So I got a fun one. The repair of an Ampeg SVT-3 Pro output stage I can’t wrap my head around. It came in with a rapid clicking output relay. A quick measurement showed the output had +65 volt on it and the relay repeatedly disengaged as protection. I suspected the output MOSFET’s, but they test good (tested on my Peak Electronics DCA75 pro). I also removed three on each side, but the issue remained. I continued to test all surrounding transistors and they all came out fine. Measured diodes, zeners, caps, resistors, all good. Voltages check out as well.
There’s a balance trimpot between the gate rails and when turned almost fully counterclockwise, the output offset goes down to milivolts, but it can’t amplify the positive part of the soundwave anymore. I don’t fully grasp the function of this trimpot or the workings of the TL074 in the feedback path.
I appreciate any fresh perspectives, ideas or similar experiences. Attached is the schematic.
There’s a balance trimpot between the gate rails and when turned almost fully counterclockwise, the output offset goes down to milivolts, but it can’t amplify the positive part of the soundwave anymore. I don’t fully grasp the function of this trimpot or the workings of the TL074 in the feedback path.
I appreciate any fresh perspectives, ideas or similar experiences. Attached is the schematic.
First thing is if the output fets are faulty, you would have a short circuit. From what I read, they are fine as I would expect. No need to even remove them.
Second the IC1 only trims the output, the bias is set with Q2 by the pre set AP1.
The centre voltage is what you need to look at: Q3 pulls positive and Q4 negative. They must be balanced to acheive zero volts DC on the output and that is acheived with the zeners (setting emitter voltage) and R19 setting the current flow through each transistor which should remain equal. This is where IC1 comes in, it trims the offset.
Check the offset voltages, dead easy to work out what to expect and then you will find your issue.
Could be a leaky transistor, faulty zener, faulty emitter resistor, etc etc.
The audio drive from the valve stage is coupled through a pair of 680n capacitors. In my experience, not likely to be the problem but, hey, we learn as we go and I am happy to learn a bit more.
Keep a level head, don't throw parts at it and keep logical. You will get there.
Second the IC1 only trims the output, the bias is set with Q2 by the pre set AP1.
The centre voltage is what you need to look at: Q3 pulls positive and Q4 negative. They must be balanced to acheive zero volts DC on the output and that is acheived with the zeners (setting emitter voltage) and R19 setting the current flow through each transistor which should remain equal. This is where IC1 comes in, it trims the offset.
Check the offset voltages, dead easy to work out what to expect and then you will find your issue.
Could be a leaky transistor, faulty zener, faulty emitter resistor, etc etc.
The audio drive from the valve stage is coupled through a pair of 680n capacitors. In my experience, not likely to be the problem but, hey, we learn as we go and I am happy to learn a bit more.
Keep a level head, don't throw parts at it and keep logical. You will get there.
Thanks for thinking with me! I got some updates and progress. After more diagnosing, I got fed up so I started replacing components section by section. Once I finally replaced Q3 and Q4 (with a MJE340/350 pair since I didn't have a 2n5415/3440 around), plus their zeners D2 and D3, things started looking better. Both halves of the waveform were once again amplified! So I suspect that even though my transistor tester gave the 'OK', either Q3 or Q4 still was defective. I will look further into that.
Although better than before, the waveform was now asymmetrical and showed massive crossover distortion. I found the negative supply voltage for IC1 was -7 instead of -16 volts and my thermal cam showed it heated to about 80 degrees C. Probably damage from the massive amounts of DC on the output. So I replaced IC1 and the waveforms became symmetrical.... but still with crossover distortion. I can get rid of it for 90% by turning AP1 fully up. I suspect that either the MJE340/350 aren't fully compatible or zeners D2 and D3 are crap (they show a drop of 2,2V instead of the quoted 3,3V). I will order proper replacements for these parts and will report back.
Although better than before, the waveform was now asymmetrical and showed massive crossover distortion. I found the negative supply voltage for IC1 was -7 instead of -16 volts and my thermal cam showed it heated to about 80 degrees C. Probably damage from the massive amounts of DC on the output. So I replaced IC1 and the waveforms became symmetrical.... but still with crossover distortion. I can get rid of it for 90% by turning AP1 fully up. I suspect that either the MJE340/350 aren't fully compatible or zeners D2 and D3 are crap (they show a drop of 2,2V instead of the quoted 3,3V). I will order proper replacements for these parts and will report back.
To wrap it up, the crossover issue was indeed caused by the two new zeners not having the zener-voltage they were advertised at (2.2v instead of 3.3v), which prevented to supply enough bias voltage to the MOSFETs. New, proper, zeners installed and everything is good now.