If it occurs with temperature then that could be an issue. Also check the continuity between the solder pin of the cap and the next component leg on the track. There could be a hair line break in the trace.
Just an update: I found out I had an issue with my wall socket having an open neutral (ground to neutral showed around a fluctuating 120v). It just happened to be the only socket defective in my house and also happened to be the one socket I plugged all my audio equipment in. Btw, It should be noted I did test my amp on another socket after I noticed the humming, and the problem remained. I don't know how long it was like that so I don't know if it's a coincidence or not, nor if it caused damage to my amp - but talk about terrible luck! The wall socket has now been changed and everything works on that end, but the problem remains with the amp so now I'm thinking there may have been damaged components of which I am still diagnosing.
Being that I am not an electrical engineer - what sort of problems would be expected when connecting an amplifier and an open neutral situation? I am going to assume pretty much what I have been experiencing, but I'll need the experts to weigh in on this one. Could my transformers be damaged as a result of this?
Being that I am not an electrical engineer - what sort of problems would be expected when connecting an amplifier and an open neutral situation? I am going to assume pretty much what I have been experiencing, but I'll need the experts to weigh in on this one. Could my transformers be damaged as a result of this?
The transformers seem to be at fault - all due to a faulty wall socket that had an open neutral. Does anyone have any reference or idea on what sort of specification of replacement toroids I would need? I'm looking at transformers from Toroidy in Poland as a replacement... Thanks again!
Well toroidal transformers are more coupled, is if you do have DC somehow then that's really going to show up. My A220 original toridal transformer hums slight but it's not worryingly bad. That was made by Toroidal International IIRC.
To test the transformers I would first test the power coming in - buzzing is generally attributed to DC on the input. Also if the transformers are getting hot (in the case of suspect 'naked' transformers don't use your hand but use a laser spot thermometer).
As a starting point - take each transformer and test out of circuit and unpowered, the DC resistance as a starting point. If the resistance varies massively between mirrored windings then something could be wrong in that winding. I don't think there's an internal fuse blown because you should have some resistance and you have power.
Also you should have some shielding bands wound into the transformer - those will have grounding, so check the grounding points for any shielding and that they have good grounding.
Next power up (not connected to the amp) and carefully measure the voltage for each winding. I would be a little concerned doing this on the mains side with a simple multimeter - I'd use clip probes incase there's any transients around. Clip, then power up, test on AC setting, switch off, move clips. Note you may get a voltage spike from the flux collapsing when you power off.
The no-load voltage will be higher than connected with a load, that's normal, but you're looking for any oddities in voltage (ie that style of amp should have two mirrored voltages approximately the same), the average vs true RMS isn't that important - we're mainly looking for obvious issues such as resistances being different or voltages being different that show something is bad in the windings.
If you have a current loop multimeter - connect that around one of the wires of each section when the amp is running to check for asymmetric current draw. Don't put your multimeter in 10A current mode in series with the transformer as the inrush will most likely blow the fuse. If not another option is to use a high power current sense resistor and just measure either side of it (or if there's a resistor in the RC section you could use that too in the same way as a shunt). Measure the voltage across the resistor using V=IR to find the current.
To test the transformers I would first test the power coming in - buzzing is generally attributed to DC on the input. Also if the transformers are getting hot (in the case of suspect 'naked' transformers don't use your hand but use a laser spot thermometer).
As a starting point - take each transformer and test out of circuit and unpowered, the DC resistance as a starting point. If the resistance varies massively between mirrored windings then something could be wrong in that winding. I don't think there's an internal fuse blown because you should have some resistance and you have power.
Also you should have some shielding bands wound into the transformer - those will have grounding, so check the grounding points for any shielding and that they have good grounding.
Next power up (not connected to the amp) and carefully measure the voltage for each winding. I would be a little concerned doing this on the mains side with a simple multimeter - I'd use clip probes incase there's any transients around. Clip, then power up, test on AC setting, switch off, move clips. Note you may get a voltage spike from the flux collapsing when you power off.
The no-load voltage will be higher than connected with a load, that's normal, but you're looking for any oddities in voltage (ie that style of amp should have two mirrored voltages approximately the same), the average vs true RMS isn't that important - we're mainly looking for obvious issues such as resistances being different or voltages being different that show something is bad in the windings.
