"Connecting one side at a time I can dial everything in. Connect em both and it's a crap shoot."
Ah, that comment brings back memories..... There is a line in the A75 article that says something to the effect of "connect one side, adjust it, completely disconnect it and hook up the other side and adjust". I spent over a month trying to adjust the pots on my A75 with both sides connected. When I would get close to turning up the bias, it would start to oscillate as I was turning the pots. Finally, I did the adjustments, one side at a time. Then I connected up both and turned it on -- and voila, as long as I had the inputs connected to something, it worked fine. Turn it on with the inputs not connected and it would oscillate. Perhaps this is part of the problem you are seeing?
JJ
Ah, that comment brings back memories..... There is a line in the A75 article that says something to the effect of "connect one side, adjust it, completely disconnect it and hook up the other side and adjust". I spent over a month trying to adjust the pots on my A75 with both sides connected. When I would get close to turning up the bias, it would start to oscillate as I was turning the pots. Finally, I did the adjustments, one side at a time. Then I connected up both and turned it on -- and voila, as long as I had the inputs connected to something, it worked fine. Turn it on with the inputs not connected and it would oscillate. Perhaps this is part of the problem you are seeing?
JJ
Help again. I've got the DC dialed in both sides, no oscillations and everything seems to be a go. 0-10 mv dc offset on both channels and about 6.5v across +/- DRV. When I hook up the scope and the signal generator with a 1khz tone from my computer the DC on one side goes +3V and on the other side -3V. not what I was expecting. I'm running unbalanced with the -In on the pcd tied to the star grd along the with the grd on the input jack. Nothing new or different that they way it's been the last 15 years. I've been over the connections again and again and have no idea what's up now. I'd post some pics if you think that might help. Not sure how but I'll figure it out.
I suppose that's possible. I'll give it a check and call it a day. 5am rolls around pretty early and I'm no young pup any more. Thanks again Bob. I think I've learned more about this amp now than when I built it. If it blows again I'm not sure I've got the patience to do another repair job. Stay tuned.....
Dan
Dan
I checked for DC on my computer and that isn't the problem. It was a long shot since with the computer connected before feeding a test tone I didn't measure any VDC. When I switched up the grounding I moved the Reg PSU grounds to the star grd which may be the cause. Gonna move them back to the main cap plate grd tonight and see if that makes a difference.
update for today. Frontend checked out, +/- 10mv offset on both sides and a good looking trace on the scope. Everything back together...output and unreg PS connected. But now I have +3vdc on one side and -3vdc on the other on the outputs. Not sure what the problem is now. The output is clean. When I measured the +/- DRV to ground before hooking up the output stage the + measured about 5V and the -DRV at 0. Didn't seem right but the DC dialed in fine. R7/8 at 1V/.8V respectively.
Any thoughts anyone?
Any thoughts anyone?
The heatsinks getting nice and toastly, but the fans keep the temps down. If I did my math right 300mv across the ouput resistors should give me class a up to about 65W with an 8ohm load.
8 ohm bias fot 100% Class A
Rail VDC35
Load Impedance 8
=4.375 A
Output Devices 6
Bias/device = (ma)300
=1.8 A
Watt Bias rms 63.0
I'm using 6 each of the IRFP140/9140's per channel. These are rated at about 180W each if I recall the specs.
Can't wait to plug back in and listen to some music.
8 ohm bias fot 100% Class A
Rail VDC35
Load Impedance 8
=4.375 A
Output Devices 6
Bias/device = (ma)300
=1.8 A
Watt Bias rms 63.0
I'm using 6 each of the IRFP140/9140's per channel. These are rated at about 180W each if I recall the specs.
Can't wait to plug back in and listen to some music.
