

Such a scenario (involving active semiconductor components) might well be more likely than any such temperature-related changes in power-supply system capacitors/inductors involved in filtering the internally-regulated power-supply voltages. While active semiconductor circuitry in the image-sensor data read-electronics could be involved, the power-supply electronics themselves might be a more likely (active semiconductor component) source ...

Could it be that the 20mm F/1.7 lens draws more (average) power-supply current that the other lenses which do not show such a problem ? The existence of a higher load current on the power-supply system (might) cause the camera body's power-supply to generate more high-frequency noise on one or more internally regulated power-supply lines.

Such a higher (average) power-supply current load on the camera's internal power-supply system (could, possibly) cause such a problem as described above to arise - whether or not the electronic circuitry/inter-connections within/to the 20mm F/1.7 lens are themselves directly involved.

Interesting ideas DM. What seems pretty clear to me at this point is that we are dealing with electric interference of one kind or another caused by components in the camera and/or lens. The other two types of hypotheses that people tossed around at an early stage (the nature of the light used, the lens acting as some kind of antenna) are, I would say, already refuted.

But I have hitherto assumed that it must be something in the lens that interferes with the body, more specifically the sensor and associated analog read-out circuitry (if the ADC are not on chip). What you point out is that it might also be camera components interfering with one another as a result of unusual external conditions, such as high power drain.

Right. I am suggesting that it (could) be that a higher quiescent (always present) or a transient (short-lived) current-draw on the part of the 20mm F/1.7 lens that happens to be higher in magnitude (relative to other, non-problematic lenses) which (might) be responsible for increased levels of high-frequency switching-noise appearing on one (or more) internally regulated DC power-supply lines (until this "thermal equilibrium" effect that you are observing is reached) - which could [as a result of possible DC to DC conversion switching-supply circuit(s) behavior following cold start-up] indirectly impact circuitry located within the camera body itself via internal power-supply lines and entry-points lacking high noise-immunity (and not necessarily specifically involving circuitry within, or interconnections to, the lens electronics themselves).

In addition, it is not at all unlikely that the 20/1.7 does indeed consume greater than average power at peak, due to its old-fashioned focusing system (no internal focus; the motor has to move the whole package of lens elements in and out). On the other hand, this is hardly continuous consumption. Since I have the camera set so as to reset the lens on startup, the AF motor is engaged briefly when I switch on the camera, moving things in and out to check that they are working and then setting focus to infinity. But this takes no more than a second or so. And even though I AF before each exposure, I don't AF during the exposure, or, more importantly, during readout. And how could high power drain at one time cause interference at another time, particularly the time during which the sensor is read?

I hear what you are saying. However, without directly measuring the lens' current-draw with a series-connected ammeter, it may be hard to know for sure what the current consumption is at these particular times. If the (overall, total, including quiescent) current draw is itself higher than other (non-problematic) lenses, then such a distinction (might) still be relevant.

One possible experiment related to the 20mm F/1.7 lens' average power-supply current load might be to set the camera up for continuous auto-focusing upon power-up - as a way of being able to (very likely further) increase the lens' average supply-current load, and observe whether the recorded banding characteristics may change (in their nature as recorded in images, and/or in the length of time following power-up from a cold state during which banding appears to persist).

It may (possibly) be that such high-frequency switching noise might (to some extent, also) appear on the battery supply-line running directly from the camera battery to the camera itself. The placing of a reasonably high bandwidth oscilloscope (with a suitable low-capacitance/inductance probe), AC-coupled, and with a reasonably high channel-gain (Y-axis scale-factor) might make the observation of such switching-noise possible.

No easy access to an oscilloscope I am afraid. But I will try your test with AF-C when I find the time. Ironically, the 20 doesn't really allow AF-C because of its old-fashioned AF system. On the G1 it was simply impossible to turn AF-C on with this lens. The camera would simply tell me to switch to AF-S. But if I remember correctly, the E-M5 is more tolerant in this respect and will allow me to push the lens to its AF limits.

It may not be that such (internal, buffered by analog regulation circuitry) high-frequency (internal, regulated) power-supply switching-noise transients would appear all the way back at the (only practically accessible) battery-terminal to camera body connections - but, due to their possible high-frequency, transient, wide-bandwidth nature, it is possible that some clues (might) be visible using a sufficiently wide-bandwidth osciiloscope (and probes) when using an AC-coupled measurement, in addition to a sufficiently high channel (Y-axis deflection) gain.

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Detail Man, Oct 19, 2012

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

After quite a bit of experimentation and testing over the past few days, I am now reasonably (if not perfectly) convinced that I have found at least one such condition. If you leave the camera turned off for at least an hour or two and then turn it on, the first shots are very likely to show strong banding, at least if you start shooting right away. But if you fire a few shots (five or so), the banding will gradually disappear.

The test images posted on top show a prototypical exemple. Both are shot at ISO 6400 and underexposed a couple of stops. In the first image, shot right after turning the camera on, there is very evident banding. However, the second, which was shot less than one minute later after firing four additional shots inbetween and briefly looking at them on the LCD, is about as free from banding as we can hope anything to be at such an ISO and level of exposure. I have repeated this test (turn camera off, let it rest for a while, turn it on, fire five to ten shots and see what happens) quite a few times over the last few days and the result has so far always been the same: evident banding at first but then gradually disappearing.

