Diy Digital Rf - Is It Possible?

[t6un]

Reading the thread backwards, looks like Brian has stolen both my ideas

 

[Frankie]

Well, Frankie, you are laying it out pretty good here. If there is a viable market for such a thing (which I doubt), then some company with the R&D money, equipment and manpower will be all over it and beat you to market!

I'm not sure, though, if there are really many folks who want to spend a grand making their old camera a limited functionality digital camera.

Funny that was precisely the counter-argument 20 years ago. Many larger companies have capital...and a brand name...thus wouldn't do it, avoiding being labelled second fiddle...

Many [practitioners] could borrow money to buy new [instrument] but wouldn't, because the contrived justification doesn't make economic sense.

In the end, it was my gang who made that happen, and launched a business, because common sense dictates you use what you got for as long as possible. Our customers, including a few outfits in Switzerland and Germany, who resisted keeping up with the Jones, just because, rewarded us again several years later when we launched an all digital system, wiping out Leica and Zeiss's illustrious history.

So far a couple RFF members seems to be interested already...wouldn't you if you know photography; doesn't need all them scene modes; shoot RAW as you do in film...daylight or tungsten only; and won't need to pay for film/processing any more?

 

[Sonnar Brian]

A lot of people buy Digital Backs for their old film cameras, and spend a lot more than a Grand on them. They are available for Medium Format cameras, and seem to sell.

 

[Sonnar Brian]

Reading the thread backwards, looks like Brian has stolen both my ideas

I hereby grant unlimited rights of this idea of a lens adapter with a built in RF mechanism to ANYONE that will actually build it and sell one to me.

 

[Frankie]

A lot of people buy Digital Backs for their old film cameras, and spend a lot more than a Grand on them. They are available for Medium Format cameras, and seem to sell.

Yep, the first one I did was a Hasselblad. The 16 Mpixel Kodak digital back would have cost $16k [had it not been an indirect Kodak gift, a long story], a 40mm Distagon another $2500 on eBay, just to overcome the crop factor of 1.4.

It worked as predicted...whereby I had also proven aerial imaging exposure based on my hunch that it was native ISO 100.

[We have to achieve a density of 0.3 above film Base and Fog, ~0.15, in aerial imaging.]

 

[Pickett Wilson]

Well, if you are confident such a back would sell, just do it! Sure money makers are hard to come by these days. It could be a boutique business like Tom A's Rapidwinders.

 

[Frankie]

Well, if you are confident such a back would sell, just do it! Sure money makers are hard to come by these days. It could be a boutique business like Tom A's Rapidwinders.

Funny you should say that. TomA and I live in the same city, and 10 years ago, even the same neighbourhood. The first time we crossed path was decades ago in my university days. [I was making a quick coupler to match my Leica ball & socket head... We also have mutual friends, since then.

[The quick coupler was based on the eccentric cam principle and made out of stainless steel...still in use today.]

TomA was the one who advised me on how to file down my new CV 40/1.4 mount claw to bring up the 35mm frame line...a couple of years ago.

I am in a financial position to enjoy early retirement and the "dM-retrofit" project might just be the thing to keep me sharp. Spending cash to make a prototype is within my means...as I have already spent $2500 for an M2 + R-D1 and things.

I have a few colleagues who would be keen to contribute...just because.

 

[Frankie]

I had since bought a late serial number M2 on eBay...cost $800...and started to think seriously...

[I had also found the instructions of how to disassemble the M2 down to the chassis...]

A month ago, I won an eBay auction on an RD-1...first hand learning in how that camera works started in earnest.

[So far no surgeries...if the idea didn't work, I have a nice M2 and a RD-1 to use on my CV 40/1.4 and CV 25/4P. ]

The M body is perfect for conversion because:

1. The removable back, once removed, should provide enough room to fit the "sensor package"...including a piggy back PCB and a data-port, mini USB or whatever plug to link up to an iPhone app for chimping.
2. A neat enough package will generate interest in the Leica market...especially if the retro-fit can be reversed or removable.

[Besides, critic cannot argue my M2d body is flimsy, the shutter no good, the RF patch flares or the RF easily knocked out of alignment... Leica did all of it. In my stereo-plotter retro-fit days, even Leica-Heerbrugg could not criticize for the equivalent reasons...they built the original instrument mechanics which was not modified.]

The Epson RD-1 uses an EP-F80 battery [3.7v 1500mAh] that almost fit within the film cartridge space... I am sure a shorter one can be found.