If you have a current loop multimeter - connect that around one of the wires of each section when the amp is running to check for asymmetric current draw. Don't put your multimeter in 10A current mode in series with the transformer as the inrush will most likely blow the fuse. If not another option is to use a high power current sense resistor and just measure either side of it (or if there's a resistor in the RC section you could use that too in the same way as a shunt). Measure the voltage across the resistor using V=IR to find the current.
So I know it's been awhile but there has been progress and I hope you guys can help along again!
It seems I may have fixed the left channel. I reflowed every component and it seems to be working - I suspect the troubled area was the bridge rectifiers. This seems to be problematic areas by Musical Fidelity on various amps, whereby the BR's tend to lack a good amount of solder above and below the circuit board resulting in poor connection; along with the fact these are 'pressed' and bolted down on the underside of the case, and the fact I was wiggling the large capacitors out which are in close proximity of the BR's, may have been the problem all along.
With that said, in my seemingly never-ending quest, the right channel seems to have either shorted by me or some other problem on the board (if it was by me, it was via me pulling out the RCA cables while the amp was on). Fuses on the right channel blew and I did notice a small amount of smoke in the vicinity of the input stage (ie. near the three smaller capacitors and input signal wiring to the board). No physical burn marks, no broken legs on any transistors, and all resistors seem to be measuring ok. All transistors on the right channel have been checked as well and seem to test ok (in circuit) - ie. show comparable values to the left channel.
Two of the three sets of PNP/NPN output transistors were toast and I have managed to get replacements. I have purchased a Variac and I have noticed that the TR16, TR22 and TR28 (PNP) start to run hot when setting the Variac is set at 45V and measuring around the 23v - 25v at the output transistor leg 1 position. All voltages seem nominal and comparable to the left channel up to this point on the variac. Any more and I see that the voltages become unstable and the heat results. The NPN output transistors however run much cooler, and seem to be okay.
Any suggestions on how to proceed? I have taken it to a shop and they had been working on it for a few days and were dumbfounded. Again, everything else measured well and it seems the transformers were ok as well. Really close to fixing this so i'm hoping you guys can help!!
It seems I may have fixed the left channel. I reflowed every component and it seems to be working - I suspect the troubled area was the bridge rectifiers. This seems to be problematic areas by Musical Fidelity on various amps, whereby the BR's tend to lack a good amount of solder above and below the circuit board resulting in poor connection; along with the fact these are 'pressed' and bolted down on the underside of the case, and the fact I was wiggling the large capacitors out which are in close proximity of the BR's, may have been the problem all along.
With that said, in my seemingly never-ending quest, the right channel seems to have either shorted by me or some other problem on the board (if it was by me, it was via me pulling out the RCA cables while the amp was on). Fuses on the right channel blew and I did notice a small amount of smoke in the vicinity of the input stage (ie. near the three smaller capacitors and input signal wiring to the board). No physical burn marks, no broken legs on any transistors, and all resistors seem to be measuring ok. All transistors on the right channel have been checked as well and seem to test ok (in circuit) - ie. show comparable values to the left channel.
Two of the three sets of PNP/NPN output transistors were toast and I have managed to get replacements. I have purchased a Variac and I have noticed that the TR16, TR22 and TR28 (PNP) start to run hot when setting the Variac is set at 45V and measuring around the 23v - 25v at the output transistor leg 1 position. All voltages seem nominal and comparable to the left channel up to this point on the variac. Any more and I see that the voltages become unstable and the heat results. The NPN output transistors however run much cooler, and seem to be okay.
Any suggestions on how to proceed? I have taken it to a shop and they had been working on it for a few days and were dumbfounded. Again, everything else measured well and it seems the transformers were ok as well. Really close to fixing this so i'm hoping you guys can help!!
Last edited:
I forgot to mention that the only other area of concern might be the 4 small orange ceramic caps located at C4, C6, C7, and C10. I am considering just replacing everything at this point, but let me know what you guys think before I start ordering parts.
If there's a large current then it's very likely to be:
a) there's a short or partial short
b) the bias is set incorrectly
Probably worth picking a set of transistor points on each then measuring R then L and comparing - that would indicate if it's a bias issue (ie voltages wrong on the transistors), then form that you can see what is causing the bias voltage to be out by tracing that back into the power/bias circuit.
Looking at the power amp circuit, pin 1 of TR15 etc (ie the power transistors) is the base. If you put a sine wave into it what does the signal look like? when it's on the variac and not burning up. It may be you're getting an oscillation which could be causing a large current flow.
