Your DMM is perfectly fine.
Have you matched the input pair yet? Q601 - Q603. If the output went to one rail, you will cause reverse breakdown in one of the diff pair transistors. THat typically changes the characteristics, not in a good way.
Somewhere on this site is a circuit that was designed decades ago for exactly matching signal transistors (works on power devices, but that wasn't the goal). @Phloodpants designed a PCB for it.
Do not use a relay output protection device. Although I seriously hate amplifiers without effective output protection from DC offset, this amp as designed shouldn't use one. You get a DC offset on start-up due to the intentional imbalance. It slowly drifts to an acceptable value, but that takes a while. If you use a relay, it either stays open for a very long time, or you'll get a "CRACK" or "Bang" when it closes through your speakers. I view this as a design defect. If I owned one, I'd redesign the input so it was naturally balanced and eliminate the servo IC.
No amplifier is safe from very damaging DC offsets at the output terminals. Components fail, that is a fact of life. The only amplifiers without speaker protection I like are the Carver amplifiers. They have an extremely effective protection setup that shuts the amp down quickly and limits the fault current to low levels. Whether you like the sound of a Carver or not, they were in fact very well designed. Fuses are not protection and generate a high amount of distortion.
Have you matched the input pair yet? Q601 - Q603. If the output went to one rail, you will cause reverse breakdown in one of the diff pair transistors. THat typically changes the characteristics, not in a good way.
Somewhere on this site is a circuit that was designed decades ago for exactly matching signal transistors (works on power devices, but that wasn't the goal). @Phloodpants designed a PCB for it.
Do not use a relay output protection device. Although I seriously hate amplifiers without effective output protection from DC offset, this amp as designed shouldn't use one. You get a DC offset on start-up due to the intentional imbalance. It slowly drifts to an acceptable value, but that takes a while. If you use a relay, it either stays open for a very long time, or you'll get a "CRACK" or "Bang" when it closes through your speakers. I view this as a design defect. If I owned one, I'd redesign the input so it was naturally balanced and eliminate the servo IC.
No amplifier is safe from very damaging DC offsets at the output terminals. Components fail, that is a fact of life. The only amplifiers without speaker protection I like are the Carver amplifiers. They have an extremely effective protection setup that shuts the amp down quickly and limits the fault current to low levels. Whether you like the sound of a Carver or not, they were in fact very well designed. Fuses are not protection and generate a high amount of distortion.
I need to look when i get home to see what fuse was blown when I got the amp. I think it was the fuse closer to the front of the amp on the bad channel.
Also do the 4.7 uF and 47 uF caps on the amp boards need to be replaced?
I was told once that fuses after the power supply in an amp are there only to protect the power supply should something in the amp fail.
I do love the sound of the Carver TFM series amps.
I'll take more measurements when I get home from work.
Could the better channel also have the same problem the bad channel has just not as bad?
Also do the 4.7 uF and 47 uF caps on the amp boards need to be replaced?
I was told once that fuses after the power supply in an amp are there only to protect the power supply should something in the amp fail.
I do love the sound of the Carver TFM series amps.
I'll take more measurements when I get home from work.
Could the better channel also have the same problem the bad channel has just not as bad?
Last edited:
Depends on run time and temperature for those capacitors. These are old now, and the capacitors don't cost much, so you could replace them.
The fuses limit damage to the amp circuits, and I guess reduce power supply damage, but the mains fuse should open before you cause capacitors to vent.
The better channel could use the same treatment and care as the bad one. The diff pairs weren't matched at the factory for one, and it's been on this earth for a while now. Go over the entire amplifier and enjoy it for years.
The fuses limit damage to the amp circuits, and I guess reduce power supply damage, but the mains fuse should open before you cause capacitors to vent.
The better channel could use the same treatment and care as the bad one. The diff pairs weren't matched at the factory for one, and it's been on this earth for a while now. Go over the entire amplifier and enjoy it for years.
I do need to disconnect one end of R605 and measure what voltage is on Q601 base.
Does the amp passing a clean sinewave mean the input transistors are ok far as testing good even though they are not matched?
What I don't understand is based on the voltage divider consisting of R609, R611 and RR613, I should not be getting -1.822Vdc at Q601 base especially with the OP-AMP producing nearly +12Vdc output.