Last edited:
I've been trying to clean up some hum and found this comment from NP on an Aleph thread back in 2007:
"That will help, but it is equally important to arrange your input
and other grounding so that they experience exactly the same
magnetic field, and towards that end, you want to avoid
surface area (the interior of any conductor loop) in your separate
ground wires. Most often the practical way is to wire your
RCA grounds together at the connectors and use that spot for
a signal star ground, and running a wire from there to the
main supply ground."
Gonna give this a try but was wondering if this star point should also be used for the output grounds? Intuitively it doesn't seem like I'd want to run the output ground to the same star point as the input grounds. Right now I have all the grounds connected at the main PSU cap plate. Worked the last 12 years this way but now it's producing a slight hum in both channels...louder in the left.
Any thoughts anyone?
"That will help, but it is equally important to arrange your input
and other grounding so that they experience exactly the same
magnetic field, and towards that end, you want to avoid
surface area (the interior of any conductor loop) in your separate
ground wires. Most often the practical way is to wire your
RCA grounds together at the connectors and use that spot for
a signal star ground, and running a wire from there to the
main supply ground."
Gonna give this a try but was wondering if this star point should also be used for the output grounds? Intuitively it doesn't seem like I'd want to run the output ground to the same star point as the input grounds. Right now I have all the grounds connected at the main PSU cap plate. Worked the last 12 years this way but now it's producing a slight hum in both channels...louder in the left.
Any thoughts anyone?
if you're interested I posted a couple pics on this thread...
http://www.diyaudio.com/forums/pass-labs/166784-pictures-your-diy-pass-amplifier-59.html#post2511646
http://www.diyaudio.com/forums/pass-labs/166784-pictures-your-diy-pass-amplifier-59.html#post2511646
Update. Last night I rearranged the input ground tieing the L/R grounds together at the input and running that ground to the main cap ground....not an ideal location I know. It's not perfect but much better than before. I only hear a slight hum with my ear right up to the speaker. Because it involved more disassembly and reassembly I didn't run the input ground connection to the chassis ground but suspect that, having now read Doug Self's article on grounding, that this would eliminate the hum completely. Was also thinking that, since I'm not using the PSU fan I could tie in to the ground connection there by jumping that trace to the ground trace. Conceptually, connecting directly to the chassis ground seems to be a better choice since it puts the earth ground 5R resistor and diodes between the input signal ground and the PSU ground. Not sure this is what Nelson had in mind.
I think it's time to replace the output caps. The other day I was fine tuning the offset with the amp in the system. My ear was next to the speaker and I noticed that *** buzz again. It would be relatively quiet one moment and the next would get louder. So I ran a little test this morning. Disconnected the AC cords on the rest of the equipment in the system, which should rule out any ground loop related problems. The buzz didn't go away so back on the bench and took a look at the DC across the main caps. It looks like a 60hz sine wave. I measure 38v across all the main caps but as soon as I power down the voltage drops within minutes to 2.5v on all 4 caps. Either there's a short to case on each bank of the output Qs (highly unlikely) or it's time to replace the main caps. Unless...the bridge rectifiers have gone bad? The transformer seems pretty loud as well.
If anyone is still following this thread I would appreciate any comments / suggestions.
Thanks again for all the help and advice.
If anyone is still following this thread I would appreciate any comments / suggestions.
Thanks again for all the help and advice.
With a class A amp I would expect the rails to collapse pretty quickly. Have you checked the ripple on your rails? Just measure AC volts referenced to ground.
I'd start the PSU buildup troubleshooting procedure again. Start with the transformer secondaries disconnected. Measure the AC current in the primaries. With a typical power transformers, expect a few VA but anything significant indicates a transformer problem. Connect the bridges one at a time but don't connect the caps yet. How is the primary current? Add the caps one channel at a time. You still should see a small primary current.
Now connect the amps one channel at a time. The primary current should increase to something that corresponds roughly to the draw of each channel.
