UPDATE: was reading 0 v on bad channel PNP board w Variac with common to the (+) speaker terminal on all resistors. Checked voltage at the white wire coming in to the PNP board and I do have voltage. As a sanity check, I tempted fate and checked the good channel the same way and I did see voltage rising as I raised the variac.
Pulled bad channel NPN board and noticed the fuse was blown. Upon visual of the PNP board, i noticed that the diode I replaced (D1) was installed backwards. 😳
I just pulled all of the 2sb554 and 2sd424 and they are all still testing good. Could the backwards diode be the cause of the issue and causing the fuse to blow on the other board?
I am still going to install all 8 new transistors and that would make a full rebuild on PNP and NPN boards on the bad channel, but wanting to know if the backwards diode was the cause after the first repair...
Pulled bad channel NPN board and noticed the fuse was blown. Upon visual of the PNP board, i noticed that the diode I replaced (D1) was installed backwards. 😳
I just pulled all of the 2sb554 and 2sd424 and they are all still testing good. Could the backwards diode be the cause of the issue and causing the fuse to blow on the other board?
I am still going to install all 8 new transistors and that would make a full rebuild on PNP and NPN boards on the bad channel, but wanting to know if the backwards diode was the cause after the first repair...
When selecting the transistors, is it recommended to more closely match the Beta or VFE?
MJ21194
B VFE
1 62 548
2 67 556
3 67 553
4 73 558
5 73 553
6 67 558
7 62 557
8 88 547
9 67 557
10 58 557
MJ21193
B VFE
1 91 542
2 87 542
3 87 540
4 90 540
5 88 542
6 89 543
7 88 541
8 91 539
9 90 540
10 87 541
MJ21194
B VFE
1 62 548
2 67 556
3 67 553
4 73 558
5 73 553
6 67 558
7 62 557
8 88 547
9 67 557
10 58 557
MJ21193
B VFE
1 91 542
2 87 542
3 87 540
4 90 540
5 88 542
6 89 543
7 88 541
8 91 539
9 90 540
10 87 541
Update: After all that, I thought I was good. 0v reading at emitter with common of multimeter to (+) speaker terminal. Dim bulb still coming on when variac is turned up. Visually scouring everything over and over to see if there's anything that jumps out.
Also taking various continuity and resistance values between good and bad channels and nothing seems different between the 2. Not sure where else to check to start narrowing it down....
Also taking various continuity and resistance values between good and bad channels and nothing seems different between the 2. Not sure where else to check to start narrowing it down....
With variac at a stable voltage, put black of DVM to speaker ground or center of the two main rail caps, with alligator clip.
With red probe check to see if each analog transistor reads >1 c Vc to Ve. Take c & e readings and subtract. Since output transistors are new, start at the predrivers Q9 Q10. voltage should be symmetric around 0, and the same as the good side. b-e should be 0.6 v or - 0.6 v. Those okay, look at the drivers Q11 Q12.
There could be shorted diodes like D5 D6 or D1 or D2.
P1 cold be 1000 ohms, the wiper open. c Q9 to c Q10 should be about 1.8 to 2.4 v.
With red probe check to see if each analog transistor reads >1 c Vc to Ve. Take c & e readings and subtract. Since output transistors are new, start at the predrivers Q9 Q10. voltage should be symmetric around 0, and the same as the good side. b-e should be 0.6 v or - 0.6 v. Those okay, look at the drivers Q11 Q12.
There could be shorted diodes like D5 D6 or D1 or D2.
P1 cold be 1000 ohms, the wiper open. c Q9 to c Q10 should be about 1.8 to 2.4 v.
Thanks @indianajo
I got a little lost at this:
"With red probe check to see if each analog transistor reads >1 c Vc to Ve"
but I pulled q11 and q12 both tested ok.
D1, D2, D5 and D6 tested ok in circuit with mk168 tester.
As far as probing Q9 and Q10 with voltage, I feel like I'm going to short something out so decided against. I will pull them tomorrow and test outside. They were new also, so unless the backwards D1 fried something, I'm thinking they should be ok. I will confirm.
