Converting projector to scanner questions

[Simon Lucas]

I'm going to convert a projector into a scanner and want to hear from people who have done this.

I know there are one or two already. Josh And Carl, for example.

I'm going for the frame-by-frame model controlled by Arduino and what I need to know at the moment is how much torque I will need in the stepper motor to drive the projector mechanism? Obviously I can find out by trail and error but it would be nice to hear from people who have already done this.

I have a 18-5 Bolex Super 8 or may get something like an Elmo, both of which I have seen used.

Thanks in advance.

[slashmaster]

Have not built one yet, but the steppers from any junk printer you find on the side of the street should work if you gear it down. Could probably use the cogs and belt from them too. Maybe even the driver. I would use that Bolex but the best film scanners have the gate filed out wider and higher. It's a little bit of a shame to have to mutilate your bolex gate like that. If you're going to buy anything, I would buy a junk 18-5 for parts.

[slashmaster]

Did you already pay for the arduino? Because if not, there is a lot of stuff on cnc forums that show how to do that part for free from junk!

[JeremyC]

That's a great challenge to get stuff from junk, my personal challenge is how to calculate going from a stepper motor's movement via gears to equal the distance between 8 mm and 16 mm sprockets. There is a maker movement group near me here in South London and a couple of the members have built 3D printers using junk bits and stepper motors which work brilliantly and way more complicated than just advancing film frame by frame using a stepper motor but i just can't get my head around it at the moment. I perhaps could get the maker people to run me up some gears once I work out the gearing ratios.

[Simon Lucas]

Have not built one yet, but the steppers from any junk printer you find on the side of the street should work if you gear it down. Could probably use the cogs and belt from them too. Maybe even the driver. I would use that Bolex but the best film scanners have the gate filed out wider and higher. It's a little bit of a shame to have to mutilate your bolex gate like that. If you're going to buy anything, I would buy a junk 18-5 for parts.

Good advice on the Bolex. I keep looking at it and not wanting to destroy it. If I can get another perhaps broken one, I will. Looking at the inside of the Bolex, it has three belts. The motor's pulley - a complex-looking mechanism is attached by screws to the original motor shaft. I wonder whether I can remove it and put it on the stepper, use the belts, all on one go.

Also, enlarging the gate. I had not considered that.

I've seen people use motors from floppy disk drives for a super-8 flatbed scanner. It's an option, but I'm now eyeing the stepper I used to test an idea for an external shutter mechanism. I can at least try that one, first.

Searching through posts, I found one where Carl Looper had identified a torque of 540g-cm for his Elmo. Thinking about this, I seem to remember that it's the ability of the motor to turn the load from stationary that counts when you are stepping through a series of discrete positions. I think this is called the Pull-in torque. 'Pull-in torque must overcome friction and inertia'. But none of the motors seem to list a 'pull-in torque', just a hold-in torque. My current stepper has a 2.3kg*cm holding torque and current of 0.33A. I watched the video of Josh Gladstone's projector conversion and he talked about a 1 amp current being drawn by his motor, so I wonder whether his stepper had a higher holding torque. - like 3kg-cm. I can only try the one I have.

[Simon Lucas]

That's a great challenge to get stuff from junk, my personal challenge is how to calculate going from a stepper motor's movement via gears to equal the distance between 8 mm and 16 mm sprockets. There is a maker movement group near me here in South London and a couple of the members have built 3D printers using junk bits and stepper motors which work brilliantly and way more complicated than just advancing film frame by frame using a stepper motor but i just can't get my head around it at the moment. I perhaps could get the maker people to run me up some gears once I work out the gearing ratios.

I've been pondering the same question. There's a much larger pulley on the shutter mechanism in my Bolex than the pulley on the motor. If the motor can do 200 or 400 discrete steps, then hopefully it will be accurate enough to rotate the shutter mechanism exactly a single turn (or is it 1/3 turn?) - as the motor will have to turn a large number of times to rotate the projector shutter once. But will it gradually get out of sync?

To be honest I have a shaky understanding of the relationship between the shutter and the film advance - at least until I take the whole projector apart and watch it moving. I somehow imagine a sprung mechanism that gets triggered with each rotation and then reset before the next rotation. That would give a discrete event that positions and momentarily locks the film, whilst the shutter opens. If I am right, that would mean that the precise rotation of the motor would not be critical. All conjecture at the moment. Can anyone explain how this really works?

[RCBasher]

Why do you want to use a stepper motor? What is the plan for your whole setup? E.g. lighting, camera, etc.?

Projectors are rather good at advancing the film a frame and have it waiting in the gate - it's what they do! Why not just use as is? Unless you plan to run at a snail's pace, a stepper motor will not do the job as getting sufficient resolution and speed is not so easy. How will you drive the film, through a sprocket wheel? If so, you will likely get vertical weave as the wheel will not have perfectly even teeth spacing and diameter where the film sits. Stepper motors will only be truly accurate once per revolution, a typical 200 steps/rev motor will not have every step at exactly 1.8 degrees - only the average step size will be 1.8 degrees over a full rotation. Think about the vertical positioning accuracy as a percentage of the frame height and keep in mind that film stretch and shrinkage will mean that a constant number of steps per frame may not hold for the whole film.

[Simon Lucas]

Why do you want to use a stepper motor? What is the plan for your whole setup? E.g. lighting, camera, etc.?

Projectors are rather good at advancing the film a frame and have it waiting in the gate - it's what they do! Why not just use as is? Unless you plan to run at a snail's pace, a stepper motor will not do the job as getting sufficient resolution and speed is not so easy. How will you drive the film, through a sprocket wheel? If so, you will likely get vertical weave as the wheel will not have perfectly even teeth spacing and diameter where the film sits. Stepper motors will only be truly accurate once per revolution, a typical 200 steps/rev motor will not have every step at exactly 1.8 degrees - only the average step size will be 1.8 degrees over a full rotation. Think about the vertical positioning accuracy as a percentage of the frame height and keep in mind that film stretch and shrinkage will mean that a constant number of steps per frame may not hold for the whole film.

It's in the title. I'm making a scanner. You need to advance one frame at a time to allow the camera to scan the frame and have it saved to hard drive. I'm not the first to try this and I assume a stepper is a better way to control motion than a standard motor. I'm not sure why the modified projector would suffer from weave any more than a projector running at 18fps would - I'm not planning on changing anything in the project other than the light source and the motor.

[RCBasher]

It's in the title. I'm making a scanner.

Yes, I can read. Furthermore, I usually manage to comprehend what I read.

You need to advance one frame at a time to allow the camera to scan the frame and have it saved to hard drive. I'm not the first to try this and I assume a stepper is a better way to control motion than a standard motor.

I do know how scanners work and you are correct, others before you have made film scanners using a projector - I have done my fair share, suggest you click on the WWW symbol (globe) in my signature, you may learn something. I would like to relieve you from your incorrect assumption that the best way to drive the film transport is a stepper motor. In fact I would go as far as to say it is the very last thing to use in a film scanner. Steppers have their place (I have built CNC machines, 3D printers, etc. and they are rather useful for these applications) but they should be avoided like the plague in a film scanner. Apart from the noise, the vibrations are not want you want when magnifying an 8mm film frame by more than 1:1 to fill the sensor on todays decent cameras.

I would suggest you take a close look at how a projector works. Most use a pull-down claw to advance the film in the gate, where it will sit stationary for quite some time. At 18 fps and assuming approximately 1/6th of a frame period is used for the pull-down (it is usually less) then the film will sit stationary for more than 45ms. In my system the exposures are around 100us - so I could theoretically take over 460 frame grabs of a single film frame if the camera, PC and storage system would allow! So I will ask again, why do you think it necessary to use a stepper motor?

I'm not sure why the modified projector would suffer from weave any more than a projector running at 18fps would - I'm not planning on changing anything in the project other than the light source and the motor.

Because you didn't explain how your proposed system is intended to work. Those before you who have tried the stepper motor route have done so thinking they can eliminate the pull-down claw mechanism and just pull the film linearly through the gate, stopping the stepper while grabbing the frame. I can advise you that this method will not be successful!

If you use the projector mechanism which is already there, then you have nothing to do but open out the film gate, add a position sensor (e.g. a magnet and Hall Effect device costing a couple of $) a good lighting system and a camera with external trigger. The Point Grey BFLY-U3-23S6-C ($500) makes for a very good and cost effective camera solution for film scanning. Sufficient speed (over 31fps in 10bit mode) and resolution, also sufficient dynamic range to eliminate the need for HDR except in very special circumstances. The key is good balanced lighting (ideally adjustable) to enable full exposures while the film sits in the gate.

If you looked, you may have seen on my website that I showed swapping out the motor in past projectors. This was for a DC motor with speed control as at the time I didn't have a fast enough camera to make high resolution, high bit depth frame grabs at normal film running speeds, so the swap out of motor was just to enable me to slow down the mechanism so the camera could keep up. Todays available cameras are significantly faster and cheaper, so now there is now no need to tinker with the transport system.

[Simon Lucas]

It's in the title. I'm making a scanner.

Yes, I can read. Furthermore, I usually manage to comprehend what I read.

Thank-you for your informative post. It is very useful to read your webpage and see the mechanism, get an idea how it actually works. As I said, I have not pulled a projector apart yet, so had nothing to go on. Your DIY scanner and software looks very impressive but I can also see that it works in a different way to what I have in mind:

I would like to a cheap dSLR or other stills camera with electronic shutter. It seems to me that a sensor in a dSLR is far more advanced than those found in machine vision cameras and can be accessed for less money. I can achieve anything from 2k to 5k scans with the right camera. Also, I will get RAW files from the camera which eliminates work at the PC end, aside from assembly in an editor like Premiere. I can also control the camera and the desired rate of capture using an Arduino and some off the shelf stop motion software which has an API for the Arduino.

I'm not clear that all the people who have used steppers have eliminated the pull-down claw. I'd like to hear from them.

Is there a way I can achieve my stated aims to re-purpose a projector without a stepper? Would a variable speed DC motor work?

I'm just starting out. I am not an engineer and the purpose of me doing this is to make a useable system with which I can get good scans for film-making, to eliminate repeated costs and inconvenience of having it done commercially. My working method will require me to shoot short sections of film, develop them and scan them – and this is not something that would work for me if I had to have it all done professionally.

[JeremyC]

Basher,

After reading up on Steppers yesterday I quickly realised why you recommend a DC brushless motor and can see that with a stepper you could end up with problems ranging from frame advance getting out of sync to possibly stretching film. I remember that the people I mentioned yesterday who use steppers on their purpose built 3D printers told me that getting rid of vibration took the largest amount of effort out of any part of getting the printer to work effectively to the resolution they need, moving a strip of film has less scope when you think about it.

[Simon Lucas]

As far as I can tell, the vibrations inherent to steppers occur during the movement phase. This is because at each step, the motor over-steps fractionally before being pulled back into position. I could speculate that this might be an issue when running a projector at normal projection speeds. But I am not clear that vibration need concern us when the projector is being run in a frame-pause-frame-pause- fashion. Here. I might speculate that, because the scanning/grabbing will occur whilst the motor is stationed at a step – in a rest position, it will not be vibrating.

[JeremyC]

As far as I can tell, the vibrations inherent to steppers occur during the movement phase. This is because at each step, the motor over-steps fractionally before being pulled back into position. I could speculate that this might be an issue when running a projector at normal projection speeds. But I am not clear that vibration need concern us when the projector is being run in a frame-pause-frame-pause- fashion. Here. I might speculate that, because the scanning/grabbing will occur whilst the motor is stationed at a step – in a rest position, it will not be vibrating.

The problem with any vibration is that it gets transmitted through what is connected to the motor and travels up and down the connection and will not stop just because the motor movement stops but takes a finite time to die down. It could end up still causing a movement within a film frame when that frame is being scanned depending on how the frame is held in place.

[Simon Lucas]

As far as I can tell, the vibrations inherent to steppers occur during the movement phase. This is because at each step, the motor over-steps fractionally before being pulled back into position. I could speculate that this might be an issue when running a projector at normal projection speeds. But I am not clear that vibration need concern us when the projector is being run in a frame-pause-frame-pause- fashion. Here. I might speculate that, because the scanning/grabbing will occur whilst the motor is stationed at a step – in a rest position, it will not be vibrating.

The problem with any vibration is that it gets transmitted through what is connected to the motor and travels up and down the connection and will not stop just because the motor movement stops but takes a finite time to die down. It could end up still causing a movement within a film frame when that frame is being scanned depending on how the frame is held in place.

Delay the scan? Has anyone seen the effect of a stepper motor's vibration on a scan?

Sorry, all I'm hearing here are hypothetical problems. Don't worry - I'll find a way.

[Pj]

I have dabbled with DIY frame by frame scanning of Super 8 and 16mm for some time now, when I first started I modified a Chinon Super 8 projector, but in the end I just found there was too much to do to the projector, the old AC motor need to be changed, the gate needed to be enlarged, the shutter disk need to be removed and then I found getting a decent diffused light source and micro-switch too tricky to fit in. I decided to build a simple film channel - a new enlarged gate that is kinder and gentler to film, I had a gate machined out of aluminium and I illuminated the frame by a simple LED light taken from a Veho film scanner. I used various machine vision cameras but earlier this year I started to experiment with CSC (compact system cameras - mirror-less) that have lens inter-changeability and use a Schneider S Componon 50mm enlarger lens [reverse mounted], I have found the results pretty good, each film frame is saved as a JPEG or in RAW on the cameras SD card. It’s a simple and easy set up and it’s great not having to tie up the computer during transfers too.

Pav