"I had full adjustment of the bias on both channels before the output stage conversion."
And now you need two more diode drops worth of bias, right?
And now you need two more diode drops worth of bias, right?
You now have a triple EF, or six drops.
Do you want to run AB, or AB+B (like the original quasi)?
Do you want to run AB, or AB+B (like the original quasi)?
You need to convert the PL-20 bias circuit to the PL-36 bias circuit and you should be good. When I convert Series II quasi-comps to full comp, I do not change the bias circuit, I just readjust it.
That's what I ended up doing. I tried some additional diode drops and was having issues. I finally decided that it would be easier to convert to the later resistor divider. I converted the right channel and it was biasing up. I will play with it some more later tonight.
Okay some updates.
After I started messing with the right channel bias circuit I am now having issues with the right channel latching up (best way I can describe it) and loading the rails down to around 10VDC when on the dim bulb tester (72W bulb). It happens with the added bias diodes I tried and with the later PL36 bias circuit and even with the original bias parts back in-circuit. The left channel is operating fine if I pull the right channels fuses. The main caps charge up fine and the left will pass a signal. I haven't touched the left channel bias circuit yet.
Once the right channel latches up, the only way I've found to get it to stop is to power up the left channel with the right channel fuses removed. If I pass a signal on the left then some times I can then put the fuses back into the right channel and it will then power up fine and pass a signal, Some times not it just loads the rails down again. It doesn't appear to be oscillation as there is no signal at all on the scope, but there is about +5.5vdc of offset on the r-chan speaker outs when it happens. When the channels come up fine there is minimal DC offset on the channels, like 10 - 20mv.
Anyhow, in my haste I pulled a bonehead maneuver this afternoon and blew the rail fuses on the right channel. The channel is still fine as I've passed a signal on it a couple of times since I blew the fuses using the L-ch fuses. I want to be able to power up both channels simultaneously though and I do not have any spare fuses in that oddball size. I went by Fry's and Radio Shack and of course they didn't carry them so now I've got to place an order with Mouser. Anyway, I'm done with it for a couple of days until the fuses arrive, I needed a break from it anyway. 😕
After I started messing with the right channel bias circuit I am now having issues with the right channel latching up (best way I can describe it) and loading the rails down to around 10VDC when on the dim bulb tester (72W bulb). It happens with the added bias diodes I tried and with the later PL36 bias circuit and even with the original bias parts back in-circuit. The left channel is operating fine if I pull the right channels fuses. The main caps charge up fine and the left will pass a signal. I haven't touched the left channel bias circuit yet.
Once the right channel latches up, the only way I've found to get it to stop is to power up the left channel with the right channel fuses removed. If I pass a signal on the left then some times I can then put the fuses back into the right channel and it will then power up fine and pass a signal, Some times not it just loads the rails down again. It doesn't appear to be oscillation as there is no signal at all on the scope, but there is about +5.5vdc of offset on the r-chan speaker outs when it happens. When the channels come up fine there is minimal DC offset on the channels, like 10 - 20mv.
Anyhow, in my haste I pulled a bonehead maneuver this afternoon and blew the rail fuses on the right channel. The channel is still fine as I've passed a signal on it a couple of times since I blew the fuses using the L-ch fuses. I want to be able to power up both channels simultaneously though and I do not have any spare fuses in that oddball size. I went by Fry's and Radio Shack and of course they didn't carry them so now I've got to place an order with Mouser. Anyway, I'm done with it for a couple of days until the fuses arrive, I needed a break from it anyway. 😕
It seems like this thread outlines what is going on with the r-channel issue. However I have added the anti-parallel diodes across C6 to supposedly eliminate this issue. I do see +.6 V (the diode drop) across C6 when this happens though.
After rereading that thread at the link I just posted I had an idea. The R-ch was latching up when I last operated it. What I decided is to bring up the L-ch (no signal) and allow the main caps to fully charge. Then I powered the amp off and completely discharged the pos cap leaving about - 75 vdc on the neg cap. I installed the fuses into the R-ch and it powered up no problem. I guess I'm not ready to take a break from it....
