K5 bounce flash overexposure - mystery solved! MUST READ!

Started Jun 7, 2011 | Discussions
dtsui
Forum MemberPosts: 86
Like?
K5 bounce flash overexposure - mystery solved! MUST READ!
Jun 7, 2011

I had my K7, yes, its K7 and AF540FGZ sent to the Pentax factory in Japan to fix bounce flash overexposure problem. After 4 months in waiting and transit, they equipment came back without anything being fixed. Factory claimed they could not find any fault. My K7 actually has the same problem as K5. When the equipment came back from Japan to CR Kennedy (Pentax's agent) in Adelaide I had the chance the test it alongside a K5 and BOTH cameras displayed the same problem with bounce flash, in fact with two separate AF540, mine and a new one. I have come to the conclusion there is bug in the firmware across both cameras. They share the same PTTL algorithm.

Those of you who own a K5 or K7 and AF540 can try the following test to see if you can replicate the problem.

1. Set camera to M 1/125 f/5.6 ISO 250

2. Point your flash up at 75 degree, make sure the ceiling is not too high and is white in color.
3. If you have a zoom lens, set to 18mm or 17mm.
4. Take a shot of a subject at about 2 to 3m and note its exposure.

5. Change the zoom setting on the flash manually and take a shot at each step and note the exposure.

6. You should notice a progressive decrease in exposure as the flash zoom angle changes in steps from 16--> 19--> 24--> 34--> 48--> 58. Note the exposure at 58. It should be the best out of the lot.
7. Now set the lens to 55mm or 70mm.

8. Change the flash zoom angle in steps from 58--> 48--> 34--> 24--> 19--> 16. You should notice a progressive increase in exposure towards 16. Note the exposure at 19 or 16. It should be the best of the lot.

If you follow the test and can confirm the same results then the logical conclusions are:

1. The K5 and K7 PTTL algorithm applies a flash exposure compensation in bounce flash in accordance with the zoom setting on the flash unit.
2. The flash exposure compensation is the opposite of what is required.

To get around this problem with bounce flash, when you shoot at 18mm, manually set your flash to 58, and when you shoot at 70mm, manually set your flash to 16.

I do not understand why Pentax tries to deny the problem and fails to deliver a firmware upgrade for such a blatant and obvious mistake in their flagship camera.

David

soheil
Senior MemberPosts: 2,136Gear list
Like?
Asking for confirmation.
In reply to dtsui, Jun 7, 2011

Maybe some of our technical fellows should take it into their consideration and if it's confirmed, send a feedback to Pentax.
--
--------------------------------
It's the singer not the song.

 soheil's gear list:soheil's gear list
Pentax Optio 330GS Pentax MX-1 Pentax K-7 Pentax K-5 Pentax K-5 IIs +18 more
Reply   Reply with quote   Complain
brecklundin
Senior MemberPosts: 1,995Gear list
Like?
Wow...if this is truly the case then it implies
In reply to dtsui, Jun 7, 2011

That Pentax has not even looked at this issue, ever. This sort of glitch should be readily apparent when examining the source code the the subsystem. Somewhere they is program logic that has a conditional reversed. Stepping through the code should have shown this at some point.

Should this really be the case I imagine someone is gonna get fired at HoyaTax...poor SOB...but still cool if the fix is this trivial, in the software I mean not the hardware.

Hmmm, gets me thinking if this is in the microcode and not firmware fixable then how would a fix be implemented? If there is hardware needing replacement because the code is not updatable and is living on a static ROM then...hmmmm. Maybe a new layer to intercept the signals, reverse them and do the same going back to the flash...might not be as simple a fix as it appears to address at the camera level. Users should be able to adapt manually but then again is that a real fix?

-- hide signature --

Real men shoot Pentax because we were born with our own Canons!!
{Ok...ok, some of use just shoot with a PnS but it always makes me happy! :D}

Reply   Reply with quote   Complain
fotobert
Contributing MemberPosts: 654Gear list
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to dtsui, Jun 7, 2011

Hi,

afaik the Metz tech support has confirmed, that there is a bug in the K5 firmware that cannot be fixed / worked around by the flash manufacturer. Search this forum or try google.

