I've acquired an Audiolab 8000A which is popping its mains fuses.
I've checked the power supply with the rails to the output stage disconnected and all is OK - voltages OK, minor ripple.
Looking in to the power rails the resistance on both channels and rails is the same and showing no shorts or low resistance so I think the output transistors are OK.
With the power connected to the power amp again and the amp powered via a DBT nothing looks untoward - the expected inrush then a steady dim glow.
However, when directly connected to the mains the transformer briefly hums as though under high load and the fuse goes.
The only anomaly I've found is that one of R495/R496 which are 4R7 2W resistors is down to 2R8 when measured out of circuit. That resistor is in series with a 100nF capacitor, which are across the output.
Is that circuit to filter high frequencies to prevent oscillation? Could oscillation be what's causing the high current draw and be due to that filter not working correctly due to the resistor's lower value? Given the amp blows its fuse so quickly it's not that feasible to detect with a scope.
It also looks like some of the output transistors have been changed - perhaps for unsuitable replacements?
I've checked the power supply with the rails to the output stage disconnected and all is OK - voltages OK, minor ripple.
Looking in to the power rails the resistance on both channels and rails is the same and showing no shorts or low resistance so I think the output transistors are OK.
With the power connected to the power amp again and the amp powered via a DBT nothing looks untoward - the expected inrush then a steady dim glow.
However, when directly connected to the mains the transformer briefly hums as though under high load and the fuse goes.
The only anomaly I've found is that one of R495/R496 which are 4R7 2W resistors is down to 2R8 when measured out of circuit. That resistor is in series with a 100nF capacitor, which are across the output.
Is that circuit to filter high frequencies to prevent oscillation? Could oscillation be what's causing the high current draw and be due to that filter not working correctly due to the resistor's lower value? Given the amp blows its fuse so quickly it's not that feasible to detect with a scope.
It also looks like some of the output transistors have been changed - perhaps for unsuitable replacements?
Remove the 2SA1216/2SC2922 output transistors and the 1N4002 diodes in parallel with them and see if the amplifier/headphone output works properly.
Thanks - I'll give that a try. Am I correct in thinking then that Q431/2 can still drive the headphones?
Also, what is the purpose of those reverse-biased 1N4002 diodes ?
Thanks to aparatusonitus for the flyback info. I assume it relates to dealing with the inductive speaker loads?
Now....
Time to admit my schoolboy error when checking the output transistors - one of the 2SC2922 is actually short collector-emitter. I had been measuring from the 44v rails to 0v which without either speakers connected or the protection relay activated are not connected so I assume when powered there's excess current flowing from the +44v to the -44v rail via the output transistors.
The 2SC2922 and the 2SA1216 on one side look to have been replaced previously with ISC-branded items which, having read a few threads on the subject, have a dubious reputation according to some.
So... I think I need to find some suitable replacements, possibly with a different package....yikes .. and see whether something other than a dubious quality component may have caused the fault.
Now....
Time to admit my schoolboy error when checking the output transistors - one of the 2SC2922 is actually short collector-emitter. I had been measuring from the 44v rails to 0v which without either speakers connected or the protection relay activated are not connected so I assume when powered there's excess current flowing from the +44v to the -44v rail via the output transistors.
The 2SC2922 and the 2SA1216 on one side look to have been replaced previously with ISC-branded items which, having read a few threads on the subject, have a dubious reputation according to some.
So... I think I need to find some suitable replacements, possibly with a different package....yikes .. and see whether something other than a dubious quality component may have caused the fault.
Output Node Catch Diodes
The output node of the amplifier must never be allowed to go beyond the power supply
rails by more than a diode drop. Such high-voltage excursions can occur as the result of
an inductive speaker load whose current has suddenly been interrupted or limited.
Such inductive kicks can damage output transistors, speaker relay contacts, and tweet-
ers. For this reason, silicon diodes are wired in a reverse-biased manner from the output
node to each of the power supply rails. They will conduct if the output voltage attempts
to go beyond the rail voltage. For amplifiers that incorporate speaker relays, it is best to
include such catch diodes on both sides of the relay. (quote from Bob Cordell's book)
That's one of the reasons you mentioned, shorted output transistors, questionable quality etc..., but before installing new pairs it's very important to test the functionality of the previous driver stage because once the output stage goes bad it often happens that they take the driver stage with them to solid state heaven. That's the main reason I suggested testing the amp as a headamp...to make sure the rest of the amp is working properly.
