More tape machine woes: crosstalk in sync only

[juanito2008]

I don't know what that last bit is about, if it's a dig at my ability to align a machine or just an unfortunate way of expression.

I don't question your ability to align the machine. I'm only wondering what can cause the problem.

 

[EmilFrid]

@juanito2008 sorry if I came off a bit dry, I've had a crap day and interpret everything in a negative light.
But yes, I'm wondering the same. I did adjust the bias traps since I swapped the head (as described) and thought that maybe I cracked the connector board when pulling and putting back channel cards, but I haven't managed to track down any cracks on the board. I even went through it with a magnifying glass looking like I was a hundred years old trying to find my own obituary in the morning paper. But no crack anywhere. I will keep looking. I will examine everything with the scope first and foremost. As the only thing that has happened is the head swap and bias trap adjustments, the problem is in a broad sense localized. Just need to find the final coordinates.

 

[RoadrunnerOZ]

I agree with you. But how he brakes a ground plane or whatever if he only swapped two connectors ?
He should have done something else in the machine. Recapping, wiring...

Old wiring can go brittle, crack, erode from oxidation and cause all sorts of problems. PCB connections can have dry joints, plugs when pulled and reinserted can cause print to crack - sometimes the cracks are invisible. I fix vintage gear every day and I find sometimes there’s board erosion under the solder mask (the green coating on circuit boards) that’s almost invisible to the eye - the copper is eaten away by either salt or moisture creep corrosion underneath the mask - the only clue is a blackening of the print traces. Sometimes you pull a cable when accessing others and don’t even notice but it causes a fault, you can break a connection where there’s corrosion by simply pulling a plug and slightly moving the connectors. You just need to use logic to look at the actual fault and think of what could be shorted/open circuit/blown and where it could be. Since this is common to all channels and affects all adjacent channels it’s likely a common cable or print trace that has failed - question is where.

 

[juanito2008]

At 10 khz the wave is even louder on the adjacent tracks than on the one in record mode, but sensitivity drops A LOT to an ok crosstalk at 1 khz.

When you generate your frequency, do you send the oscillator to one input at a time or do you send to all inputs at the same time ?
Like you will do on a console "Oscillator to group"

If you send the oscillator in one input compare to all inputs, do you have a level change in the crosstalk level ?

 

[EmilFrid]

I patch the oscillator to one channel only.
Sounds like a good idea. I will check this

 

[map2112]

I patch the oscillator to one channel only.
Sounds like a good idea. I will check this

When you first did the head swap and did the bias setup, do you recall if those adjustments were very minor, or were any substantial adjustments?

Later when you tried to revert back to the original head did you also readjust the bias for that head?

I am not familiar with the bias adjustment for the machine, only trying to help understand “what else” may have changed along the way, and how to “walk back” the changes to the beginning.

BTW I particularly enjoyed your illustrative comment “I even went through it with a magnifying glass looking like I was a hundred years old trying to find my own obituary in the morning paper.”

 

[EmilFrid]

When you first did the head swap and did the bias setup, do you recall if those adjustments were very minor, or were any substantial adjustments?

Hi! The changes weren't substantial, and when I changed back to the old head I didn't readjust the bias. Perhaps I should've done that now that I think about it, although the changes weren't huge.

I am not familiar with the bias adjustment for the machine, only trying to help understand “what else” may have changed along the way, and how to “walk back” the changes to the beginning.

I'm grateful for that. I feel like I get "stuck" in my own head and become a bit blind from time to time while working on this, so a few fresh sets of eyes are invaluable to me.

BTW I particularly enjoyed your illustrative comment “I even went through it with a magnifying glass looking like I was a hundred years old trying to find my own obituary in the morning paper.”

Lol. Thanks.

 

[RoadrunnerOZ]

I would also look at the crosstalk on the scope and see what it comprises - is it only the fed signal or is it overlaid with any other frequency

 

[EmilFrid]

I did look at the crosstalk on the scope and it shows a pure sine wave. Very loud and clear at 10 khz, much much weaker but still clean at 1 khz

 

[RoadrunnerOZ]

Worth probing around with the scope looking at any supposed to be grounded metalwork for traces of the injected signal with scope grounded at mains earth. Check also the mains earth integrity of the machine itself back to the wall.

 

[EmilFrid]

I'll follow your advice. This far, I've only done continuity checks with my multimeter, but the scope might tell other stories... We'll see if I have time to work on it tonight. Back at the university tomorrow so I can't unfortunately stay up all night.

 

[Voyager10]

Could you do the crosstalk test with the bias oscillator not running? If the problem still exists then we can basically rule the bias traps (and their adjustment) out of the picture.

The bias oscillator is on the "system control (driver)" PCB:


Lifting one end of R172 would disable the oscillator, if you can do that safely.

 

[EmilFrid]

I'll see if I can do that. I have tested the crosstalk with talk radio instead of sine waves as well though, and no channel shows a constant level as would be the case if the bias traps were maladjusted. Or am I missing something perhaps?

 

[EmilFrid]

Didn't find time to work with the scope today but did a quick ohm test on the GNDs of the repro and rec cards. I haven't done these specific measurements before; I simply measured the GND trace on each board (repro and rec cards independently) and read the resistance between them. What I found was that the resistance between the GNDs increase the further away the cards are. So between card one and two there is basically no resistance, but the resistance builds up to a point were between channel one and channel 16 I see 15 ohms. It's the same with both sets of cards. Maybe a stupid question, but this is odd, right? I've tried wiring them together and it didn't change the crosstalk issue, but perhaps this is a find, still, that points in the right direction? I guess my scope will tell me more during the week.

 

[RoadrunnerOZ]

What exact point are you measuring the ground(s) at

 

[EmilFrid]

From pin B on the repro board to pin B on another, and from pins 6,7,8 (those are tied together) on the rec board to the corresponding pins on another rec board