Cheers
Albert

 fotobert's gear list:fotobert's gear list
Pentax K-5 Tamron SP AF 70-200mm F/2.8 Di LD (IF) MACRO Pentax smc DA 18-135mm F3.5-5.6ED AL [IF] DC WR Sigma 70mm F2.8 EX DG Macro Sigma 8-16mm F4.5-5.6 DC HSM +10 more
Reply   Reply with quote   Complain
dtsui
Forum MemberPosts: 86
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to dtsui, Jun 8, 2011

To add insult to injury, I discovered the K7 has developed a temperature dependent lockup problem after the service at Japan. When I first turned on the camera at the agent's place, it would not respond to the shutter or any of the controls, the camera simply froze. We all fumbled for a while thinking it was the lens that I put on it that was at fault so we changed to a different lens and after a while it worked again. I have since taken dozens of shots with it with no problem at all.

Last night I left the camera on the table in a room that was about 16 degrees and it locked up again when I turned it on this morning. I immediately suspected the problem is temperature related. To test my theory, I put the camera in the fridge for about 20 minutes to bring its temperature down and retested it again. It locked up as I expected. I then put it in front of the gas heater to let it warm up for about 10 minutes and it returned to normal. I repeated this for a second time and the same happened.

In hind sight, I believe what got the camera working again at the agent's place was the body heat from the hands that handled it.

I just could not believe my luck!

Reply   Reply with quote   Complain
matthis
New MemberPosts: 20
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to dtsui, Jun 8, 2011

David, which firmware did you use on the K-7?

Reply   Reply with quote   Complain
dtsui
Forum MemberPosts: 86
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to matthis, Jun 8, 2011

FW 1.11 which is the latest

Reply   Reply with quote   Complain
soheil
Senior MemberPosts: 2,136Gear list
Like?
No cold problem with me
In reply to dtsui, Jun 8, 2011

I used my k7 at -12C with no problem.
--------------------------------
It's the singer not the song.

 soheil's gear list:soheil's gear list
Pentax Optio 330GS Pentax MX-1 Pentax K-7 Pentax K-5 Pentax K-5 IIs +18 more
Reply   Reply with quote   Complain
Ishpuini
Veteran MemberPosts: 6,061Gear list
Like?
sure you took everything into account?
In reply to dtsui, Jun 8, 2011

I read your post and this thought ran through my mind. Don't have my camera and flash at hand so couldn't test.

Couldn't you just be seeing the impact of the change of width of the flash beam as you change the focal length setting on the flash?

  • at 58mm: narrow and concentrated beam, i.e. too small for the FOV of a wide angle lens and thus not impacting most of the frame.

  • at 16mm: wide and much more diffused beam, i.e. majority of the light being sent outside the FOV of a tele lens.

I suppose even with bounce the width of the beam would make a difference? Both cases illustrate why setting the opposite focal lengths in lens and flash may cause the frame to (partially) receive less light than anticipated and thus reduce the effect of the faulty overexposure.

I'm not saying the flash system is correct though. I use a diffuser always and that reduces the light as well, compensating for the overexposure. Also I've learned to play with ISO settings as this seems to be another way to enhance or reduce flash impact. I'm guessing your manipulation of the flash zoom setting is another way of compensating an error, rather than the supposed flash zoom setting's wrong interpretation by the camera being the cause of the overexposure as such.

Wim

-- hide signature --

Belgium, GMT+1

 Ishpuini's gear list:Ishpuini's gear list
Pentax K-5 Pentax Q Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited Pentax smc DA* 16-50mm F2.8 ED AL (IF) SDM +14 more
Reply   Reply with quote   Complain
dtsui
Forum MemberPosts: 86
Like?
Re: sure you took everything into account?
In reply to Ishpuini, Jun 8, 2011

I thought about that as well but there are no signs of uneven illumination within the frame when the flash zoom setting is changed.

The other confirmation that uneven illumination cannot be the cause is that this works perfectly with the K10D and AF540FGZ. The K10D shows no change in exposure when the zoom setting on the flash is changed.

I have posted a similar thread back in Feb and one poster confirmed that his K7 and AF360FGZ displayed no change in exposure when the zoom setting on the flash is changed. His K7 was also loaded with firmware v1.11, same as mine. This is the puzzling part, if the problem only lies in the firmware, how can you explain two different behaviors on different K7 bodies. (forgetting his flash is a AF360 for the time being). I am beginning to suspect there may be two different hardware baselines for the K7, a later baseline that shares some parts with the K5 and an earlier batch. Mine probably belonged to the later batch as I bought it about a year ago, close to the end of the production run after the K5 was released. This can probably explain why the K5 behaves in a similar manner to my K7 in so far as bounce flash is concerned.