I got a little lost at this:
"With red probe check to see if each analog transistor reads >1 c Vc to Ve"
but I pulled q11 and q12 both tested ok.
D1, D2, D5 and D6 tested ok in circuit with mk168 tester.
As far as probing Q9 and Q10 with voltage, I feel like I'm going to short something out so decided against. I will pull them tomorrow and test outside. They were new also, so unless the backwards D1 fried something, I'm thinking they should be ok. I will confirm.
Tests of transistors and diodes at 2 v with a toy are useless. Except for the totally dead ones, which are not 100% of the damage. Totally dead ones should have showed up already with in ciccuit DVM diode scale test.
If your hand shakes, cover red probe of DVM with heat shrink except for last 1/16". Or buy a Pamona grabber and use instead of a DVM probe.
If your vision is bad, buy reading glasses. Increase the local lighting on the circuit board.
Transistors are not the only components that short or go open. Resistors and capacitors also can be damaged. I have found some of each by finding stupid voltages with the DVM. PCB lands can be broken, or solder joints can be intermittant. I found a bad solder joint from the factory on an input op amp by pushing on the joint with the probe, that had baffled 2 or 3 previous techs. Big pop on the speaker when I pushed there. Warning, use a trash speaker from the junkyard and protect it with back to back >1000 uf electrolytic caps with voltage rating greater than the amp rail. One tech's solution was to install a label on the speaker jack "DO NOT USE CHANNEL A". Speaker terminal would whang off to 180 vdc at random times.
If your hand shakes, cover red probe of DVM with heat shrink except for last 1/16". Or buy a Pamona grabber and use instead of a DVM probe.
If your vision is bad, buy reading glasses. Increase the local lighting on the circuit board.
Transistors are not the only components that short or go open. Resistors and capacitors also can be damaged. I have found some of each by finding stupid voltages with the DVM. PCB lands can be broken, or solder joints can be intermittant. I found a bad solder joint from the factory on an input op amp by pushing on the joint with the probe, that had baffled 2 or 3 previous techs. Big pop on the speaker when I pushed there. Warning, use a trash speaker from the junkyard and protect it with back to back >1000 uf electrolytic caps with voltage rating greater than the amp rail. One tech's solution was to install a label on the speaker jack "DO NOT USE CHANNEL A". Speaker terminal would whang off to 180 vdc at random times.
Last edited:
You don't move the variac control. With the voltages of Q9 Q10 c, b, e you should report the rail voltages V+ and V-. Voltages at card connectors +D and -D at same time would be interesting.
Note arranging the driver card so you can access these points is a problem left up to you. I found some of the amps I work on are arranged where left and right driver boards can be swapped to put the bad channel on top. Or driver board can be turned upside down with insulator board preventing the down side to touch any dangerous voltage. Only people that have owned or worked on this particular amp can help you with the test arrangement.
Note arranging the driver card so you can access these points is a problem left up to you. I found some of the amps I work on are arranged where left and right driver boards can be swapped to put the bad channel on top. Or driver board can be turned upside down with insulator board preventing the down side to touch any dangerous voltage. Only people that have owned or worked on this particular amp can help you with the test arrangement.
Hi wiz1,
With a dim bulb in circuit anywhere, you simply do not have any stable voltages.
MATCH BETA !!!!!!!! The best meter for this is a Heathkit IT-18 or similar tester. They make more expensive ones, but these measure leakage as well. Modern cheap digital testers are not sensitive enough for leakage and they don't really show beta or how it varies with temperature. The ones built into multimeters are a joke, but better than nothing I guess.
Geez, everyone wants to take the cheap and easy way out. This simply isn't correct or a good way to do anything. Dim bulb when you should be using a variac - alone! Match transistors measuring beta, not Vbe (easier and not representative of the degree of match). Yes, you should have instruments for matching transistors, and you should have a variac (which you now do - good). Just do it right. It saves time and avoids future failures.
Do not swap anything between channels of an amplifier. That is the best way to end up confused with two non-functional channels. Follow the schematic, use logic, write down measured values. If you are looking to see where the current is going, measure voltage drops. You do not need to compare to other channels normally. Use common sense, don't jump to conclusions.
