Diy Digital Rf - Is It Possible?

[RFH]

Deterred by the high prices of the digital RF cameras on the market, I have been pondering the possibility of constructing my own device. Alas, as often when I conjure up some zany DIY project, I am amazed to find that no one has done it before and posted the instructions already. Much to my surprise, I have not yet been able to find anyone online who has built themselves a digital RF. The closest I came was this thread, but apparently the experimenter in question never followed up with any results.
My idea for the prototype is to modify a FED-2 to house the innards of a digital camera minus the lens and flash assembly. My working hypothesis goes like this:

As long as the CCD retains its connection to the PCB and other mandatory control assemblies, it should not care what kind of lens focuses on it.

If the CCD is placed at the location of the film plane, any given lens should be able to focus onto the CCD, with a crop factor of X% (the crop factor is a technicality that can be ignored in this theoretical consideration). A APS-C sensor as found in consumer DSLR cameras (and in the Epson R-D1) would be nice.

The 'shutter' in a digital camera is defined by the time the CCD switches on/off, and is controlled by the timer chip on the PCB, so the FED body may not need its analog shutter mechanism for this purpose. The curtains will of course have to stay open during recording.

The CCD can record without electronic feedback from the lens. I suppose this is what happens when you set a DSLR to AV mode and use adapted lenses.

The CCD can record without being coupled to the flash unit.

I assume that the digital unit will need to be electrically isolated from and / or grounded in the housing in order to avoid frying the circuits. On this topic, I'm pretty clueless, though.

I imagine that one should be able to mount the CCD where the film plane is; mount the PCB on the inside of the FED hatch next to the CCD and drill the hatch with holes to accommodate the menu controls; mount the LCD on the back, just opposite the film plane / CCD location and run the cables through slits made in the hatch. You would need to find room for the release button from the top of the digital camera you just gutted. With not much room to spare on top of the FED, perhaps this could be put on the front or rear panel instead. Also, a solution for the memory slot and USB port as well as the battery compartment must be found. With no film in the FED compartment, there might just be room for these things.

Let's figure out if this is possible or not. If any of you have the technical know-how and are willing to share, please add your suggestions and insights on how to accomplish this mission.

 

[chris00nj]

There's a lot of electronics involved in controlling the CCD, hooking it up with a memory card, software to export the image to the memory card. It's all really beyond me.

 

[Pickett Wilson]

"The 'shutter' in a digital camera is defined by the time the CCD switches on/off, and is controlled by the timer chip on the PCB, so the FED body may not need its analog shutter mechanism for this purpose."

This is actually not the case. Digital cameras, even P&S, still have shutters.

 

[RFH]

Okay, thanks for clearing that up. I'm no expert on this subject, so I'm just thinking aloud.

 

[Bugleone]

I have wondered about this too. The Epson RD-1 had a APS-C sensor that was also used in at least three other makes of DSLR, namely Nikon, Minolta and Pentax. These are now getting to be worth realtively low prices and some 10 meg sensor models are nearly as cheap.

Providing the sensor does not fill the 35mm frame and give wide-angle problems then it should be straightforward to fit the guts of an obsolete DSLR into the film gate of a RF camera. It would be good to use a RF that has a detacheable back and access to a reange of lenses including moderate wide-angles. The 'Kiev' models are amonmg the most well made and nicest cameras but lens aquisition is much more difficult that the L39 types.

Since the Epson camera uses a conventional shutter between teh sensor and hte lens, it looks like there would be not tricky electronics to negotiate, just the physicality of fitting the DSLR guts into the RF body, hopefully allowing the DSLR back to replace the detacheable back of the RF..... One of the areas that would make things slightly tricky is the need to incorporate the top lcd display with it's vital info about the recorded files etc.

In the case of Pentax and Minolta cameras, the sensors were mounted in anti-shake modules. This would be particularly useful for a typical RF shooter! A good candidate for canbilisation would be the pentax km (or k2000 in USA) which is just being deleted from production and is available new for only 250 pounds including a zoom lens to put back on ebay... The camera is very small, has a10 meg sensor and the anti-shake works vey well,...power requirement is 4 'AA' batts....And, the km does NOT have a top lcd.....

 

[RFH]

Keep the comments coming. Soon it should be possible to acquire a used DSLR for not too much money. It would be very difficult to make the back from a DSLR fit a FSU camera. My guess is that it would be better to take all the necessary parts from the digital camera and reposition them in the FED body. The downside is that the PCB on a DSLR is quite large, particularly hieght-wise, so that may pose an engineering challenge as well. I realize that one might have to sacrifice some functions in order to make everything fit. I could live with an internal memory and no slot for the card if necessary - you can always use the USB connection if that is easier to accommodate in the body.
Here's a breakdown of a simple digital camera.
I have also found this, which seems to be relevant to the discussion. Apparently, this is what happens when you don't have an actual shutter in the camera. Tough stuff, lesson learned. I guess I would have to figure out how to synchronize the mechanical shutter with the CCD clock time - probably by letting the FED shutter button activate the one on the digital control board.
But considering all the R&D and production work hours spent doing this maybe it would be cheaper just to earn the money get a M9 in the first place... but that's not the point of DIY now, is it?
Back to the lab.

 

[BillP]

http://www.rangefinderforum.com/forums/showthread.php?threadid=82438

Deja-vu all over again. My points from yesterday's thread stand.

Why don't you two get together?

What's driving all this sudden "creativity" I wonder? Is it the price of the M9?

Regards,

Bill

 

[Bugleone]

Some points,..in no particular order;....

If you really wanted to do this then it has to be a straightforward hack,...most ordinary people, even with technical background, don't have the facilities to do more unless they can use their work environment,..or, they can invest in this so that a limited production of conversions can be made commercially (not impossible or uninteresting)

looking at the internals of digicams is not likely to be too helpful for such a project as this since digicams use a different architecture and have more in common with phones and webcams than with APS camera sensors.

A way forward on this is to have a good look at how the Epson RD-1 negotiated the various corners of putting a proprietry Sony sensor into essentially the 'Bessa' rf camera. It was widely rumoured at the time that the development and design elements of RD-1 were minimal and it was the idea of one (or a small number) of design staff and was essentially 'done on the cheap',...which is probably why the camera has never been developed of even properly suported by Epson. This may mean that the job of putting together the rD-1 was a relatively simple one of just joining up bits from the shelf........rangefinder already existed as a component set,..sensor already available 'off the shelf'.....essential camera body parts such as lens mount etc already existed.

There are various places around the world that service and repair RD-1 independantly of Epson. These people would be in a key position to advise on the internal layout.

 

[sojournerphoto]

If you really wanted to do this then it has to be a straightforward hack,...

I think that will be the first problem. It really is not going to be straightforwards.

Good luck none the less - to think I get frustrated at scanning film when the digital worflow is so much faster🙂

Mike

 

[40oz]

A) why not start by taking a cheap P&S apart and mounting it in a FED and taking a picture. What's the real hold-up here? You are simply using the FED for lens holder and shutter. If you trigger the P&S shutter button at the same time as the FED's, the P&S "brain" doesn't care where the illumination comes from, it simply processes the data captured by the sensor.

B) why do people always start by posting on an internet forum instead of by showing what they've cobbled together?

 

[Sonnar Brian]

You could start with one of these:
http://cgi.ebay.com/KODAK-DCS-420C-...wItemQQptZDigital_Cameras?hash=item19b7d9a44a

Hack the electronics to mimic the signal the N90 gives through the databack connections. Those are the only connections between the Digital Back and the camera.

Kodak made a lot of digital backs for film cameras.

http://cgi.ebay.com/KODAK-DCS-PRO-P...ViewItemQQptZFilm_Cameras?hash=item1e57c222a5

 

[Sonnar Brian]

Starting to make an M9 look like a real bargain.

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=CYII4SM014K-EVAL-ND

http://download.cypress.com.edgesuite.net/design_resources/datasheets/contents/ibis4_14000_8.pdf

14MPixels. Full Frame.

 

[Bugleone]

40oz,

"....what's the real hold-up here?..."

OK,..like i mentioned above, to use the sensor/guts of a P&S digicam is not a practical feasability becaus of the nature/architecture of digicams,......the sensor is usually incorporated inot the circuit board (or it's modern encapsulated equivalent) and anyway, is minute in size being smaller than the fingernail on a childs little finger. If you did put the circuit with sensor into, say a FED or similar, then yo would be stuck for lenses since the shortest leica/L39 lenses would be much too long in focal length,...if you incorporated the original lens from the digicam then it would be 'difficult' to retain rangefinder focussing,..and, anyway, why bother,..just keep the digicam...

Taking an obsolete/discontinued DSLR and removing the sensor, electronics and back screen to fit into the FED is a mch more realistic and interesting project which must be at least potentially feasible since Epson have already done it with the RD-1,..in effect.

One of the biggest hurdles is how to re-wind the shutter and re-tension the main spring,...in the RD-1, Epson basically ducked the issue by keeping the lever wind. Incorporating a suitable electric micro motor and power suply is problematic as i found when trying to build my high speed 'Kiev' hybrid....there is no room at the top of the camera which means if has to be at the bottom and there has to be a drive shaft up thru the camera.

If such a project is going to cost as much, or nearly as much, as a M8 then there is no point,...hoever, if you are going to use 'good junk' such as a pentax dslr and a FED or similar then for much less than half the cost of a seconhand RD-1 you could potentially have a camera with better features,...AND,...a smack in the eye for the Jap camera industry since not all of us want a black plastic blob with 29 auto-focus points and 200 menu options.........

