Dodgy mount registering/not registering 6 bit coding

[thighslapper]

Advertisement (gone after registration)

I also concluded it was the actuator.

 

The sensor will register almost anything that is not a full dense black as 'white' ..... so if the problem is misalignment it is logical to assume it results from part of the 'non-black' bits being registered by an adjacent sensor as 'white' and producing a coding that does not fit with what the camera expects.

 

A while ago, as an experiment, I took my rogue 75/2 and put black marker over EVERYTHING .... including before and after the 6 bit codes....plus in between the pits.... and reduced the white bits to 1mm central dots.

 

It still registers as 75/2 ..... but as before, ONLY when pushed as far as it will go clockwise, and even the tiniest movement backward will de-register it. As a result I conclude it cannot be the alignment of the sensor array/6 bit coding that is at fault.

 

I have hunted high and low and can find no diagrams of how this works or whether it is adjustable .... even Mark Nortons disassembly of the M8 is unhelpful.

 

If anyone out there knows how it works .... please let me know !

[ModernMan]

Ok, I'll try this: take a 6-bit lens flange (cheap version) for 90mm, which has the largest frame line actuator and then iteratively reduce the size of the actuator by say 0.1mm until it is the size of the 50mm actuator, each time observing the behavior of the lens recognition.

 

If we're right then we would expect to observe correct behavior when the actuator is somewhat (0.2mm perhaps) larger than "correct" for 50mm, and also observe the problematic behavior when the actuator is slightly less than nominal for 50mm. The 90mm actuator bayonet lobe is about 0.8mm larger than the 50mm version (which is about 1.0mm larger than the 35mm version).

 

Even if this confirm the hypothesis the larger question remains of how to correct it in the camera body.

[furcafe]

As I mentioned in this thred (post #5), I started my own thread soon after getting my 240 re: problems I've had w/it properly recognizing my 28-35-50 Tri-Elmar when set to 50mm:

 

http://www.l-camera-forum.com/leica-forum/leica-m-type-240/300473-m240-6-bit-recognition.html

 

I'm in the U.S., so I'm going to wait until Leica NJ is qualified to work on the problem.

 

I have read this before but it was to do with not reading a 6 bit code in 3rd party lenses. However, I have noticed this annoyance with my Noctilux that if you mount the lens as usual and lock it in, it comes up with the Noct profile fine, but if you 'slightly' twist it (even adjusting the focus) can do this.. it then becomes 'uncoded'... but again, slight twisting it back brings the Noct profile back up. Although not a game breaker, it is very very annoying.. is it because of the weight of the lens? :S

[ModernMan]

From Mark Norton's classic post on M8 Anatomy, pictures of the frame line actuator position sensor coupling switches! The mechanism on the M240 may be a bit simpler since there's no preview lever.

 

 

...

 

The final surprises come when you remove the battery case. The front of the battery case and the lens throat are a single plastic moulding and underneath the battery case is the circuit for the lens coding detector.

 

 

In this picture, the lens code sensor connects bottom right and there are six (black) transistors, presumably to amplify the signals from the photodiodes in the sensor. There is also a pair of tiny toggle switches which are activated by the frame selection lever.

 

Clearly, repairing this part of the camera if it should fail is complex, involving as it does almost complete dismantling of the camera.

 

 

In this case, the upper switch is activated. When the lever is in the centre position, neither switch is activated. In the lower position, the lower switch is activated. These switches are used to determine the selected focal length when the Normal Tri-Elmar is mounted on the camera.

 

That’s it. I hope you’ve enjoyed this tour of the Anatomy of the Leica M8.

[ModernMan]

If suspicions are correct and the issue lies with the tramline actuation switch coupling, then a firmware fix is possible. The firmware can just ignore the frame lines, as the 6-bit code is sufficient to determine the lens type in (almost) all cases. The MATE is an example where the camera needs to consider the frame line input to determine the correct focal length to put in the exif, but for the most part, if the camera says uncoded because of a frame line / coding mismatch, it is just being unnecessarily fussy.

 

(still waiting for my test flange to arrive in order to validate the predicating suspicion)

[ModernMan]

Ok, I completed my investigation into electronic reading of the frame line "information" from the lens, and here's what I found.