Reply   Reply with quote   Complain
kikivrany
Senior MemberPosts: 2,812
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to dtsui, Jun 8, 2011

Hello dtsui

Those of you who own a K5 or K7 and AF540 can try the following test to see if you can replicate the problem.

1. Set camera to M 1/125 f/5.6 ISO 250

2. Point your flash up at 75 degree, make sure the ceiling is not too high and is white in color

I tried your test with my Pentax 360 and Metz 58 AF-1

Clear, you asked for a AF540, but I hoped that my AF 360 is maybe similar to yours and the Metz is similar to the power of your.

I got other results with this flash.
Only to get sure, let me point this out :
I do not post to questioning your finding.

I did my test in a room like you indicated. It was at daylight and with your setting for the shutter / aperture / ISO I got a ca. -5... -4 underexposed metering (clear without the flash).

Then I repeated the test with a highter ISO, so that (without a flash) I had have only a ca. -2...-1 underexposing metering.
I coudn't get your result.

I had my best results, if I used via manual setting the same zoom-position, which the flash would use in the auto-zoom mode.

BTW: maybe it's interesting
The different between the Pentax 360 and Metz 58 in the auto zoom mode is

  • If I bounce my Metz 58 the zoom-reflector went always to ca. 75mm and stay there, independently to the zoom-position of my lens from 28 to 70mm

( same on my ist DS)

  • If I bounce my Pentax 360, the zoom-reflector still changed it's position corresponding to the zoom position of my lens from 28 to 70mm

(same on my ist DS)

My K5 have a production date: 8.02.2011 and the latest software.

best regards kikivrany
--
carpe noctem

Reply   Reply with quote   Complain
dtsui
Forum MemberPosts: 86
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to kikivrany, Jun 9, 2011

kikivrany wrote:

I did my test in a room like you indicated. It was at daylight and with your setting for the shutter / aperture / ISO I got a ca. -5... -4 underexposed metering (clear without the flash).

Then I repeated the test with a highter ISO, so that (without a flash) I had have only a ca. -2...-1 underexposing metering.
I coudn't get your result.

What is ca. -5...-4 underexpose metering? Are you referring to the EV indicator at the bottom of the viewfinder? If you are talking about this then I will shoot at lower ISO to make the flash the main lighting source.

BTW: maybe it's interesting
The different between the Pentax 360 and Metz 58 in the auto zoom mode is

  • If I bounce my Metz 58 the zoom-reflector went always to ca. 75mm and stay there, independently to the zoom-position of my lens from 28 to 70mm

( same on my ist DS)

  • If I bounce my Pentax 360, the zoom-reflector still changed it's position corresponding to the zoom position of my lens from 28 to 70mm

(same on my ist DS)

Now this is a very interesting finding, the two flashes do behave differently in so far as zoom is concerned, did you notice any difference in exposure between the two flashes, say when you shoot at 28mm?

Reply   Reply with quote   Complain
Ishpuini
Veteran MemberPosts: 6,061Gear list
Like?
Re: sure you took everything into account?
In reply to dtsui, Jun 9, 2011

dtsui wrote:

I thought about that as well but there are no signs of uneven illumination within the frame when the flash zoom setting is changed.

Perhaps the bounce diffuses it just enough for it to be difficult to see in a picture? But if you took this into account and checked for it, I suppose the conclusion would be that the effect of the flash beam's width is probably annulled by the bounce.

I am beginning to suspect there may be two different hardware baselines for the K7, a later baseline that shares some parts with the K5 and an earlier batch. Mine probably belonged to the later batch as I bought it about a year ago, close to the end of the production run after the K5 was released. This can probably explain why the K5 behaves in a similar manner to my K7 in so far as bounce flash is concerned.

Could be, because your observation didn't seem to correspond to my experience with my K-7 which was one from one of the first batches. I should revisit my pictures from the last holiday season (with my first stained K-5) to check for this, but I don't remember anything "unexpected"...

Wim

-- hide signature --

Belgium, GMT+1

 Ishpuini's gear list:Ishpuini's gear list
Pentax K-5 Pentax Q Pentax K-3 Pentax smc DA 15mm F4 ED AL Limited Pentax smc DA* 16-50mm F2.8 ED AL (IF) SDM +14 more
Reply   Reply with quote   Complain
kikivrany
Senior MemberPosts: 2,812
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to dtsui, Jun 9, 2011

Hello dtsui

kikivrany wrote:

I did my test in a room like you indicated. It was at daylight and with your setting for the shutter / aperture / ISO I got a ca. -5... -4 underexposed metering (clear without the flash).