For minimum hassle but maximum cost, check out https://www.diyaudio.com/community/threads/sanken-mt-200-2sa1216-2sc2922.412935/post-7687207
If you are skilled in a little drilling and tapping, I suggest converting from MT-200 to TO-264 package. Check out https://gb.profusion.uk/uk/audio/transist/audio-transistors as well as https://www.onsemi.com/products/dis...1ZX4yfiF+TGFzdCBTaGlwbWVudHN+IX5PYnNvbGV0ZX4= because you can buy complementary output transistors cheaply there, and even get them for free if you register and request samples.
A bit of progress but further advice needed if possible.
As suggested I removed the four main output transistors and their associated flyback diodes. I checked over the transistors, diodes and resistors and none appear blown or reading wrongly although in the protection circuit had cracked away from the pcb which was repaired. This was with a view to check the amp would drive headphones.
However, when the amp was powered up (with no speaker load) via the DBT, after the initial inrush and dimming the DBT started to glow bright again after a couple of seconds and a few whisps of smoke appeared from somewhere in the earlier stages of the left power amp (I have disconnected the 44v rails from the right channel) before I quickly disconnected the power.
It's not really clear to me what path any large current flow might be taking but I assume it's the 2SC4382 or complement.
NOTE. It's a late model 8000A for which a service manual does not appear to be available so I'll try to describe the differences to the diagram below.
Given it's a DC-coupled amp I wondered if there was some DC offset at the input causing this and given the timing of the increasing current after the initial inrush I was drawn to what I think is a soft mute formed by the 3.9k resistor, 47u cap and 4V7 zener on the bottom of the input differential.
So, with the 3.9k resistor disconnected the circuit stayed on mute and the current draw subsided as expected after the inrush so it would seem a DC offset is being introduced when the mute comes off. With that resistor reinstated I disconnected the 1M resistor going to the TL072 but that made no difference - the excess current draw comes back - so it's perhaps not the automatic offset control causing this.
On the diagram is my crude arrow on the output which is where output relay contacts sit and which actually connect to the headphone out until the protection circuit operates the relay. When the mute was disabled I measured about 9mV at the output which satisfied the protection circuit to operate the relay.
However, I'm now confused as to what might be causing the input to drive the output hard over. Would the 9mV at the output be likely to do that?
NOTE 2. On my amp the differential pair is a 2SK389 FET pair package rather than the BC546 pair.
As suggested I removed the four main output transistors and their associated flyback diodes. I checked over the transistors, diodes and resistors and none appear blown or reading wrongly although in the protection circuit had cracked away from the pcb which was repaired. This was with a view to check the amp would drive headphones.
However, when the amp was powered up (with no speaker load) via the DBT, after the initial inrush and dimming the DBT started to glow bright again after a couple of seconds and a few whisps of smoke appeared from somewhere in the earlier stages of the left power amp (I have disconnected the 44v rails from the right channel) before I quickly disconnected the power.
It's not really clear to me what path any large current flow might be taking but I assume it's the 2SC4382 or complement.
NOTE. It's a late model 8000A for which a service manual does not appear to be available so I'll try to describe the differences to the diagram below.
Given it's a DC-coupled amp I wondered if there was some DC offset at the input causing this and given the timing of the increasing current after the initial inrush I was drawn to what I think is a soft mute formed by the 3.9k resistor, 47u cap and 4V7 zener on the bottom of the input differential.
So, with the 3.9k resistor disconnected the circuit stayed on mute and the current draw subsided as expected after the inrush so it would seem a DC offset is being introduced when the mute comes off. With that resistor reinstated I disconnected the 1M resistor going to the TL072 but that made no difference - the excess current draw comes back - so it's perhaps not the automatic offset control causing this.
On the diagram is my crude arrow on the output which is where output relay contacts sit and which actually connect to the headphone out until the protection circuit operates the relay. When the mute was disabled I measured about 9mV at the output which satisfied the protection circuit to operate the relay.
However, I'm now confused as to what might be causing the input to drive the output hard over. Would the 9mV at the output be likely to do that?
NOTE 2. On my amp the differential pair is a 2SK389 FET pair package rather than the BC546 pair.