I'd like to think there was a problem with Q601, Q603 or Q605, however shorting Q601 base to ground increases the DC offset to +70.2mVdc and I did get a clean sinewave out of the bad channel and when the amp did go into clipping, both halves of the sinewave clipped evenly.
Once I get the amp working properly I may look into purchasing one of Phloodpants power supply boards for this amp. Might also get one for my 545II as well.
Does the amp passing a clean sinewave mean the input transistors are ok far as testing good even though they are not matched?
What I don't understand is based on the voltage divider consisting of R609, R611 and RR613, I should not be getting -1.822Vdc at Q601 base especially with the OP-AMP producing nearly +12Vdc output.
I'd like to think there was a problem with Q601, Q603 or Q605, however shorting Q601 base to ground increases the DC offset to +70.2mVdc and I did get a clean sinewave out of the bad channel and when the amp did go into clipping, both halves of the sinewave clipped evenly.
Once I get the amp working properly I may look into purchasing one of Phloodpants power supply boards for this amp. Might also get one for my 545II as well.
Last edited:
I'm increasingly suspicious that the input pair may be at fault. Excessive base current (i.e. low Beta) can be the cause of the negative bias at the PA + input. Measuring bias voltage across R607 should reveal with better clarity.
I don't disagree with any of anantech's comments. I'm not a fan of the unipolar bias servo--- seems a poor cost saver.
I don't disagree with any of anantech's comments. I'm not a fan of the unipolar bias servo--- seems a poor cost saver.
Yes, just a cost saver. To save one supply voltage to the servo.
Best thing to do is get your transistor testers together. Build the matcher and simply test your signal transistors and match the diff pairs. The Heathkit IT-18 is a wonderful meter for testing bipolar transistors. Low tech, but it simply works very well and measures real leakage. It is a lot more sensitive than most modern "transistor checkers" for leakage currents, and you need that.
Best thing to do is get your transistor testers together. Build the matcher and simply test your signal transistors and match the diff pairs. The Heathkit IT-18 is a wonderful meter for testing bipolar transistors. Low tech, but it simply works very well and measures real leakage. It is a lot more sensitive than most modern "transistor checkers" for leakage currents, and you need that.
I read somewhere about Adcom amps where on semi manufactured transistors are better matched and would be a better choice.
Would it be possible to order suitable subs from them?
Would be awhile before I can gather the necessary equipment to test the transistors.
That said if the transistors don't cost much I can order a batch and test them using the equipment you mentioned.
The same would apply to the 545II, right?
Would it be possible to order suitable subs from them?
Would be awhile before I can gather the necessary equipment to test the transistors.
That said if the transistors don't cost much I can order a batch and test them using the equipment you mentioned.
The same would apply to the 545II, right?
Same applies to the 545 II.
No, they do not have the time to really match transistors. No easy / cheap way out here. You could buy dual transistors from Linear Systems, those are very good. Not sure on pinout or specs.
You really want to order a bunch and match them. You have a bunch of amplifiers. You can match more closely than duals, but dual transistors track temperature better (same die). The IT-18 is the best way to test the other transistors. I have three of them and use them daily, I have worn a couple of them out. I'm using one right now on the bench for a Yamaha PC-2002 repair.
I'll always advise people on the best, most effective way to approach service. I'll never suggest something I haven't found to be the best way and I do myself.
No, they do not have the time to really match transistors. No easy / cheap way out here. You could buy dual transistors from Linear Systems, those are very good. Not sure on pinout or specs.
You really want to order a bunch and match them. You have a bunch of amplifiers. You can match more closely than duals, but dual transistors track temperature better (same die). The IT-18 is the best way to test the other transistors. I have three of them and use them daily, I have worn a couple of them out. I'm using one right now on the bench for a Yamaha PC-2002 repair.
I'll always advise people on the best, most effective way to approach service. I'll never suggest something I haven't found to be the best way and I do myself.
Now it is possible for me to buy these matched transistors.
https://hoppesbrain.com/product/matched-input-transistors-2-pairs/
That might be a solution for now until I can save up the money to get the necessary equipment to match my own transistors.