 

As discussed above (so safely skip this and the next paragraph if you've already read the thread), while the 6-bit code is read from the markings on the back of the lens, the frame line selection is "read" from the size of the bayonet lug at the 3 o'clock position (when looking at the sensor). That bayonet lug comes in three sizes and actuates the mechanical stuff that brings one of the three pairs of frame line masks into position in the optical viewfinder. In the M240 (and also the M8 and M9) that lug is also mechanically toggles a pair of microswitches enabling the camera electronics to know the frame lines selected by the lens. The frame line lug comes in three sizes (small: 35/135, medium: 50/75, and large: 28/90, each differing by about 1mm).

 

For most Leica-M lenses, the lens model and consequently the focal length is known simply from the 6-bit code (tri-elmars being an exception to this). The camera "validates" the 6-bit code with the frame lines selected and so long as there is a match the lens type is automatically detected. However, if the 6-bit code (say 110001 for the 0.95/50) is a mismatch for the frame lines detected (say 35/90, rather than 50/75) then the camera reports "uncoded".

 

The problem is the "dodgy behavior" reported here for lenses actuating the 50/75 frame lines, only with the M240, such lenses being detected correctly on the M9. The hypothesis being tested here is that in the case of 50 and 75mm lenses, the electronic frame line detection malfunctions on the M240 and hence the "dodgy" lens detection behavior.

 

Here's how I investigated:

 

(1) got a cheap 28/90 6-bit lens flange on ebay

(2) code it 110001 for 0.95/50

 

This flange ended up as collateral damage in the investigation of dodgy lens detection.

 

(3) mounted it on the M240, and as expected the camera showed "uncoded" and 28/90 frame lines

(4) reduced the size of the lug by 1mm in several steps, progressing from 28/90 size to 50/75 size

 

Lug reduction happening.

 

(5) at -0.4mm still showed uncoded, 28/90 frame lines displayed

(6) at -0.5mm showed 0.95/50, but 28/90 frame lines displayed (!)

(7) at -0.75mm (nominally 0.25mm greater than the correct size lug for 50/75) showed 0.95/50, 50/75 frame lines displayed (no dodgy behavior when mount play is exercised)

(8) at -1.0mm (nominally the correct size lug for 50/75) showed 0.95/50, 50/75 frame lines displayed (dodgy behavior when mount play is exercised)

(9) at -1.1mm, showed uncoded, 50/75 frame lines displayed

 

My conclusion is that (at least in my camera) the electronic frame line detection (using the switches) borders on being out of calibration. The mechanical frame line actuation is fine.

 

I think the fix has to be an internal adjustment or parts replacement inside the camera. (Maybe I'll send it in when I get one of those new Sony's so I'm not without for the repair duration -- probably a trip to Solms).

[ModernMan]

Advertisement (gone after registration)

Pics for post above.

 

This flange ended up as collateral damage in the investigation of dodgy lens detection.

 

Lug reduction happening.

[jdlaing]

Pics for post above.

 

This flange ended up as collateral damage in the investigation of dodgy lens detection.

 

Lug reduction happening.

 

keep trying

[ModernMan]

keep trying

 

Indeed...

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

This flange ended up as collateral damage in the investigation of dodgy lens detection.

 

 

Lug reduction happening.

[algrove]

modernman

 

Maybe you should avail yourself to Leica so they can start to get things right. Thanks for all the hard work. Love the drill bit shot.

[jdsheepdog]

This may be too basic a question, but as someone who has a bag of old glass, the only exception being the coded 28 mm Elmarit asph, and who is contemplating plonking down a large chunk of cash on the M240, can the lens selection on the M be set manually?

[ModernMan]

This may be too basic a question, but as someone who has a bag of old glass, the only exception being the coded 28 mm Elmarit asph, and who is contemplating plonking down a large chunk of cash on the M240, can the lens selection on the M be set manually?

 

The "lens detection" can be set to manual and then a lens type can be selected from the menu. That will determine what gets put in the exif info and what lens corrections get applied. Unlike automatic lens detection, with manual lens selection the frame lines need not agree with the lens selection.

