Anatomy of the Leica M8

[4season]

Advertisement (gone after registration)

Mark,

 

I think you've just helped Leica to sell a few more M8s: They should reprint your article! Very impressive-looking innards: Looks like a clean design with no obvious last-minute kludges like hand-installed jumpers or cut PC traces. Amazing how totally new the camera is, yet it's still unmistakably an M.

 

Jeff

[marknorton]

Jeff, yes, it's impressive and most of all, there are just 6 wires in the whole camera.

[scott kirkpatrick]

Mark, an afterthought question -- you found an optical sensor that allows monitoring correct operation of the shutter cocking mechanism, and two toggle switches that read the frame lever. Were there any sensors, optical or mechanical, that would allow the camera electronics to know the focal distance that the rangefinder mechanism has determined? I can see pieces of flex going into the viewfinder area -- were those only to drive the LED information readout?

 

scott

[marknorton]

Scott

 

There are two sensors for the shutter wind, one to detect motion and the second to detect the position of the shutter lock.

 

The rangefinder is completely independent of the rest of the camera. Its only inputs are the roller lever and frame lever position. The flex you're referring to makes its way down the front of the camera to the 6 bit sensor and frame lever position switches, there's nothing in the camera to let the camera electronics know the position of the roller and hence focusssing cam/focussing distance.

 

As you know, the roller only moves about 3-4mm between closest focus and infinity, so sensing that position to, say, 0.01mm would be quite tough.

[barjohn]

Mark,

I can't remeber if I added my applause to that of the others earlier or not, but you did a fantastic piece of reverse engineering, especially in identifying the various components and their functions. Dissassembly and reassembly is one thing, figuring out how it works is quite another. It really makes me wish I had the spare time to do the same to the software. I could then add many of the features that people have been asking Leica to do. Just looking at the code with a hex editor leads me to believe it wouldn't be too dificult. Most hardware engineers these days have a pretty good knowledge of firmware, are you in that category?

 

If so are you aware of where one might find the assembler and other SDK tools for the processors used in the camera?

 

Thanks again for a fantastic job!

[marknorton]

John

 

As I explained in my opening post, I've been careful to limit what I've published to what can be gleaned by simple observation and interpretation.

 

I'm first and foremost a software engineer but I have drawn the line at disclosing anything specific about the software and detailed hardware design which I regard as strictly Leica's Intellectual Property.

 

The idea of the thread was to celebrate the excellence of Leica's engineering, not lay bare their IP.

[barjohn]

Advertisement (gone after registration)

Mark,

 

I dare say you are being too modest. I seriously doubt that very many people could have provided as much insight into what each circuit board, chip or component did from mere observation without a significant amount of engineering knowledge (software & hardware). As to whether you revealed or did not reveal any Leica IP would be hard to know withut knowing what Leica considers there IP to be. A quick look at their code told me which processors were used but not to the detail revealed y your dissassembly and comments. Was that a Leica secret? I don't know.

 

If you buy an author's book and write explanatory notes in the margins have you violated the author's copyright? I don't think anyone would think so. (Legally, I know you wouldn't have). Adding functionality to Leica's code is no different. You are not depriving Leica of revenue as it is useless unless you own a Leica camera and have already paid for the use of the IP.

 

While I can't speak to UK law, Leica has protected there IP in one of several ways, copyright, patent or trade secret. The first two are legal protections and the latter is considered fair game to reverse engineering in the USA. It is not fair to be given access to the IP and then disclosing it. If the IP is patented it must be disclosed so you would not be violating their IP and copyright by its very nature is a disclosed thing, music, pictures, books, software, etc.

 

If what you are saying is that you may have discovered some of their trade secrets and you choose not to disclose them, that is of course your perogative but it is a personal decision to which I place no moral value. In any case, I don't believe my question as to your knowledge of the location for the SDK tools for commercially available processors was asking you to violate any IP at all.

[marknorton]

John, each of the three processors - Intel, Analog Devices Blackfin, Renesas has their own commercially available SDKs and it looks like the Intel processor updates the other two - and the Xilinx FPGA - from the firmware UPD file.

 

Incidentally, the Analog Devices Blackfin costs just $32.95 in single quantities, $25 if you're buying 500. Something of a bargain considering what it's doing...