 

[Pickett Wilson]

Folks are still asking several hundred dollars for some of the old Kodak DCS cameras. Must be a collectors market for the antiques! 😉

 

[Sonnar Brian]

I bought -and sold- a pair of DCS420c's for under $100 not long ago. Less than for a N90s film camera at the time. The 6MPixel DCS460 goes for a few hundred. It was originally $30,000.

The DCS200 is less common, goes for about the same. I paid $12,400 for mine when new, almost 17 years ago. I picked up the Nikon E3 for $200 about 3~4 years ago. Originally $7,000 in 1998. I still use it as it works with the SB-29, and is full-frame. All One-Point-Three MILLION Pixels.


Hint: the PCMCIA drive goes with the later DCS420 and DCS460. The DCS-200 has a built in 80MByte 2.5" SCSI drive. Holds FIFTY images.
http://cgi.ebay.com/NIKON-N8008S-AF...emQQptZLH_DefaultDomain_0?hash=item53dbe4e00e

If anyone buys it, I have the TWAIN drivers for Photoshop 3.0.

 

[Pickett Wilson]

That looks like fun to play with. Hmmm.

 

[Sonnar Brian]

You will need a SCSI card for the DCS200, and a cable with the DB25. The Jaz Traveller has the correct cable adapter, and I probably have a couple in the basement.

 

[Frankie]

I have been think about a universal digital conversion kit since 2000 when I saw the first digital aerial camera in ISPRS Congress in Amsterdam. Those cameras starts at $1.5 million++.

I had soon built a 16Mp Hasselblad for aerial experimentation and help built and implemented stereo display software for another, a couple of years later...a dual 22Mp unit.

I had posted "think-aloud" pieces in RFF a few times but always stomped on by the expert...

Bottom lines:

• A sensor of FF or larger size is available from Kodak or Dalsa...they don't care what you do with it.
• A buffering-and-writing-to-memory-card software system is very simple.
• A CCD is nominally ISO 100, with latitude on the under exposure side by about 4 stops...12 bits, sound familiar?
• A battery and memory card hardware is cheap...ask any OEM supplier at CES.
• An iPhone app can be written for post-exposure reviewing or chimping.

I had even offered the bodies of my many F2s and a Leica as surgery subjects.

When you look at various cut-away pictures...from R-D1 to D3X...the sensor package is only~5mm thick and could well fit into the space now occupied by the pressure plate. A multi-layer circuit board can be made to fit into the take-up spool cavity. A battery fits well into the film cartridge cavity...cutting a hole somewhere to fit a mini USB plug for the iPhone is easy.....

Best of all, the frame format, film cartridge size and required cavities are always available and long standardized...no?

So I won't have 52 spot matrix metering...bracket then; I will only have ISO 100~1600 sensitivity range...good enough; I don't have instant chimping...I don't care, there is no live-view in an M9 either; I will have vignetting...I don't use extreme wide angle lens, only a CV 40/1.4 [or use a couple of the extra bits to even it out]; I could only shoot Raw and post-process at a computer...so what, SD cards are cheap.

All Leica did in the M9 was to guarantee a large order of the previous-generation leftover Kodak 6.8 micron chip...likely cut down from the standard 22Mp offering. The micro lens layer was nano- optical engineering long after the fact...which I will skip. The RF is a 50 year old design since the M3...all it does is to focus the subject within the depth-of-field tolerance onto the CCD [film] plane, you never know if it didn't anyway.

Of course, such a digital retrofit unit will be look-down upon by the latest-tech mafia... But then, I had the personal experience of retrofitting over 400 analogue Leica/Zeiss/Kern/Jena stereo plotters, world-wide, for digital mapping and stalled the sales program of their new analytical plotter offerings for 5 years at $250k++ each...and then truncate their product live-cycle all together via an all digital system now commonly known as "softcopy photogrammetry".

Imagine the same now happens to struggling Solms. The pricing of an M9 is begging for someone to take a chapter from this history book.

 

[Sonnar Brian]

No one is saying that it cannot be done. I look at the Kodak back for the N8008s, N90s, and Hasselblad and know it can be done. It's time, money, and effort. Anyone with an electronics background and access to a modern electronics/machine shop can do this. The shop that does my custom computer boards and embedded systems could do it. Again, the design, development, and low-volume manufacturing of the data back can be done. It's not something I could pull off on my kitchen table. It's more involved than hacking a Sonnar onto a J-3 LTM mount. It's not more involved than projects I've done at work.

 

[Pickett Wilson]

Brian, I have the SCSI cards and cables. No problem there. We'll see if the thing will even fire up 🙂 Should be fun.