Wolverine-Hawkeye Telecine

[David Brown]

Bruce I will be printing those files.

Stan I will be glad to wait for the V12 board. I had the impression that you had just received and tested 20 V11 boards and they would be going to waste.

Yesterday I installed a new lens and let it capture for a few minutes. To answer your questions Stan on jitter, I have a V6 board and 2 speeds. The fast speed has no obvious restraint on the takeup. The slow speed does make small increments in sync with the claw, but it moves more than one frame. If I restrain the takeup for 10 cycles, it will take up the slack in 2. I'm only using the first and last rotating pulleys.


This contact sheet easily shows frame positioning errors (jitter).



Just captured the clip again in fast mode, then in slow. I placed a piece of masking tape on my monitor at the top of the film perforation and a thin strip across the gate.


I couldn't watch both so I put a camera on the gate.


https://drive.google.com/open?id=18i...-q0gVBW8USMvZR


Fast mode had a couple of jumps, but the video camera's battery died and I lost that file.

Slow mode was solid as a rock.

Again in fast mode, solid as a rock. Well there's random for you.

I was watching the takeup and did not see any takeup pull thru the gate. I might have missed it. I will try this again later, it takes a lot of time and is unpredictable.

It's hard to imagine what could do this besides the takeup. My shutter speed is 1/20 and if the trigger was early there would be motion blur. Maybe the claw didn't make a complete cycle and stopped short? It's difficult to identify if the positioning error is late or early.

[Stan Jelavic]

Thanks for the input David. I will send you the V12 and to the rest of people on the V11 list.
The setup that you have will be definitely susceptible to jitter. The takeup pulse is too long in all of the boards shipped so far and it cannot be made shorter in FW. V12 should solve this issue because the takeup pulse can be very short. I have to do more testing to see how short I should make the pulse. If too short the takeup may not be able to keep up.
Here is a bit more info on pre V12 takeup timing:
- the claw starts going up - the camera trigger goes high on pre V11 boards
- the claw engages with the film - the trigger goes low the takeup short pulse is sent to the PS. The PS turns takeup on
- a few hundred ms later the PS turns the takeup off. The motor keeps spinning due to inertia for much longer.
- The claw at the end of travel. Starts going down and disengages the film. The takeup motor stops but possibly pulls on the film a bit while claw disengaged.
- The claw is down and travels back. Once it gets back to the trigger point it triggers the camera but the film is still disengaged and moved a bit. And that is where you get the jitter.

With V12 the voltage to the motor goes through the board and is shorter. Some inertial still present but not enough to move the film at the end because the pulse is shorter. Hope this explanation helps.

And yes one more thing. The trigger pulse is now at the point where claw goes down allowing much more time for the camera to take a stable picture. This does not help with jitter but fixes the smearing problem at 2FPS... just wanted to mention it.

[Chris LaRocque]

Hello all

Yes the libraries have transforms for correcting a variety of sins. The trick is to have a good reference so I need the edge of the film and sprocket holes in frame.

I have a first pass script that detects and crops to a frame. Next step is adding some qualification code to assure a robust discrimination of the frame structure.

Maybe next weekend I'll have something to share. It'll likely operate on a folder full of frames initially.

Regards

C.

[Stan Jelavic]

Thank you Chris. Do not want to sound negative but there is an issue with that approach. If you need to keep the sprocket hole for reference the auto exposure will be thrown off. You will have to use manual exposure or implement auto exposure in your code or take another approach. Still doable but more work.
V12 update:
Redesigned the front panel to include the takeup forward/reverse switch. Here it is what it will look like. Let me know if you want to implement any changes.

[Bruce Davis]

DAVID hold off until I make the files dated March 9th available.
Sorry posted a little prematurely.
Rgards - Bruce

[Bruce Davis]

Stan, have printed the S8 & N8 capstan pulley's with corrected dimensions and they work as expected, will post the modified tomorrow. When testing the S8 jitter yesterday I had a half captured reel of film in the Hawkeye (half the film on the feed reel and half the film on the take up reel). The jitter was very minimal.

When testing the N8 capastan pulley, put a new full feed reel on and an empty take up reel. Found a problem in that the pull strength (torque) decreases as the diameter of the take up real increases and this causes the frame to slowly shift position in the gate, this means having to stop the capture and re-adjusting the framing. This was not noticeable with half empty/full reels as the diameter does not change as quickly as with an empty take up reel.

Being able to change the pulse time is good but how does one take into consideration the change in diameter of the take up real. As diam increases the less pulse time may be required.

Regards - Bruce

[Stan Jelavic]

How are you routing the film Bruce? Is the capstan on the first location? What type of pulleys are you using? All roller? I will test it today.

[Stan Jelavic]

Ordered V12 boards qty 15.
Updated the qty sheet.
https://docs.google.com/spreadsheets...it?usp=sharing

[Kamel Ikhlef]

Hello Everybody,
While checking the dimensions (inspired by Bruce), I just noticed that there was an error on my part
on the short version of the old standard curved pulleys (0.405mm too close to the metal wall of the wolverine).
The film is therefore slightly misaligned. So I fix that and I will post my 3D files.
(Which will leave a wider choice with Bruce's files (Thank you Bruce for sharing) and mine.
(With brass insert or not).

Thank you Chris for your contribution, I look forward to your feedback (automated cropping can be very interesting).
Any means to improve the system (and the jitter) is welcome. Thanks to all.

Stan, would have to scan the hawkeye in photogrammetry to have it in 3D printing !...

BTW : Here the new STL Files Update.
https://drive.google.com/file/d/1zbK...ew?usp=sharing
Have fun !...

[David Brown]

Hey Stan On the list, what is a complete unit? Is that include a Wolverine? If I add "include capstan", does that mean the motor and pulleys?

The capstan motor is still unavailable on Amazon. It comes and goes. I had to order one from China that took so long, Amazon put them back on line. I ordered one from Amazon and was still able to cancel the China order!

My motor is ok. I would be interested in the pulleys.

This might need mention in the manual, if I just leave a note here, It will get lost over time. BE VERY GENTLE with the capstan set screws. I have a heavy hand. Before mounting the back pulley on the motor, put the BP on a 2mm or 5/64" drill bit to get the screws even and close, then do final tighten on the motor.

I just stripped the threads on one of the new set screws. I've been moving things around. It was my third time mounting it. I've have even used too much pressure with a set screw with insert. The insert will come out!

It was quick printed with PLA. I don't know how much more durable PETG would be, it has more tensile strength. I planned to reprint anyway.

[Stan Jelavic]

Funny that you mention it Kamel. At one point I wanted to do a quality 3D capture of Hawkeye. Maybe once we are all done with it Thank you for your hard work.
David, I am still struggling with what to provide for capstan. As you said the motors are is short supply and long delivery and I just cannot find low cost good quality prints. The industrial grade are asking $10 per part approximately. Will see how that goes.
I will add your note to the manual. Possibly stronger material may help.
Bruce, I ran the 400 ft reel at the end and there is no problem with the takeup.
Rewound the reel to the end (reverse takeup works great):
https://photos.app.goo.gl/Ut6MH5fCGy2rvbB38

Ran slow and fast mode. Takeup still works:
https://photos.app.goo.gl/rtQsDG8YUWS4KnYM7

IC Capture - pretty low jitter. Actually the same as no takeup at all:
https://photos.app.goo.gl/XRoPvAxUYrShzAap8

[Bruce Davis]

Hi Stan Kamel David Chris, Stan I see you have moved the capstan to the third pulley location, mine is still in the first and am using all roller pulleys (although they do not seem to roll). Your results look fantastic, love what you have done with the film rewind.

Kamel when I originally measured the distance between the inner edge of the S8 film and the Wolverine metal plate it was 3.8mm. Have since re-checked and it was wrong, it is 3.4 to 3.5mm, depending on film position in the gate guide as there is some sideways clearance between gate guide and the film. The dimensions I will use are S8 = 3.5mm & N8 = 3.1mm. With the capstan in the third location (as per Stan's demonstration video) this dimension is not going to be as critical. Thanks for the revised STL files.

David 3mm dia set screws are already very small, with 2mm dia there is very little thread depth to grip the plastic (not sure if you are using inserts). PLA has a very low softening temperature, the friction that is created when screws are threaded into plastic creates considerable heat, the faster the rotating speed the more heat so stripping happens easily. I find that if a lubricant is not used (spit or water) the plastic can strip and clog the metal thread root with plastic causing poor thread formation.

Regards - Bruce

[Kamel Ikhlef]

Aha ! OK Bruce,
All the modifications I made in my last update are based on a distance of 3.8mm (metal / inner film edge).
I have serious vision problems (I work with magnifying glass and zoom) and I have a lot of trouble measuring this distance.
It is for this reason that I started from your dimensions.

If you are sure of your new measurement, I will modify my design (no problem, it is better to be very precise).
We have to take every chance to avoid jitter and It should not take much time...
Thank you again for your feedback.

Finally, I put my wife to work !
Here is the reading of his measurements at the vernier :
from the Metal Panel to the outer edge of the film : 11.80 mm / 11.85 mm (Min/Max)
(according to the lateral movement in the film guide gate).
Width of my film S8 = 7.98 mm (not 7.90 mm)
Inner edge = 11.80 mm - 7.98 mm = 3.82 mm min and 11.85 - 7.98 = 3.87 mm max
My last update should therefore be correct if we take into account the lateral movement tolerance.

[Stan Jelavic]

Thank you Bruce. The reason I moved the capstan to the 3rd position is to avoid any jitter caused by capstan misalignment. The roller type pulleys are required to enable to pull the stuck film from the 3rd position. The roller pulleys do roll in in my installation. They also make the takeup work better. But the V12 is required for the setup with the takeup cct. Otherwise you could get excessive jitter.

[Kamel Ikhlef]

Thanks Stan for the v12 progress videos.
It looks very promising. Can't wait to test it !