I would be measuring the voltages across the resistors such as the output devices (ie between pin 3 and pin 5), and the bias current resistors (ie the 1K resistors such as R22, R23, R30, R31). Plus R16 and R14 to see how the current mirror is working, and look at D3 to see if it's conducting and voltages. Check the diode chain D1 and D2.
You could have fried a resistor, or a diode etc. In the end I'd look at those as caps aren't normally a big issue and ceramics are normally ok (you could DMM them to see if they are a short).
Check C6 - this is the bypass for the NFB resistor.
Basically looking for shorts and odd current flows through resistors.
a) there's a short or partial short
b) the bias is set incorrectly
Probably worth picking a set of transistor points on each then measuring R then L and comparing - that would indicate if it's a bias issue (ie voltages wrong on the transistors), then form that you can see what is causing the bias voltage to be out by tracing that back into the power/bias circuit.
Looking at the power amp circuit, pin 1 of TR15 etc (ie the power transistors) is the base. If you put a sine wave into it what does the signal look like? when it's on the variac and not burning up. It may be you're getting an oscillation which could be causing a large current flow.
I would be measuring the voltages across the resistors such as the output devices (ie between pin 3 and pin 5), and the bias current resistors (ie the 1K resistors such as R22, R23, R30, R31). Plus R16 and R14 to see how the current mirror is working, and look at D3 to see if it's conducting and voltages. Check the diode chain D1 and D2.
You could have fried a resistor, or a diode etc. In the end I'd look at those as caps aren't normally a big issue and ceramics are normally ok (you could DMM them to see if they are a short).
Check C6 - this is the bypass for the NFB resistor.
Basically looking for shorts and odd current flows through resistors.
Preliminary testing: R22, R23, R30, R31 and R16 R14 all mirror the left channel and are consistent. Same with D3 and D1 and D2. I checked all the other diodes and seem to be all mirroring the same values.
I could not check C6 as the value was too low for my multimeter (I confirmed this with the working left channel and could not get a reading) but I did check the other ceramics and they seem to be at correct values - so no need for replacement.
The shop I took it to seemed to have checked all of the transistors on the right (broken) channel as they came back bent in a direction so that readings could be taken with a meter.
I will have to get back to you regarding the sine wave at low voltage as I do not have an oscilloscope at the moment. In the meantime I can relay what the shop told me and they said that everything looks great at low power and is reading and functioning correctly up until 25v. They were also using just a single pair of output transistors and even then it was drawing too much current. Increasing the voltages beyond that 23-25v is when voltages start going erratic at the output transistors and the heat picks up significantly.
It should also be mentioned that they released the amplifier to me at their loss - in other words, they put many more hours into it then what they charged me, so it was in their interest to get it fixed properly.
Could the main relay be at fault by chance? Perhaps a fried potentiometer? It just seems so odd that nothing is reading or looks physically out of the ordinary on the board... or at least so far.
I could not check C6 as the value was too low for my multimeter (I confirmed this with the working left channel and could not get a reading) but I did check the other ceramics and they seem to be at correct values - so no need for replacement.
The shop I took it to seemed to have checked all of the transistors on the right (broken) channel as they came back bent in a direction so that readings could be taken with a meter.
I will have to get back to you regarding the sine wave at low voltage as I do not have an oscilloscope at the moment. In the meantime I can relay what the shop told me and they said that everything looks great at low power and is reading and functioning correctly up until 25v. They were also using just a single pair of output transistors and even then it was drawing too much current. Increasing the voltages beyond that 23-25v is when voltages start going erratic at the output transistors and the heat picks up significantly.
It should also be mentioned that they released the amplifier to me at their loss - in other words, they put many more hours into it then what they charged me, so it was in their interest to get it fixed properly.
Could the main relay be at fault by chance? Perhaps a fried potentiometer? It just seems so odd that nothing is reading or looks physically out of the ordinary on the board... or at least so far.
The low power - high power problem could be:
a) the operating point (are the transistors replaced precisely the same part and not a lower wattage/voltage version?) this means at the same bias and B+ voltages cause too much current leading to thermal runaway.
b) thermal conduction causing a thermal runaway on the BJT (I destroyed a pair of BJTs on my hybrid amp due to too high voltage and thermal runaway - BJTs as they get hot allow more current that cause more heat.. and thus uncontrolled thermal increase and current until the device fails).