The 4.7 uf resistor being cooked to 2.8 ohms indicates possibly ultrasonic oscillation. The 100 n cap may be cooked also. Lacking a capacitor meter, I would replace both.
If you don't have an oscilloscope, the cheap way to check for ultrasonic oscillation is an analog VOM with a 20 vac or 50 vac scale. These can be had for $20. You need a .047 uf >200 v cap series the black probe to analog ground )bottom of 100 nf cap, or center of the 2 rail caps. Use two alligator clip leads. If you have AC voltage at the top of the 4.7 resistor (speaker before relay disconnect) while the amp is silent (input shorted) you have ultrasonic oscillation.
Some basics for newbies so to not kill or injure yourself. Never use 2 hands to probe any circuit with >12 v. >25 v across your heart from one hand to the other can stop it. Never wear jewelry or metal on hands wrists or neck. 1 v at high current through metal can burn your flesh to charcoal. Wear safety glasses, or reading glasses. Parts can explode, solder splashes into eyes, especially desoldering.
Another check to make with DVM, when dim bulb lights again, what is DC voltage at top of 4.7 resistor or junction of 2.7 k and 47 p?
I don't know about the drivers driving the headphone. There appears to be no connection between MPSA42 e and MPSA92 C and the output. Many amps willl drive the speaker with the driver transistors, but not this one. Also personally I think driving output transistors with TO92 transistors is cheap.
Your checks "output transistors and diodes and resistors not blown" may not prove them good. I suppose each transistor reads 550 to 700 b to e and b to c, and 9999 or ---- backwards? That means they are okay at 2 v of the diode scale, but does not prove them good at 42 v rail voltage. I have found that transistors that pass that test, but leak significant current at 12 v are still bad. I hook up 12 v battery charger through 47 k resistor and microamps current scale of DVM. + to C of npn, - to C of pnp. End of circuit is the e. If current is 12/47000 or 255 uA, the transistor is trash. Current under 5 uA, transistor might be okay.
Thanks Indianajo - I put my scope on the output (centre of the 2 0R22 resistors ) and sure enough when the DBT glowed again there was a signal of about 600kHz and about 30v pk-pk (take these numbers with a pinch of salt as I only had a moment to read before switching off due to a puff of smoke).
I've not been able to determine where the smoke comes from or whether that's a symptom or cause - probably the former?
The Zobel network of the 4R7 and 220nF cap have now been changed (it says 100nF on the diagram but mine is a 220nF) but that made no difference.
I just noticed you suggested grounding the input - I assume that's the direct input to the power stage and not whichever pre-amp input is switched through ?
So, what other options should I look at (apart from testing with the input grounded) and if it is oscillation what is the likely cause?
I've seen various posts on other forums that power supply instability might be a cause. The +ve 44v rail has about 60mV of ripple on it but the -ve is ripple-free. The +ve rail feeds other parts such as the tone circuit and phono preamps. The voltage at the 12v zeners in the power amp differential stage also looked clean.
I suggest ground the input at the selected input. I suppose tuner input is most convenient. I take an RCA cable and use alligator clip lead to connect center to rim.
I'm no expert on killing oscillation, but that long feedback line looks like a good place for oscillation to feed back to a high gain point. Check that the feedback transistor of the long tailed pair is oscillating. If so, I would start with changing the 47 p cap paralleling the 2.7k feedback resistor to 100 pf or 220 pf. If that doesn't kill it, put a series 6 turn inductor series both. I salvage these from PCAT switcher supplies or flat screen TV's. If that does not do it I would try 51 pf caps c to e on the predrivers, the MPSA 42 and 92.
If not, then the emitter follower output is oscillatiing by itself. Try putting resistors series the base of the drivers and or the output transistors. So called base stopper resistors.
BTW I was wrong about driving the headphone with drivers. What is shown as 2sc4382 2sa1668 are drivers that do connect to the speaker out line.
I'm no expert on killing oscillation, but that long feedback line looks like a good place for oscillation to feed back to a high gain point. Check that the feedback transistor of the long tailed pair is oscillating. If so, I would start with changing the 47 p cap paralleling the 2.7k feedback resistor to 100 pf or 220 pf. If that doesn't kill it, put a series 6 turn inductor series both. I salvage these from PCAT switcher supplies or flat screen TV's. If that does not do it I would try 51 pf caps c to e on the predrivers, the MPSA 42 and 92.