Interesting. From my background in electronics design - and just some guesses - I might wonder if a power supply circuit (possibly using DC to DC conversion using switching techniques to generate certain power-supply DC voltages) may go through a thermal stabilization cycle when started "cold" that places related high-frequency switching-noise on one or more power-supply lines, and then improves once having had a few minutes to thermally stabilize.

While digital switching circuitry has high power-suppy noise-immunity, potential entry-points involving analog circuitry, and grounds/ground-planes/shields, having some finite inductance, allow voltage-drops across themselves.

Such a scenario (involving active semiconductor components) might well be more likely than any such temperature-related changes in power-supply system capacitors/inductors involved in filtering the internally-regulated power-supply voltages. While active semiconductor circuitry in the image-sensor data read-electronics could be involved, the power-supply electronics themselves might be a more likely (active semiconductor component) source ...

Could it be that the 20mm F/1.7 lens draws more (average) power-supply current that the other lenses which do not show such a problem ? The existence of a higher load current on the power-supply system (might) cause the camera body's power-supply to generate more high-frequency noise on one or more internally regulated power-supply lines.

Such a higher (average) power-supply current load on the camera's internal power-supply system (could, possibly) cause such a problem as described above to arise - whether or not the electronic circuitry/inter-connections within/to the 20mm F/1.7 lens are themselves directly involved.

Interesting ideas DM. What seems pretty clear to me at this point is that we are dealing with electric interference of one kind or another caused by components in the camera and/or lens. The other two types of hypotheses that people tossed around at an early stage (the nature of the light used, the lens acting as some kind of antenna) are, I would say, already refuted.

But I have hitherto assumed that it must be something in the lens that interferes with the body, more specifically the sensor and associated analog read-out circuitry (if the ADC are not on chip). What you point out is that it might also be camera components interfering with one another as a result of unusual external conditions, such as high power drain.

Right. I am suggesting that it (could) be that a higher quiescent (always present) or a transient (short-lived) current-draw on the part of the 20mm F/1.7 lens that happens to be higher in magnitude (relative to other, non-problematic lenses) which (might) be responsible for increased levels of high-frequency switching-noise appearing on one (or more) internally regulated DC power-supply lines (until this "thermal equilibrium" effect that you are observing is reached) - which could [as a result of possible DC to DC conversion switching-supply circuit(s) behavior following cold start-up] indirectly impact circuitry located within the camera body itself via internal power-supply lines and entry-points lacking high noise-immunity (and not necessarily specifically involving circuitry within, or interconnections to, the lens electronics themselves).

In addition, it is not at all unlikely that the 20/1.7 does indeed consume greater than average power at peak, due to its old-fashioned focusing system (no internal focus; the motor has to move the whole package of lens elements in and out). On the other hand, this is hardly continuous consumption. Since I have the camera set so as to reset the lens on startup, the AF motor is engaged briefly when I switch on the camera, moving things in and out to check that they are working and then setting focus to infinity. But this takes no more than a second or so. And even though I AF before each exposure, I don't AF during the exposure, or, more importantly, during readout. And how could high power drain at one time cause interference at another time, particularly the time during which the sensor is read?

I hear what you are saying. However, without directly measuring the lens' current-draw with a series-connected ammeter, it may be hard to know for sure what the current consumption is at these particular times. If the (overall, total, including quiescent) current draw is itself higher than other (non-problematic) lenses, then such a distinction (might) still be relevant.

Thanks for the additional detail outlining your ideas. I am certainly listening with interest. I have a multi-meter of course and know how to use it. But I am afraid measurement is nevertheless a bit complicated in this case.

My own best guess currently is otherwise interference from the aperture actuator, which, as far as I can tell, is operative during the read-out phase. I have in some tests seen indications that aperture movement might be responsible (no banding if I shoot wide open so that it won't have to open up again after exposure but banding over a vertically increasing area of the frame with increasingly smaller apertures). However, in other cases, I have seen banding with the lens wide open too so perhaps this is a blind alley, a partial truth, or a valid theory but with some important bits still missing. We'll see.

As to the focus mechanism, you can sometimes see something looking very much like the banding you get in actual images when watching live view while focusing. Furthermore, it is possible (I have to check again) that this is more apparent right after turning the camera on after having left it off for a couple of hours, just like the banding in actual images. On the other hand, this live view banding stops immediately when you stop focusing. So how could it affect the exposure itself or the read-out of the exposure.

One possible experiment related to the 20mm F/1.7 lens' average power-supply current load might be to set the camera up for continuous auto-focusing upon power-up - as a way of being able to (very likely further) increase the lens' average supply-current load, and observe whether the recorded banding characteristics may change (in their nature as recorded in images, and/or in the length of time following power-up from a cold state during which banding appears to persist).