One other thing I noted here
https://docs.google.com/spreadsheet...LfF0qF3U7KpEdxA9CVT1gf1Zk/edit#gid=1202533542
C611 and C612 are replaced with a Wima MKS2 cap and C629 and C630 (filter the supply for the servo) are replaced with a 470uF cap with the explanation that's what they are in the 565 and it won't hurt to use the larger cap in the 535II.
However, I need to ensure the amp is working properly first before doing any sort of mods to it.
The amp has been on about an hour. I'll take some measurements and post them here.
https://hoppesbrain.com/product/matched-input-transistors-2-pairs/
That might be a solution for now until I can save up the money to get the necessary equipment to match my own transistors.
One other thing I noted here
https://docs.google.com/spreadsheet...LfF0qF3U7KpEdxA9CVT1gf1Zk/edit#gid=1202533542
C611 and C612 are replaced with a Wima MKS2 cap and C629 and C630 (filter the supply for the servo) are replaced with a 470uF cap with the explanation that's what they are in the 565 and it won't hurt to use the larger cap in the 535II.
However, I need to ensure the amp is working properly first before doing any sort of mods to it.
The amp has been on about an hour. I'll take some measurements and post them here.
Last edited:
Well if your input transistors are actually bad enough to cause this offset, they must be defective. You could try an un-matched pair just to see if the problem goes away. The servo should null out the offset even if they're bad matches. It would be a good idea to make sure it basically works before investing in matched transistors.
C629 and C630 are power supply filters for the servo power supply. 4.7uF always seemed a ridiculously low value for this purpose, I suspect it was a decimal-place typo. The 555 II (not the 565) uses the exact same circuit with a 470uF cap.
C611 and 612 are bypass caps across the bias-spreader, poly is a nice upgrade that will never wear out.
Neither of these things affect your issue you probably know.
C611 and 612 are bypass caps across the bias-spreader, poly is a nice upgrade that will never wear out.
Neither of these things affect your issue you probably know.
By clean I meant the sinewave looked like a sinewave on a scope with no clipping whatsoever. That don't mean no distortion is present though.
Q601 base with R605 disconnected is -2.605Vdc.
Q601 emitter is -3.194Vdc.
So I'm going to suspect Q605 may be my problem and Q604 in the other channel could be failing as well.
Q601 base with R605 disconnected is -2.605Vdc.
Q601 emitter is -3.194Vdc.
So I'm going to suspect Q605 may be my problem and Q604 in the other channel could be failing as well.
Yep, a good pair of loosely matched transistors can help troubleshoot.
Your voltage change across the bias spreader is next to zero. Cap type doesn't really matter here.
Changing the 4u7 cap to 470 uF is a very bad idea. Look at the time constant (R691). 4u7 is adequate, it's there to reduce HF impedance. You could install poly or P-A there if you wanted, but there is zero point in increasing it. Increasing this cap will probably make your startup DC offset higher.
Increasing capacitor value is generally a bad idea. You have to look at the entire picture, and realise what that capacitor is doing.
Your voltage change across the bias spreader is next to zero. Cap type doesn't really matter here.
Changing the 4u7 cap to 470 uF is a very bad idea. Look at the time constant (R691). 4u7 is adequate, it's there to reduce HF impedance. You could install poly or P-A there if you wanted, but there is zero point in increasing it. Increasing this cap will probably make your startup DC offset higher.
Increasing capacitor value is generally a bad idea. You have to look at the entire picture, and realise what that capacitor is doing.
I may see if I can find a transistor to temporarily replace Q605 to see if the problem goes away. It could also be Q601 as well.
If so then I'll order new electrolytics for the amp board, the matched transistors for Q601 and Q603 and a transistor for Q605 and do the same for the other channel.
If so then I'll order new electrolytics for the amp board, the matched transistors for Q601 and Q603 and a transistor for Q605 and do the same for the other channel.
Q605 is your tail current generator. Measure across R619 to see what current you're sinking. It should be about one diode drop.
Q601 is matched to Q603. Either one could be bad depending on the offset polarity when it failed. Or one may simply have gone defective. So replace both to check, then install a closely matched pair.
That sounds very reasonable.
Q601 is matched to Q603. Either one could be bad depending on the offset polarity when it failed. Or one may simply have gone defective. So replace both to check, then install a closely matched pair.
That sounds very reasonable.