Thanks Mooly, I'll try it in the morning. Why though is one amp working with no problems ? Both channels are identical in every respect.
Fair doos with a cheapish op-amp, I'm just working with what I've got here without spending any more money. I've got OPA627AP, OPA604AP and somewhere I've got some 741s. I've got some TLxxx but most of them are duals.
It's just showing how critical component choice can be sometimes ?
At least it works so that it can be measured without smoking.
In production the Quad 606 would not have had this problem ??
It would have made mass production very expensive to fault find each amplifier.
It would have made mass production very expensive to fault find each amplifier.
Maybe, maybe not... its a clone, not an original and so something may have got lost in translation.
The OPA604 should be OK but it can draw up to 7ma and so the 10k needs to be nearer 3k9 to cover that. Everything is as it is for a reason, 10k and the TLC type opamp. Any component change needs following through to make sure it will be OK.
Well I've tried two suggestions each had no effect on the oscillations:-
1. 10pF between TR1 (b) and TR3 (e).
2. 1nF between TR12 (b) and TR12 (c).
Now I'm lost.
1. 10pF between TR1 (b) and TR3 (e).
2. 1nF between TR12 (b) and TR12 (c).
Now I'm lost.
I've just tried a 100nF directly across IC1 but still no change.
I have noticed that there is slight oscillation at TR1 (b). The I/P is short cct.
I would have thought that C1 and C2 would kill any HF oscillation.
I have noticed that there is slight oscillation at TR1 (b). The I/P is short cct.
I would have thought that C1 and C2 would kill any HF oscillation.
It won't be the opamp. That can't do anything with it configured as an integrator. If the boards are truly identical then maybe its worth swapping the boards and seeing if the problem transfers to the good board, and if so indicating a wiring issue.
These things are difficult to diagnose at a distance.
These things are difficult to diagnose at a distance.
I've already done that, it is a board issue.
The boards are not in any chassis yet so I can swap them onto the PSU at will. The second transformer has only just arrived.
The boards are not in any chassis yet so I can swap them onto the PSU at will. The second transformer has only just arrived.
Last edited:
I'm tempted to fit a huge cap between TR1 (b) and GND to see if I can see where the oscillation is starting.
I've just ordered a couple of TLC271 to see what changes.
I've just ordered a couple of TLC271 to see what changes.
OK. Well don't overlook or assume anything. Check visually that each part and device type number is identical between boards. Its worth doing a DC check of all the operating points, in particular making sure that current flows are similar (volt drops across resistors) around the front end. Are the coils a bit of an unknown ? Are they fitted identically with regard to orientation, although not polarity sensitive they are not symmetrical with regard to how they are wound. So fit them so the windings are the same on each channel.
The opamp just provides a DC bias voltage the value of which changes slightly to keep the offset at zero. You should see zero oscillation on its output.
Now that I've got the second 800VA transformer I can rig them up as a stereo pair which might make comparative testing a bit easier.
If there is doubt over the coils then swap them all with the other board keeping the same orientation as now.
Ignoring the DC output offset can the amp be run without the IC ?
I am seeing oscillation at its output but that might be being reflected back through R5.
That was my rational behind fitting 100nF bypass to IC1.
Last edited:
I'll eat my hat if it's one of the coils being 180 degrees out, I'll try it and see what happens.
Yes you can run the amp with fixed bias to test. Measure the DC voltage at the opamp output and then rig up a pot to give a similar voltage at the wiper and feed that into R5 (having removed the opamp).
(Don't connect speakers while you initially set it)
(Don't connect speakers while you initially set it)