It may (possibly) be that such high-frequency switching noise might (to some extent, also) appear on the battery supply-line running directly from the camera battery to the camera itself. The placing of a reasonably high bandwidth oscilloscope (with a suitable low-capacitance/inductance probe), AC-coupled, and with a reasonably high channel-gain (Y-axis scale-factor) might make the observation of such switching-noise possible.

No easy access to an oscilloscope I am afraid. But I will try your test with AF-C when I find the time. Ironically, the 20 doesn't really allow AF-C because of its old-fashioned AF system. On the G1 it was simply impossible to turn AF-C on with this lens. The camera would simply tell me to switch to AF-S. But if I remember correctly, the E-M5 is more tolerant in this respect and will allow me to push the lens to its AF limits.

It may not be that such (internal, buffered by analog regulation circuitry) high-frequency (internal, regulated) power-supply switching-noise transients would appear all the way back at the (only practically accessible) battery-terminal to camera body connections - but (due to their possible high-frequency, transient, wide-bandwidth nature, it is possible that some possible clues (might) be visible using a sufficiently wide-bandwidth osciiloscope (and probes) when using an AC-coupled measurement in addition to a sufficiently high channel (Y-axis deflection) gain.


 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 19, 2012

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Skeeterbytes
Veteran MemberPosts: 5,236
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 19, 2012

Despite the fact that the level of geekiness in this thread is making my head spin, I applaud your tests and reporting! Because I have the camera and lens, I'm eager for any tricks leading to better performance, regardless of whether I understand what the ghost in the machine may be up to. Thanks!  Rick

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Skeeterbytes, Oct 19, 2012

Skeeterbytes wrote:

Despite the fact that the level of geekiness in this thread is making my head spin, I applaud your tests and reporting! Because I have the camera and lens, I'm eager for any tricks leading to better performance, regardless of whether I understand what the ghost in the machine may be up to. Thanks!  Rick

None of us can really claim to have understood the "ghost in the machine" yet, Rick. We're just speculating about what's happening behind the stage. But at least there's some pretty down-to-earth stuff to try that might possibly help. Please check it out to see if it works in your case too and report back. Looking forward to hear about your experiences, whether good or bad.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Detail Man
Forum ProPosts: 15,000
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 19, 2012

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Detail Man, Oct 19, 2012

Detail Man wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Happy to see you drop in again DM. I would have asked you to do so if you hadn't volunteered. I think I am pretty decent tester of pretty much anything. But in terms of theorizing, I am far off home base here. My knowledge of electronics on the hardware level is ... ahem ... spotty. So I much appreciate your input about likely causes. This is your field after all.

Supposing for the time being that my particular copies of lens and camera are not unrepresentative: What would be your take on what's going on here?

Supposedly, we are dealing with three main components on the lens side (apart from the lens elements themselves and the plastic barrel, both of which I think we can acquit right away): the PCB, the focus motor, and the aperture actuator. The focus motor seems like an unlikely bet to me since it supposedly isn't active during the critical phase (sensor read-out). Remains PCB and aperture actuator. Could the PCB have capacitors or something like that, which might cause strange ("spiky") power drain? As I said, the aperture actuator would be my favorite culprit unless I knew I can get banding in a "cold" state even when shooting the lens wide open where the actuator (supposedly) wouldn't need to go into action at all.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Detail Man
Forum ProPosts: 15,000
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 20, 2012

Anders W wrote:

Detail Man wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Happy to see you drop in again DM. I would have asked you to do so if you hadn't volunteered. I think I am pretty decent tester of pretty much anything. But in terms of theorizing, I am far off home base here. My knowledge of electronics on the hardware level is ... ahem ... spotty. So I much appreciate your input about likely causes. This is your field after all.

Supposing for the time being that my particular copies of lens and camera are not unrepresentative: What would be your take on what's going on here?

Supposedly, we are dealing with three main components on the lens side (apart from the lens elements themselves and the plastic barrel, both of which I think we can acquit right away): the PCB, the focus motor, and the aperture actuator. The focus motor seems like an unlikely bet to me since it supposedly isn't active during the critical phase (sensor read-out). Remains PCB and aperture actuator. Could the PCB have capacitors or something like that, which might cause strange ("spiky") power drain? As I said, the aperture actuator would be my favorite culprit unless I knew I can get banding in a "cold" state even when shooting the lens wide open where the actuator (supposedly) wouldn't need to go into action at all.

Capacitors (in general), and as involved in any filtering of the power-supply lines (acting as a parallel-connected shunt impedance) are going to charge up quite quickly upon lens power-up. Ditto for any series-connected inductance components where it comes to magentic flux levels.

Any additional power-supply current drain that such capacitor charging represents would likely be very short-lived (relative to the time you report of recording several shots, etc.).

Such a current-drain would be relevant only in the case of my originally stated idea - that a high level of current-consumption (might) be causing power-supply circuitry locates in the camera body itself to have increased high-frequency switching-noise on various DC-voltage power-supply lines - until (for some reason), the active semiconductors in the camera's power-supply regulation circuitry itself thermally stabilize.

The above scenario is perhaps a bit of an "outside" scenario - as a well designed power-supply regulation system (should not) require such a long time period to thermally stabilize.

