Here is my background in the making of film scanners.
I first started by modifying old projectors to be able to make film scans. The captures were made with an industrial camera, frame by frame, the projector working continuously and an optical fork ensured the synchronization of the capture.
On the whole the result was good, but it was a little painful to have to change the projector to digitize another projector to digitize another format. The installation of the camera and its lens could be complicated depending on the available space. The vibrations required weighting and stiffening the fixation of the elements.
In short, this formula was not very practical.
I started to build my own scanner, about three to four years ago.
The specification was to be able to scan all film formats on one machine, with a minimum of time spent between format changes.
It had to be possible to scan all roll diameters.
It also had to be possible to build the machine without using a drill press or a metal lathe, which I don’t have.
Simple technical solutions had to have top priority.
By analogy with 3D printers, I chose a 40x40 aluminum profile for the chassis.
This type of construction is very rigid and modular. That is to say that you can make modifications on an element (motor, tensioner, film guide, lighting, etc…)
without having to rebuild your entire machine.
The arms of the reels are foldable to make the machine more compact.
Moreover, these elements can be assembled without any problem and everything is easily available on the market, for a reasonable price.
Great build, thanks for sharing!! I (and I’m sure others) would be interested in more details of your scanner. I assume that your setup is still frame by frame. How do you manage scanning different format film reels on your scanner? Do you have to make any changes or adjustments when switching between an 8mm to a 16mm or 35mm? I would like to see more pictures of the gate and the film path. If you could share any details about the light source and control that would also be mich appreciated.
thank you for sharing this amazing build, I’m getting back to this picture over and over and really want to build exactly the same machine. It’s quite easy to understand the tension profile and mechanics for most parts, but I can’t decompose the structure of film gate and how it’s conformed between 8/16mm formats. Can you please add some extra photos or descriptions on how it’s made? I’d appreciate a lot if it’s possible. Thanks.
Actually, I hesitated before giving any more details.
The development of this machine took me more than 4 years and hundreds of hours of work.
In fact, I’ve never been able to stop improving it, which doesn’t mean it’s perfect :-).
Let’s be crazy, with a more qualitative manufacturing of parts and a few modifications, this machine is marketable and would offer, I think, a replacement to MovieStuff.
But, if a hobbyist shares or reproduces this machine for his own use, I’d be very happy.
Great build. I like Andriy have been drooling over the photos of your machine for a while. Can you provide more details on things like Tension control ,Software ,Capture times etc. Im a few months into my journey and the Kinograph community has been so helpful. Any information is greatly appreciated.
Wow, thank you so much for clarifying those details and sharing the video. I totally get your concern, I’d be that nervous also. The funny part here that I imagined sort of the same idea on multiformat film gate structure, simple yet effective. Have you ever encountered abrasive scratches on film rebates with such design?
I’m not too tech-savvy so my idea is probably frame by frame machine with mirrorless camera, but the prettiest part of your design is
modular structure, so it’s possible to upgrade over time. In any case I’ll need laser registration pin, need to learn coding at some degree.
After all, it would be slightly different machine, and now I have everything to start engineering process. Will update on my progress and issues that will arise for sure))))
I didn’t need to check the film tension. When the film is scanned, the spool is disengaged from the rewind motor. There is therefore no stress on the film at this point.
The film is held laterally by pressure in the door over a fairly long distance, which is enough to stabilize the translation. The traction motor is a stepper motor that runs continuously, so here too, we achieve a smooth feed.
The motor speed is regulated by the laser pulse detection frequency.
I use the capture software supplied by the Basler camera manufacturer, which works well and allows a wide range of adjustments.
Capture is continuous, based on the flash-scan principle, making it easy to work at 18i/s.
My door system has its limits.
If your film is badly damaged, it will come out of its guides, and this design makes it difficult to put back in place.
I have over a hundred kilometers of film experience with this machine, and I’ve only had to give up scanning one 16mm film, which was too damaged for my machine.
The 9.5mm films are often twisted, but they go through without a hitch.
Collages also go through without a hitch, provided they’re not made with a piece of tape that goes around the film three times…
I’ve never torn a film, and damaged perforations are no problem.
The film passes through a polished rail, and you can feel it sliding along without catching.
I always clean my films with iso alcohol and never dry, even on the scanner, never dry.