The relay (ie the output relay) could have a short - you should be able to test that with continuity/resistance. If there's a short between the pins then that would be easy to spot. Also check the DC resistance across L1 matches the other channel - you don't have an inductance LCR meter so a simple test is what we can do to check it.
If all the parts are correct in terms of mirroring until a certain point then:
a) operating points - bias/parts/adjusted bias pot (and that's not broken?).
b) breakdown at higher voltages - partially damaged/fake parts/cold solder joint opening due to thermal expansion.
c) oscillations - these shouldn't be a problem if the other side is working but if one of the components isn't functioning it could be causing an oscillation causing higher load.
As you've tested the voltages across the resistors (V=IR thus current I = Vacross/Reistance), then the current through are matching.
Going back to what you have said.
If the output SAPs are getting hot and one is cold - this could be that one is dead (cold) and causing an imbalance on the others or one of the other transistors is dead/failing.
a) the operating point (are the transistors replaced precisely the same part and not a lower wattage/voltage version?) this means at the same bias and B+ voltages cause too much current leading to thermal runaway.
b) thermal conduction causing a thermal runaway on the BJT (I destroyed a pair of BJTs on my hybrid amp due to too high voltage and thermal runaway - BJTs as they get hot allow more current that cause more heat.. and thus uncontrolled thermal increase and current until the device fails).
The relay (ie the output relay) could have a short - you should be able to test that with continuity/resistance. If there's a short between the pins then that would be easy to spot. Also check the DC resistance across L1 matches the other channel - you don't have an inductance LCR meter so a simple test is what we can do to check it.
If all the parts are correct in terms of mirroring until a certain point then:
a) operating points - bias/parts/adjusted bias pot (and that's not broken?).
b) breakdown at higher voltages - partially damaged/fake parts/cold solder joint opening due to thermal expansion.
c) oscillations - these shouldn't be a problem if the other side is working but if one of the components isn't functioning it could be causing an oscillation causing higher load.
As you've tested the voltages across the resistors (V=IR thus current I = Vacross/Reistance), then the current through are matching.
Going back to what you have said.
If the output SAPs are getting hot and one is cold - this could be that one is dead (cold) and causing an imbalance on the others or one of the other transistors is dead/failing.
So in regards to the replacement output transistors, there is nothing to indicate that they are different aside from the date codes and Musical Fidelity seems to have placed a red dot on the top left corner of each while the replacements do not have this. There is one original pair that I will solder back in to confirm they run hot and act the same as the replacements.
Thermal conduction runway is a possibility as I am running the outputs without the heatsink attached. However, the left working channel is fairly cool to the touch while the right can get to the point where you cannot even touch them (again, no problem when voltages are below the 23-25v threshold). When I mentioned the NPN side of output transistors run cooler than the hot PNP side, that still seems to be the case, but they are still much warmer than all of the outputs compared to the left working channel.
I have checked the relay and continuity exists. I did check L1 but that seems to be a bit more difficult to measure as it has a coil around it - it still measures the same as the opposite channel.
I have not touched the bias pot as of yet, nor do I know if it has been affected by the short. I am not too familiar how to properly adjust this and they do seem to be very sensitive to changes from what I've read. I assume this will be the next step?
Thermal conduction runway is a possibility as I am running the outputs without the heatsink attached. However, the left working channel is fairly cool to the touch while the right can get to the point where you cannot even touch them (again, no problem when voltages are below the 23-25v threshold). When I mentioned the NPN side of output transistors run cooler than the hot PNP side, that still seems to be the case, but they are still much warmer than all of the outputs compared to the left working channel.
I have checked the relay and continuity exists. I did check L1 but that seems to be a bit more difficult to measure as it has a coil around it - it still measures the same as the opposite channel.
I have not touched the bias pot as of yet, nor do I know if it has been affected by the short. I am not too familiar how to properly adjust this and they do seem to be very sensitive to changes from what I've read. I assume this will be the next step?
Another update:
I have reflowed every single component on the right channel and rechecked all electrolytics for any failure - they all tested ok and good spec.
When I powered up with variac, I noticed that right channel TR16, TR22, TR28 (PNP) output transistors still getting much warmer and faster than the NPN equivalents on the same channel.