But (if, as you say), the in-lens aperture-actuator as well as focus-motor appears to not be involved, and given the information below about capacitors, it should not be ruled out ?

Connecting your multi-meter as a series-connected ammeter would indeed be a rather laborious procedure - and might only show relatively small differences in battery-cuttent consumption levels. Some of the change that you see might (also, to some extent) be related to any decreasing of the Li-Ion cell voltage potential (that takes place after a cold-start of the camera body, or perhaps after connecting a lens that draws more surrent from the power-supply).

Connecting a high-bandwidth oscilloscope and probe to the battery might be more instructive - (if) this is a matter of high-frequency switching-noise ...

Regarding the ability of capacitors to effectively function as filters of the power-supply lines. Their effectiveness in doing so is not related to whether or not they have yet fully charged-up to some particular DC-voltage value - and the temperature-coefficient of their value of capacitance (even in the case of large capacitance valued electrolytic capacitors) is small (on the order of a fraction of a percent per degree Celsius).

The (seemingly reported) intermittent nature of the banding (even when "cold") makes it seem even more mysterious - and (might) tend to point in the direction of some mechanical cause ?

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Detail Man, Oct 20, 2012

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Happy to see you drop in again DM. I would have asked you to do so if you hadn't volunteered. I think I am pretty decent tester of pretty much anything. But in terms of theorizing, I am far off home base here. My knowledge of electronics on the hardware level is ... ahem ... spotty. So I much appreciate your input about likely causes. This is your field after all.

Supposing for the time being that my particular copies of lens and camera are not unrepresentative: What would be your take on what's going on here?

Supposedly, we are dealing with three main components on the lens side (apart from the lens elements themselves and the plastic barrel, both of which I think we can acquit right away): the PCB, the focus motor, and the aperture actuator. The focus motor seems like an unlikely bet to me since it supposedly isn't active during the critical phase (sensor read-out). Remains PCB and aperture actuator. Could the PCB have capacitors or something like that, which might cause strange ("spiky") power drain? As I said, the aperture actuator would be my favorite culprit unless I knew I can get banding in a "cold" state even when shooting the lens wide open where the actuator (supposedly) wouldn't need to go into action at all.

Capacitors (in general), and as involved in any filtering of the power-supply lines (acting as a parallel-connected shunt impedance) are going to charge up quite quickly upon lens power-up. Ditto for any series-connected inductance components where it comes to magentic flux levels.

Any additional power-supply current drain that such capacitor charging represents would likely be very short-lived (relative to the time you report of recording several shots, etc.).

Such a current-drain would be relevant only in the case of my originally stated idea - that a high level of current-consumption (might) be causing power-supply circuitry locates in the camera body itself to have increased high-frequency switching-noise on various DC-voltage power-supply lines - until (for some reason), the active semiconductors in the camera's power-supply regulation circuitry itself thermally stabilize.

The above scenario is perhaps a bit of an "outside" scenario - as a well designed power-supply regulation system (should not) require such a long time period to thermally stabilize.

But (if, as you say), the in-lens aperture-actuator as well as focus-motor appears to not be involved, and given the information below about capacitors, it should not be ruled out ?

Regarding the ability of capacitors to effectively function as filters of the power-supply lines. Their effectiveness in doing so is not related to whether or not they have yet fully charged-up to some particular DC-voltage value - and the temperature-coefficient of their value of capacitance (even in the case of large capacitance valued electrolytic capacitors) is small (on the order of a fraction of a percent per degree Celsius).

Well, it's a nice mystery isn't it? We both have trouble coming up with any reasonable hypothesis. If the focus motor and aperture actuator aren't involved, we have only the PCB left. And why should it need "warming up". Can a "spiky" load (which might be there) as opposed to a strong load (which seems unlikely to be there at the relevant, i.e. readout, time) cause problems here?

Unless you have a better idea, my next experiments will try to separate keeping the lens "warm" (having it mounted with the camera on) from shooting it. If it's just about keeping the lens "warm" without shooting it, we're down to the PCB, wouldn't you say?

Interesting times, aren't they. Gentlemen, start your lenses please!

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Detail Man
Forum ProPosts: 15,000
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 20, 2012

Anders W wrote:

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Happy to see you drop in again DM. I would have asked you to do so if you hadn't volunteered. I think I am pretty decent tester of pretty much anything. But in terms of theorizing, I am far off home base here. My knowledge of electronics on the hardware level is ... ahem ... spotty. So I much appreciate your input about likely causes. This is your field after all.

Supposing for the time being that my particular copies of lens and camera are not unrepresentative: What would be your take on what's going on here?

Supposedly, we are dealing with three main components on the lens side (apart from the lens elements themselves and the plastic barrel, both of which I think we can acquit right away): the PCB, the focus motor, and the aperture actuator. The focus motor seems like an unlikely bet to me since it supposedly isn't active during the critical phase (sensor read-out). Remains PCB and aperture actuator. Could the PCB have capacitors or something like that, which might cause strange ("spiky") power drain? As I said, the aperture actuator would be my favorite culprit unless I knew I can get banding in a "cold" state even when shooting the lens wide open where the actuator (supposedly) wouldn't need to go into action at all.