I have also bought a couple of micro-SD readers...8 Gb HC is cheap AND small...fits right over the film transport, or easily within the take-up spool cavity. A clip-on plastic pillow-block is simple enough to make [I have a Unimat lathe/milling machine.]

The M2 body is 33mm thick [not counting the back which sticks out another 2mm]. The M-mount flange to film plane distance is specified at 28mm (27.95mm)...

If I want to keep the M profile, I will have 7mm thickness limit for the sensor/PCB package. If I emulate the Epson out-rigger style [as used to mount the LCD], I mighty have ~10mm.

[The Epson total body thickness is 39mm [as is the M8/9], and would have 11mm to house both the sensor package and LCD. But the RD-1 M-mount protrudes another 3mm, making the net thickness ~14mm.]

The next task: figure out how to fit the package within a space of 45mm wide x 90mm long x 7~10mm deep. Imagine a digital M2/3/4/6/7...

This Md-Package is basically a dedicated micro computer mother board, with a CCD as the only input, no monitoring except a plug, and output only to the SD card RAW...

Cosina manufactured the RD-1 for a MSRP of ~$3500. It would have been decent business if they earn $1000 each piece. That $1000 earnings must also cover CCD, design/development or outsourced PCB...plus manufacturing tooling/labour cost.

KISS.

The first function to support is RAW+daylight only. No JPEG-H or N, straight capture and record onto the micro-SD. Let PhotoShop or whatever is your favourite do the post-processing.

The Nikon D3S sports a CCD with ISO 200~12,800. A good 2 stops better than the M9 ISO 200~3,200...2-bits worth of headroom for pixel-vignetting correction. [Lens vignetting might be variable, but pixel vignetting pattern is constant...regardless of lens mounted.]

I never thought I would hit Leica again, 20 years later...sorry Solms.

Brian Sweeney had posted elsewhere about his Kodak DCS upgrade, with pictures worth a thousand words. http://rangefinderforum.com/forums/showpost.php?p=1225329&postcount=74

I have added a few things to illustrate my post...I hope you don't mind Brian, and you are entitled to a 10% discount if and when I got this thing built.

 

[Sonnar Brian]

TEN PERCENT DISCOUNT! I am ONBOARD WITH THIS!

DCS200c's go dirt cheap. Pick it, or a DCS420c up to play with.

 

[Frankie]

TEN PERCENT DISCOUNT! I am ONBOARD WITH THIS!

DCS200c's go dirt cheap. Pick it, or a DCS420c up to play with.

I fondly recalled a week before I departed for Zürich to install a digital orthophoto system [that killed Leica's $750K optical wonder a year earlier in 1991] that Leica sent me a letter by hand demanding to be granted exclusive distributorship in Switzerland.

My customer shared a parking lot with the Heerbrugg gang. To add insult to injury, SWISS AIRPHOTO had also ordered a system during the same trip...

I will now start shopping for a DCS on eBay. I also have a few contacts at Kodak and Dalsa. They too know who we are.

 

[RFH]

Still Going Strong

Still Going Strong

It's nice to see the revival of this thread. I am glad that someone else is taking the idea seriously and has the know-how to figure out the specifics. I am still playing with the idea, but unlike Frankie, I have no spare funds for this kind of experiment at the moment. But I'm on a constant lookout for cheap spare parts to toy around with.

 

[Sonnar Brian]

My DCS200ir was $12,400 in 1993 or so. The last one I saw on Ebay was ~$50. And M8 owners complain.

 

[aniMal]

Since you clearly know how to do this, and it also seems to be no major obstacles, would it not be a good idea to put up a small website? You could then offer proof of concept, some schetces, and means of putting in a pre-order. There would be a lot of interest when it goes from pure discussion of the possibilites, to a project that is up an running. I use an M8, and would just love to have a ´DMR´ back/bottom to fit on an old M. It would make the perfect backup for my M8! I would not really mind if it was on the size of a winder, as long as I could mount if on and off without too much work. I imagine it would also be much more silent than an M8, or even an M9?

 

[Frankie]

Back in 1992/3 when I managed the building of the world's first roll film scanner [that sent Leica, Vexcel, Wehrli and Zeiss all back to the drawing board], a digital camera based on a linear CCD array had to be built first.

