Diy Digital Rf - Is It Possible?

I'm not sure how you would determine the actual market for such a product. Many folks would say, "Sure, I'd like something like that." How many would actually plunk down money to make a dumbed down digital camera out of an old Leica is another matter.
 
I'm going to let Frankie handle the personnel end! I have a feel for how long it would take, comparing with digital projects of similar complexity.

On sensors- CMOS and CCD's have a lot of unique timing that the board has to generate. To design a board to handle sensor upgrades means using programmable/synthetic clocks rather than a simple clock source. It's adds complexity and cost. For re-using the fixture: Noise Isolation. The mechanical layout must support the electronics board. The requirement to isolate analog and digital sides of things means the fixture has to be designed around the electronics. So it's a trade-off. You end up revving the mechanical design to support a new generation of sensor and electronics, or run into degraded performance. Ask me how I know.

You could "probably" design a digital back with a socket for the CCD and change it out for color/monochrome/infrared version of the sensor. Store the raw image from the sensor, and let the computer-based software take care of the conversion to a displayable image.

the market- there was not a market for a lot of things. The Nikon SP-2005 and S3-2000 were done "because". And not because there was a big market that was going to make the product profitable. If someone wants to put the NRE into this, limited production of electronics boards and mechanical fixtures is not expensive these days. Miracle of modern limited-production manufacturing. Software development, test, and debug is time consuming. But less time and much easier than rebuilding a Jupiter-3.

So: the most important question is: What processor are we going to use? And a dedicated PIC processor for the Shutter/Sync circuit would make a lot of sense rather than having the main processor having to monitor the camera. Use an analog comparitor for the PhotoDiodes and digital latch for the Flash Sync socket, Interrupt the main processor. I'm not sure how much lead time the Sensor takes to wake up. At worst, use a wake-up button for the digital back as you do for the light meter of a Nikon FE2.

And I forgot to add: Thank God for the 20ms curtain travel time on a Leica M!
 
Last edited:
OK - I will never ask for interchangeable sensors again here! 🙂

This is an outsider, but one that I am sure has good chances. I have met so many collectors and believers in the Leica Myth the past 15 years, and I know several of them would just order it straight away out of sheer interest and need-to-have.

Add some design that fits with a Leica, not copying, but rather building on the design like an architect uses new material to blend in with older buildings, and it would really make people drool... And please make the outer part of the casing in metal! 🙂

As for using iphone as a screen, would it be relatively easy to put bluetooth in it? Then one could use the iphone as a standalone - even making for some really interesting extra uses. (height photography for one.)

A wake-up button would be OK for me, but why not use the bulb-synch of older Leicas? Worst case would then be that M6-7 will need to use a wake-up, while the older ones could have direct sync... (I cannot remember how early it fires, but it is before the release of the shutter to get an even colour from the bulb going off...)

Is there not also access to the release somehow under the cover at the bottom of Ms? It is too long since I have worked on them, but if a small modification/hole in the inner bottom plate is needed most users would not mind.
 
You want to power-down most of the electronics to conserve power. Ideally, have them wake-up as the shutter trips. There is no linkage between the shutter speed dial and the back, so I would look at using a photodiode on each side of the film gate to detect the shutter speed. Basically, like a built in shutter-speed tester, measure it in realtime. Use a low-power dedicated processor for this function. This "interface Controller" would wake the main processor and electronics. Wake up time has to be fast, 0.2ms would mean 1% of the shutter travel.

I like the idea of an embedded processor based system. A Custom ASIC means more NRE and development time. FPGA's use more power and produce more heat than an embedded processor. So- an embedded processor for lower-power, and shift the development burden to software. I need to look at the wake-up time for some of the ARM based and other processors. I just do not know them offhand, need to look at spec sheets. I'm used to getting sub-microsecond interrupt response time out of MIPS R3000 and R4000 based processors. And that was a long time ago. The embedded processor based system would mean slower frame rates. But faster than a double-stroke M3.
 
I have put a lot of think-time into the dM-retrofit project, since 2001 as a matter of fact, and as evidenced by my l-o-o-o-ng posts; as did Brian Sweeney...while we both have full-time jobs; mine perhaps a little more demanding because I had long become an evangelist in my own niche-of-a-niche of a field.

I can say from experience the from-scratch camera needed as a component in building a scanner in 1992/3 was developed in ~6 months by a single dedicated electronics engineer...while I only supplied money and kept him on track, and I picked the lens...an 80mm Apo-Rodagon.

I can also say many of the watershed software system we have launched was written by a single BRAIN who prefers to work late nights alone. We had no capital then, so we paid him in company IOU's...shares.

All the specialized hardware we have to make was designed by a U.S. Navy and Zeiss Jena trained optical/mechanical engineer now retired. He was indoctrinated to believe design failure [in battle] means death to users. He will serve again immediately should I called...we have a history earned in battle.