You have no idea how many products I get for service that are the result of improper previous service. I even get pro power amps where the fix was to simply remove the blown outputs (left out) along with completely inappropriate replacement components. Big shock they fail again. Please. Relax, follow the current flow and examine the schematic. Use logic and you will sort it out. Use real devices and match the parallel outputs. Check emitter and base resistors. As I say (not joking to some), beer after it is fixed and running.
With a dim bulb in circuit anywhere, you simply do not have any stable voltages.
MATCH BETA !!!!!!!! The best meter for this is a Heathkit IT-18 or similar tester. They make more expensive ones, but these measure leakage as well. Modern cheap digital testers are not sensitive enough for leakage and they don't really show beta or how it varies with temperature. The ones built into multimeters are a joke, but better than nothing I guess.
Geez, everyone wants to take the cheap and easy way out. This simply isn't correct or a good way to do anything. Dim bulb when you should be using a variac - alone! Match transistors measuring beta, not Vbe (easier and not representative of the degree of match). Yes, you should have instruments for matching transistors, and you should have a variac (which you now do - good). Just do it right. It saves time and avoids future failures.
Do not swap anything between channels of an amplifier. That is the best way to end up confused with two non-functional channels. Follow the schematic, use logic, write down measured values. If you are looking to see where the current is going, measure voltage drops. You do not need to compare to other channels normally. Use common sense, don't jump to conclusions.
You have no idea how many products I get for service that are the result of improper previous service. I even get pro power amps where the fix was to simply remove the blown outputs (left out) along with completely inappropriate replacement components. Big shock they fail again. Please. Relax, follow the current flow and examine the schematic. Use logic and you will sort it out. Use real devices and match the parallel outputs. Check emitter and base resistors. As I say (not joking to some), beer after it is fixed and running.
The proper use for a dim bulb is to prevent fires on the bench when testing unvetted amplifier designs for the first time. Voltage that drops under load is exactly what you want then. Even then it needs to be used with a variac to get the rail voltage up ll the way. More often than NOT amplifiers will “work” at reduced voltage and you only see stability problems at full bore. Vcc goes up, Cob goes down, presto - destructive local oscillation in an output stage. The variac alone results in blown outputs and drivers. Measuring the actual frequency of oscillation is often needed to assess the fix. On the DBT it will sit there and flash on and off, going “zzzzzhhhmmmp zzzzhhhhmmpp zzzzzhhhmmmp” out of the speaker.
Guess I've been lucky then. Never lost an output stage while using a variac, not even with a new design.
I'll be honest, if someone replaces output transistors with another type you can get full bore oscillation - after you excite the amplifier with a signal. That's the definition of surprise. Or after it warms up without signal.
Properly used, a variac is safe. Fuse it lower if you want, but you'd be further ahead fusing the DC supplies after filtering since the supply caps can store a lot of energy. Disconnecting or limiting the primary won't help you then. If the amp is conditionally stable, you can still troubleshoot at reduced voltages to control the dissipation or current flow.
I honestly can't see any use for a dim bulb, except in cases where you don't want to take the time to work carefully. Changing AC mains voltage introduces all kinds of variables I'd rather not fight. I use a variac, or a Keysight 6812C that has current limiting. Even then, just fuse your rails to some low current. You honestly do not need or want full current available during initial testing.
I'll be honest, if someone replaces output transistors with another type you can get full bore oscillation - after you excite the amplifier with a signal. That's the definition of surprise. Or after it warms up without signal.
Properly used, a variac is safe. Fuse it lower if you want, but you'd be further ahead fusing the DC supplies after filtering since the supply caps can store a lot of energy. Disconnecting or limiting the primary won't help you then. If the amp is conditionally stable, you can still troubleshoot at reduced voltages to control the dissipation or current flow.
I honestly can't see any use for a dim bulb, except in cases where you don't want to take the time to work carefully. Changing AC mains voltage introduces all kinds of variables I'd rather not fight. I use a variac, or a Keysight 6812C that has current limiting. Even then, just fuse your rails to some low current. You honestly do not need or want full current available during initial testing.