Capacitors (in general), and as involved in any filtering of the power-supply lines (acting as a parallel-connected shunt impedance) are going to charge up quite quickly upon lens power-up. Ditto for any series-connected inductance components where it comes to magentic flux levels.

Any additional power-supply current drain that such capacitor charging represents would likely be very short-lived (relative to the time you report of recording several shots, etc.).

Such a current-drain would be relevant only in the case of my originally stated idea - that a high level of current-consumption (might) be causing power-supply circuitry locates in the camera body itself to have increased high-frequency switching-noise on various DC-voltage power-supply lines - until (for some reason), the active semiconductors in the camera's power-supply regulation circuitry itself thermally stabilize.

The above scenario is perhaps a bit of an "outside" scenario - as a well designed power-supply regulation system (should not) require such a long time period to thermally stabilize.

But (if, as you say), the in-lens aperture-actuator as well as focus-motor appears to not be involved, and given the information below about capacitors, it should not be ruled out ?

Regarding the ability of capacitors to effectively function as filters of the power-supply lines. Their effectiveness in doing so is not related to whether or not they have yet fully charged-up to some particular DC-voltage value - and the temperature-coefficient of their value of capacitance (even in the case of large capacitance valued electrolytic capacitors) is small (on the order of a fraction of a percent per degree Celsius).

Well, it's a nice mystery isn't it? We both have trouble coming up with any reasonable hypothesis. If the focus motor and aperture actuator aren't involved, we have only the PCB left. And why should it need "warming up". Can a "spiky" load (which might be there) as opposed to a strong load (which seems unlikely to be there at the relevant, i.e. readout, time) cause problems here?

If the lens presents a time-varying current-load to the camera-body generated DC power-supply voltage lines that it connects to, that could cause voltage-transients and/or higher levels of high-frequency switching-noise to appear on such DC power-supply voltage lines.

It may (also) be that the PCB of the lens itself (also) uses (further) voltage-regulation circuitry on-board - in order to attempt to isolate the effects of electrical loads of focus and aperture stepper-motors from sending transients and noise back towards the camera body's DC power-supply voltage lines. That would not surprise me - as motors/actuators (representing high inductances seeing pulsed drive-currents with high rates of change in time) are the most likely generators of such unwanted voltage-spike "garbage" on DC power-supply lines.

Unless you have a better idea, my next experiments will try to separate keeping the lens "warm" (having it mounted with the camera on) from shooting it. If it's just about keeping the lens "warm" without shooting it, we're down to the PCB, wouldn't you say?

Interesting times, aren't they. Gentlemen, start your lenses please!

See my edited (most recent) post for some additional thoughts in general.

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Detail Man, Oct 20, 2012

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Happy to see you drop in again DM. I would have asked you to do so if you hadn't volunteered. I think I am pretty decent tester of pretty much anything. But in terms of theorizing, I am far off home base here. My knowledge of electronics on the hardware level is ... ahem ... spotty. So I much appreciate your input about likely causes. This is your field after all.

Supposing for the time being that my particular copies of lens and camera are not unrepresentative: What would be your take on what's going on here?

Supposedly, we are dealing with three main components on the lens side (apart from the lens elements themselves and the plastic barrel, both of which I think we can acquit right away): the PCB, the focus motor, and the aperture actuator. The focus motor seems like an unlikely bet to me since it supposedly isn't active during the critical phase (sensor read-out). Remains PCB and aperture actuator. Could the PCB have capacitors or something like that, which might cause strange ("spiky") power drain? As I said, the aperture actuator would be my favorite culprit unless I knew I can get banding in a "cold" state even when shooting the lens wide open where the actuator (supposedly) wouldn't need to go into action at all.

Capacitors (in general), and as involved in any filtering of the power-supply lines (acting as a parallel-connected shunt impedance) are going to charge up quite quickly upon lens power-up. Ditto for any series-connected inductance components where it comes to magentic flux levels.

Any additional power-supply current drain that such capacitor charging represents would likely be very short-lived (relative to the time you report of recording several shots, etc.).

Such a current-drain would be relevant only in the case of my originally stated idea - that a high level of current-consumption (might) be causing power-supply circuitry locates in the camera body itself to have increased high-frequency switching-noise on various DC-voltage power-supply lines - until (for some reason), the active semiconductors in the camera's power-supply regulation circuitry itself thermally stabilize.

The above scenario is perhaps a bit of an "outside" scenario - as a well designed power-supply regulation system (should not) require such a long time period to thermally stabilize.

But (if, as you say), the in-lens aperture-actuator as well as focus-motor appears to not be involved, and given the information below about capacitors, it should not be ruled out ?

Regarding the ability of capacitors to effectively function as filters of the power-supply lines. Their effectiveness in doing so is not related to whether or not they have yet fully charged-up to some particular DC-voltage value - and the temperature-coefficient of their value of capacitance (even in the case of large capacitance valued electrolytic capacitors) is small (on the order of a fraction of a percent per degree Celsius).