Further, I noticed that the voltages when taken on leg 1 of each transistor are lower than comparing on the left working channel. For example, when taking a reading on TR16 on leg one, it will read 22v (and getting hot) whereby checking TR116 on leg one, it will show over 25v and cool to the touch. I'm going to assume this is indication of voltage droop from a short or perhaps a bias problem.
Looks like oscilloscope will have to be ordered and more head scratching is in order...
I have reflowed every single component on the right channel and rechecked all electrolytics for any failure - they all tested ok and good spec.
When I powered up with variac, I noticed that right channel TR16, TR22, TR28 (PNP) output transistors still getting much warmer and faster than the NPN equivalents on the same channel.
Further, I noticed that the voltages when taken on leg 1 of each transistor are lower than comparing on the left working channel. For example, when taking a reading on TR16 on leg one, it will read 22v (and getting hot) whereby checking TR116 on leg one, it will show over 25v and cool to the touch. I'm going to assume this is indication of voltage droop from a short or perhaps a bias problem.
Looks like oscilloscope will have to be ordered and more head scratching is in order...
Since I have tested all the resistors and they measure and seem to function just like the left channel - along with the repair tech who checked all of the transistors... Just curious, how likely could it be the choke regulator causing the issue? Would it be 'safe' to remove the choke regulator and test how the output transistors perform? Or would I have to have one in circuit thereby swapping the left channel for the right one in order to test this?
btw - I'm sorry if this is a stupid question but I'm running out of idea's
p.s - I did swap in the original working pair of SAP15's and they function the same as the replacements. So I doubt the replacements are incorrect/fake.
EDIT nevermind - just swapped the CR's and it didn't change anything. The output transistors still climb to about 23v and if you keep raising the voltage, they get a bit higher and then start to drop - all the while getting very hot and I have to shut down. The search continues.
btw - I'm sorry if this is a stupid question but I'm running out of idea's
p.s - I did swap in the original working pair of SAP15's and they function the same as the replacements. So I doubt the replacements are incorrect/fake.
EDIT nevermind - just swapped the CR's and it didn't change anything. The output transistors still climb to about 23v and if you keep raising the voltage, they get a bit higher and then start to drop - all the while getting very hot and I have to shut down. The search continues.
Last edited:
If it has a choke for each channel you should see if the voltage output is low for DC.
Have you checked the volume pot output to check both channels are equal?
Have you checked the volume pot output to check both channels are equal?
I checked DC on the right channel and it does seem low - hovering around the 0v mark
I have not checked the volume output and I will add it to the list although I have adjusted and measured the bias... seems to be functioning although the adjustment didn't lower/raise voltages all that much. Btw, the SAP15's do read a little high on the (broken) right channel, PNP side. It should be set and read around 0.8v from what I gather, but they are at the 1.2v range. Left channel reads fairly consistent and is at the 0.7 - 0.9v (measurements taken between 4 and 5 on transistor leg)
The strangest problem is setting the variac to the 40v-45v mark where the output transistors hold steady at 20v. If I were to put the voltage up to 60-70v, the output on leg 1 voltage reading is erratic and jumps from 20v to 100mv to -15mv and back up again etc. When I bring it back down to 40v, it stabilizes and runs at normal temperature. It's when I bring up the voltage so that the output on leg 1 is around 23v mark that it gets very hot (albeit, stable voltage) and beyond that, it gets erratic again.
If it's not resistors (all of them were checked and double checked) and it doesn't seem to be transistors (repair shop confirmed they checked all of them in circuit) and it's not the output transistors and it seems power supply is fine (all caps check out) and the transformers seem to be just fine and I can confirm the relays click on.... How is this amplifier not working? oh lord, what a pain - haha, some strange electrical gremlin it seems
I have not checked the volume output and I will add it to the list although I have adjusted and measured the bias... seems to be functioning although the adjustment didn't lower/raise voltages all that much. Btw, the SAP15's do read a little high on the (broken) right channel, PNP side. It should be set and read around 0.8v from what I gather, but they are at the 1.2v range. Left channel reads fairly consistent and is at the 0.7 - 0.9v (measurements taken between 4 and 5 on transistor leg)
The strangest problem is setting the variac to the 40v-45v mark where the output transistors hold steady at 20v. If I were to put the voltage up to 60-70v, the output on leg 1 voltage reading is erratic and jumps from 20v to 100mv to -15mv and back up again etc. When I bring it back down to 40v, it stabilizes and runs at normal temperature. It's when I bring up the voltage so that the output on leg 1 is around 23v mark that it gets very hot (albeit, stable voltage) and beyond that, it gets erratic again.