UPDATE: with Variac set to output approx 45V, I took the following readings with common probe to chassis:
Q9 - E 8.4mv C 21.3mv B 27.5mv
Q10 - E 12.5mv C 51mv B 28.5mv
Q11 - B 21.5mv C 16V E 26.9mv
Q12 - B 51mv C 16.3V E 28.5mv
Voltage at main filter caps was reading 15-16V
At the speaker terminals I am reading 38mv and 35mv with the same 45V applied from Variac.
Q9 - E 8.4mv C 21.3mv B 27.5mv
Q10 - E 12.5mv C 51mv B 28.5mv
Q11 - B 21.5mv C 16V E 26.9mv
Q12 - B 51mv C 16.3V E 28.5mv
Voltage at main filter caps was reading 15-16V
At the speaker terminals I am reading 38mv and 35mv with the same 45V applied from Variac.
Last edited:
Proper connection of the black probe of the DVM is to the common terminal of C8 & C9 of one channel, or C6 and C7 of the other channel. The two main cap centers appear to be connected together and to line, which perhaps might be chassis. You should verify this zero ohm connectioin with the dvm diode scale.
Base of Q12 same as C of Q10. Base of Q11 same as collector of Q9. So R17 & R19 okay. Collector of Q11 should be minus rail, presumably -16. Emitter of Q11 is same as emitter of Q12. So R21 is okay. Both Q11 Q12 b-e readings should subtract to -0.6 v or 0.6v Perhaps R20 and R19 are shorted. Perhaps C5 is shorted.
Base of Q12 same as C of Q10. Base of Q11 same as collector of Q9. So R17 & R19 okay. Collector of Q11 should be minus rail, presumably -16. Emitter of Q11 is same as emitter of Q12. So R21 is okay. Both Q11 Q12 b-e readings should subtract to -0.6 v or 0.6v Perhaps R20 and R19 are shorted. Perhaps C5 is shorted.
Last edited:
Hi wiz1,
Take output stage readings from emitter and bases referenced to the hot output lead terminal (speaker + in this case). That way you get true, stable readings.
Take output stage readings from emitter and bases referenced to the hot output lead terminal (speaker + in this case). That way you get true, stable readings.
C of Q10 should be about 1.6 to 2.2 v higher than E of Q9. If C5 is not shorted, perhaps P1 is turned down to low. Perhaps T2 is turned down too low. C voltage of Q10 derives from collector of Q7. Perhaps it or R6 is open or bad solder joint. Emitter of Q8 should be at same voltage as plus of rail cap.
If none of those is wrong, Q9 Q10 could be shorted. Even though they test okay on your 2 v transistor tester.
Don't know what package Q9 Q10 are in. Do not fall for temptation to buy exact match part # from ebay, aliexpress, amazon. Such are usually fakes. A TO92 part such as MPSA06/56 from an authorized ON distributor should be okay, but those are EBC viewing from flat side left to right. If you have ECB pinout, try to buy KSC1845 KSA992.
If none of those is wrong, Q9 Q10 could be shorted. Even though they test okay on your 2 v transistor tester.
Don't know what package Q9 Q10 are in. Do not fall for temptation to buy exact match part # from ebay, aliexpress, amazon. Such are usually fakes. A TO92 part such as MPSA06/56 from an authorized ON distributor should be okay, but those are EBC viewing from flat side left to right. If you have ECB pinout, try to buy KSC1845 KSA992.
Shorted parts will test short at low voltage. Leakage is another matter.
That's when a good semiconductor tester is invaluable. Curve tracers are too expensive and a pain to use.
That's when a good semiconductor tester is invaluable. Curve tracers are too expensive and a pain to use.
Boy, you guys have the patience of saints. I would have given up a while back if it looked like someone butchered the amp. These bigger older Adcoms are hard to fix for long term reliability. The biggest issue is THE PCBs. They're usually cooked from the smoldering drop resistors cooking away whch.provide voltage drop to the zeners on the driver board. Once the fiberglass resin layer is discolored, it turns to carbon and starts conducting. The only solution is to replace.the PCBs.
