Capstan Servo Duty Cycle / Jitter

[bjoneson]

I'm in the process of restoring a Scully 280B 2 Track, and am running into a few issues with the DC Servo capstan.

The motor is mechanically in good shape, but is exhibiting some "jitter", and I'm seeing a duty cycle of around 75/25, as opposed to the specified 50/50.

The documentation for these machines is a bit thin, and I have been in touch with a Scully specific user group on Yahoo, but it seems there are very few people with knowledge of this DC Servo Capstan which was an optional accessory for these machines.

I've attached the manual, which includes a schematic of the servo control board, and I'm hoping some folks may be willing to offer some circuit analysis skills.

It seems other decks of this vintage from Otari and Ampex often had trimmers to dial in the duty cycle, and jitter. There is one trimmer on the Scully control board, but it is undocumented in the manual. I believe it may be a gain trim for the drive amplifier.

I did a bunch of rudimentary stuff when I got the machine including replacing electrolytic already, but trying to narrow down the issue with the duty cycle and jitter.

Any help is greatly appreciated!

 

[pvision]

I don't have any experience of this machine but, based on my experience with car ABS systems, it might be worth checking the output from the speed detector to make sure it's a noise-free, stable signal

If the system is failing to stabilise properly I would think it's either a mechanical problem, amplifier / servo problem or a bad  sensor

If the motor doesn't have a big flywheel, how does it damp quick speed variations - what I imagine you are calling jitter? There must be a mechanism to do this

Does the jitter occur in play only? Are there any mechanical problems with the feed & take-up spools?

Just read a bit of the PDF - the speed sensor is optical. Is it clean?

Nick Froome

 

[bjoneson]

Thank you for the reply. The tach signal (provided by a marked rotating disc with optical sensor in the motor housing), seems good. The reference signal appears good as well.

The motor does "lock" to speed, but has a very slight jitter. It's present even without tape loaded in the machine, when I close the tape break arm to run the capstan.

It manifest itself as high frequency flutter (introducing periodic phasing) in recordings. The frequency of the flutter increases with tape speed.

In addition to the mechanical jitter, checking the output of the phase comparator in the circuit with a scope (TP3 in the schematic), shows jitter that visually aligns to the mechanical jitter I can hear from the capstan. The duty cycle of the phase comparator output, also is showing about 75/25 whereas the manual calls for 50/50 (symmetrical).

I'm at work currently, but plan to take a large series of measurements around the circuit with a scope and will post some videos this evening.

I don't expect anyone will have a "magic fix", but looking for some extra eyes and insights as I try to analyze this thing.

Thanks again!

-Bob

 

[JohnRoberts]

Yup, servo 101... look for a clean tachometer signal... If that looks good, follow the tach signal through entire servo loop...  Schematic even has typical waveforms.

I can't recognize some of those logic PN or find more info in the BOM

For global interference that could affect motor speed, look for clean regulated PS rails, maybe a tired cap somewhere.

JR

 

[bjoneson]

Agreed. I'll do a bunch of measurements this evening and see if any anomalies pop up.

I've been dealing with some of the same issues, unable to identify all of the ICs, and find data sheets.

Sort of like trying to repair an alien space ship here. But the good news is, the basic servo controlled capstan concept us pretty ubiquitous (was used heavily in VCRs I understand), so I've got a reasonable expectation of what the circuit is "supposed" to do. Diagnosing why it's not is another story.

Thanks again!

 

[pvision]

There has to be a time constant somewhere to damp the jitter. Would a bad cap be a likely culprit?

I assume the duty cycle should be 50/50 at nominal tape speed. If the mechanical side is OK, and the jitter & duty cycle error are there when the capstan is run unloaded, logically there has to be an electronic reason for the duty cycle being off

Nick Froome

 

[Walrus]

I had a similar problem once on the opto tach of a Neumann VMS80 lathe turntable. Although I was getting a tach pulse Ok, it was lower level than it should be. Turned out the opto LED was low emission hence weak signal.
Maybe your "photon coupler" is dying and the reason the duty cycle is off is because the  opamp hasnt got enough signal to reliably switch?

 

[bjoneson]

I had a similar problem once on the opto tach of a Neumann VMS80 lathe turntable. Although I was getting a tach pulse Ok, it was lower level than it should be. Turned out the opto LED was low emission hence weak signal.
Maybe your "photon coupler" is dying and the reason the duty cycle is off is because the  opamp hasnt got enough signal to reliably switch?

That makes sense. Hoping to get some measurements in the next few hours. Appreciate the thoughts / ideas!

 

[radardoug]

Optos definitely a possibility. R24 will control gain, may affect duty cycle too. Check out the Ampex 440 servo diagrams, also used in the MM1100/1200. They have two pots, one for gain, one for duty cycle. Its quite important to get them set right. Too much gain will cause jitter.

 

[bjoneson]

Thanks for the leads.  Unfortunately,  in recklessly probing around, I shorted the supply in the servo drive assembly, and blew the bridge rectifier. Got a new one enroute from digi key, and some spare fuses, should be here friday.

I did notice R24 seemed to affect the jitter a bit, though did not change the duty cycle. Will try to do some research against the Otari / Ampex implementations while I'm waiting for the parts to arrive.

Thanks again guys for all of the leads / insights!