Kinoton Basement Reader SR-D video signal issues

[Tiago João Silva]

Hi.
Thank you to everybody who tried to help.
Eventually I got some new LEDs from Boston Light and Sound (http://www.blsi.com/) and replaced the one inside the original optical block (BLSI also sells the full block, and I should have asked for that instead because field-soldering stuff in narrow places is very much out of my comfort zone, but that will be a separate post). After some very fiddly adjustments, I think I got the soundhead as good as it is ever going to get.

But as you can see, I still have sporadic uncorrected block errors. Whereas before readjusting the soundhead most films were on a 3-4 error rate baseline, now it's between 0 and 1+. But there are still error bursts. It can stay at 0-1 for tens of seconds but then jump to 3...2...1...0 (several more seconds at 0), jump to 4 (and "Fault" LED)...3...2...1...0 (several more seconds at 0) or straight to "F" (uncorrected block) - attached video, "error-bursts.mp4".

But just I noticed that with the DA20 powered on but the soundhead powered off, I have spurious signals on the "sync" LEDs of the Cat.No.721s. Do your DA20/CP500 do the same? (video "idle-sync-lights.mp4").

Thank you.

[David Ferguson]

With my DA20 on, and cat 702 reader on, if I'm not running film through it, the sync/corners found LEDs are never lit. So that is an oddity that should be investigated. Those LEDs indicate that a corner of the SRD block has been found, and as those corners are special markers (cross multiplied barker codes), it's very odd that it's finding them if film isn't being run.

Remember that the SRD reader is an analog video reader, and it sends analog video signal back to the DA20 which decodes it. If you're using a too long cable run, or the wrong cable, or it's not properly shielded, it can pick up interference which would degrade performance. Radio stuff is far beyond what I know about, but maybe this could be the cause? Do you have another known good cable and/or DA20 (or CP500/CP650) that could be swapped out, to see if that is the issue?

I also know that Dolby claimed the metal chassis of the DA20 was pretty critical for interference, and said not to run it with the front unscrewed. But afaik, this was because of signals leaking out, rather than in (and was more prominent for the DA10).

[Tiago João Silva]

The sync LEDs blink with the soundhead's light source off. When the light source is on the sync LEDs are off.

The cable was our first suspect, since it seemed the easiest thing to replace. In the end, we were lent an original projection booth cable, an absolute hose of a unit (CAROL (R) 24AWG C0926 - 75't E111240-6 CM(UL) C(UL) CMH OR AWM STYLE 2493 MADE IN USA) almost 5 meters long that performed just as well as our original 1.5 meter cable made out of AES-EBU strands.

I took one reel to the main projection booth of Cinemateca that has a CP650 and has just been aligned by the installer. I only had one uncorrected block instead of the 6 I had on the same reel read on my reader, and the other statistics were worse than my reader, except Focus.
I'm trying to locate a CP650 to test with.

[Henry Atkinson]

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).

Not to derail, but I've had 0 issues with WinDRAS on Windows 10 64-bit with native or FTDI serial ports. I've yet to try it in VMs, but on bare metal, I've yet to be unable to run it.

[Tiago João Silva]

Whatever it is, doing all this troubleshooting (better cable, new LED and optical alignment, putting all the covers back on) removed the noise around it and I think I've finally found where the issue is.
On a cold start of the DA20, I can have 20 or more read errors on a single reel.
After warming up the DA20 for six hours, I have 0-2 errors per reel.
Does it sound like capacitors for you too?

[David Ferguson]

I would be checking the PSU of the DA20 - with a multimeter, are all the voltages in spec? My guess would be that on initial power up, it's not in spec, but it gradually drifts into spec as it warms up.

There is a Dolby bulletin about the power supply connectors of DA10s and DA20s, that may be interesting for you to read over. Most relevant is it contains a chart showing the acceptable voltage ranges of the power supply, which I've included below. Measure it, and see if you're in spec. If not, then checking the PSU connector, and maybe a power supply replacement (or more likely, recap) will likely fix it. If you are in spec... then it could be caps elsewhere, but that's getting into analog video processing and beyond my area of knowledge.

Screenshot 2024-10-29 at 13.48.31.png

[Tiago João Silva]

Thanks for reminding me. I had found that Field Bulletin in the Warehouse, tested the voltages and they were in the tolerance zone, but I think I tested them warm, not cold. But now that it's finally stable, I have a backlog of films to read. I'm not shutting this DA20 down unless the power goes out. No new troubleshooting until next week, I'm afraid.
Thank you to everybody who tried to help!

[Ryan Gallagher]

Slightly off topic, being that I have Dolby 702 readers. This is the most active SRD thread currently. But where does one begin to isolate/improve the "vertical jitter" number? It is our only indicator that bounces between yellow and red on WinDras.

For my understanding, can one correlate vertical jitter to lateral "stretch" of the signal on the oscilloscope too? (Am I understanding the time-base of the scope x-axis correctly there?). Similarly it would then correlate to the 702 tension roller compensating arms stability? So our goal should be to get those to compensators to calm down entirely, which is an easy thing to look for.

I've read in other threads that some additional tension ahead of the reader (such as kelmar cleaners etc) potentially help. But we have nothing ahead of the reader, not even a guide roller on the reel arm.

My prior instincts were having to do with reel/core/spindle condition, which can certainly play a role. But we recently ran 24" 6K 8" core Film-Tech reels that were minty and round, but still had some vertical jitter drift between yellow and red, same as our other reels.

Our feed tension was backed off significantly recently in an effort to keep it well below our non-adjustable rewind bench tension and not in the cinch danger zone. But the vertical jitter was also present before at those higher feed tensions.

My next instinct is to service the arm clutches, which no doubt have not had that done in a decade. Perhaps re-oiling the clutch pads, if ours are designed to be oiled, will smooth things out. We don't have many problems reading the soundtrack, error rates are typically 1-6 depending on the print. It's just one of those things to improve upon eventually. It's hard to notice at the feed arm tensions, but the take-up being much higher you can feel the "start/stop" stuttering friction behavior in the clutch when turning a reel slowly by hand.

[Steve Guttag]

Vertical jitter is the stability of the film through the soundhead. So, the flywheel, bearings, castings that the bearings sit in, and how well isolated the previous is from outside mechanical sources will affect it. Also know, printing can affect it too so some may be baked into the film you are using.

If you have an upper reel or a platter that is putting irregular tension on the film, I've seen the jitter get pretty bad. If on reels, make sure the clutch is set up well and if a platter, particularly Christie, either remove or well-lubricate its back tension spring. Also, any out of round rollers will cause irregular movement.

Dampers can have an effect on it if both under or over damped. In penthouses, too many people reached for the airpot to "fix" things when really they were causing more harm by not fixing what was causing the jitter (fix it at the source). Century soundheads have bearing receivers that are over-bored and will allow the drum shaft to orbit some. Component Engineering had a fix by changing where the loading spring goes...remove the factory one, push the bearing all of the way in, then use a loading spring on the inner race (all on the gear-side, not the film side). That did wonders for Century soundheads. Naturally, make sure that the bearings are in good order without drag or jitter due to the balls not spinning freely internally.

Christie has numerous issues...they start the jitter with the gear belts they used and then combined it with an inadequate flywheel and rounded it off with their damping. Over the years, you could see them address their design issues, one-by-one. So, their last generation of "GP" projectors had usable basement readers. The flywheel got massive and the gearbelts became round-tooth, which among other things, stopped the cogging. If you want to see what is going on in your projector's gear train or film path, get a strobe and try to "freeze" the moving part of interest and see just how irregular its motion really is (be careful beause the stuff is really spinning as before but it will look stopped).

Simplex gearbox soundhead didn't have many issues with jitter (They had good flywheels and such and the loose loop system does a good job of isolating the sound...plus they had an extra sprocket to isolate the take up from the sound). The 5-star shipped with too tight a return spring. You can re-bend it to get it better (reduce the force). I found that if you held the roller (removed from the projector) such that the roller was parallel to the floor, if you pulled the roller to the stop pin, it should have just enough return spring tension to return to the rest pin (with the damping "fluid/gel/goop" applied). The film should pucker out from the pinch roller when up to speed, if you have it right. This will cut way down on jitter.

The Dolby firmware does well with jitter but, like all of the parameters, the less it has to do the better the results.

[Tiago João Silva]

IT'S FIXED!

Well, almost. For now.

Getting a new DA20 finally got the system working properly. Now when it errors out, it does it always in the same place, and the playback statistics look very nice.
This DA20 is not without its issues - the 5V rail's LED is off and it measures 4.87, so we have to check if the Film Bulletin was applied and trim the PSU output if there's the margin for it. But so far, so good.

Thank you for the help, everyone, and if you need anything, just ask.

New processor.jpg

[Steve Guttag]

There is a margin and I don't believe it is ever below 5V. A big problem was the molex connector on the power supply. The pins would burn up. They went to a three sided connector to address the issue but the field-fix was to solder the wires directly to the supply's output.

That said, given the age, it is worth ensuring that the power supply itself doesn't have issues with its capacitors.

[Tiago João Silva]

Thanks, Steve. You don't happen to have a picture or a SKU/PN for the new connector, by any chance? "Dolby part number 83268" isn't getting me much.

Yes, I hope we get to do preventative maintenance on this decoder before it breaks on me. But the previous one has definitely something bad somewhere, and this one is hanging in there (probably by a thread) for the moment.

Just read a 1900ft reel, 0 block errors . Now, why 0 and not the usual (about) 100 where the Dolby start is loading code and metadata between the START mark and the twopop? Because some nimrod at the original lab (28 years ago) spliced leader until 7 frames before the twopop, so DA20 and DRAS could only pick up the stream somewhere after the twopop. To find out how if any frames were lost I'm going to have to digitize the SVA track and match... What fun.

[Tiago João Silva]

I've just realised I didn't update this thread with what we've found out was causing the errors on the first DA20: it was one of the Cat.No. 671 cards that was misbehaving.

It was from the later revisions (serial 10###, Rev.1) while the other three cards across our two DA20s were serials ### (Rev.0). So I don't know if it was something specific to that card, or if cards from different revisions just don't work together. But replacing the Rev.1 on the "bad" DA20 for another Rev.0 solved the issues immediately, and putting the Rev.1 on the "good" DA20 reproduced the issue.

In fact, since the PSU on the original "bad" DA20 is actually in better shape than the replacement, we're back to using it as our main machine with two Rev.0 671 cards.