Then I repeated the test with a highter ISO, so that (without a flash) I had have only a ca. -2...-1 underexposing metering.
I coudn't get your result.

What is ca. -5...-4 underexpose metering? Are you referring to the EV indicator at the bottom of the viewfinder? If you are talking about this then I will shoot at lower ISO to make the flash the main lighting source.

yes, I mean the EV value.

here I used -1 EV for that the flash must add a '1 EV flashlight' for a correct exposed picture.

BTW: maybe it's interesting
The different between the Pentax 360 and Metz 58 in the auto zoom mode is

  • If I bounce my Metz 58 the zoom-reflector went always to ca. 75mm and stay there, independently to the zoom-position of my lens from 28 to 70mm

( same on my ist DS)

  • If I bounce my Pentax 360, the zoom-reflector still changed it's position corresponding to the zoom position of my lens from 28 to 70mm

(same on my ist DS)

Now this is a very interesting finding, the two flashes do behave differently in so far as zoom is concerned, did you notice any difference in exposure between the two flashes, say when you shoot at 28mm?

I see no big differents. The only thing is that if I use -5 EV or lower EV values, I must check if my Pentax 360 and the Metz 58 have the power to compensate this, if I bounce it, depending on the room.

best regards kikivrany
--
carpe noctem

Reply   Reply with quote   Complain
soldbear
Contributing MemberPosts: 804
Like?
No problem with K7 and 540
In reply to dtsui, Jun 10, 2011

I did a test last night with my K7 (firmware version 1.11) and a copy of 540 flash. I could not reproduce the problem.

The lenses I used were (1) DA* 16-50mm at 16mm and at 50mm and (2) DA 70mm Ltd.

I set the camera as instructed: M mode, 1/125, F/5.6, ISO 250. I tested all metering modes: spot, center-weighted, and matrix. 540 was in p-TTL mode, manual zoom, head tilted 75 degrees toward an 8-ft white ceiling.

The exposures were practically identical (almost perfect), regardless the zoom setting of the flash head.

Unfortunately, I bring the wrong SD card with me to work, so I can't post the photos now.

Reply   Reply with quote   Complain
SungiBr
Regular MemberPosts: 109Gear list
Like?
Re: K5 bounce flash overexposure - mystery solved! MUST READ!
In reply to dtsui, Jun 10, 2011

it does make sense to me.

I noticed that flash exposure gets better when you reach the longer focal lenghts of one given lens.

On mine, i came to this because when setting Flash ev to -1.7, when i got longer than 35mm on the 18-135, flash start to pop weaker, as going to tele end, on both direct and bounce (have a af 360)

 SungiBr's gear list:SungiBr's gear list
Pentax K10D Pentax K-5 II Pentax smc DA* 16-50mm F2.8 ED AL (IF) SDM Pentax smc FA 31mm F1.8 AL Limited Adobe Photoshop Lightroom 3 +1 more
Reply   Reply with quote   Complain
dtsui
Forum MemberPosts: 86
Like?
Re: No problem with K7 and 540
In reply to soldbear, Jun 11, 2011

soldbear, can you tell me what the serial no. of your K7 body is?

Reply   Reply with quote   Complain
dtsui
Forum MemberPosts: 86
Like?
Re: No problem with K7 and 540
In reply to soldbear, Jun 11, 2011

soldbear, can you also tell me roughly when did your purchase your K7?

I am very vexed by this problem, same firmware, same camera, same flash but different behaviour, how can you explain that? And this is down to some very fundamental operational difference, not like a switch is easily broken in one unit but not in another unit. Perhaps the only logical explanation is that there are different hardware baselines for the body (and/or the flash) with subtle differences that leads to such difference in behaviour.

If this is true, it begs the question about the competency of Pentax in configuration and quality control.

Reply   Reply with quote   Complain
soldbear
Contributing MemberPosts: 804
Like?
K7 S/N and test photos
In reply to dtsui, Jun 11, 2011

K7 S/N 34055xx, bought from Willoughby's (via Amazon.com) December 2009.

Here are the photos from the test, all arranged (left to right, up to down) in the FL settings of the flash head, from 16mm to 58mm.

DA* 15-50 at 16mm:

DA* 15-50 at 50mm:

DA 70mm Ltd.:

Reply   Reply with quote   Complain
soldbear
Contributing MemberPosts: 804
Like?
Correction
In reply to soldbear, Jun 11, 2011

DA* 16-50, not DA* 15-50.

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads