Leicina 8 Crystal Sync Update

[Aapo Lettinen]

I started to modify my new Leicina S to real Crystal Sync as soon as it arrived today.

There seems to be a demand for sync sound Dual 8 cameras so the update, when ready, will become available for order one way or another. I will do a lot of prototyping and testing in the following months and will post the results here ?

The first stage was to disassemble the motor side of the camera to see what the possibilities are to attach a speed sensor to the original motor and how much there will be space available for the circuit boards. I also calculated the reference frequencies needed and did some drafting. 

The next stage will be to manufacture a custom speed sensor assembly for the motor and then hook it up to my existing Crystal motor controller prototype and start to run sync tests. 

I can already say at this stage that this WILL work correctly. Release schedule is TBA at the moment but I would say "pretty soon" at this stage because I already know how to do this conversion and have most of the electronics design ready  ?

 

Edited June 29, 2020 by aapo lettinen

[Simon Wyss]

I’d like to make a suggestion, that you change to 9 V batteries.

[Aapo Lettinen]

7 hours ago, Simon Wyss said:

I’d like to make a suggestion, that you change to 9 V batteries.

I'll test how that would work!

I am probably testing some type rechargeable Li-Ion batteries as well. The main goal is to have a small battery so that it does not take much space and leaves more room for the control boards.

[Simon Wyss]

Perfect!

[Aapo Lettinen]

Today I made a very basic magnetic encoder disc for the first tests I am going to run soon. I had these axle fittings from another project so was fastest this way... I just put some neodymium magnets inside the bolt holes and sealed them with tape after that. This adapter is made for slightly larger axle but the fitting works OK for the first tests after I added some tape on the motor axle ? 

I am going to design a better encoder system later but this is great for tests and demonstration purposes. I am going to use self made Hall sensor circuit for reading the magnetic field changes which are used for generating the speed measurement signal after treated in the self made preamplifier. 

Everything extra is stripped from the camera body so that I can get good access to the motor. Easier to work that way. I am probably going to keep the original battery voltage meter so it can stay for now ? 

[Aapo Lettinen]

I purchased two extra Leicina S cameras as well to get more lenses for my own use and to be able to disassemble one completely to reveal how the exposure control could be connected to the new electronics boards I'm going to make. Will also get some spare parts that way :)  I am going to run some Crystal Sync tests this weekend if I have time. As said I already know it is going to work (my sync system works and the motor itself works and the encoder works as well. so all the components of the system work correctly and I just need to connect them together, do some basic adjustments and shoot a video out of it :D  

Then I can start designing the final circuits and the circuit boards. The exposure system needs to be figured out first so I need to disassemble one of the extra cameras to see how to control it efficiently. If there is a possibility for it, I will maybe try to replace the exposure meter battery with an internal voltage regulator which takes power from the same battery than the rest of the electronics. So it is some prototyping now and then lots of circuit design :)

 

[Aapo Lettinen]

I was so curious about the Leicina's auto exposure system that I already opened the front part of the camera to check it out. It needs a little bit more work to get it working it seems and there will be two different options how the exposure system of the camera can be modified: 

A.  Two extra wires are attached inside the camera body to run stabilised power for the light meter. So the additional light meter battery is not needed anymore at all ( the light meter battery compartment is left original looking so that the outer body of the camera is not altered. Automatic exposure works like before. No manual exposure. Probably no additional calibration made to the exposure system though I will adjust the supply voltage of the light meter so that it is as close as possible to the target to get proper readings. This light meter power replacement feature will be developed in any case and will be attached in the cameras which have the crystal sync update.

 

B. Light meter battery replaced with internal wires so that you don't need separate light meter battery just like in the update A. But a Manual Exposure Dial is manufactured and attached to get full manual exposure approximately calibrated in T stops ( I will work on this only if it gains enough interest. It needs lots of testing and calibrating, film tests and some machining job done. Price will be thus also higher and this will thus be an additional feature in any case. So let me know if you are interested in this manual exposure feature. Otherwise this feature will be delayed a lot or not developed at all.)  . With this update,  Automatic Exposure is limited or not possible at all (will know if there is enough interest on this feature so that I can start working on it) .

---------------------------

Feature list of the camera update I am currently working on: 

- all features powered by one battery. No separate light meter battery needed 

- similar exposure system than the original camera has (let me know if you want full manual exposure. It is possible but it is lots more work and I need to know if it's needed or not or otherwise I will not work on the manual exposure feature for now) 

- battery options TBA. I will probably add some type of voltage regulation so that one can use a range of different batteries from 6 or 7 v upwards. I would like this to work on LP-E6 batteries for my own use so I will most likely work on that option among others. I will test the 9v batteries as well how they would work with the camera mod.

- FRAME RATES!  I will add at least 16fps, 24fps and 25fps for sync sound shooting. All very accurate crystal speeds. But does someone need 18fps? let me know at early stage because it will change the circuit boards and I will need to know early on.  

I am unsure if the camera body can handle more than 30fps. maybe I could test it because I have a additional camera body which is just for spare parts and can be sacrificed. Not promising anything though at this stage. 

- Start - Stop button on the side of the camera so that the handle does not need to be used for start and stop of filming. Also easier to start the camera on tripod. 

- Speed selector on the side of the camera. I have to check what would be most convenient. Simon originally proposed that the outer body of the camera should not be altered to keep the classic original look but I will try to make all the Crystal speeds internal and easily selectable so the selector is necessary to install. 

---------------------------

Very early to ask but how many persons would be interested in this Leicina S camera update with that feature list? 

The Leicina seems to be very good camera so after the update it will be really useful for indie films, music videos and similar type of material. There is of course lots of work to do these updates and the camera needs to be serviced as well to be in good mechanical condition (Simon can arrange CLA for these Leicinas easily) . But it will be a real workhorse after these electronics updates and CLA. So it will be absolutely worth it ? 

[Simon Wyss]

https://www.filmvorfuehrer.de/topic/31948-die-leicina-ist-60/

First publication of my article, a translation is being prepared.

It’s going to be the ONLY automatic camera with crystal control to the motor for Double-Eight film. A Paillard-Bolex H-8-4 with the ESM offers crystal speed accuracy, too, but is way bigger, heavier, and more expensive.

The Leicina 8 S has a low optical axis, a straight-pull claw, a ground glass can easily be added to the finder for critical focusing, and the built in 15 mm lens is very good.

[Aapo Lettinen]

I made a "proof of concept" type of test today and ran the camera on crystal sync with all the basic components in place. It works correctly but my camera is not in very good mechanical condition and proper overhaul would greatly enhance speed stability. 

https://www.youtube.com/watch?v=Y0cEqkdUnaQ

But like I said, it works. This is not the final system in any way and I am using different circuit design for the final one. I just wanted to show what I am doing to the camera right now and how I am testing stuff with it. The commentary is pretty interesting if anyone is interested in how these systems actually do the speed stabilization. Check out my other crystal sync test videos as well :) 

Btw I secured some basic Leicina S kits from eBay for my tests. But I have enough of these cameras to build one or two final systems out of them and Simon can make proper overhaul to them to restore them in the best possible condition. If anyone is interested in this type of option, let us know :)   

[Aapo Lettinen]

On 7/3/2020 at 3:17 PM, aapo lettinen said:

- FRAME RATES!  I will add at least 16fps, 24fps and 25fps for sync sound shooting. All very accurate crystal speeds. But does someone need 18fps? let me know at early stage because it will change the circuit boards and I will need to know early on.  

The frame rates are by current plans:

- 25.00 fps 

- 24.00 fps 

- 18.00 fps 

- 16.00fps 

- one lower fps which is either 12fps or 8fps.  I have one unused divider so it is not extra work to add the 5th speed. But let me know which one of these would be more practical?

[Aapo Lettinen]

On 7/4/2020 at 2:06 AM, aapo lettinen said:

Btw I secured some basic Leicina S kits from eBay for my tests. But I have enough of these cameras to build one or two final systems out of them and Simon can make proper overhaul to them to restore them in the best possible condition. If anyone is interested in this type of option, let us know ?   

Included lenses are at least the 9mm Dygon and the original lens hood for the built-in 15mm lens. Both in excellent condition. One of the original carrying bags is included as well. The original carrying strap for the camera is included too.

When the rest of the cameras arrive I will look if I can throw anything extra in. I have to purchase two or three camera kits to build up one great working and nice looking final one so that is why I'll have some extra lenses and bags available.

So it looks like we have two kits available for preorder if anyone is interested ?  the exact price needs to be calculated but if someone is already seriously interested you can PM me and I will find out how much a kit would cost.

[Aapo Lettinen]

I am making optical encoder tests too! 

a prototype encoder made out of a cheap plastic gear. The final encoder disc will be CNC machined from better materials to ensure proper balance but this kind of hand made one can be easily used for all testing purposes.

I am working on a self made optical sensor circuit which reads the light changes caused by the rotating encoder disc and converts them to electric pulses which can be used by the Crystal Sync system. Needs some prototyping and experimenting :)  

the sensor needs to be integrated to the same circuit board than the rest of the control system which is why it is easier to design my own sensor system than trying to adapt an existing factory made sensor+preamp board to fit (most of the factory made ones don't even fit inside the camera so easier to just make your own then)

[Aapo Lettinen]

I purchased couple of Leicina 8S camera bodies for being modified and for spare parts. Additionally I have one of the Leicina SV models (the one on the left which has zoom lens) which I am going to modify exclusively for my own use. It has different motor system so it is more work to modify it (a lot more expensive) so I am not planning on selling them but if someone is seriously interested in it I can maybe arrange something :) 

Some of the cameras had original bags too. The idea was that we can cherry pick the best working ones out of these to get two or three working S-model cameras which are overhauled to perfect condition and then use the rest for spare parts. 

 

So the project goes on pretty well so far. I have the needed camera bodies and parts and everything planned. Control system design is pretty much completed and now I will need to build the first crude circuit boards to test the final control system in action.

There is one serious issue though. I am heavily relying on customer feedback when designing new stuff and I haven't got any responses from potential customers so far. I am sure there is people on this forum who are interested in this project but I cannot know without someone commenting or responding in some way to this thread. 

 

What do YOU think about this project? 

Is it a good idea in the first place? 

Did I forget some features which would be absolutely needed?

Now that I threw that zoom model in, is there someone who would only be interested in the project if getting a camera which has a zoom lens?

The fixed lens S-model was chosen because there is more of them available in relatively good condition and the built in prime lens is very good. The motors are easier to modify as well and it is possible to get spare ones more easily. The zoom model (8SV) is much more work to modify and repair and will thus be more expensive but I can investigate how to make it work if someone would like to order one. I am using that one in the picture for my own use :)

[Aapo Lettinen]

I checked the Leicina 8 SV model (the one with the zoom lens) and it is possible to modify to crystal sync as well. But will be more expensive like I said in the previous post because I will need to use some additional parts to make it work compared to the fixed lens S model which needs very little extra parts. 

I should be able to do the same 4 crystal speeds modification to the SV model as well. If someone is really experimental and wants to have lots of special features I could make a custom system which uses external control box via cable. Then a display and other features and lots more speed presets would be possible (I have for example external box systems available which have from 15 to 20 speed presets and other type of boxes can be made as well).

Edited July 17, 2020 by aapo lettinen

[Aapo Lettinen]

I made various tests with the exposure system of these cameras. 

I can easily make the automatic exposure to compensate for different framerates. As a byproduct the original button cell is not needed for the exposure system anymore. All the functions are powered by the same main battery. I can recalibrate the exposure system so that the speed ratings and markings are correct again even when the photocell has aged a lot. These features will be included in all the cameras.

It is relatively easy to make SWITCHABLE MANUAL EXPOSURE function for this camera as well. But it needs an additional selector and a bit more work and I have to calibrate them for every camera so it is an additional feature for preordered cameras. Not very expensive but will only be done for users who really need the function.

Let me know what you think. 

Now the camera has basically everything one can need when shooting Dual8 film. It has multiple speeds, crystal accuracy, reliable exposure system and for users who need it, I can add an additional manual exposure function for little extra. Even zoom lens is possible if one really wants it :)

[Friedemann Wachsmuth]

Nice work, Aapo!

In the video it sounds as if your Leicina controller is overshooting when getting started, are you compensating for that later? I think phase-locking alone is not enough, you should also keep the impulse *count* in sync if you want synced sound. 

A quick hint for optical sensor: Lookup QRE1113GR. It costs pennys and does exactly what you need — two resistors and optionally lowpass/schmitttrigger and you got perfect TTL impulses with minimal space consumption. 

Allow me one more thought: Why are you dividing your pulses at all? Syncing up higher frequencies is much faster and allows "intraframe" controlling of the motor.

Last but not least, the 8 bit PWM resolution of the Atmega isn't very much and might cause some jitter, no? I'd consider a broader PWM chip or a 12-bit I1C-DAC like the MCP4725 with a high impedance op-amp to drive the motor more accurately. That's what I did on my synced film viewer driver at least.

Keep up the cool hacking and don't hope too much on customers. It is hard to swallow but actual buyers are almost non-existent. Do this for fun and make it Open Source, it's worth it!

Edited July 17, 2020 by Friedemann Wachsmuth

[Aapo Lettinen]

11 hours ago, Friedemann Wachsmuth said:

Nice work, Aapo!

In the video it sounds as if your Leicina controller is overshooting when getting started, are you compensating for that later? I think phase-locking alone is not enough, you should also keep the impulse *count* in sync if you want synced sound. 

A quick hint for optical sensor: Lookup QRE1113GR. It costs pennys and does exactly what you need — two resistors and optionally lowpass/schmitttrigger and you got perfect TTL impulses with minimal space consumption. 

Allow me one more thought: Why are you dividing your pulses at all? Syncing up higher frequencies is much faster and allows "intraframe" controlling of the motor.

Last but not least, the 8 bit PWM resolution of the Atmega isn't very much and might cause some jitter, no? I'd consider a broader PWM chip or a 12-bit I1C-DAC like the MCP4725 with a high impedance op-amp to drive the motor more accurately. That's what I did on my synced film viewer driver at least.

Keep up the cool hacking and don't hope too much on customers. It is hard to swallow but actual buyers are almost non-existent. Do this for fun and make it Open Source, it's worth it!

Thanks! 

Keep in mind that I am using one of my old circuit prototypes for running the camera in the video. The final design is completely different and it has other style of encoder as well. I did not want to show it yet because I want to do crude circuit boards of it first. 

That QRE sensor looks very useful. It is not available from my regular suppliers though but there is similar style of sensor available. It would require additional parts to the boards though because of the encoder alignment so I may need to stick with the fork style sensor or the self made one.

Pulses are divided for getting exactly 50% duty cycle without needing precise tuning of the encoder disc and sensor alignment by 0.5mm or so. It was necessary for the first prototype because the encoder is not cnc machined. there is a gear ratio between the motor and the camera movement which is why smaller slot count is OK. I don't want to go into detail how my phase locking works but the impulse count is taken care of too among other things.

If you watched closely I used an op amp based analog pwm circuit for driving the motor in the test video. I have another much more advanced design which is used for the final system. I have never had any problem with 8-bit pwm when driving motors, it all depends on how you set the compensation settings and what the approximate sampling rate is.

Frankly speaking, this is a pretty challenging modification for 99.5% of the end users to do and I don't think they could make it even if having detailed instructions, constant phone support and readily available cad designs for the boards which someone else has completed. There is just too much different stuff to take care of. The end users just want to shoot with the cameras, it would be a nightmare for them to try to assemble this kind of system by themselves because there is tons of mechanical stuff to do and manufacture as well which needs lots of different skills (I am not doing all these by myself either) and the circuit boards are tiny and pretty complicated so working with them is not easy for anybody. Testing the final system is another matter and it is not easy for them either. There is things like light meter calibration which can be pretty complicated as well because it needs to be partially hardware and partially software based tuning.

I could do an open source project out of this of course but there is no use for an open source system which only me and you and two or three other people in the world are able to put together. So I would not want to waste time to a system which works great and is everything one could need from a camera but is so complicated to put together that no one can really actually use it in the end. That is why this needs to be a commercial project to some extent to even make it happen

 

Edited July 18, 2020 by aapo lettinen
typos

[Friedemann Wachsmuth]

Crossing fingers that you will succeed with this as a commercial project! I guess once you have your final PCB worked out and it all fits snugly into the camera, interest might go up. 

[Aapo Lettinen]

On 7/19/2020 at 1:22 AM, Friedemann Wachsmuth said:

Crossing fingers that you will succeed with this as a commercial project! I guess once you have your final PCB worked out and it all fits snugly into the camera, interest might go up. 

Thanks!  most of the PCB is worked out but there will be lots of other costs as well so I would like to have more funds for making the final system. But I will probably make a single speed version first for my own use and then the interest will definitely go up :)

the disadvantage of low interest projects is that the final price of the system will go up signifiantly and may not come down again even if the final product gains more attention in the end because the damage is already done (from 50% to double the price it would otherwise cost. so people are getting the same system in the end but if they want it being affordable that is not possible anymore) but that is how this camera stuff seems to work so what can one do :/  

[Simon Wyss]

Thinking over the speed question. Double Eight is the poor man’s film format. It still is the cheapest system, if it’s about black and white. Ektachrome is too expensive, regardless of which size. Sorry, Kodak!

8-R history is 88 years old now. It has, over decades, experienced a variety of SEPMAG sound systems that have something in common: high-fidelity sound at moderate costs. Sound couplers work a rougher or finer synchronism between a projector and a reel-to-reel tape recorder. There have been zebra stripes on the back of magnetic tape to allow a stroboscope adjustment with a portion of the projection beam reflected on it. Other couplers control the projector motor’s power supply directly off the tape speed, the tape rolling a smooth or a sprocketed drum. Again other couplers keep tape and film at an average constant speed. One projector runs picture film and magnetic film 8 strictly parallel at 16 or 24 fps.

At 16 fps 25 foot of 8mm film lasts two minutes and five seconds. At speed 24 it’s only a minute and 23 seconds. I think we should choose speed 16 to begin with. All 8mm projectors have the 16 fps speed but not all go up to 24 or 25. Also, film should continue to be enjoyable as projected film, so the decisions ought to be made traditionalist. Last, the camera makes least noise at 16 fps.

[Aapo Lettinen]

12 minutes ago, Simon Wyss said:

Thinking over the speed question. Double Eight is the poor man’s film format. It still is the cheapest system, if it’s about black and white. Ektachrome is too expensive, regardless of which size. Sorry, Kodak!

8-R history is 88 years old now. It has, over decades, experienced a variety of SEPMAG sound systems that have something in common: high-fidelity sound at moderate costs. Sound couplers work a rougher or finer synchronism between a projector and a reel-to-reel tape recorder. There have been zebra stripes on the back of magnetic tape to allow a stroboscope adjustment with a portion of the projection beam reflected on it. Other couplers control the projector motor’s power supply directly off the tape speed, the tape rolling a smooth or a sprocketed drum. Again other couplers keep tape and film at an average constant speed. One projector runs picture film and magnetic film 8 strictly parallel at 16 or 24 fps.

At 16 fps 25 foot of 8mm film lasts two minutes and five seconds. At speed 24 it’s only a minute and 23 seconds. I think we should choose speed 16 to begin with. All 8mm projectors have the 16 fps speed but not all go up to 24 or 25. Also, film should continue to be enjoyable as projected film, so the decisions ought to be made traditionalist. Last, the camera makes least noise at 16 fps.

hmm  would a single speed system be a better choice then?  I could make a system which only has one speed and if someone preorders the camera one could then choose which speed it would be. Then I could use simpler circuit board (cheaper starting costs so much faster to get the first cameras out) and it would save some modification work. 

so this would be a camera with: 

- only one single Crystal speed. one could choose which one it would be (16fps, 24fps or 25fps) but it needs to be set "in the factory" and cannot be changed by the user at all.

- automatic exposure which takes the power from the main battery like in the other designs. Meter re-calibrated for the chosen "hard-wired" filming speed.

- price would be a lot cheaper.

[Heikki Repo]

Before you make any final decisions on the pricing or the speeds, you really should show a test made with vision3 50D and scanned with good quality. As I see it, one of the fundamental issues is that for home hobbyists who are shooting to project, the old clockwork non-crystal cameras are good enough. However, the crystal sync becomes interesting as soon as one is trying to do something more soundtrack critical work. That pretty much means color neg and scanning.

Currently Youtube & Vimeo are full of Super-8 clips. If one is interested, there are references available. However, regular 8mm doesn't really have such references available - and the ones that are available, have some issues. For example this beautiful Bolex D8L footage by D. Toeppen (who is active on these forums as well) is only 12 fps  due to the way it was scanned:

Unless the real potential of regular 8mm as a professional specialist tool (and a good workflow) is demonstrated, I think it'll be difficult to find customers.

[Aapo Lettinen]

1 hour ago, Heikki Repo said:

Before you make any final decisions on the pricing or the speeds, you really should show a test made with vision3 50D and scanned with good quality. As I see it, one of the fundamental issues is that for home hobbyists who are shooting to project, the old clockwork non-crystal cameras are good enough. However, the crystal sync becomes interesting as soon as one is trying to do something more soundtrack critical work. That pretty much means color neg and scanning.

Currently Youtube & Vimeo are full of Super-8 clips. If one is interested, there are references available. However, regular 8mm doesn't really have such references available - and the ones that are available, have some issues. For example this beautiful Bolex D8L footage by D. Toeppen (who is active on these forums as well) is only 12 fps  due to the way it was scanned:

Unless the real potential of regular 8mm as a professional specialist tool (and a good workflow) is demonstrated, I think it'll be difficult to find customers.

Carefully shot test material with professional 4k scanning is of course possible to make and it would be great for the project to demonstrate the capabilities of Regular8 this way. 

Unfortunately making this type of tests is relatively expensive compared to the other costs of the project so it will cause delays to the project (eats up the portion of the budget which could be used for making prototypes or final cameras) . I am a little hesitant making that move because it could mean that the multi-speed version would not happen in time or at all.

Maybe I could ask if the film suppliers want to make a demo video and I could cover some of the costs that way

[Friedemann Wachsmuth]

Sorry, Aapo, but if extra speeds mean extra costs, I think your approach is somehow fundamentally wrong.
You didn't like my Timer-ISR suggestion for (almost) arbitrary reference clock speeds derived from one crystal. Your Atmega or Attiny has multiple of these timers built in already.

If you don't like that approach, you could also take a Si5351 which generates up to eight frequencies of your choice between 2kHz and 200 MHz at once. 

Don't get me wrong, I just want to help you to keep the BOM low. Supporting more speeds should nowadays not bump the cost anymore. 

[Aapo Lettinen]

On 7/22/2020 at 5:35 PM, aapo lettinen said:

Carefully shot test material with professional 4k scanning is of course possible to make and it would be great for the project to demonstrate the capabilities of Regular8 this way. 

I will probably shoot some b/w material with one of the cameras next month when testing the exposure meter calibration. Maybe the 4k scanning test can be arranged then ?  my camera is not yet fully serviced though so can't know how stable the image would be but it would be interesting to test nevertheless!