All product decisions in company history were made dictatorially...by me, alone...after I have listened to the gang over beer and sushi, every Wednesday night. We called that the Sushi Board meeting.

I was both the DDI and DDO in the company's CIA...I read, acquire intelligence in the field; and analyze data and remember if not yet fully understood while doing so.

If I have to convince an investor as to the size of the potential market, all I know are [and what do you care if no money solicited]:

More than 220,000 M3 were made [data from a Cameraquest article], far more than perhaps all M2/4/6/7 combined...say 3000,000 all told. Many sat in estates, some in show cases, a lot was consumed in action...perhaps some 200,000 still alive and in action.

I always believe that maybe 10% will eventually buy, when they realize there are no alternatives. [Film will still be here, but digital is the future and with NO RECURRING COST. The price of $1000 means 1000 exposures worth of film/processing cost...for B/W. 20,000 units at $1000 each is $20 million.

Many argued CCD sizes, all the time, and pined for bigger ones... I believe when FF is reached, a truce or a sense of satisfaction would have been achieved. The frame-window size was long fixed fixed...by Oskar.

Pixel sizes may get smaller, albeit at a far slower pace. A 24 M-pixel 6 micron CCD is the state-of-art. I am SURE Nikon, Sony, Canon and anybody else would love to agree...thus avoiding a protracted upgrade war.

I will also say don't worry about how this or that can be achieved. That is what engineers are for...so long as they are kept on track to build this "make it sturdy, keep it simple" product. Kodak's experience in DCS where Brain knows much first hand is a major lesson to be learned. Kodak did it well before Nikon, Canon...and Leica. The Yellow Father was not stupid.

As aniMal had said, this is a "need-to-have" product, even if you can afford an M8/9. What if we have read the mind-set of Leica users [not collectors] correctly while Solms was high in the ivory tower?

AND what if we can make the MIS/KIS back user removable without a visible trace?
 
Last edited:
I love seeing the enthusiasm and momentum behind this thing. Cadeaux to Frankie for his persistence and know-how. I'd love to help with the project in any way I can, but the to-do-list seems way beyond my skills and technical knowledge. Mind you, my humble ambition was to do a simple hack out of curiosity.
I work as an author, freelance journalist and translator, with ambitions of adding photography to my professional resume. Hence, I'm trained in research and communication, and I am quite good at mental analysis of nonspecific processes and systems. Don't hesitate to let me know if I can be of any assistance.

If a full-frame digital back were available for my M6 and priced below $2000, I'd seriously consider saving up for one. And if cutting some needless functions were necessary to lower the price, I'd be happy to sacrifice them and save my money.
 
This is all doable, but I'm concerned about the economies of scale required to make it viable at such a price point. My background is circuit design and manufacture (mostly using PowerPC cpus) and these are not trivial circuits to build. One must also factor in manufacturing fallout, for one. Unless the volume is large enough, it's very difficult to get the best manufacturers (i.e., the ones with the experienced personnel and top-quality assembly equipment) to build the product.

A low-volume product (which this certainly would be) would be relegated to 2nd or 3rd tier manufacturing partners. Such firms are capable of building the product, but usually won't have the in-circuit test equipment to debug boards that do not pass. Such boards will have to be sent through a manual inspection process, which can often resurrect boards, but often does not, at relatively high labor cost.

A non-working board costs a fair amount, it's not easy to reball such chips. I am assuming the sensor uses a fine pitch ball grid array.

When I was building dual CPU PowerPC boards, if a board failed and couldn't be resurrected, it usually meant throwing away two $300-$400 chips. The rest of the parts were minimal but that was additional loss.

How much is a typical sensor chip going to cost?
 
I know I'm chiming in late in the conversation, but I have some ideas that might help you with your digital back for analog cameras.

1) Only cameras with winder options or swing open backs can be used.
2) A 2 piece digital back will be required.
3) Reliable connections between these two pieces are required.
4) A larger result must be accepted as adding these two pieces will surely look like a MD-12 plugged onto a FE2.

Now for the reasons.
1) Film bodies with winder options have a cam on the bottom of the body for the winder to wind, but it also winds as you use the normal advance lever. This can be used as an indicator for the CCD to be charged for the next shot. A swing open back also makes it easier to just place the CCD on the back and swing it closed onto the film plane. The sprocket inside the body can also be used to send a signal to the brain, but it will have to be small enough to squeeze in when the back is closed. Imagine a small barrel with film sprocket holes matting to this along with IR sensors like in a PC mouse.
2) Trying to house all the components into a small form factor initially requires too much engineering and EE work. With this we can also have upgradeable backs for the sensors only; probably why major manufacturers won't do it; it's shooting themselves in the foot for new camera sales.
3) With micro electronics such as sensors, the connections are vitally important to communicate with the digital brain connected below.
4) When the Leica DMR first came out, I knew it was the only solution for an analog body being retrofitted with a digital back. But I didn't realize it would be that big and heavy. I already foreseen the expensive proposition as it was for it was body specific for a non mainstream body, and the fact it came from Leica.

