First off I'd really like to thank Nelson PAss for sharing all of his designs with the DIY community we all appreciate his work and efforts.
I just finished building ver 2(non-inverting version) of the "L" and powered things up only to discover that the operating voltages are not even close to what the schamatic calls for. Power supply is a rock solid 59.6 volts. All parts were checked before construction began. Everything is 1% or better on tolerance. Although both channels did pass audio it was somewhat distorted indicating that either the MOSFETs are not matched(should they be in this circuit, as I did not bother to, I only verfied that they were good?) or that bias was too low for these particuluar transistors. I ended up lowering the value of R110/210, 22.1K to 6.8K to increase the bias on the gate of Q101/102 to about 6 volts and low and behold everything cleaned right up. Each channel draws about 40ma now which seems more realistic to me. Things get warm but not excessivly so. The schematic also calls for 30vdc on the drain of Q101/210 and I am getting more like 54 volts. Although all seems well something is strange here. Myself not being an engineer, would anyone have any suggestions here as to what might be happenning? I can reply with any measurements you might need taken.
Thanks!
Mark Gulbrandsen
Salt Lake City, UT
[Edited by Mark A. Gulbrandsen on 09-26-2001 at 02:07 PM]
I just finished building ver 2(non-inverting version) of the "L" and powered things up only to discover that the operating voltages are not even close to what the schamatic calls for. Power supply is a rock solid 59.6 volts. All parts were checked before construction began. Everything is 1% or better on tolerance. Although both channels did pass audio it was somewhat distorted indicating that either the MOSFETs are not matched(should they be in this circuit, as I did not bother to, I only verfied that they were good?) or that bias was too low for these particuluar transistors. I ended up lowering the value of R110/210, 22.1K to 6.8K to increase the bias on the gate of Q101/102 to about 6 volts and low and behold everything cleaned right up. Each channel draws about 40ma now which seems more realistic to me. Things get warm but not excessivly so. The schematic also calls for 30vdc on the drain of Q101/210 and I am getting more like 54 volts. Although all seems well something is strange here. Myself not being an engineer, would anyone have any suggestions here as to what might be happenning? I can reply with any measurements you might need taken.
Thanks!
Mark Gulbrandsen
Salt Lake City, UT
[Edited by Mark A. Gulbrandsen on 09-26-2001 at 02:07 PM]
Hi Gray,
I believe that the resistor I changed only affects the bias of the input device...increasing the gate voltage increases the source voltage, etc. It does not seem to affect the 2nd stage device. That is biased from fixed 10 Ohm(Source) and 1K ohm (gate) resistors. Before I changed R110/210 the circuit would draw only about 12 ma and would clip really early on. I wonder if adding say a 10k trimmer pot across R110 would be as good an alternative to matching devices, although thats not hard to do either. This circuit is a bit different than the other single ended "L" as its biased by just the resistors in the circuit and not an extra current source sort of thing. The first version of the "L" is designed more like a minature Aleph amp whereas this is a two stage job that is resistance biased.
I'll go through and check all my devices to see where I'm at, it sure couldn't hurt at this point to have them matched...everything else is. Give me a day or two on this but if they're not close enough I'll have to order more. Luckily, I did get 12 of each device though.
Thanks!
Mark Gulbrandsen
SLC, UT
I believe that the resistor I changed only affects the bias of the input device...increasing the gate voltage increases the source voltage, etc. It does not seem to affect the 2nd stage device. That is biased from fixed 10 Ohm(Source) and 1K ohm (gate) resistors. Before I changed R110/210 the circuit would draw only about 12 ma and would clip really early on. I wonder if adding say a 10k trimmer pot across R110 would be as good an alternative to matching devices, although thats not hard to do either. This circuit is a bit different than the other single ended "L" as its biased by just the resistors in the circuit and not an extra current source sort of thing. The first version of the "L" is designed more like a minature Aleph amp whereas this is a two stage job that is resistance biased.
I'll go through and check all my devices to see where I'm at, it sure couldn't hurt at this point to have them matched...everything else is. Give me a day or two on this but if they're not close enough I'll have to order more. Luckily, I did get 12 of each device though.
Thanks!
Mark Gulbrandsen
SLC, UT
Mark,
Let me apologize...I was thinking of the differential preamp circuits, not the L. You are quite correct, the devices will not need to be matched. In any event, you would have a hard time matching N- and P-channel devices to one another. I plead lack of sleep.
Okay, let's see if I can reboot my brain and try this again, albeit with bleary eyes.
I'm suspicious of that 54V on Q101's drain.
Check R113 to make sure you got a 1k resistor in there. Likewise, verify that R111 is 10 ohm. I know I'm prone to swapping 1001s, 1000s, and 10R0s when I get tired.
I'm assuming that you are using IRF610s and 9610's, yes?
You might try going back to the original R110 value to see how Q102 reacts, as the two MOSFETs are interdependent. The more current Q101 draws, the more voltage will appear across R113, which will drive Q102 harder.