If not, then the emitter follower output is oscillatiing by itself. Try putting resistors series the base of the drivers and or the output transistors. So called base stopper resistors.
BTW I was wrong about driving the headphone with drivers. What is shown as 2sc4382 2sa1668 are drivers that do connect to the speaker out line.
Last edited:
OK, tried with the input grounded both at the CD input and directly on the power amp input - no change - so this looks like oscillation.
Before I make any of the suggested modifications I assume it should be fine as designed? If so then something has gone out of tolerance so any suggestions on where to look? I'm a little stumped now.
Also, even without any power the circuitry seems to be picking up what I think is RF. This can be as much as 10mV and around 90MHz. This is even on the main power rails. Ok, the covers are not on so maybe it's lack of screening but could that be what sets things off?
Before I make any of the suggested modifications I assume it should be fine as designed? If so then something has gone out of tolerance so any suggestions on where to look? I'm a little stumped now.
Also, even without any power the circuitry seems to be picking up what I think is RF. This can be as much as 10mV and around 90MHz. This is even on the main power rails. Ok, the covers are not on so maybe it's lack of screening but could that be what sets things off?
I asked you to check the collector of the feedback transistor for oscillation. You checked the power supply. Have it your way.
What is the Ft of the replacement output transistors? What was the Ft of the specified transistors? Replacement > original, wrong. Unspecified Ft, whonose?
You can fix the oscillation, or you can pitch the product in the bin. Or pay the electronics recycle fee if your bin is inspected by the dustmen.
I find Peavey products, even though "dangerous" emitter follower design, do not oscillate. Another brand I observed lately, had zobels installed at 2 places as the model numbers increased. You have observed your example has a more expensive zobel than what is on the base schematics. What does that tell you? I observe from the used equipment for sale, Peavey had market share about 10-20 times that of audiolab. What does that tell you?
See this thread for an extensive discussion of oscillation, its causes and corrections, in adcom amps. https://www.diyaudio.com/community/...e-impact-of-this-signal-anomaly.421059/latest
What is the Ft of the replacement output transistors? What was the Ft of the specified transistors? Replacement > original, wrong. Unspecified Ft, whonose?
You can fix the oscillation, or you can pitch the product in the bin. Or pay the electronics recycle fee if your bin is inspected by the dustmen.
I find Peavey products, even though "dangerous" emitter follower design, do not oscillate. Another brand I observed lately, had zobels installed at 2 places as the model numbers increased. You have observed your example has a more expensive zobel than what is on the base schematics. What does that tell you? I observe from the used equipment for sale, Peavey had market share about 10-20 times that of audiolab. What does that tell you?
See this thread for an extensive discussion of oscillation, its causes and corrections, in adcom amps. https://www.diyaudio.com/community/...e-impact-of-this-signal-anomaly.421059/latest
Last edited:
The main output transistors are still out of circuit - I didn't want to fit them until I could get the amp stable without.
This amp has been worked on in the past and I think I need to check if the driver transistors are original or not - they're covered by heatsinks that also cover another couple of transistors (for temperature stability/compensation I assume ) so I can't see their designation without some significant dismantling. It's a real PITA to get to them but I think I may have to go there in case, as you suggest, they are unsuitable replacements.
Given it's cheap to do I may also try replacing a number of decoupling capacitors and those in the differential stage. There seem to be remarkably few for the power amp.
This amp has been worked on in the past and I think I need to check if the driver transistors are original or not - they're covered by heatsinks that also cover another couple of transistors (for temperature stability/compensation I assume ) so I can't see their designation without some significant dismantling. It's a real PITA to get to them but I think I may have to go there in case, as you suggest, they are unsuitable replacements.
Given it's cheap to do I may also try replacing a number of decoupling capacitors and those in the differential stage. There seem to be remarkably few for the power amp.
If the feedback transistor, the one at the left end of the feedback line from the output, is oscillating, then one path of modification is indicated. If the driver transistors alone are oscillating, another path of modification. Drivers in my experience can have 30 mhz Ft without problems. In fact, I found highs rather dull on an amp I built where I used 6 mhz Ft drivers TIP31c./32c. That amp originally had 20 mhz 2n5320/5322 and the surviving ones sounded fine on MJ15003 outputs. 20 mhz MJE15028/29 sounded much better on highs as drivers. I test on Steinway grand top octave tracks, also cymbals and tinkly bells.
If there was a meltdown from the output transistors back, yes some decoupling capacitors in the front end could be blown. In a PV-1.3k that I put 140 parts in, there were some 10 uf tantalum caps near the input op amp that worked a lot better after I replaced them. 50 v rated caps, 180 vdc from the output transistors through the base line to the input parts when the amp burned out. 140 parts is not an economic repair, but I learned a lot on that amp.
If there was a meltdown from the output transistors back, yes some decoupling capacitors in the front end could be blown. In a PV-1.3k that I put 140 parts in, there were some 10 uf tantalum caps near the input op amp that worked a lot better after I replaced them. 50 v rated caps, 180 vdc from the output transistors through the base line to the input parts when the amp burned out. 140 parts is not an economic repair, but I learned a lot on that amp.
OK, even more confused now...
The situation is:
All output transistors and their flyback diodes in both channels are out of the board.
The driver transistors are the originals.
No load - no speakers or headphones.
Running via DBT.
Both channels go into oscillation. That seems odd to me.
The only common part is the 44v rails supply to both sides.
Since I still had oscillation with the preamp disconnected which shares the common 30v supply I assume that's not involved.
Could something odd with the filter caps cause this or am I missing something?
The situation is:
All output transistors and their flyback diodes in both channels are out of the board.
The driver transistors are the originals.
No load - no speakers or headphones.
Running via DBT.
Both channels go into oscillation. That seems odd to me.
The only common part is the 44v rails supply to both sides.
Since I still had oscillation with the preamp disconnected which shares the common 30v supply I assume that's not involved.
Could something odd with the filter caps cause this or am I missing something?
You seem fixated on the rail voltage and caps. These are not usual suspects of oscillation. If dried up they cause low power out or lack of bass.
The oscillation could be coming from the input, the feedback transistor, the DC bias op amp, the drivers and predrivers as an emmiter follower. Only you, with the scope, can eliminate these sources. The input, because the amp was not designed to be in a room with a radio, ie your cell phone. Short the inputs with alligator clip leads. The op amp could have open or shorted local bypass caps. Or just bad layout. You need to probe the op amp output with the scope. The feedback transistor I have asked you twice to probe the collector to see if the oscillation is there. You have not yet. If none of the above are oscillating, then the predriver and the drivers are oscillating as an emitter follower. I suggest no modification to the predrivers and drivers unless all other possible sources are sliminated.
The oscillation could be coming from the input, the feedback transistor, the DC bias op amp, the drivers and predrivers as an emmiter follower. Only you, with the scope, can eliminate these sources. The input, because the amp was not designed to be in a room with a radio, ie your cell phone. Short the inputs with alligator clip leads. The op amp could have open or shorted local bypass caps. Or just bad layout. You need to probe the op amp output with the scope. The feedback transistor I have asked you twice to probe the collector to see if the oscillation is there. You have not yet. If none of the above are oscillating, then the predriver and the drivers are oscillating as an emitter follower. I suggest no modification to the predrivers and drivers unless all other possible sources are sliminated.
I'd hoped I'd documented trying many of your suggestions throughout the thread but was probably not clear about what I have scoped-
I've tried both with the inputs grounded and the pre-amp disconnected from the power amp. Made no difference. There is a lot of RF which I could pick up at numerous points in the circuit even with the power off. That was 90Mhz which may be coming from a FM transmitter about 7 miles away.
I'll go back and check these points again.
Question - if this is feedback due to a loop how do I eliminate cause and effect throughout the loop at various stages? To my mind loop oscillation is a function of the interaction of the components so having a "source" didn't seem to sit well.
Also, given this problem is common to both channels I was drawn to components common to both sides or maybe environmental. I'm working on the assumption that unless this amplifier has had the problem from new then something has changed and the same change in both channels (there are no mods apparent) would seem unlikely.
I've tried both with the inputs grounded and the pre-amp disconnected from the power amp. Made no difference. There is a lot of RF which I could pick up at numerous points in the circuit even with the power off. That was 90Mhz which may be coming from a FM transmitter about 7 miles away.