[barjohn]

Mark,

 

Thanks for the info. I just had my power switch fail, any clues from what you have seen as to what might cause this? It seems software related as pushing the shutter release turns it on but I never had to that before today. I had the camera with me in my car and the outside temp was a 101 today so I don't know if it is a temp sensitivty issue or not.

[Guest sirvine]

I seriously doubt it's an IP violation unless someone takes this information and uses it to commercial purposes. In that case, I would think that the profiteer would be the infringer. None of us signed confidentiality terms when we bought the camera, so simply disassembling and posting the inner bits (without sharing any privileged information) is probably fine. (Obligatory disclaimer: I'm not an expert in this area, though, so don't rely on my advise to your detriment.)

[R.Morrison]

This is the most impressive posting that I have seen since I've been a member of the forum. You have expended much effort with your macro-photography (I've done enough of it myself to know). You've done us a great favour.

 

Thank you,

Robert Morrison, M4-P, etc.

[Stevez4]

Thank you Mark. I have a much deeper respect for my M8 as a result of reading your astounding study.

Cheers

Steve

[jlm]

Mark: any more informed idea of what the blue dot is for?

[marknorton]

John, the picture of the light metering circuit shows that the blue dot sensor and the main light metering sensor are handled by the same chip, probably an amplifier with possibly a logarithmic characteristic. Take the difference between them and you get a signal which is proportional (and approximates to) the lens aperture.

 

The photo-diode for the blue dot sensor is at the back of a light "tube" which will give it a relatively narrow field of view meaning it will be metering, approximately, the centre of the image.

 

So, it looks like the camera is using the blue dot to estimate the shooting aperture (which may be relevant for image correction); the camera is also using the ambient light level to set the brightness of the viewfinder LED display.

 

Scott has identified that another input to the image processing appears to be the sensor temperature which may be used in noise reduction. The hotter the sensor, the noisier the image.

[eronald]

 

The photo-diode for the blue dot sensor is at the back of a light "tube" which will give it a relatively narrow field of view meaning it will be metering, approximately, the centre of the image.

 

Hmm - the blue dot has tunnel vision

 

Can it also serve as a flash metering sensor, I wonder ?

[gogopix]

Amazing

 

Will you try to reassemble it?

 

I believe Mark is a forensic pathologist.

 

The 'bodies' are not required to 'work' after the autopy

 

:-)

 

Good show, Mark!

[marknorton]

Hmm - the blue dot has tunnel vision

 

Can it also serve as a flash metering sensor, I wonder ?

 

Interesting question, I wondered if it was taking over the role of the sensor in a conventional automatic flash gun which shuts off the flash when sufficient light has been reflected back but trying it with the SF-24D set to auto, I could see no difference when the sensor is covered and when it's not. Doesn't mean it's never used of course but the blue dot remains something of a mystery.

[scott kirkpatrick]

Interesting question, I wondered if it was taking over the role of the sensor in a conventional automatic flash gun which shuts off the flash when sufficient light has been reflected back but trying it with the SF-24D set to auto, I could see no difference when the sensor is covered and when it's not. Doesn't mean it's never used of course but the blue dot remains something of a mystery.

 

Here's another experiment to try. Cover the blue dot and see what sort of AWB values you end up with. Sometimes it doesn't matter, sometimes it does, and I haven't figured out what it depends on. But then the AWB is like that with the blue dot uncovered, too.

 

scott

[scott kirkpatrick]

I believe Mark is a forensic pathologist.

 

The 'bodies' are not required to 'work' after the autopy

 

 

But in fact, it did. This unit must have a future as Mark's heavy rain camera, or for covering mud-wrestling.

 

scott

[Advertisement]

Hallo,

Look here: Leica M

[eronald]

Interesting question, I wondered if it was taking over the role of the sensor in a conventional automatic flash gun which shuts off the flash when sufficient light has been reflected back but trying it with the SF-24D set to auto, I could see no difference when the sensor is covered and when it's not. Doesn't mean it's never used of course but the blue dot remains something of a mystery.

 

Maybe this function is there to be enabled with a future firmware for a future flash unit.

 

If it has flash-metering functionality, then odds are there will be an integrating circuit sitting behind the sensor.

 

Edmund