Time to let go?

[Joseph S Wisniewski]

Thanks to that louse Joe, I have to eat crow. Do you have any good
recipies?

Joe's a vegetarian. Eat brussel's sprouts

--
Normally, a signature this small can't open its own jumpgate.

Ciao! Joe

http://www.swissarmyfork.com

 

[Chato]

my idea posted below.

A small latch, activated by the camera turning on, that locks the door release, and deactivates when the camera shuts off?

Dave

As far as it only being a switch, it's a part that has weight,
needs to be soldered (often at considerable expense, because
switches often can't go through reflow soldering with the rest of
the components, and require hand assembly), and needs to be
accounted for in system reliability calculations (the infamous FMEA
that we engineers hate so much: failure modes and effects analysis).

All well and good. But I would like to give you a second chance to
prove me a moron.

A micro switch is one of the simplest gadgets I know of. And when I
say this I am referring to a Good micro switch. And while I'm not
an engeenier, I am a licensed heating Installer, which is closely
related to photography... It's also a big thing in NYC. Sort of
like a master plumber. (Proof of course that plumbers as well as
heating instlaller are idiots).

But plumbers can blame lead poisoning.

So can we painters

Why could this not be embeded in some sort off board? In this case
the up position of the lever would make the switch, and pushing it
down break it. Ok, so it would cost them a dollar for the switch.
It would cost them nothing above that if it was embedded. Moreover,
it's not as if it's the camera mirror. You activate the switch
three or four times a day.

Every component on a board has a total cost that's determined by
its reliability (chance of failure * cost to fix under warranty)
and how much it costs to assemble into the system.

Micro switches in heating and cooling are all hand soldered.
Components in devices like a camera are assembled automatically:
glue drops are screened onto the board to hold components in place,
"pick and place" machines put the component into position, and
reflow solder machines melt all the solder for all the components
at once, soldering them in simultaneously. Components with large
thermal mass compared to the others (like a switch, compared to a
resistor, diode, or chip) won't get hot enough, fast enough to be
soldered well. So whether it's put on a separate board or the same
flex circuit as the main processor, there's still a step that's
going to have to be done by hand.

Are you really saying that you cannot visualise such a switch being
intergrated into a board, so that this hand soldering couldn't be
eliminated?

Yes. And bringing a board into position to get the microswitch to
where it needs to be to contact the CF eject mechanism also adds
cost. Either you make a flex circuit more complex (one more
"tentical" extending from the main board to connect to the switch
in it's final resting place) or you add a second board, which mans
one more board, connectors on this board and the main board, and a
wire with connectors on both ends to connect the switch board to
the main board.

Look at it this way, to help keep cost in the D50 down, Nikon
eliminated the front command dial. That's one such assembly
(switches, sensor board, connectors, and wire). To justify adding a
different board, connectors, and wire for the eject button, they
would have to view protecting the memory card from premature
ejection as being important than a second command dial.

--
Normally, a signature this small can't open its own jumpgate.

Ciao! Joe

http://www.swissarmyfork.com

 

[dinoSnake]

I still have not understood the reasoning behind this...

• Canon: You open the door, camera shuts down and stops writing to

the card

• Others: You open the door and remove the card, camera stops

writing to the card (obviously)

Why would shutting down the camera be better? At least you can not
remove the card by accident...

It has been suggested that the Canon system stops writing in a safe
way, planned manner wheres the other systems only stop writing when
the physical contact with the card is broken.

But that in itself - breaking CF card contacts - can be done in a (somewhat) controlled manner.

CompactFlash is an extension of PCMCIA with built-in ATA compatible interfaces. A CF card communicates with the outside world in exactly the same manner as a ATA HDD inserted into a PCMCIA interface - 16 bit parallel data channels, power and ground.

If Nikon simply does the expedient of having the power and ground pins shorter than the data lines, the CF card will be powered down a few fractions of a second during the eject cycle before the data lines lose their connections.

This simply expedient will "assure" that the CF card will be powered down in the proper organized manner - power, then data.

AFAIK Nikon does indeed shorten the power pins, and AFAIK this is indeed standard in the CF interface.

Also, from where have you gotten the silly idea that a drive using
the FAT file system would get all it's data corrupted because of
this?