In the model I have the differential pair is a Toshiba dual FET package. I'll check the sources again. I know that that stage is involved as when I forced it to stay muted by removing the mute timer's charging resistor the problem went away.
This was "disabled" by removing its 1M input resistor but, of course, it could still be having an effect. Maybe one to try.
This has been scoped .
I'll go back and check these points again.
Question - if this is feedback due to a loop how do I eliminate cause and effect throughout the loop at various stages? To my mind loop oscillation is a function of the interaction of the components so having a "source" didn't seem to sit well.
Also, given this problem is common to both channels I was drawn to components common to both sides or maybe environmental. I'm working on the assumption that unless this amplifier has had the problem from new then something has changed and the same change in both channels (there are no mods apparent) would seem unlikely.
If the source or drain of the input transistor is showing the oscillation on the scope without modification, then you have loop oscillation.
I have to say removing the 1 meg resistor to the minus input of the op amp is a dingbat way of disabling it. A high gain point with no termination but a land on a pcb is a perfect oscillator.
One thing that has changed since 1985 is the amount of RF in the room. Do you put your cellphone out of the room when you are making these tests? And the cellphone companies have been building towers within 1 mile of everybody's home or business.
I have to say removing the 1 meg resistor to the minus input of the op amp is a dingbat way of disabling it. A high gain point with no termination but a land on a pcb is a perfect oscillator.
One thing that has changed since 1985 is the amount of RF in the room. Do you put your cellphone out of the room when you are making these tests? And the cellphone companies have been building towers within 1 mile of everybody's home or business.
Hi thinman 65, Sorry I wasn't able to reply earlier (work is draining me and taking up my free time you know).
Anyway, can you upload a high-resolution image of the output channels so we can all see what we're dealing with...do nothing in the meantime.
Anyway, can you upload a high-resolution image of the output channels so we can all see what we're dealing with...do nothing in the meantime.
Hi aparatusonline,
Whilst I'm here, a few thoughts:
I'm running exclusively via a DBT at present. Given both channels oscillate could the DBT be the cause? I've read some posts stating that amplifiers can become unstable when the rail voltages are below the intended voltages. I'm a little worried about removing the DBT. I assume that given I have no load connected and that the output transistors are absent that the current flow under oscillation is rail-to-rail through the driver transistors? If so there's two 47R resistors in that path so it's about 80/94 = 850mA hence a fairly strong glow on the DBT but maybe something those transistors could endure for a short period if testing without the DBT?
Secondly, there's an old post on the maker.pro site by a N_Cook maker.pro that appears very similar in subject, He found that in the preamp there's a RC timer circuit for what appears to be a start-up mute and also a zener circuit that does some kind of mute. However, he states that a side effect of these is to inject a 3-5V DC pulse into the power amp for about 100ms around 4 seconds after power-on and he felt this might instigate oscillations. I've not yet checked mine but the timing looks about right.
He also advocated fitting 22uF filter caps across the 4 x 12v Zeners diodes that provide the 12v supply to the power amp differential stage. I have seen that once oscillation starts those rails are a mess - about 1v p-p of noise.
Could I clarify please? Is this a photo of the board, a circuit diagram or a scope trace?
Whilst I'm here, a few thoughts:
I'm running exclusively via a DBT at present. Given both channels oscillate could the DBT be the cause? I've read some posts stating that amplifiers can become unstable when the rail voltages are below the intended voltages. I'm a little worried about removing the DBT. I assume that given I have no load connected and that the output transistors are absent that the current flow under oscillation is rail-to-rail through the driver transistors? If so there's two 47R resistors in that path so it's about 80/94 = 850mA hence a fairly strong glow on the DBT but maybe something those transistors could endure for a short period if testing without the DBT?
Secondly, there's an old post on the maker.pro site by a N_Cook maker.pro that appears very similar in subject, He found that in the preamp there's a RC timer circuit for what appears to be a start-up mute and also a zener circuit that does some kind of mute. However, he states that a side effect of these is to inject a 3-5V DC pulse into the power amp for about 100ms around 4 seconds after power-on and he felt this might instigate oscillations. I've not yet checked mine but the timing looks about right.
He also advocated fitting 22uF filter caps across the 4 x 12v Zeners diodes that provide the 12v supply to the power amp differential stage. I have seen that once oscillation starts those rails are a mess - about 1v p-p of noise.