[Aerial films are in rolls, and no end-user, especially governments would allow cutting into frames. Frame-by-frame format was possible only if the negatives were reproduced as diapositives first...an additional cost that equals scanning with degradation of resolution. However, no one built roll-film compatible scanners, despite all contact printers in existence accepted roll film...I had to be the first to say the obvious. ]

The basic architecture was: a CCD, a frame grabber, a buffer, and finally writing
-to-disc. No one in the industry then use JPEG compression at scanning. Capturing/preserving all 8-bits was difficult enough.

[We need at least 10 micron pixel size across a 23cm x 23cm frame, and the best linear array available then was only 4,000 pixels...so we have to do 6 passes...and still maintain <1/3 pixel of X/Y RMSE.]

Today, we have CCD's at all sorts of pixel size. Kodak's latest is a 6 micron anti-blooming array [already 2 years old, while Leica stuck itself at the far older 6.8 micron CCD because of the self-celebrated micro lens...]. 6 microns means 24 M-pixel, at least keeping up with Nikon/Sony.

I believe capturing and writing only RAW is the best first approach...SD cards are now so cheap. Besides, everyone who had posted here seems to believe RAW first, post-processing [with PhotoShop, LightRoom or whatever] later.

To make a dedicated computer mother board with lots of buffer memory [also cheap] and basic firmware is not much in R&D. There might be more work in taming the pixel vignetting [fortunately predictable, based on incident ray angles from lens principal axis]. The equalization I would implement will only be making the whole frame even [removing only pixel vignetting].

[The research I need to do is to quantify the rear nodal distances from image plane of various super-wide lenses. I have a hunch that supporting classes of vignetting, rather than individual lens' vignetting is all that is necessary. A mathematic model will soon tell.]

I will leave lens vignetting for their makers to defend. If Leica made such good lenses, then they will have no problems...right? Likewise for CV, Zeiss, Konica, Canon or Nikon.

I firmly believe in a plug-in external LCD, or an iPhone app. Plug a new and better one into the system when available. The display resolution need not be much better than 800 x 600, given the screen size.

Epson offers R-D1s firmware upgrade via a down-load written onto an SD card, then up-load onto the camera ROM chip...simple enough...so long as the on-board processor has enough muscle to handle new tricks. Again, Better ROM chip or the cost is not a real factor here.

So what if my first version is not perfect in Leicaphiles' eyes. There are enough evidence in these forums that few worship Leica software anyway...as nobody did in my field either.

 

[aniMal]

For this kind of project, I think a very simple and straightforward version would be accepted as perfect. Just a back to slap on, as small as possible - and using as little power as possible. Capture in DNG preferrably, but no need for any compensations for vignetting and such. (unless it would actually be possible to buy something like the M-9 sensor from Kodak).

Since something like this would be far too complicated if it is geared to emulate an M9, why not go along with all the people who wants a "digital film" that is dead simple?

What would the price be for something like that, given a production of 100 units? Or 500?

 

[Sonnar Brian]

The main cost will be the CCD/CMOS sensor, which can run several $K. I looked at replacing the ancient KAF-1600 with a newer KAF-1603, and it would have been $2,000. The cost of buying a development kit, full-frame CCD and support electronics from Digi-Key is several thousand.

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CYII4SC014K-EVAL-ND

SO: step 1 is to find a cheap-enough Sensor, Step two is to get the development kit to do the embedded software and proof-of-concept camera, step 3 is to design a custom board for the electronics/firmware/interfaces, step 4 is a fixture designed around it, step 5 test and debug, step 6 set up for liited production. I still think it is amazing that Leica kept the M9 to $7K.

 

[Frankie]

For this kind of project, I think a very simple and straightforward version would be accepted as perfect. Just a back to slap on, as small as possible - and using as little power as possible. Capture in DNG preferrably, but no need for any compensations for vignetting and such. (unless it would actually be possible to buy something like the M-9 sensor from Kodak).

Since something like this would be far too complicated if it is geared to emulate an M9, why not go along with all the people who wants a "digital film" that is dead simple?

What would the price be for something like that, given a production of 100 units? Or 500?

Kodak or Dalsa sells CCD to anyone for whatever applications. They don't care.

However, the M8/9 off-set micro lens is said to be proprietary...likely also made by Kodak.

Pixel vignetting is mostly a function of hooding in CCD pixels, the effect is a function of incident ray angles from the lens rear nodal point [located along the principal axis].

Lens optical vignetting is best corrected at the lens. Many such lenses were supplied with AV filters ranging from Filter Factor of 1.4~2.

Personally, I am not a super-wide fan. I am more than satisfied at 28mm focal length. I also believe that most M users stop at around that focal length, or Leica would have made lower VF magnification.