It's not a silly idea - the FAT file system stores files in a
series of blocks spread all over the card. Somewhere on the card
resides the File Allocation Table describing which blocks belong to
which file. The FAT is updated every time a file is written, and if
interrupted could leave you with no FAT and thus no files.

The FAT file system can be in 2 versions - FAT16 and FAT32. In FAT16 the entire FAT structures begin on cylinder 0, side 1 sector 2 directly after the MBR. FAT16 has 2 copies of the FAT, variable in length to match the number of allocated clusters, and each write to the drive edits both FATs - the main and the redundant backup.

FAT32 adds extended FAT entries into the disk, which can be at almost any location as those FAT entries are in the form of extended data strings, which the OS IFS interprets as a FAT.

FAT16 is "relatively" easy to recover (I do it by hand), FAT32 much less so.

While CompactFlash does not have “physical” cylinder, heads and sectors the CF’s internal ATA interface segregates the solid-state memory to match the ATA standard.

-------------------------------------------------

Just for edification.

Good luck,
Snake

 

[dinoSnake]

The FAT file system can be in 2 versions - FAT16 and FAT32. In
FAT16 the entire FAT structures begin on cylinder 0, side 1 sector
2 directly after the MBR.

OS boot sector, not MBR. Sorry - no edit available in the forum.

 

[PhotoTraveler]

If it cost them a dollar to do, it's way to much. Electronics work in fractions of cents. A dollar buys you a ton of stuff, or is adding a lot to the cost of the item.

Also increased component count increases manufacturing cost by a lot.

 

[Chato]

If it cost them a dollar to do, it's way to much. Electronics
work in fractions of cents. A dollar buys you a ton of stuff, or
is adding a lot to the cost of the item.

Also increased component count increases manufacturing cost by a lot.

i.e. a latch activated by turning the power on to prevent the door from opening, is probably cheaper to implement than the Canon solution. No connection to the camera electronics at all.

And no doubt Nikon could then use this as an advertising gimick...?

Dave

 

[Tim in upstate NY]

..........In 2 1/2 years and several thousands images shot with a 300D and then a 5D, I've never once 'needed' to open the CF door without first turning off the camera. It's no flaw for me and most others as well I'd think.

 

[Steven Noyes]

i.e. a latch activated by turning the power on to prevent the door
from opening, is probably cheaper to implement than the Canon
solution. No connection to the camera electronics at all.

Who wants to shut the camera off when you want to take the CF/SD card out. Likewise:

Who wants to wait for the buffer to flush when you REALLY need to put a fresh card in because you know what is coming is better than what is IN the buffer. You really need a method to abort the buffer and the CF card door is it.

Steven

--
---
2006 Southern Arizona Monsoon Wildflowers
http://www.pbase.com/snoyes/images_summer_2006_ii

Summer 2006:
http://www.pbase.com/snoyes/gallery/images_summer_2006

 

[Leonard Migliore]

Since the supposition is incorrect, AND since implementation would
cost peanuts, what's the big deal. Just what does a simple micro
switch cost anyway. I've bought them for a dollar a piece in three
packs on Canal Street. I would bet Nikon and Canon could do
better...

Dave

I used to buy all sorts of electrical stuff on Canal but I thought all those stores went away in the '60's.

--
Leonard Migliore

 

[markeast]

... interesting discussion. I shot that off too quickly.

Perhaps I (personally) would have worded it a little differently ... "... Phil seems to keep on about this issue ... is it an issue to anyone else?..."

But I would bet that without the jarring catch line ...no discussion would have ensued!
--
markeast

 

[Aaro]

If what you're saying is correct then it would be equally true of
Canon. Canon has a micro switch that shuts the juice. Nikon has a
micro switch that cuts the juice.

Nikon has nothing of the sort...

Once again, if you have the schematics I'll take your word for it.
Otherwise you are just speculating.

I have the parts lists, and I've torn down Nikon, Canon, and Fuji
cameras. They use standard CF sockets, no microswitch in the eject
mechanism.

Standard CF sockets do however have shorter and longer pins. Shorter pins are shortoff a replacement for the microswitch. Ofcourse one must have wires connected to them and some sw.