The other problem is the availability of original spec transistors. Anything on ebay is more than likely counterfeit. I remember buying a whole crap load of 2SJ56/2SK176 and MJ15022/23/24/25 before they all were discontinued.
A big problem with these older monster sized amps was the incosistemt assembly. The semi point to point wiring made each unit perform slightly different. Reliability was iffy, as some of the instability quirks were caused by this ie. stray capacitance and inductance.
I do love old Haflers which suffered similar issues once they were of a certain vintage and the PCBs were toasted to a crisp from excessively hot running components. I finally gave up with these and started buying Parasound amps. These have been far more reliable and accurate sounding. My favorite Adcom was the 5802.
I still have a Hafler XL600 that needs a recap, so that may be the last American amp I ever work on and own. They're just a labor of love at this point. Not surprisingly, they command big money on the used market these days.
The other problem is the availability of original spec transistors. Anything on ebay is more than likely counterfeit. I remember buying a whole crap load of 2SJ56/2SK176 and MJ15022/23/24/25 before they all were discontinued.
A big problem with these older monster sized amps was the incosistemt assembly. The semi point to point wiring made each unit perform slightly different. Reliability was iffy, as some of the instability quirks were caused by this ie. stray capacitance and inductance.
I do love old Haflers which suffered similar issues once they were of a certain vintage and the PCBs were toasted to a crisp from excessively hot running components. I finally gave up with these and started buying Parasound amps. These have been far more reliable and accurate sounding. My favorite Adcom was the 5802.
I still have a Hafler XL600 that needs a recap, so that may be the last American amp I ever work on and own. They're just a labor of love at this point. Not surprisingly, they command big money on the used market these days.
Well, there are a lot of old amplifiers well worth repair, and they hold their own today.
Marantz 500. Way ahead of it's time and absolutely worth rebuild and improvement. Then we have many unstable designs out there. Phase Linear for one as an example. I found Adcom amplifiers to be reliable, and I would know since I was a warranty station for them in the Toronto area. You're right in a way. Once a PCB gets cooked, things can get interesting. Some Adcom amps had issues with capacitor electrolyte (really ugly situation).
Marantz 500. Way ahead of it's time and absolutely worth rebuild and improvement. Then we have many unstable designs out there. Phase Linear for one as an example. I found Adcom amplifiers to be reliable, and I would know since I was a warranty station for them in the Toronto area. You're right in a way. Once a PCB gets cooked, things can get interesting. Some Adcom amps had issues with capacitor electrolyte (really ugly situation).
And PL series 2 like to cook the PCBs. Another reason you don’t like them - firing one up on a variac is taking your life into your hands. Dim bulb all the way. What they do have going for them is the best “cool factor” in the business, even though they NEED fans. They can be stabilized, but as stock they are ticking bombs.
Ain’t nothin you can do about the heat but put a fan on it, though. Or be gentle. Yeah, right.
Ain’t nothin you can do about the heat but put a fan on it, though. Or be gentle. Yeah, right.
lol!
For sure. A couple Pioneer amps suck extreme current until the voltage is higher. Stupid design.
Fans. Yeah, they get clogged with dust and never cleaned, or stop turning. I've seen people disconnect them because they got noisy. Yeah, there's a fix. A fan = maintenance, simple.
I have two SAE P500's on the bench. Not sure I want to fix them. The output transistors are not rated for the supply voltage. Clip without a load and they may pack it in. That and they have multiple ground to case connections and the one I have working as very high hum and 120Hz ripple. Filter caps are fine. These give me the creeps.
For sure. A couple Pioneer amps suck extreme current until the voltage is higher. Stupid design.
Fans. Yeah, they get clogged with dust and never cleaned, or stop turning. I've seen people disconnect them because they got noisy. Yeah, there's a fix. A fan = maintenance, simple.
I have two SAE P500's on the bench. Not sure I want to fix them. The output transistors are not rated for the supply voltage. Clip without a load and they may pack it in. That and they have multiple ground to case connections and the one I have working as very high hum and 120Hz ripple. Filter caps are fine. These give me the creeps.