Well, it's a nice mystery isn't it? We both have trouble coming up with any reasonable hypothesis. If the focus motor and aperture actuator aren't involved, we have only the PCB left. And why should it need "warming up". Can a "spiky" load (which might be there) as opposed to a strong load (which seems unlikely to be there at the relevant, i.e. readout, time) cause problems here?

If the lens presents a time-varying current-load to the camera-body generated DC power-supply voltage lines that it connects to, that could cause voltage-transients and/or higher levels of high-frequency switching-noise to appear on such DC power-supply voltage lines.

It may (also) be that the PCB of the lens itself (also) uses (further) voltage-regulation circuitry on-board - in order to attempt to isolate the effects of electrical loads of focus and aperture stepper-motors from sending transients and noise back towards the camera body's DC power-supply voltage lines. That would not surprise me - as motors/actuators (representing high inductances seeing pulsed drive-currents with high rates of change in time) are the most likely generators of such unwanted voltage-spike "garbage" on DC power-supply lines.

Unless you have a better idea, my next experiments will try to separate keeping the lens "warm" (having it mounted with the camera on) from shooting it. If it's just about keeping the lens "warm" without shooting it, we're down to the PCB, wouldn't you say?

Interesting times, aren't they. Gentlemen, start your lenses please!

See my edited (most recent) post for some additional thoughts in general.

Thanks DM. I'll do some further experiments to see if we can get any further towards isolating the cause. In the meantime, I am happy that I (and hopefully others too) may be closer to some practical solution that we can live with.

With regard to intermittent behavior in your previous post: I think I am pretty close to some kind of predictable pattern now. Thinking on the basis of stochastic as opposed to deterministic evidence is awful until you have a clue. Once you do, a little random error is not the whole world.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
NeroMetalliko
Regular MemberPosts: 220
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 20, 2012

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Hello,

Olympus claims for the E-M5 an operating temperature in 0-40°C. I can assume the same should be valid for the Panasonic 20mm.

It will be interesting to try some cold/warm temperature operating test (with no warming-up time obviously).

Using a common refrigrator and a open oven, with some care, it could be theoretically possible to put the lens and body at 4-6°C and at 25-30°C and perform a set of lens/camera shoots like for example:

- cold lens/cold body

- warm lens/warm body

- warm lens/cold body

- cold lens/warm body

The first two combinations are more easy to do, the two "mixed" ones are more tricky because a fast action is required to join the two items and shoot the test image in very short time.

The main problem I see will be the condensation that will surely occur, so I'm NOT recommending to anyone to try these kind of test.

And... ehm... I have sold my 20mm lens, so "unlucky" I cannot try with my camera...  ehm...

Ciao

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to NeroMetalliko, Oct 20, 2012

NeroMetalliko wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Hello,

Olympus claims for the E-M5 an operating temperature in 0-40°C. I can assume the same should be valid for the Panasonic 20mm.

It will be interesting to try some cold/warm temperature operating test (with no warming-up time obviously).

Using a common refrigrator and a open oven, with some care, it could be theoretically possible to put the lens and body at 4-6°C and at 25-30°C and perform a set of lens/camera shoots like for example:

- cold lens/cold body

- warm lens/warm body

- warm lens/cold body

- cold lens/warm body

The first two combinations are more easy to do, the two "mixed" ones are more tricky because a fast action is required to join the two items and shoot the test image in very short time.

The main problem I see will be the condensation that will surely occur, so I'm NOT recommending to anyone to try these kind of test.

And... ehm... I have sold my 20mm lens, so "unlucky" I cannot try with my camera...  ehm...

Ciao

Thanks for these additional testing suggestions. But why stay at these moderate temperatures? Luckily, I have easy access to an oven capable of pyrolysis temperatures (approaching 500 C). And then deep freezing of course.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
NeroMetalliko
Regular MemberPosts: 220
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Anders W, Oct 20, 2012

Anders W wrote:

NeroMetalliko wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Hello,

Olympus claims for the E-M5 an operating temperature in 0-40°C. I can assume the same should be valid for the Panasonic 20mm.

It will be interesting to try some cold/warm temperature operating test (with no warming-up time obviously).

Using a common refrigrator and a open oven, with some care, it could be theoretically possible to put the lens and body at 4-6°C and at 25-30°C and perform a set of lens/camera shoots like for example:

- cold lens/cold body

- warm lens/warm body

- warm lens/cold body

- cold lens/warm body

The first two combinations are more easy to do, the two "mixed" ones are more tricky because a fast action is required to join the two items and shoot the test image in very short time.

The main problem I see will be the condensation that will surely occur, so I'm NOT recommending to anyone to try these kind of test.

And... ehm... I have sold my 20mm lens, so "unlucky" I cannot try with my camera...  ehm...

Ciao

Thanks for these additional testing suggestions. But why stay at these moderate temperatures? Luckily, I have easy access to an oven capable of pyrolysis temperatures (approaching 500 C). And then deep freezing of course.

Yeah, you are right, that's exactly what we need for the ultimate test bench!

More seriously,

I was thinking a lot regarding the aspect you have observed/explained related to the decrease of the banding after a warm-up of the lens.