What are you reading at the gate of Q101 now, and with R110 at 22.1k?
I'll try to get adequate snooze tonight. Perhaps once my brain is at full voltage, I'll be able to think of something.
Grey
Let me apologize...I was thinking of the differential preamp circuits, not the L. You are quite correct, the devices will not need to be matched. In any event, you would have a hard time matching N- and P-channel devices to one another. I plead lack of sleep.
Okay, let's see if I can reboot my brain and try this again, albeit with bleary eyes.
I'm suspicious of that 54V on Q101's drain.
Check R113 to make sure you got a 1k resistor in there. Likewise, verify that R111 is 10 ohm. I know I'm prone to swapping 1001s, 1000s, and 10R0s when I get tired.
I'm assuming that you are using IRF610s and 9610's, yes?
You might try going back to the original R110 value to see how Q102 reacts, as the two MOSFETs are interdependent. The more current Q101 draws, the more voltage will appear across R113, which will drive Q102 harder.
What are you reading at the gate of Q101 now, and with R110 at 22.1k?
I'll try to get adequate snooze tonight. Perhaps once my brain is at full voltage, I'll be able to think of something.
Grey
Hi Grey,
It appears that I found the problem! The input caps were leaky to the tune of about 1.5 meg. These were brand new caps too rated at almost 3 times the circuits maximum voltage. Both output caps are fine though. Its odd as they are all the same type(they were temporary ones though). I changed them out for some old Rel-Caps I had laying around and now things seem normal. I went back to the 22.1 K value for R110 as well and the Voltage readings are as follows.....
Q101 Gate 13.1, Drain 53.9, Source 10.2
Q102 Gate,53.9, Drain 29.5, source 57.9
Also there is a 4 volt drop across the 1K resistor R113 and a .4 volt drop across the 10 ohm, R111. It now appears that things are more normally biased than they were and both channels pretty much read identical.
I think the markings on the Pass Labs schematic are perhaps a bit mis-leading in indicating where these readings actually are. Let me know what you think.
Also on another note what type of caps would one reccommend for this type of circuit. What type of caps are used in the originals? I've always favored film caps but I was thinking about using Black Gate for the coupling caps in this project.
Mark Gulbrandsen
Salt Lake City, UT
[Edited by Mark A. Gulbrandsen on 09-28-2001 at 01:51 PM]
It appears that I found the problem! The input caps were leaky to the tune of about 1.5 meg. These were brand new caps too rated at almost 3 times the circuits maximum voltage. Both output caps are fine though. Its odd as they are all the same type(they were temporary ones though). I changed them out for some old Rel-Caps I had laying around and now things seem normal. I went back to the 22.1 K value for R110 as well and the Voltage readings are as follows.....
Q101 Gate 13.1, Drain 53.9, Source 10.2
Q102 Gate,53.9, Drain 29.5, source 57.9
Also there is a 4 volt drop across the 1K resistor R113 and a .4 volt drop across the 10 ohm, R111. It now appears that things are more normally biased than they were and both channels pretty much read identical.
I think the markings on the Pass Labs schematic are perhaps a bit mis-leading in indicating where these readings actually are. Let me know what you think.
Also on another note what type of caps would one reccommend for this type of circuit. What type of caps are used in the originals? I've always favored film caps but I was thinking about using Black Gate for the coupling caps in this project.
Mark Gulbrandsen
Salt Lake City, UT
[Edited by Mark A. Gulbrandsen on 09-28-2001 at 01:51 PM]
Mark,
Excellent news! Glad you found the problem. Leaky caps are a nuisance.
If you want a sympathetic ear regarding things on the schematics, talk to Mark Finnis. He built an Aleph with the wrong number of output devices (too few) due to an acknowledged error (since corrected) in the schematics. Not having built an L, I'm having to do this on the fly. That's what led me to make a fool of myself above. I had to print out a schematic and look it over, rather than try to rely on memory. I fear that my mind has reached a state where it's more leaky than your caps...
Speaking of caps, I'm a confirmed film cap boy, myself, but the Black Gate caps are well thought of. I seem to recall that there are two or three other lines that folks seem fond of, but can't remember which ones at the moment. You might try looking over Michael Percy's inventory, as he usually tends to carry pretty ritzy parts.
Grey
Excellent news! Glad you found the problem. Leaky caps are a nuisance.
If you want a sympathetic ear regarding things on the schematics, talk to Mark Finnis. He built an Aleph with the wrong number of output devices (too few) due to an acknowledged error (since corrected) in the schematics. Not having built an L, I'm having to do this on the fly. That's what led me to make a fool of myself above. I had to print out a schematic and look it over, rather than try to rely on memory. I fear that my mind has reached a state where it's more leaky than your caps...
Speaking of caps, I'm a confirmed film cap boy, myself, but the Black Gate caps are well thought of. I seem to recall that there are two or three other lines that folks seem fond of, but can't remember which ones at the moment. You might try looking over Michael Percy's inventory, as he usually tends to carry pretty ritzy parts.
Grey
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.