Also with a properly designed software it is possible to achieve a file system that only looses the last file written what ever is removed (the CF card or the power).

Aaro

--
Normally, a signature this small can't open its own jumpgate.

Ciao! Joe

http://www.swissarmyfork.com

--



http://www.lumisoft.fi/gallery

 

[Chato]

I used to buy all sorts of electrical stuff on Canal but I thought
all those stores went away in the '60's.

Actaully most didn't vanish until the early eighties.

Two years ago, there were three stores in operation. Now? Who knows

There are still three blocks of cheap clothing stores on Broadway above canal that haven't been replaced by boutiques selling garbage for ten times their value.

Dave

--
Leonard Migliore

 

[Chato]

I turned to the Nikon manual. Nikon, not Canon. It states, turn the camera OFF before pulling the card...

However, I must say, that the Nikon is instant on, so this doesn't strike me as that much of a problem. The problem, as elucidated by so many is that you lose your buffer when you pull the card, and Canon solution is to shut the camera, Nikons is to allow you to commit suicide.

I'm not happy with either, but Canon's is certainly better, and is hardly a "fault."

Dave

i.e. a latch activated by turning the power on to prevent the door
from opening, is probably cheaper to implement than the Canon
solution. No connection to the camera electronics at all.

Who wants to shut the camera off when you want to take the CF/SD
card out. Likewise:
Who wants to wait for the buffer to flush when you REALLY need to
put a fresh card in because you know what is coming is better than
what is IN the buffer. You really need a method to abort the
buffer and the CF card door is it.

Steven

--
---
2006 Southern Arizona Monsoon Wildflowers
http://www.pbase.com/snoyes/images_summer_2006_ii

Summer 2006:
http://www.pbase.com/snoyes/gallery/images_summer_2006

 

[Joseph S Wisniewski]

If what you're saying is correct then it would be equally true of
Canon. Canon has a micro switch that shuts the juice. Nikon has a
micro switch that cuts the juice.

Nikon has nothing of the sort...

Once again, if you have the schematics I'll take your word for it.
Otherwise you are just speculating.

I have the parts lists, and I've torn down Nikon, Canon, and Fuji
cameras. They use standard CF sockets, no microswitch in the eject
mechanism.

Standard CF sockets do however have shorter and longer pins.
Shorter pins are shortoff a replacement for the microswitch.
Ofcourse one must have wires connected to them and some sw.

All the longer pins do is insure that the power connects before the data lines when inserting a card, and that the data lines break before the power lines when removing a card. This prevents a very annoying effect, current flow through "parasitic diodes", that makes enough current to actually power a device flow through the data lines. This current flow can damage the device, or the computer (or camera) it's plugged into.

You'll see similar things on SD and xD cards, too.

Also with a properly designed software it is possible to achieve a
file system that only looses the last file written what ever is
removed (the CF card or the power).

How does your "properly designed software" assure that the directory or FAT table are not being written when the power or the card are removed at a totally random time?

--
Normally, a signature this small can't open its own jumpgate.

Ciao! Joe

http://www.swissarmyfork.com

 

[Joseph S Wisniewski]

If what you're saying is correct then it would be equally true of
Canon. Canon has a micro switch that shuts the juice. Nikon has a
micro switch that cuts the juice.

Nikon has nothing of the sort...

Once again, if you have the schematics I'll take your word for it.
Otherwise you are just speculating.

I have the parts lists, and I've torn down Nikon, Canon, and Fuji
cameras. They use standard CF sockets, no microswitch in the eject
mechanism.

Well, fortunatly for my reputation my first post on this thread
pointed out that I am an idiot. Thus there is a finite limit to how
low my reputation can sink.

Such being the case I have to really apologise to Phototraveler and
Matt. But NOT you.

For you, I'm reserving a special place in Islamic Hell for infidels
who turn out rigth. Bad Joe, bad Joe. If you happen to see a 120
pound White Wolf in your neighborhood, he's actually my flunky
coming to get you...

My spirit guide is the hawk. My current neighborhood, by pure coincidence, has a huge reserve of them within visual range of stately Wisniewski manor. Your wolf won't surprise me.

But them, by stupidity revealed for all to see, what the heck IS up
with Phil and his calling this Canon safty feature a fault?

I have absolutely no idea, unless he simply doesn't want to admit he's wrong...

--
Normally, a signature this small can't open its own jumpgate.

Ciao! Joe

http://www.swissarmyfork.com

 

[Kendall Helmstetter Gelner]

Others: You open the door and remove the card WHILE the directory
is being written and you POTENTIALLY have lost every image on the
card.

In the Sigma DSLR series, when you open the door the rear LCD lights up and warns you not to remove the CF card. It's very noticable.

This is a simple compromise that works really well, as it has stopped me every time I opened the CF door before the camera was off or finished writing.

Each one of these times, I would have been very unhappy had the camera discarded my buffered images. Not all of us are studio shooters and there are many times when you are not able to take that same shot again

Using the Sigma system, I have never once pulled the CF card before the camera was properaly shut down in many years of use, so I have never experienced CF card corruption or dropped images either.

It amazes me Phil does NOT understand this simple design.

Mandatory data loss seems a poor design choice to me when there are other things that can be done, and the images dropped may well be irreplacable. Yes the card is full of equally ireplacable images but even there you have a good change of recovery even if you did do something wrong, and had formatted the CF card to start with (as I always do).

It is not a silly idea. It is partially tied into the CF block
writes of data to the FAT table and is VERY timing specific. You
have to erase the FAT before you update it. This is different than
a HD. This has actually happened to me once on the old 1D based on
the "bad" design of non-controlled shut-down.

I agree a controlled shutdown is important, but let the user have a second change to re-think a stupid action - perhaps a second CF card door behind the first, and that one drops the buffer?

--
---> Kendall
http://InsideAperture.com
http://www.pbase.com/kgelner
http://www.pbase.com/sigmasd9/user_home

 

[Steven Noyes]

When your buffer is almost 512 MB this is can be a serious issue when shooting fast moving situations. It would drive me nuts to sit and wait while I watch the shot I really wanted simply go away.

Steven

--
---
2006 Southern Arizona Monsoon Wildflowers
http://www.pbase.com/snoyes/images_summer_2006_ii

Fall 2006:
http://www.pbase.com/snoyes/images_fall_2006

 

[orlandoom]

3 plus years with 10d, 30K plus pictures and not once have I opened cf door without power off.
Should not be an issue.
don.

 

[Matt105483]

... interesting discussion. I shot that off too quickly.

Perhaps I (personally) would have worded it a little differently
... "... Phil seems to keep on about this issue ... is it an issue
to anyone else?..."

But I would bet that without the jarring catch line ...no
discussion would have ensued!
--
markeast

No worries .. and yes the wording could have been polished

Another idea I just had .. Canon could install a loud beeper that beeps once you open the CF door ... but then its not all that much needed and just ads cost..

Matt

 

[Aaro]

Hi Joe

I think for start one must read the specs before making any hw or sw.

CF specs can be found at
http://www.compactflash.org/

About the pins.
There are three leghts used:

Quote from cf specs:
Description Pin Number L ± 0.10
Power 1, 13, 38 & 50 5.00 [.197]
General All other pins 4.25 [.167]
Detect 25, 26 3.50 [.138]

CD1, CD2 (Detect) pins are at the opposite sides of the card and are generally used to inform the software that the card is not inserted properly or that it is being removed. After the Detect pins lose connection there is still enought time to do many things.

As for the software:

After getting familiar with the way CF cards manage the writing and the household routines there are few ways to prevent data loss.

1. Prepare the directory for the next file beforehand using the max filesize as the lenght and leaving the file status as erased. This needs one sector write and if it is terminated premature nothing is changed.

2. Start writing the file. If it is terminated premature there is 'carbage' at the position of 'erased' file, All Ok.

3. If the file write was successfull, the directory entry and fat tables are changed and adjusted to accordingly. If this last task fails we still have the file on cf as erased and knowing that it is easyly recoverable.

This is only one way that I have used for data collection if the cf card needs to be compatible with computers.

Main cause for loosing all data on CF cards is the fragmentation of the data area caused by deletion of the older (earlier) pics from card. The new files are now written at noncontinuous space and this leads to big troubles even if only one file gets corrupted. In an other thread there was some advice for the case.

Best
Aaro
--



http://www.lumisoft.fi/gallery