There are some items that are still not clear to me:

- the warm-up should be "active" (you have to shoot a certain number of pictures before getting some "less banding" pictures) or it's enough to wait some time without shooting (maybe only playing in the menu in order to keep the camera awake) and then even the first shot is less banding affected?

- are we speaking of a warm-up time of one-two minutes or more?

Please, let me know if I missed something.

Ciao

Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to NeroMetalliko, Oct 20, 2012

NeroMetalliko wrote:

Anders W wrote:

NeroMetalliko wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Hello,

Olympus claims for the E-M5 an operating temperature in 0-40°C. I can assume the same should be valid for the Panasonic 20mm.

It will be interesting to try some cold/warm temperature operating test (with no warming-up time obviously).

Using a common refrigrator and a open oven, with some care, it could be theoretically possible to put the lens and body at 4-6°C and at 25-30°C and perform a set of lens/camera shoots like for example:

- cold lens/cold body

- warm lens/warm body

- warm lens/cold body

- cold lens/warm body

The first two combinations are more easy to do, the two "mixed" ones are more tricky because a fast action is required to join the two items and shoot the test image in very short time.

The main problem I see will be the condensation that will surely occur, so I'm NOT recommending to anyone to try these kind of test.

And... ehm... I have sold my 20mm lens, so "unlucky" I cannot try with my camera...  ehm...

Ciao

Thanks for these additional testing suggestions. But why stay at these moderate temperatures? Luckily, I have easy access to an oven capable of pyrolysis temperatures (approaching 500 C). And then deep freezing of course.

Yeah, you are right, that's exactly what we need for the ultimate test bench!

More seriously,

I was thinking a lot regarding the aspect you have observed/explained related to the decrease of the banding after a warm-up of the lens.

There are some items that are still not clear to me:

- the warm-up should be "active" (you have to shoot a certain number of pictures before getting some "less banding" pictures) or it's enough to wait some time without shooting (maybe only playing in the menu in order to keep the camera awake) and then even the first shot is less banding affected?

- are we speaking of a warm-up time of one-two minutes or more?

Please, let me know if I missed something.

Ciao

No, I don't think you missed anything. I still don't know whether it's just about waiting some time or whether it is about excercising some particular functions of the camera and/or lens. I don't pretend to know anything more than I have actually tested and reported. At the moment, it's just a "cook-book" recipe: put the lens on, shoot 5-10 frames before you start, and the banding will be gone. With a bit of further testing, I hope to get further than that but right now I don't have more than a recipe, the latest ingredient of which is that warming up the camera doesn't suffice. You have to warm up the lens too.

Hope to be back after I have done some further testing to sort things out.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Anders W
Forum ProPosts: 17,838Gear list
Like?
Re: Banding with the E-M5 and the 20/1.7: A potential work-around
In reply to Detail Man, Oct 23, 2012

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

Detail Man wrote:

Anders W wrote:

A little update to my original post: Clearly, the banding depends on whether the equipment is properly "warmed up" before use. But is it the camera or the lens that needs to be warmed up?

Tonight, I ran a first test to answer that question. Rather than keep the 20/1.7 mounted on the camera throughout, I mounted the 14-45 before turning the "cold" camera on. I then repeated my standard procedure of shooting 5-10 frames so as to "warm up" the camera. Only then did I mount the "cold" 20/1.7. If only the camera had to be "warmed up" to get rid of the banding, we would now expect clean images right away. If, however, the lens required the same treatment, it would take a few shots to get there as usual.

Result: Clear banding on the first trials, then gradually disappearing as usual. So clearly, the lens requires a bit of idling too before being ready for use.

Sounds like good news, really. The problem (as you observe in the case of your copies of the camera body and particular lens, anyway) recedes relatively quickly - (perhaps) constituting a fairly straightforward work-around.

It sounds like the thermally-related nature of what you are seeing likely rules out any mechanical electrical connection issues, and rules out issues of adequate electrical shielding existing - as these things are not temperature-related. However, in the scenario that the lens' is (for this relatively brief time period) radiating electro-magnetic fields which are then inducing voltages/currents via entry-points within the signal-processing sections of the camera body itself, then I suppose that the overall quality of electrical shielding (in the lens, as well as in the camera body itself) could be a relevant issue.

In a previous thread (that gollywop started, regarding continuously adjusted Shutter Times and F-Numbers in the E-M5 when in Aperture Priority or Shutter Priority shooting modes, respectively), I located and referenced information which appeared to indicate the likelihood that the mechanical position of the mechanical shutter mechanism in these lenses is controlled by a stepper-motor (probably not in itself being highly suspect).

(Perhaps) it (might) come down to temperature/friction related issues in the lens' mechanical-shutter mechanism itself (as that relates to and interacts with the E-M5's electronic-shutter operation) ? I know very little about such subjects myself - but you and others likely may know more ?

Happy to see you drop in again DM. I would have asked you to do so if you hadn't volunteered. I think I am pretty decent tester of pretty much anything. But in terms of theorizing, I am far off home base here. My knowledge of electronics on the hardware level is ... ahem ... spotty. So I much appreciate your input about likely causes. This is your field after all.