Following a very good remark from cpixip who mentioned that to improve image stabilization during scanning, it would be more effective to trigger the shot using the perforation that received the camera claw as a reference.
So, I redesigned the laser pointer mount to bring it closer to the camera lens, leaving only two frames between the pointer and the lens. For larger film formats, this won’t be a problem; it will simply be a matter of shifting the pointer to the left.
I don’t really like the appearance of the part; there are too many visible screws. I’ll change that in the next version.
Another modification was made by replacing the laser focus lens that produced a 0.4mm beam with one that produces a 0.1mm beam. In practice, I have no idea if this will actually improve precision, but I think that since the beam is narrower, the trigger decision will be shorter and perhaps more accurate. We’ll see as we conduct more tests.
These modifications aren’t really essential, as the image stability is already more than sufficient for me, since it can easily be corrected with software stabilization. But we all tend to always look for better results… and sometimes end up with worse.
I’ll digitize a few reels and keep you updated if you’d like.
The laser holder has been rebuilt, as explained above.
In reality, the part is less beautiful, I should have used PLA instead of ASA.
Reducing the laser diameter to 1/10 mm instead of 4/10 seems like a good idea.
I’ve tested these two modifications on a dozen long 8mm and 9.5mm films, and the stability is good to very good.
Hi Roland, I just finished my 16mm double perf film scanner https://youtu.be/Binn5WTQ4Nk and now would like to build a small 35mm film scanner from scratch using the DS8 software to scan, film will stop when Rpi HQ takes a picture and I will be using 3xsprocket similar to the 16mm scanner to move the film. I’m now looking for a way to mount my prototype. Could you provide some details on the frame that you’ve used?
It’s an aluminium profile that I’m ordering from Aliexpress, but it must also be available from other online shops.
I’m using 4040 Black T-slot Aluminum Extrusions (40mm x 40mm).
Smaller sections are available, but for good rigidity, I prefer this size.
The best thing is to choose the right lengths when you order, if you don’t have a circular saw.
There are lots of accessories that allow you to assemble the elements together, but it’s best to choose large squares, again for rigidity.
Otherwise, when I need axial mechanical movement, I use these rails that attach to the aluminium profile.
To guide the film, I just use the rail (MGN5), which I polish beforehand.
It’s almost a shame to build a whole scanner just for one film format, isn’t it?
Many thanks Roland, this looks promising and useable. For 8mm, 9.5 and 16mm I’ve modified projectors to scanners. A 35mm projector is a bit large and I’ve only got old 35mm films (limited length, no colour and sound) so first start with that. My goal is to find a solution that I can use also for other formats, ultimately I would like to build a simple 9,5mm scanner (a bit like the 8mm Reflecta design) with DS8 software, so much better output quality. Users can use it to view and or digitize the material… but first 35mm! Yesterday I’ve made a first prototype of film sprocket, support with bearings and mount of GT2 pulley, the support is already 40x40 mm.
It’s a beautiful piece. I’ve never been able to print the sprockets properly with my printer, at least not that well.
It’s a nice project to design, although my preference is to capture it without stopping the film. I don’t think it’s as easy to position the film accurately if you stop the scroll at each frame.
Thanks Roland, I’ve used a simple 3D printer, a 4 year old FlashForge Finder, still going strong. Used PLA Plus filament. I probably need to go to a smaller sprocket. This one has room for 7 frames. Probably need to design one for 4 frames. I will open another topic to keep yours clean. Continuous movement is nice, I’ve done it with some Bauer projectors adding a trigger to the main axis. Somehow I don’t like the image quality of the FLIR camera used and I don’t think that the prices of these kind of industrial camera’s with global shutters are very affordable. I would like to stick to my Rpi HQ camera. So far DS8 has done a great job in all the projectors that I’ve modified to scanners. Accuracy of frame position is very well, much better than the T-Scann8 that I’ve used in the past. Ofcourse sprocketless transport has advantages when transport holes on the film are damaged.
My experience is much the same as yours. I modified 4 projectors ( 8, S8, 9.5, 16mm), one for each format.
While retaining the film advance gripper, an optical trigger was used to control the camera.
This system meant that I had to move the camera to the right projector depending on the format of the film to be digitised.
But with this system, where the projector took care of image stability, there was no problem.
On the other hand, with this multi-format scanner, if I’d had to stop the film to capture an image, I wouldn’t have been able to do it without software stabilisation during capture.