So in summary, I think the Bessa R2/3m would be the best candidate for a digital M. Since the aperture and focus is manual, the CCD only needs to be ready to accept a new image every time you wind the film advance lever, and since it accepts a trigger winder there's a coupling to attach to for the communication for the brain to get ready; being cheaper than an Leica M or Zeiss ZM helps too. I haven't worked out the buttons or the LCD if one needs one, but this is the gist of what I am hoping for from any maker out there that can produce one. 🙄
 
It's easy to build ladders to heaven on an Internet forum. There are no real world constraints and we are all experts. Houses look easy to build. It's all those nails and 2x4's that make it hard. 🙂
 
It's also possible to implement a 20-year old idea using modern design. Kodak's first Digital Camera was the DCS back for the F3. The DCS-200 was much more simple. The hard part was developing the actual CCD. It's no longer my job to design and build digital imagers. It used to be, a long time ago. maybe when i retire.

I absolutely agree that the job of gaining support and funding for such a project is the hard part. Technically, you are recreating something done over 20 years ago by Kodak. Development-wise, as long as you do not allow feature-creep to destroy the project cost and timelines, you are on good grounds. As far as the final product is concerned, I do not know how many people would buy it. Personnaly, I think the Monochrome and Infrared Sensors would find a good niche market for those that own an M9 or M8, or those that want monochrome and rejected the color-only cameras as offered by the mainstream. And don't reject off markets such as amateur astronomers that want a good digital option for their telescopes. I've known a couple that built their own.
 
Last edited:
Regarding the mechanical design, what is the plan? Again, everything is tied to volume (or lack of it.) There are many methods of production (hydroforming, injection molding, etc.) Every camera would need a different mold. This is price-prohibitive (it's probably about $50-100k per mold, assuming each initial mold is perfect and there are no re-do's.) Looks to me like the back/housing would need to be CNC machined.
Do you have an ME who is skilled at design and can do all the precise measurements for the various cameras and complete the CAD? Do you have a machine shop in mind to build prototypes of each type, and then production?

Circuit board production will need to be at least 100 at a time (absolute minimum) in order to get the interest of manufacturers. Many manufacturers will kit the parts (assuming Digital Back, Inc. is economically viable) but they will not if DB, Inc. is a start-up and poorly funded. In this case, DB, Inc. will need enough cash to buy all the parts for each production run, and will need manpower to handle the logistics of product ordering.

Test fixtures also need to be designed. These can often cost more than the design of the production circuit board.

Lastly, will the CCD manufacturer be able/willing to supply parts in low volume (say 100 at a time?) Is there a major OEM on which DB, Inc. can piggy-back orders? What happens if that OEM discontinues their product and no longer requires the CCD? How many units will the OEM require to be purchased at that point (relates to the cost of wafer fabs.)

Bottom line: this will require expertise, and a significant amount of money. I can put a team together to provide the former, if someone can provide the latter. 🙂
 
“Feature creep” and “ladder to heaven” are already popping up all over, never mind “pet features” and “unrealistic wishes”.

My original intent was to build a slim box [not even a back] that fits within the body cavity after the M-body back door is removed. That nets an external volume of 45 x 90 x ~7mm, if I insist on maintaining the body thickness. [Lay a straight edge onto the protruding rims of the camera back door and you will see immediately what I mean.]

[If the body cavity is not enough room for the electronics, then a base-mounted box can be used, Leicavit or Rapid Winder style and still keeping the body profile.]

The battery goes into the film cartridge space and the SD or Micro-SD writer mounted on a plastic pillow block into the film take-up space…all linked up via flexible circuits instead of soldered wires.

Obviously, the base plate will have to be kept and the manual film advance must be used for cocking the shutter…classic M or R-D1 style. No high speed motor drive, thus limiting the buffering/writing-to-SD within human speed…

Shutter cocking can also be exploited as a trigger to power-up the CCD. Learning the habit of cock-aim-fire, just like safe usage of a bolt action rifle cannot be that tough.

Such a retrofit would not look any different from an unconverted M, except the round film ISO reminder will be missing. Buy a stick-on one on eBay if you must.


I have no personal interest in an LCD…I do not believe in histogram chimping after the fact, too late anyway. Perhaps a plug could be exposed to link up to an iPhone app...later, not unlike the RS232 plug in early DCS’s.