 

However, unlike other M cameras including the M9, the M240 has no frame line selection lever and the frame lines will be determined by the lens (the bayonet flange of the lens).

 

Interestingly, none of my "old" Leica glass, (now 6-bit coded) including 50 and 75, suffer from the dodgy mount problem (where "old" means that the 6-bit coding required the 5-screw flange, the one where the 6th screw is omitted because its int he middle of the 6-bit code).

[thighslapper]

Thanks MM.

 

Nice to see suspicions/logic proved by experimentation !

 

From those that have sent lenses/cameras back I think Leica just replace the flanges (camera and/or lens) and hope the 'tolerance' in the build allows it to work ok. From Marks original M8 strip-down it would suggest a complete dismantling of the M would be required to correct the misalignment of the board the switches are mounted on.......

 

I was Leica I would replace tha camera and recycle the 'old' one ....

[wlaidlaw]

I feel part of the issue is the change in the design of the lens locking pin. In earlier M’s, this is a bar with a rounded end at the throat side of the locking pin. In the M240, it is now a rounded pin. I have had problems with third party adapters (e.g. Leitax R to M and a Fotodiox R to M) failing to lock in place properly and also, some older Leica lenses seem to have too much radial movement when locked. This means that hand coding switches on and off.

 

Wilson

[ModernMan]

I feel part of the issue is the change in the design of the lens locking pin. In earlier M’s, this is a bar with a rounded end at the throat side of the locking pin. In the M240, it is now a rounded pin. I have had problems with third party adapters (e.g. Leitax R to M and a Fotodiox R to M) failing to lock in place properly and also, some older Leica lenses seem to have too much radial movement when locked. This means that hand coding switches on and off.

 

Wilson

 

Yes, when things are so precariously borderline such very slight differences can make the difference between apparently all ok, and dysfunction.

[Ecar]

I feel part of the issue is the change in the design of the lens locking pin. In earlier M’s, this is a bar with a rounded end at the throat side of the locking pin. In the M240, it is now a rounded pin. I have had problems with third party adapters (e.g. Leitax R to M and a Fotodiox R to M) failing to lock in place properly and also, some older Leica lenses seem to have too much radial movement when locked. This means that hand coding switches on and off.

 

Wilson

 

Same here - unfortunately...

[pico]

[...] The electronic actuation is presumably by means of a three position switch which is actuated by the same mechanical interface that sets the frame line mask.

 

That would be poor design, and I doubt that is how it works. Why would Leica double the possible fail points? The frame-lines do not disappear or change when you wiggle the lens, do they? With the exception of the toggles for the TriElmar, the mechanics that set the frame-lines is simple enough and has enough degree of tolerance that it works when the 6-bit lens type sensor does not.

[bocaburger]

Had a chance to try a couple of my DIY coded lenses on an M last weekend, they were "recognized" perfectly. One was a Type IV 35 Cron and the other an M-mount 21 f/4 Voigt. I only code the black ones, and my pits don't look anything like the factory ones. On the 35 I have a single pit surrounding a mount screw, and filled with glossy black touchup paint. The 21 has a single narrow oblong pit running almost the entire width of the flange (from OD to ID) filled with flat-black model airplane paint. Both lenses locked in solidly on the M with no radial play. I realize this is a sample of one body and 2 lenses, but just as statistically reliable/unreliable as the rest of these anecdotal reports.

[ModernMan]

That would be poor design, and I doubt that is how it works. Why would Leica double the possible fail points? The frame-lines do not disappear or change when you wiggle the lens, do they? With the exception of the toggles for the TriElmar, the mechanics that set the frame-lines is simple enough and has enough degree of tolerance that it works when the 6-bit lens type sensor does not.

 

From the M8 internal pictures above (courtesy of marknorton) it seems clear that's how it works. I think you are correct that it is poor design (or poor implementation). The mechanical actuator driven by the lens bayonet sets both the frame-lines and the switches, trouble is it doesn't do both exactly simultaneously. Usually close enough but not always.

[Advertisement]

Hallo,

Look here: Leica M

[lawman]

Nova burger - please post a picture of your mounts.