Kinoton Basement Reader SR-D video signal issues

[Tiago João Silva]

Hi!
I work at a film archive where we're capturing from film materials the 5.1 sound of movies that have lost their MODisk. We've been using a Kinoton Basement Reader pulled from an FP30D and we have random block errors on the output, in random places (so it's not physical damage) and in different amounts from print to print.

Below is the signal taken from the test points of the Dolby DA20 we're using (running film, and running leader - so only the perforation signal in the latter). Vertical scale is 500mV.


Two issues, as I see it:
1 - less than 2V amplitude, even though the minimum is 2.7 and the recommended is 4 to 5V.
2 - irregularity of illumination (not a flat top)
Does anyone notice anything else I may be missing?

The Kinoton Basement Reader manual warns that incorrect LED alignment would cause "scalloping" of the signal - which I guess is what I'm seeing there. I tried adjusting the position of the LED source, but while I could get the perfo signal to go flat(ter), the signal got so low that the DA20 went to solid F.

So, is it worn LEDs? Is that more or less likely than an issue with the PSU of the Basement Reader? Or the optics? Or the CCD board?
And if it's the LEDs, does anyone know of a source for replacements in Europe? It's Kinoton part# 0080 500 32063

Thank you.

[Steve Guttag]

You want the whiter-than-white (perforation) burst to be 4Vp-p (not 4-5) for optimal response. It should track fine down to 2Vp-p. If you cannot obtain 4Vp-p, with the LED aligned and its power turned to max, then the LED needs to be changed. Absolute flatness is not critical but you don't want a truly spiked response either. There are two rows of segments in the LED. If you get the vertical alignment right it will look flatter (not flat but flatter). You'd be better served by using an analog phosphor scope when aligning a digital reader. You want to look at the signal of the digital bits under the whiter-than-white curve and adjust focus until it is its tallest but also dimmest. A digital scope will not let you see that sort of detail.

You should also obtain a copy of Win DRAS (and there are probably restrictions on what computers (and serial adapters) can run it...as in nothing too modern...it will run on WinXP but I don't know if I've tested it on Win 7 or above...Win98 was in its heyday). Furthermore, you'll need measured film to utilize its numbers. It can tell you about the azimuth, lateral position and vertical jitter. Without measured film, you'll end up setting it to how your sample of film was made and trust me, they can be all over the place. We used to buy test film by the 1000' roll with a QC chart to get it more precise from system to system.

[Tiago João Silva]

Thank you very much, Steve. We have DRAS, this is what it looks like.
DRAS10_20240826_180624.jpg
As for the phosphor scope, we've noticed and I agree, so we'll ask an electronics engineer that works with us to lend us his scope when it comes to do the final alignment
Our biggest SR-D test film is a Cat.No.69P-50, so it's just 50 feet and if it came with lab measurements, we've lost them in the meantime... Some source for a big roll of Cat.No.69T would be much appreciated as well, we only have a small loop of that and it's shot to pieces, the SR-D track is unreadable.
We can't be _too_ precise with the alignment, because we read both negatives and positives, and perfect focus/position for one isn't the best for the other.

[Steve Guttag]

At first glance, it doesn't look too bad on DRAS (aside from all of the red, of course and the 2.9 error reading shows that. The vertical jitter may be a bit high and the LED level looks low (the histogram should have flattened peaks (not heavily but like the first peak...so should the second one (they represent the digital 0 and 1 and a flattened peak is a more solid determination of the bit).

Without the analog O-scope (20MHz minimum but 100MHz does have a better display when you are looking at focus), you can use the "Focus" in DRAS to optimize it a bit. Magnification should be 100% but, as I recall, 99% is a greater penalty in response than 101%. Azimuth is VERY film dependent. The reader will track well within -3 to +3 but that does not mean one should take any film and set it to 0. If you cannot get measured film...try using a collection of disparate film and go with the average of all of them (but you want them to come from different sources/vintages...etc.) You might also have to throw out any that might, clearly, be away from the norm. if you find that 10 pieces of content measure -1 to -3. But one measures +3....I'd say that the +3 was the oddball and made poorly. So with the 10 measuring between -1 to -3, presume your reader is at -2 and and adjust it so the average is closer to 0 so you get the full +3/-3 range for the all (or nearly all) of your content. Lateral is less critical. The blocks just have to stay within the CCDs ability to see it.

[Ryan Gallagher]

Can confirm, Win-DRAS still works on windows XP, 10, and 11. I tried a couple serial adapters before I got it working though, the FTDI chipset ones seem the most reliable to me.

Does your station allow you to configure it to run a loop of film? That is an answer to your short spool of test film problem. If you just need a bunch of content to test with, the cheap trailers for films no one cares about anymore on ebay are a good resource. (also a good resource for that analog scope).

Analog scope definitely helped when I was tuning my 702s. My scope now lives on a high shelf always plugged into the 650, and sometimes use it during booth tours to better explain how that audio data is getting read off the film.

[Marco Giustini]

Bom dia Tiago!

It's been a while, I do have one of those readers a home.

What Steve said of course.

I was also saddened when I realised my brand new digital scope could not show SRD readings properly It's a bit unclear on your scope but I'd say the level is more 1.5V? Anyways, it should be a bit higher. 4V or 3V won't change much but 1.5 might be a problem.

There are more experts than myself on this forum when it comes to Kinoton but some boards/power supplies had different gain settings. Start from the SRD reader. You might have a jumper to set the sensitivity. If you set it to "high", it will double the output level. The PSU I have on mine has a setting for gain as well - mine was set too high so that head had been clipping on 5V since its installation and nobody noticed. I wonder whether maybe yours might be set the other way round? I found an old picture of that just for fun - vertical scale is 2V so video level is almost 6V and it's all clipped

I'll try finding the old documents but maybe you can send some clear pictures of the SRD PCB and the power supply?

Mechanical alignment is of course super-critical and one adjustment affects all the others. But the Kinoton head is well made and adjustments are "simple". Focus adjustment is critical, you can do that with DRAS and - ideally - a film running, not a loop.

Make sure your rollers are all working well, the drum spins freely and the roller holding the print on the roller is also cleaned and lubricated properly (I know Steve disagrees with the usage of cardan oil there!).

One final comment: why are you dismissing random print damage? I had that all the time with some poorly-printed films back in the days. My V5s tended to slowly scrape off the emulsion on the SRD. Where the damage was getting larger, the processor would briefly "F" for a moment. Does the DA20 agree with what the DRAS says in terms of error rate? I remember an old version was quite misleading on that.

[Tiago João Silva]

Marco, I'm dismissing physical damage because the errors appear in different places in each playback - when there are splices or similar issues the errors are consistently in the same place.
Are these the trimmers you were talking about? Aren't they pegged to the top already?
PSU_trimmers.jpg

[Tiago João Silva]

Oh, it looks like I can't edit posts after a while. I missed Marco asking about the error rate: yes, actually DRAS can stay at Error Rate 0 while the DA20 is showing 4; happens with most movies but not all. Also, most of the time the Histogram panel doesn't show anything. If anyone knows how to fix that as well (without moving to WinDRAS) thanks in advance.

BTW, might as well leave this documented here. I don't know if it is common knowledge (it wasn't for me) but DRAS can reconfigure the DA20 (and CP500, not sure about CP650) to read sound negatives with some command-line switches. These were found out for us by Simon Lund from Cineric; after asking around, he got a message from Mr. Thomas G. Kodros (from Dolby) explaining how.
This only works with the DOS versions of QC and DRAS, hence why moving away from DOS to Windows isn't an option for us. Is this common knowledge or should I leave that information here in the forum (maybe making a new thread for it with a more appropriate title)?

[Marco Giustini]

It's been a while, I cannot find those documents. Also, I've never been a Kinoton guy so apologies if I say rubbish.

I found the attached document which says that if you have the new cat 954R you can move the jumpers J3/J4 to HIGH to double the sensitivity. Obviously you need your sound processor unpowered before you do that.

The same document explains how to modify the PSU in case the PSU is outputting too much voltage/current - which was my case but not yours I think. Unless you have the "low output" power supply with a "low gain" CAT and lens.

This also depends on what kind of lens you have - the SRD lens.

Those trimmers on the PSU should adjust the voltage/current output to the LEDs. You want to make sure you're not overdriving the LEDs - making sure you have the best optical alignment as well. But in the end, if your Video Output is low and your mechanical adjustment is ok, all you can do is turn up those trimmers and/or swap the LED if maxed out already.

DRAS should read what your DA20 reads. I've seen that discrepancy before and I don't know why that happens. I think you might want to test the latest DRAS version as a test.

Please do leave that "negative" option here, at least I am not aware of that.

Errors in different places on each playback: I think you should give WinDRAS a quick test to have a better idea of what happens. Trust the DA20: is the error rate suddenly jumping to F from 4 in random places or is it getting to F while reading higher values like 5 or 6? The graph on WinDRAS might help you understanding better what's going on. But you need to fix that video level anyways first

Also, inspect the print carefully. Many years ago I had to deal with improperly printed prints which were literally peeling off the base. That would result in the CP reading 4 and jumping to F apparently randomly.

Did you clean/lubricate the roller which applies pressure on the drum? Did you check that the drum can spin freely?

Sorry, the above is a bit of a brainstorming and might be unclear But I hope it helps.

[Ryan Gallagher]

On the DRAS / WinDRAS subject. Does Dolby still host these installers somewhere officially? Trying to get them installed on our domain controlled laptops and our IT person has made a request to find an official download location, rather than the film-tech hosted copies.

All in the goal of working towards retiring this win7x laptop in the booth that can't even run the current version of spotify anymore. Which is probably the bigger security risk. LOL

*sigh*

[Marco Giustini]

There is no official public location where to download those files. Dolby should be able to supply them but either on a support ticket (through a dealer) or if you have access to their support website (if you are a dealer).

[Tiago João Silva]

OK, that was very helpful. The digital board in this reader is a 654R and following Dolby's FB224 (on the Warehouse) it had the jumpers on the original position, which now I've put on the 2x position, hopefully increasing the signal level. I don't think this is an updated reader because the Geprüft date is 12-01-99 - so it should have both the older lens and the original PSU.

On the PSU the trimmers are the ones I noted, the digital LED trimmer is the rightmost one in that picture, and it's already as high as it can go (the signal moved, but only downwards.

Now I have around 2.3/2.4V on the perforation signal, as shown (scale is 1V/div now)
Perfo_Cat.No.654R_preamp-x2.jpg
Still not enough, so it seems I really do need a source for some replacement LEDs.
Thank you!

[Marco Giustini]

2.4V is a good starting point indeed, well done! As Steve said, that should be "good enough" to give you a good result. I wonder whether the PSU you have is a "low current" one, assuming they even exist. This is a question for a Kinoton expert. To clear all doubts you could wire your multimeter in series and check the current being drawn - make sure you wire it appropriately and, most importantly, you wire it back to "non current reading" afterwards!

While you look for a diode, I'd make sure your mechanical alignment is the best you can have and then try to improve the focus - it looked a bit low on your previous screenshot. Maybe you can install WinDRAS on a random computer which is not IT managed just to test?

You never confirmed whether you looked into the cleaning and lubrication of the roller - also whether your rubber o-ring is in good shape.

This reminds me I shall run my "home" FP20 a bit, it's been sitting idle for too long.

[Tiago João Silva]

It took a bit of time to get a CRT scope, and what a disappointment: the scope had a malfunction on Channel 2 and the Ext.Trigger input couldn't catch the clamp signal from the DA20's video acquisition board, so this was the best we could do.
CRTscope.jpg
No matter how much I focused the lens, I couldn't see a difference in the area of the digital data, all I could do was (eventually) drop the signal level. If I moved the LEDs to remove the signal reflections, the DA20/DRAS showed a straight F. Anything less than perfect LED alignment would make "Corners" drop from 95-98 to 80-ish or less. The best focus I was able to achieve (measured in DRAS) was 72.

I've pulled another PSU from another projector, which came with the trimmer at about half-power, but that actually seems to be the low-current one, because when I connected it the LEDs were noticeably dimmer, I had almost no signal and even after pushing the trimmer to max all I got was less than 1V (and the jumpers are still at 2x...). So I would think that the basement reader on that other projector is the later high-efficiency one with the better lens, where the whole chain works at lower power to eek some more lifetime from the LEDs.

So I'll try to source another oscilloscope and repeat. At least now I know if something's wrong with the PSU I'll have to change everything, because the replacements I have available can't mix-and-match.

Marco, you keep reminding me about lubrication. Everything seems to run freely, nothing's dragging or grinding as far as I can tell. What should I be paying attention to, how would I tell if the mechanics need lubrication?

I've been directed to a source for some replacement LED's, but they're in the USA. Are there really no sources for replacement parts like these in Europe?

[Marco Giustini]

What is the scale of what I see on the picture?

For the focus, ignore the scope and trust DRAS Get the highest level possible and that's it. Make sure the lens is clean - you'll have to use a cotton bud and bend it at the end (don't touch it with your fingers) to reach the lens. Ideally, if you want to be thorough, you could remove the lens and clean it on the other end as well - that will require a full re-alignment though.

I'd use some lens cleaner as a fluid, whatever you use, make sure the lens is wiped dry at the end.

I cannot find the manuals anymore but the roller which is applying pressure on the drum is critical in the smoothness of the movement. First, with no film, spin the drum while holding the spring roller away and make sure it spins freely for quite some time.

The spring roller is supposed to have some viscosity, not to spin freely - but I know Steve disagrees with Kinoton. I've never been a Kinoton person so you might want to listen to him and not me
Anyways, you can remove that roller, wash the bearing - which should be open if memory serves. I always use white spirit to wash bearings. Then add lubricant - this is where Kinoton and Steve disagree. Kinoton says to add "cardan oil" which is very viscous. The idea is that the roller should not spin freely and the oil is causing some natural friction.

Finally, the o-ring on the roller itself is the one which is actually applying pressure on the film so you're supposed to make sure it's in a good shape - remove it before washing the bearing and... it might snap So get a replacement first!

Finally, get WinDRAS for a test. We still don't know what's causing this issue and - from I can see here - your reader should be capable of giving you a good reading, 4V is not really required to read correctly. So watching an actual histogram might reveal something. Intermittent "Fs" can only be

- Bad print
- Damaged print (won't take much)
- mechanical issue
- borderline alignment

I'll see if I can find the manual with the instructions about lubrication.

(some indications here: https://www.manualslib.com/download/...n-Fp-30-D.html)