No LCD idea was immediately attacked by experts in another thread. They seem to have forgotten the bracketing old days. All I care to say, paraphrasing Henry Ford II in a Ford AGM, was: buy [an M8/9] and complain to [Leica] then.

Elsewhere, many expressed the desire of a simple B/W digital retrofit. That idea is doable. All is need is to have CCD manufacturers not affix the Bayer filtering layer. Dalsa does it all the time for aerial camera building. However, it is simpler to just convert a colour image into B/W in post-processing. However, a B/W-only version is not ruled out.

[I know, I know you get more pixels in monochrome-only captures. In aerial cameras, such a design is called a panchromatic master frame, now up to 196 M-pixel in size where 4 far smaller R/G/B/Ir only auxiliary cameras add data to colourize the image in post-processing…and is called pan-sharpening.]


As to firmware features, the answer is even simpler: none…just like films; and thus no need for LCD set-up access either.

Simply: ISO 200 equivalent – with 3~4 stops latitude in the underexposure side just like transparency films. AND:

  • Daylight only, use filters for other WB…or not needed at all in B/W version.
  • Adobe DNG only…SD cards are cheap.
  • No date/hour and memo imprint…you can remember if it is important.
  • No scene modes…you should know what to do being a Leica aficionado.
  • No GPS link…Bing Map can wait.
  • Use you own anti-vignetting filter on old or non- retro-focus super-wide’s.
This back-to-basics make-it-sturdy and make-it-simple design is totally reversible…just remove the package, battery and SD-reader… Replace the standard back-door.

Yes, I know about the fine M9 shimming…to set the CCD exactly at film plane, blah, blah, blah. Could I not rely on the highly precise, diamond-lapped Leica film guide-rails being accurate enough?

So, what we have to implement is a simple mother board for the frame grabber, adequate buffer and a write-to-SD chip all controlled by a Processor with the firmware residing in a good ROM chip, piggy-backed onto the CCD/chip-set board.

And, the works will be housed in a CNC machined T6 aluminum tray/lid, with the flat lid machined to a tolerance so that the CCD protruding through the frame window meets the image plane within the depth-of-focus tolerance [27.95~28.05mm from front flange]. Achieving a 50 microns or 0.002” tolerance is standard machining guarantee.

I have considerable experience using a local machine shop that could make any quantity of such simple pieces from an AutoCad file just-in-time, corner-rounded, bead-blasted and hard-anodized…in black or other fashion colours.
 
Last edited:
Sounds good to me. On the original DCS back, the Kodak sensors were available in Color, Monochrome, and Infrared after I called Kodak and talked to them about it. The sensors were all pin-compatible, fit into the same back. The software in the computer applied the processing for color.
 
Brian Sweeney said:
Sounds good to me. On the original DCS back, the Kodak sensors were available in Color, Monochrome, and Infrared after I called Kodak and talked to them about it. The sensors were all pin-compatible, fit into the same back. The software in the computer applied the processing for color.
Click to expand...

Now I wonder how many B/W devotees would stick to B/W when colour is available for no additional cost.
 
Unlike Leica playing catch up 10 years too late, they must claim some first-time achievements...as if only Leica could.

Leica cannot contemplate retrofitting even an M7 or MP, never mind older discontinued models. And if I were the CEO, I fire the first person who would even open his mouth.

Leica knows, as do we, evenness in illumination will be the problem...what to do? This problem was their own doing, started in 1925.

Once realizing my threat, Leica-Heerbrugg had to say to me, saving face, that they had built the instruments too good. Imagine us retrofit an 1952 A8 in 1989. The same Swiss gentleman had also quietly wished the cast iron green machines would self-destruct, so that new ones can be sold.

We on the other hand, had no face to save. Those 200,000 dead bodies are reserved for us to exploit.

What I need to know soon, is: are the back doors of M1/2/3/4/4-2/4-P/6/7/MP interchangeable?

Someone in these forums must already know?
 
Last edited:
Frankie said:
Now I wonder how many B/W devotees would stick to B/W when colour is available for no additional cost.
Click to expand...

I think the answer might surprise you. Especially if you can do this for $1,000. The Mosaic filter robs 50% of the light and adds an extra layer in front of the sensor. That is in addition to the 50% that the IR filter robs from the sensor. A lot of people pay as much to get a DSLR converted to Infrared as you are projecting the cost of the Back. When the DCS200 was new, I special ordered one in Infrared for an extra $4,000 and we also ordered a color camera. Combined cost for the two was over $20,000. How times have changed.

Getting rid of the Mosaic filter also does away with the need for Anti-Aliasing hardware and software.
 
Last edited:
Just to add- A dead F36 motor drive would be perfect if you want to do this for your Nikon F. I told my Boss that in 1988 when he asked what I planned on doing when my Chosen Profession made my favorite hobby obsolete.
 
You must log in or register to reply here.
Back
Top Bottom