Indeed, a Mk-I model could be that simple.

Furthermore, I have no interest in old optics. The digital age entails some sacrifices.

I also agree that aDNG is good enough [SD cards are so cheap. 4-pack 2Gb Sandisk Ultra is available at Costco for $35...for my new R-D1]. I said RAW only to avoid needless argument at this stage [there have been some who stomped on many of my KISS posts...just because].

I intend to fit the works in a package no different in size and looks than the removable M back door. That should give me ~7mm of depth, times 45 x 90mm in overall dimension. If an OLED thin enough can be found...perhaps add another 1~2mm out-rigger style. [I am not a fan of histogram chimping. I prefer bracketing while shooting, as if I am using free film/processing. ]

Such a Digital Film package could be nominally rated at ISO 200 with much latitude in the underexposure side, just like transparency films [whether you meter or not]. It could be as basic as supporting only 5600- and 3200-degrees Kelvin [that was the only film types one can buy anyway...yes, I ignored Photo Flood, so use a filter]. All sorts of post-processing software system support auto-leveling, including the $20 Microsoft Digital Image Suite [so buy one].

Sourcing a large capacity battery is least of the problems, also a SD reader circuitry [a whole reader can be purchased on eBay for 99 cents].

In terms of volume/cost, I am a little more ambitious. It is widely believed that at least 200k M2/3/4/6/7 bodies exist and still functioning. Selling to just 10% of those "dead bodies" is not too wild a dream.

[I had the experience of doing just that in retrofitting stereo plotters...at a price of ~$10k each. That venture netted total sales of 450 systems [out of ~5000 surviving instruments] and retarded Leica/Zeiss' sales of new analytical plotters at $250k each...for 4 years.

This dM-retrofit might do just that to M8/9/10 if I set prices at a comparable ratio. Material cost is the least of my concern, personal time in campaign commitment is...and I might even be expelled from RFF for it. ]

I am finishing a project [my final game now at the final act] and in a position to take early retirement. The dM-retrofit might well be my fun project next. Spending a few thousand dollars prototyping is within my means.

In my 35 year long career, I have collected support sources that defy logic. For example, a local machine shop that could produce 1 or 1000 pieces all based on a CAD pattern...the admin charges was sometimes waived, they know my history.

 

[Frankie]

The main cost will be the CCD/CMOS sensor, which can run several $K. I looked at replacing the ancient KAF-1600 with a newer KAF-1603, and it would have been $2,000. The cost of buying a development kit, full-frame CCD and support electronics from Digi-Key is several thousand.

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CYII4SC014K-EVAL-ND

SO: step 1 is to find a cheap-enough Sensor, Step two is to get the development kit to do the embedded software and proof-of-concept camera, step 3 is to design a custom board for the electronics/firmware/interfaces, step 4 is a fixture designed around it, step 5 test and debug, step 6 set up for liited production. I still think it is amazing that Leica kept the M9 to $7K.

Dear Brian, you read my mind.

Step 1 and 2 is only money.
Step 2 and 3 can be done by a proven member of my old gang...with joy.
Step 4 I and another willing retired old gang member can do.
Step 5 I and/or interested RFF member [you?] can do.
Step 6 can be funded via pre-orders or simply capital.

And, I can certainly do the marketing, from web site to demos.

 

[aniMal]

I must say that by now, I am sitting quite upright and reading every word here carefully...

Although I have not got the technical skills that you have, nethertheless I am that type of person that has been repairing and making whatever for years. Also modern design philosophy sometimes makes me almost explode, and I would love something like this - where the moving parts are well proven traditional mechanics, and the electronics are all simple and serviceable parts.

I think this could/will be a major hit for the first ones who do it! I have mentioned in many threads here at RFF that I would love to buy a digital FM-2, and this would make that possible! Or even a digital R6, what a thought!

So, if you need any assistance I would be glad to help... What I can offer is:

- webdesign/hosting from my company
- industrial design from a fellow Leicaphile
- helping in getting a good number of pre-orders here from Norway
- enthusiasm in general!

I really see and believe that you have the knowledge, compared to so many other wannabees. Then it is just doing it, which is always more work than expected?

Well, you really got me fired up by now - gonna make a call to the design guy now, who has a range of contacts...

 

[Jinky]

I know the crop factor may not be to everyones likes but would the innards of an Olympus E420 fit into a rangefinder body? All you need to do then is fix the internals at the register distance for an M lens and you're good to go.