To ease the pain of buying expensive outputs, only install the T03 driver and one output per channel until you are certain the amp is "debugged".
In addition, when troubleshooting the main PCB remove the rail fuses and load and only have your scope attached. This way you can power up the PCB and ensure you have a clean sine wave and the main PCB is working correctly.
The PCB will power up with no rail fuses.
In addition, when troubleshooting the main PCB remove the rail fuses and load and only have your scope attached. This way you can power up the PCB and ensure you have a clean sine wave and the main PCB is working correctly.
The PCB will power up with no rail fuses.
Nah... try "Freedom of Choice", or better yet, "Gates of Steel". When the amp is finally working with speakers, "Mr. B's Ballroom".
"Whip It" may be good, but the rest of that album is _d*mn_ good. Brings me back to my Phase Linear days.
I was starting to feel that way for sure yesterday. I started getting my second wind today!
I finally figured out what was going on with the r-ch latchup and I have a tip for all those that may do upgrade work on their PL amps. Don't leave your dummy load attached and try to power the amp with the DBT. You'll end up whacking your head on the bench repeatedly for hours on end trying to figure out why the amp will not work properly. If for some reason you want to use your dummy load with the DBT then I suggest you pull the rail fuses and power on the amp allowing the main caps to charge up. Turn the amp off and then discharge the positive bank capacitor, leave the neg bank cap charged up. Now reinstall the fuses for the channel you wish to test and then power the amplifier back on. It should come up without latching up.
So that issue is fixed..mmmmm well maybe not fixed but figured out anyway!
So I get both channels working fine with no latch up. I completed the PL36 bias circuit mod for both channels and verify that the bias will now adjust with the DBT in place. I found that a value of 200 ohms for R19 on the PL20 board will put the 1k ohm bias pot just about perfectly center of its range.
I checked the outputs of each channel with the DBT in place and a bias setting of about 5mv drop Vre (.33r). The sine waves were now clean, no distortion, no oscillation all the way out to 100khz. I took the DBT out of the circuit and tested each channel individually. The right channel looks great into the dummy load. I then check the L channel and its oscillating without the DBT. 🙁
I pulled the rail fuses and checked the signal from the control pcb, clean, no problems. I then pulled all of the outputs save one (luckily)... yeah this isn't gonna end well....
So I go back to start checking and started working my way through the board and dropped my bare tip probe, right on the **** (insert whatever word you like there) control board...... of course it landed perfectly and a rail fuse blows! Checked the rail and I have a short, the 21196 output is blown! Luckily the 21196 driver is still fine. But I also lost the L channel 2n5415, its shorted. That's what I was going to check when the mishap occurred. So I installed a new 5415 and tested the control board again, clean sine wave. I install the L-ch drivers with no outputs and put the fuses back in and I have oscillation. Nasty oscillation too, its like a half volt of fuzz rising on the entire sine wave. I started moving the board changing wiring positions etc and suddenly the oscillation completely goes away. Move it slightly and it comes back. First I thought it was a bad wiring connection, but if I put my finger on the ground trace on the board that also clears it up, completely. The right channel is completely unaffected by any of this.
Time to call it a night!
Anyhow, I am slowly but surely getting it isolated now. If it weren't for the burnt components I'd be feeling pretty good right now. That's the first active component I've killed in probably 5 years now. Three complete DIY amps with no mishaps. 🙁
So that issue is fixed..mmmmm well maybe not fixed but figured out anyway!
So I get both channels working fine with no latch up. I completed the PL36 bias circuit mod for both channels and verify that the bias will now adjust with the DBT in place. I found that a value of 200 ohms for R19 on the PL20 board will put the 1k ohm bias pot just about perfectly center of its range.
I checked the outputs of each channel with the DBT in place and a bias setting of about 5mv drop Vre (.33r). The sine waves were now clean, no distortion, no oscillation all the way out to 100khz. I took the DBT out of the circuit and tested each channel individually. The right channel looks great into the dummy load. I then check the L channel and its oscillating without the DBT. 🙁
I pulled the rail fuses and checked the signal from the control pcb, clean, no problems. I then pulled all of the outputs save one (luckily)... yeah this isn't gonna end well....
So I go back to start checking and started working my way through the board and dropped my bare tip probe, right on the **** (insert whatever word you like there) control board...... of course it landed perfectly and a rail fuse blows! Checked the rail and I have a short, the 21196 output is blown! Luckily the 21196 driver is still fine. But I also lost the L channel 2n5415, its shorted. That's what I was going to check when the mishap occurred. So I installed a new 5415 and tested the control board again, clean sine wave. I install the L-ch drivers with no outputs and put the fuses back in and I have oscillation. Nasty oscillation too, its like a half volt of fuzz rising on the entire sine wave. I started moving the board changing wiring positions etc and suddenly the oscillation completely goes away. Move it slightly and it comes back. First I thought it was a bad wiring connection, but if I put my finger on the ground trace on the board that also clears it up, completely. The right channel is completely unaffected by any of this.
Time to call it a night!
Anyhow, I am slowly but surely getting it isolated now. If it weren't for the burnt components I'd be feeling pretty good right now. That's the first active component I've killed in probably 5 years now. Three complete DIY amps with no mishaps. 🙁
What have you tried so far to resolve the oscillation issue?
Is the conversion to full comp complete?
Is the conversion to full comp complete?
Yes, the FC conversion is complete. The right channel looks great, no sign of oscillation. Remember that was the channel that oscillated before the conversion.
About the only thing I've done so far to cure the oscillation is to try and isolate where in the circuit it starts to show up. It appears on the collector of Q5. Prior to Q5 the signal looks clean. Like i was saying in my previous post just moving the board at one point last night resulted in a light switch on/off of the oscillation. I plan to try a few things with the ground to see what effect that has if any. I also plan to remove each wire connection and resolder at the board for the Lch to make sure there is no cold solder joints causing me grief.
From my experience "stock" PLs don't like Silpad transistor insulators, they cause oscillations. It was in the service manual to remove them if installed. Did you use Silpads?
Craig
Craig
Yeah I bought the new case, to-3 outputs and silpads from White Oak. That would be easy enough to test. There is no oscillation present with the control pcb only. Add the drivers and the left ch oscillates. I can swap the silpads on the drivers with mica and grease and see what it does.
I don't know if the WO pcbs have the stability problem of the stock layout or not. Does WO normally use Silpads? If so it may not be an issue. I've tried Silpads on several vintage amps without much luck. The last was an old Quatre amp, it didn't care for them at all. I just use mica and compound now, it's more time consuming but I never have any problems with them. I always torque the mounting screws to 6 in/lbs, ON Semi's spec. for TO-3 and TO-220.
Craig
Craig
Sil-pads have a nasty habit of cutting through - sometimes after thermal expansion occurs. Amp running along just fine and BAM!!! With an EF OPS, it usually won't hurt the transistor, but you have to de-burr it before you use it again.
A TO-3 transistor case, insulator, and grounded heat sink form a small value VERY HIGH Q capacitor. One trick I've used to de-Q it is a 0.01 uF Z5U ceramic connected with spade lugs directly from one TO-3 mounting screw and the other to the heat sink. Swamp the parasitic cap out with a larger value, low Q (but low inductance) cap and the oscillation often goes away.
A TO-3 transistor case, insulator, and grounded heat sink form a small value VERY HIGH Q capacitor. One trick I've used to de-Q it is a 0.01 uF Z5U ceramic connected with spade lugs directly from one TO-3 mounting screw and the other to the heat sink. Swamp the parasitic cap out with a larger value, low Q (but low inductance) cap and the oscillation often goes away.