I'd start the PSU buildup troubleshooting procedure again. Start with the transformer secondaries disconnected. Measure the AC current in the primaries. With a typical power transformers, expect a few VA but anything significant indicates a transformer problem. Connect the bridges one at a time but don't connect the caps yet. How is the primary current? Add the caps one channel at a time. You still should see a small primary current.
Now connect the amps one channel at a time. The primary current should increase to something that corresponds roughly to the draw of each channel.
I went back as you suggested ran tests on the power supply measuring the voltage drop across TH1. Since I have the bridge connections soldered except for the connections for the reg supply I didn't disconnect the bridge. Here are the measurements:
1. just the bridges - 245mv AC
2. with the reg supply and front end connected - .78v
3. with the left output bank supply - 1.10v
4. the right output bank - 1.12v
5. both banks (everything connected) - 1.00v (would have expected this to be higher with both sides drawing but could be because TH1 is heating up now)
and the final measurement...ac ripple on the main caps:
both pos rail caps measured .006/.007 vac
the neg rail caps measured .240 and .200vac. that seems awful high, especially in comparison to the pos rails. is that common for there to be such a big difference?
I should probably disconnect the caps and take a look at the voltages across just the bridges. Put the scope on the bridge and see what it looks like.
1. just the bridges - 245mv AC
2. with the reg supply and front end connected - .78v
3. with the left output bank supply - 1.10v
4. the right output bank - 1.12v
5. both banks (everything connected) - 1.00v (would have expected this to be higher with both sides drawing but could be because TH1 is heating up now)
and the final measurement...ac ripple on the main caps:
both pos rail caps measured .006/.007 vac
the neg rail caps measured .240 and .200vac. that seems awful high, especially in comparison to the pos rails. is that common for there to be such a big difference?
I should probably disconnect the caps and take a look at the voltages across just the bridges. Put the scope on the bridge and see what it looks like.
Update. I went back and disconnected the output from the main caps and remeasured the AC ripple, this time with much better results.
Pos rail - .006/.000 vac
Neg rail - .001/.007 vac
I guess this means the ripple I was measuring before with the output connected is oscillation on the output Qs or frontend. When I measured the AC on the frontend +/- driver outputs ...guess what...
Left channel - .017 / .015 vac
right channel - .025 / .022 vac
not much but I suspect enough to modulate the output drivers. I've already increased the value of the frontend gate Rs to 200R in lieu of using C9/10. I had them in before but figured while I was doing a tune-up I'd up the gate Rs and do away with C 9/10. Guess they're going back in. When I tested the front end it tested ok but I guess this small AC is all it takes to oscillate the whole amp once the output is connected....at least that's much assumption right now. At least I now know what to look for. If I get the frontend squared away and still have a problem it'll be time to look at the output.
Thanks again Bob for pointing me in the right direction. You have been a huge help. This site is an incredidle resource for the DIYer. I for one would be lost without your help and the help of all the other experts here.
If I keep at this long enough I might even get to the point where I understand what I'm doing.
Stay tuned...
Dan
Pos rail - .006/.000 vac
Neg rail - .001/.007 vac
I guess this means the ripple I was measuring before with the output connected is oscillation on the output Qs or frontend. When I measured the AC on the frontend +/- driver outputs ...guess what...
Left channel - .017 / .015 vac
right channel - .025 / .022 vac
not much but I suspect enough to modulate the output drivers. I've already increased the value of the frontend gate Rs to 200R in lieu of using C9/10. I had them in before but figured while I was doing a tune-up I'd up the gate Rs and do away with C 9/10. Guess they're going back in. When I tested the front end it tested ok but I guess this small AC is all it takes to oscillate the whole amp once the output is connected....at least that's much assumption right now. At least I now know what to look for. If I get the frontend squared away and still have a problem it'll be time to look at the output.
Thanks again Bob for pointing me in the right direction. You have been a huge help. This site is an incredidle resource for the DIYer. I for one would be lost without your help and the help of all the other experts here.
If I keep at this long enough I might even get to the point where I understand what I'm doing.