Supposing for the time being that my particular copies of lens and camera are not unrepresentative: What would be your take on what's going on here?

Supposedly, we are dealing with three main components on the lens side (apart from the lens elements themselves and the plastic barrel, both of which I think we can acquit right away): the PCB, the focus motor, and the aperture actuator. The focus motor seems like an unlikely bet to me since it supposedly isn't active during the critical phase (sensor read-out). Remains PCB and aperture actuator. Could the PCB have capacitors or something like that, which might cause strange ("spiky") power drain? As I said, the aperture actuator would be my favorite culprit unless I knew I can get banding in a "cold" state even when shooting the lens wide open where the actuator (supposedly) wouldn't need to go into action at all.

Capacitors (in general), and as involved in any filtering of the power-supply lines (acting as a parallel-connected shunt impedance) are going to charge up quite quickly upon lens power-up. Ditto for any series-connected inductance components where it comes to magentic flux levels.

Any additional power-supply current drain that such capacitor charging represents would likely be very short-lived (relative to the time you report of recording several shots, etc.).

Such a current-drain would be relevant only in the case of my originally stated idea - that a high level of current-consumption (might) be causing power-supply circuitry locates in the camera body itself to have increased high-frequency switching-noise on various DC-voltage power-supply lines - until (for some reason), the active semiconductors in the camera's power-supply regulation circuitry itself thermally stabilize.

The above scenario is perhaps a bit of an "outside" scenario - as a well designed power-supply regulation system (should not) require such a long time period to thermally stabilize.

But (if, as you say), the in-lens aperture-actuator as well as focus-motor appears to not be involved, and given the information below about capacitors, it should not be ruled out ?

Regarding the ability of capacitors to effectively function as filters of the power-supply lines. Their effectiveness in doing so is not related to whether or not they have yet fully charged-up to some particular DC-voltage value - and the temperature-coefficient of their value of capacitance (even in the case of large capacitance valued electrolytic capacitors) is small (on the order of a fraction of a percent per degree Celsius).

Well, it's a nice mystery isn't it? We both have trouble coming up with any reasonable hypothesis. If the focus motor and aperture actuator aren't involved, we have only the PCB left. And why should it need "warming up". Can a "spiky" load (which might be there) as opposed to a strong load (which seems unlikely to be there at the relevant, i.e. readout, time) cause problems here?

If the lens presents a time-varying current-load to the camera-body generated DC power-supply voltage lines that it connects to, that could cause voltage-transients and/or higher levels of high-frequency switching-noise to appear on such DC power-supply voltage lines.

It may (also) be that the PCB of the lens itself (also) uses (further) voltage-regulation circuitry on-board - in order to attempt to isolate the effects of electrical loads of focus and aperture stepper-motors from sending transients and noise back towards the camera body's DC power-supply voltage lines. That would not surprise me - as motors/actuators (representing high inductances seeing pulsed drive-currents with high rates of change in time) are the most likely generators of such unwanted voltage-spike "garbage" on DC power-supply lines.

I find your idea that the PCB of the lens itself might be equipped voltage-regulation quite realistic. I am also thinking quite a bit right now about the possibility that this voltage regulator might itself be the culprit (the source of the interference) or at least the ultimate culprit (if there are some intermediares on the way to the sensor).

Although I am still far from done with my attempts to test exactly what in the lens needs warming up for the banding to disappear (this is rather complicated and time-consuming since I can only reliably reproduce the banding when things are completely cool), I have some indirect indications that it might be this voltage regulator (if indeed there is one).

There are really only two other things in the lens that can reasonably be suspected: the AF motor and the aperture actuator. But the AF motor is off during the critical phase (read-out) and it is possible to produce at least some banding even when the aperture actuator is presumably off as well (lens shot wide open). However, the in-lens voltage regulator is presumably on whenever the lens is on, even if the lens is doing nothing except waiting for instructions. Hence, it fits the bill in terms of timing. Moreover, the voltage regulator might be equipped with switching circuitry and a clock, possibly with frequencies of the kind that might give rise to the banding. The question is whether it could do so even if it is not under any heavy or "spiky" load at the critical time (the time when the sensor is read) but merely sitting there doing next to nothing (but not quite nothing). Would be interested in hearing your thoughts about this.

BTW: I have been looking for images of what the 20/1.7 looks like when disassembled. I haven't found any but the ever-curious Roger Cicala (I like this guy) has "unmade" two other pancakes that are likely to have a very similar anatomy: the Olympus 17/2.8 and the Sony 16/2.8. See here:

http://www.lensrentals.com/blog/2012/03/unmaking-pancakes

Perhaps looking at those images, especially the PCBs and their location, might give you some additional ideas. Just taking my wild chances here of course. Not enough of an electronics expert myself to really know what other people can see or not.

Unless you have a better idea, my next experiments will try to separate keeping the lens "warm" (having it mounted with the camera on) from shooting it. If it's just about keeping the lens "warm" without shooting it, we're down to the PCB, wouldn't you say?

Interesting times, aren't they. Gentlemen, start your lenses please!

See my edited (most recent) post for some additional thoughts in general.

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +21 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads