Wolverine-Hawkeye Telecine

[Stan Jelavic]

Ah yes, this can get confusing. Looks like these lenses defy the optics laws but they do not. Here is my thinking.
The sensor in both cases is withing the max circle otherwise you would see dark corners.
The 1/2" lens is at 2f with optical magnification close to 1:1 based on the size of the image. The 1/2.5" is not. Its focal center is not at 2f, it is further away and thus lower optical magnification. Here are the calculations:

As you can see both lenses' backs are at 65mm from the film but the lens center of the 1/2.5" lens is more than 32" from the film.
With the black housing like that, you do not know where the center is unless you find the info from the manufacturer.
What you can try is bringing the 1/25"lens closer in and then you will get the same image size and there should still be no black corners.

[David Brown]

I do like to understand how that works, thanks Stan.

My sensor has arrived! A month early, they were wise to under promise. I chose the 65mm distance for the 16mm lens only because that's about where I imagine the lens should be and it was only to compare the image sizes from a common point.

Reading thru the older posts, looks like sensor to film distance has been less than 50mm.

I can begin to reassemble the Hawkeye. Though I thought I would make more progress than I did. I waited for a long time before dismantling the Wolverine, and now I remember why! Even though I bagged each screw and part that I expected to reassemble, this will not go fast!!

On the upside, the camera works.

Kamel, I have printed the M12 lens extensions. It has been a challenge to print clean threads. I'm printing with a 0.4mm nozzle, should I go smaller?

I used your camera test mount and I found I had slightly cross threaded the lens into the mount. It was not aligned with the film holder. Small threads cannot be cleaned or sanded, they must be printed clean.



Did anyone ever test the extended base for the M12 mount? This spacer would permit the use of a std metal lens mount.



Lots to do.

[Kamel Ikhlef]

David,
I print with 0.4 or 0.2 nozzle, it dépends. But the most important in this case is the layer size.
Here my S3D settings :. layer : 0.1, shell :2, infill : 100%, temp : 230°, bed: 75°. ABS Filament, Fan 0%.no support. speed 30mm/s
For me it works very well. The threads are good. But I prefer design a lens holder with a specific size (not adding an extension tube.).
Let me if you need a specific lens holder size. Else, I block the lens with a ring (for the focus if I need).
For the camera support, there is an orientation for the camera board with lens holder on (cut in in the 3D design).
I hope that help you. Sorry for my english(google translation).

[Stan Jelavic]

Let me know if you do not understand my calculations David. It looks like the 1/2.5" lens should still work but you might have to bring it in closer to the film. But the proof is in the pudding as they say.
It is good that you are taking time to reassemble. Very easy to mess up. Just talking from my own experience.
I just received a package from Digikey and hope these are the usb connectors so that I can finish the V8 tests and order the V9.
Still some front panel work to do. Ultimately plan to use Kamel's but will get some simple ones for testing for now.

[David Brown]

Kamel C'est magnifique! Juste ces 2 ou 3 changements ont fait une grande différence pour l'imprimante. Il y a plus de paramètres que je dois lire. Google fait un très bon travail de traduction pour vous. Je remarque rarement votre anglais. Merci de votre aide. Pardon my French.

Stan I do understand your calculations. Lens calcs have been a little daunting, as I don't recognise some of the terms. I have usually just tested with what I might already have on hand.

I have replicated your excel sheet, but I need to get the labeled diagram ( I've saved it somewhere) so I can identify the variables.

Are there new options you are adding in V9, or just a home for the capacitors? Can I send some money for a V9 blank, or even a V8 blank? You bought several, I believe. You have expenses for parts developing this beast. Maybe we all could share to some degree?

[Stan Jelavic]

Here is more info:

The V9 at this point would just have the caps added. The board is getting pretty much packed so not much real estate left.
Thank you for the offer David to share some of the cost. I indeed overall lose money with this project but may be a tax write off possibly.
The board cost is $2 for 5 boards and $16 for shipping, $18 total. That comes to $3.60 per board plus domestic shipping which is < $4.
Actually pretty good. That is why people go with China. These guys know what they are doing. They also offer board assembly. Will look into that and possibly
significantly cut the cost down, making the boards available for everyone. Local assembly is too expensive for small quantities. Need at least qty 50.

I still have V7 and V8 boards.

[David Brown]

Did I get this right? In the equation there is a small f. Is that the same as capital F in the diagram?

Just to be clear is the distance 'a' or 'b' measured to the front element of the lens or to center or nodal point? I know how to find the nodal point of a lens.

Maybe it's not critical as these lenses are so small.

Apologies, the A should be 'a' and the B, 'b'.

[Stan Jelavic]

Ah sorry about that. Yes The small f is same as capital f in the diagram. a and b are measured to the nodal point. In your diagram A+B = F that represents the distance from the film to the sensor but that is not the focal length of the lens. I call that d in my spreadsheet.

[Stan Jelavic]

V8 USB test status.
Miswired pin 4 of mini usb. Used a jumper to test and it works. Also, the usb adapter cable orientation is towards the back cover which will put a strain on the connector.
Will mount the connector on the bottom. That way the cable orientation will be towards the front. Fixing it in V9.


V9 with USB fix plus frame sync caps added.

[Kamel Ikhlef]

Stan,
I see you are still working on the V9.
For USB. Maybe it would be better to put a vertical connector? what do you think ?
I'm going to rework this weekend on the front panel.
Can I take as reference your last diagram for the speed connector with a turbo button?
Would it be possible to have the latest V9 version of the schematic and the MSP code?
I agree with David. It would be normal to participate financially in the costs that you assume for the research and development of this project.
I agree to participate. Of course, I am aware that all the hours spent are priceless! Thanks again for your sharing.

Here a 3D vertical connector Simulation for example :

[Stan Jelavic]

Not sure what type of google translate you are using Kamel but must admit it is remarkably good. Thanks for offering to share some of the costs and for your support. I will go one more time through the design and then I will post both the design and MSP code online some time today. I thought to use the vertical connector as well but already have qty 10 horizontal connectors so wanted to reuse them. But maybe I can lay out the board to fit ether. The mounting holes is the common reference and the vertical mounts on the top and horizontal on the bottom.

Added the vertical mount. Now we have two type A USB mount possibilities.

[David Brown]

Kamel I wish your 3d model of the circuit board could be printed and run!

Stan I've compared the two sizes of your last test capture, as best I can. Not much difference in detail. Did you resize the original down to 720, or render a second version at that size?

Based on the zoom's range here, I'll guess this was among the best S8 cameras available. It would follow that this is also some of the best detail we'll see on S8 film.





The noticeable detail difference between the 1280 vs the 720 (enlarged 178% to match) is in the color noise. A solid color is noisy because of a bayer color sensor. Here it's because the 720 image is enlarged. That would likely be true on a TV or monitor if full screen is used. This would never be noticed on a phone or tablet.

[Stan Jelavic]

Thanks David for looking into this. The original capture was done with ICCapture at 1280x1024. The 720x576 resize was done by FredVideo script using the Lanczos4Resize(W,H) plugin. The 1280x1024 is close to the original but dues to cropping the resize is used to bring it back to the original size. The plugin is very good and has very little digital artifacts but still I assume that the resize close to 1:1 would have better quality.

Kamel, yes the switches can be as shown previously with the regular speed switch and a turbo switch.

Power LED test:


Not very happy with it because do not want to drive the LED hard. Concerned about the 5V regulator getting hot on the PS board.
The colors are washed out. Could use 3 LEDs instead of one, just a question of space on the front panel.
12V - 10K resistor
5V - 3K resistor
3.3V - 1K resistor

[Kamel Ikhlef]

David, Aha! Print the pcb no problem! but it will not run. It would be too great ! ...
(Just for fit space tests, this can be useful). Maybe someday close!

David, the quality of latest Stan's scans are frankly as good, or even better than some that I had digitized on a HD memory at $ 30,000,
by a professional Parisian Laboratory.
We should not be ashamed of our results. We are very close to get a very good telecine (once all the little details have been settled!).
If your basic film has a good quality, the final scan will be very correct.
As for the size of the captured images... Actually, for me, the ideal would be 1440x1080 (4/3) to integrate directly into HD 16/9.
in waiting for 4K !
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Stan, Thanks for all your great Work. (Thank you a thousand times as we say in France).
OK, for the "speed" connector...
I will try to add 3 small separate LEDs (R,V,B) for the start-up diagnostic (power LED) on the command front panel.
(I still have a little space by just connecting a 4-pin connector from your Interface Board to the cmd panel).
I will post the shema as soon as finished.
Yes, the google translation works not too bad (not too much nonsense at the moment, I hope !)

[Stan Jelavic]

Thank you Kamel. Yes, agree the quality is remarkable. Matches and sometimes beats the professionals.
The 4-pin speed connector is all that you need. One pin controls the slow speeds (0.5FPS and 1FPS) and the other controls turbo (0.5 becomes 1 FPS and 1 becomes 2 FPS). Rewind is a separate pin and pin1 is GND.