Stay tuned...
Dan
Ok. I went back and added c9/10 on the frontend and retested.
+drv - .010/.014 vac
-drv - .011/.015 vac
Output - .013 / .013 vac
Connected the left channel output and measured ac ripple at the PS Caps.
Pos = .06 vac, Neg=.185vac ( a little better than before, was at .240)
The issue is only on the P devices. Th Ns don't have the same issue and as I turn the bias down the ripple does right down with it. Looks like I'm going to have to take the output apart and do some work.
Should I increase the gate resisitors? Remeasure the VGS on each device again and check for matching?
+drv - .010/.014 vac
-drv - .011/.015 vac
Output - .013 / .013 vac
Connected the left channel output and measured ac ripple at the PS Caps.
Pos = .06 vac, Neg=.185vac ( a little better than before, was at .240)
The issue is only on the P devices. Th Ns don't have the same issue and as I turn the bias down the ripple does right down with it. Looks like I'm going to have to take the output apart and do some work.
Should I increase the gate resisitors? Remeasure the VGS on each device again and check for matching?
Hello Dan
Here is some data from my A75, maybe this can help:
PS cap ripple with outputs connected (bias current 1,4 A, PS Caps 66.000µF) is +0.060/-0.062 VAC.
15 minutes after mains power off, voltage on the main caps is still +/- 14 VDC.
UGS of FETs used:
IRF630 (N): 3,43-3,44V (L) 3,50-3,59V (R)
IRF9630 (P): 4,21-4,22V (L+R)
IRF610 (N): 3,97V (L+R)
IRF9510 (P): 3,96V (L) 3,98V (R)
IRFD110 (N): 3,49V (L) 3,47V (R)
IRFD9110 (P): 3,68V (L) 3,69V (R)
Less ripple when turning down the bias is normal since less current is drawn from the PS caps. The larger Neg. ripple voltage could be effected by higher bias current, defective PS caps, bridge with one diode not working or defective output transistors on Pos. side.
Regarding the gate resistors, I don´t think that increasing them will help.
If i dare make a suggestion, it´s better to check for ripple voltages and oscillation using a scope to be sure what´s going on.
To me this seems to be more of a component problem, since your amp worked fine for many years. I would also suggest to check for defective solder joints and bad electrolytic capacitors. Also there could be leakage current paths in the power amp, PSU etc. due to damaged insulation of transistors mounted to the same chassis/heatsink.
If you need comparison measurements, i could do some on my A75.
Best wishes,
Norb
Here is some data from my A75, maybe this can help:
PS cap ripple with outputs connected (bias current 1,4 A, PS Caps 66.000µF) is +0.060/-0.062 VAC.
15 minutes after mains power off, voltage on the main caps is still +/- 14 VDC.
UGS of FETs used:
IRF630 (N): 3,43-3,44V (L) 3,50-3,59V (R)
IRF9630 (P): 4,21-4,22V (L+R)
IRF610 (N): 3,97V (L+R)
IRF9510 (P): 3,96V (L) 3,98V (R)
IRFD110 (N): 3,49V (L) 3,47V (R)
IRFD9110 (P): 3,68V (L) 3,69V (R)
Less ripple when turning down the bias is normal since less current is drawn from the PS caps. The larger Neg. ripple voltage could be effected by higher bias current, defective PS caps, bridge with one diode not working or defective output transistors on Pos. side.
Regarding the gate resistors, I don´t think that increasing them will help.
If i dare make a suggestion, it´s better to check for ripple voltages and oscillation using a scope to be sure what´s going on.
To me this seems to be more of a component problem, since your amp worked fine for many years. I would also suggest to check for defective solder joints and bad electrolytic capacitors. Also there could be leakage current paths in the power amp, PSU etc. due to damaged insulation of transistors mounted to the same chassis/heatsink.
If you need comparison measurements, i could do some on my A75.
Best wishes,
Norb
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.