If it's not resistors (all of them were checked and double checked) and it doesn't seem to be transistors (repair shop confirmed they checked all of them in circuit) and it's not the output transistors and it seems power supply is fine (all caps check out) and the transformers seem to be just fine and I can confirm the relays click on.... How is this amplifier not working? oh lord, what a pain - haha, some strange electrical gremlin it seems
Last edited:
For bias refer to page 2 of the attached Allegro Sanken SAP15P/N datasheet.
The built-in emitter resistor between pins S & E is 0.22Ω ref. page 1, Line up
TR27 SAP15NY NPN TO-3P5 (should be the same as TR21/15 below, but confirm for this Darlington)
TR21 SAP15NY NPN TO-3P5 measure across pins 4 & 5 S --> E
TR15 SAP15NY NPN TO-3P5 measure across pins 4 & 5 S --> E
TR16 SAP15PY PNP TO-3P5 measure across pins 2 & 1 S --> E
TR22 SAP15PY PNP TO-3P5 measure across pins 2 & 1 S --> E
TR28 SAP15PY PNP TO-3P5 (should be the same as TR16/22 above, but confirm for this Darlington)
Iq quiescent current is 40mA, so the voltage between S & E (NPN) should be 0.0088V (8.8mV).
I set bias with 8 ohm load attached as per the diagram on page 2. I used 2x 4ohm 100W resistors in series.
I'd carefully check and confirm those bias measurements and settings.
When you mention, "any more [on the variac] and I see the voltages become unstable ... can you provide more information around the anomaly with voltage instability?
I expect the +VE and -VE rails will be close to +60V and -60V under normal operating conditions.
MF use a 'capacitance multiplier' with TR5 ZTX653 in shunt and C13 100uF on its base to reduce ripple +ve rail.
TR4 /C14 -ve rail for the R/H channel.
L/H channel it is TR105/TR114 in the same configuration.
Might be worth comparing voltages up to the safe point of 45V on the variac.
The built-in emitter resistor between pins S & E is 0.22Ω ref. page 1, Line up
TR27 SAP15NY NPN TO-3P5 (should be the same as TR21/15 below, but confirm for this Darlington)
TR21 SAP15NY NPN TO-3P5 measure across pins 4 & 5 S --> E
TR15 SAP15NY NPN TO-3P5 measure across pins 4 & 5 S --> E
TR16 SAP15PY PNP TO-3P5 measure across pins 2 & 1 S --> E
TR22 SAP15PY PNP TO-3P5 measure across pins 2 & 1 S --> E
TR28 SAP15PY PNP TO-3P5 (should be the same as TR16/22 above, but confirm for this Darlington)
Iq quiescent current is 40mA, so the voltage between S & E (NPN) should be 0.0088V (8.8mV).
I set bias with 8 ohm load attached as per the diagram on page 2. I used 2x 4ohm 100W resistors in series.
I'd carefully check and confirm those bias measurements and settings.
When you mention, "any more [on the variac] and I see the voltages become unstable ... can you provide more information around the anomaly with voltage instability?
I expect the +VE and -VE rails will be close to +60V and -60V under normal operating conditions.
MF use a 'capacitance multiplier' with TR5 ZTX653 in shunt and C13 100uF on its base to reduce ripple +ve rail.
TR4 /C14 -ve rail for the R/H channel.
L/H channel it is TR105/TR114 in the same configuration.
Might be worth comparing voltages up to the safe point of 45V on the variac.
Protegimus, yes you are correct, I did mean to say that I was aiming for around the 8.0mV and not 0.7-0.9v as I originally stated. I did only use a 100w 8ohm load however, I suppose 200watt load might give more accurate results. In any case, I will check this again.
I can confirm that the voltage rails seem to be acting properly at their proper load and no load specifications (at reduced power and comparing both channels). I think you may be on to something in checking the transistors, despite them showing consistent readings in circuit using my multimeter transistor-test setting. I am in belief now that when power is increased on the variac to a certain point, the increase in current causes the faulty transistor(s) to increase load through a (partial) short. This may account for the steady voltage readings on the output transistors, but beyond a certain power level, the partial short causes large current draw causing the overheating, to finally going into a full shorting of the circuit which is causing the unstable/erratic readings at power levels beyond 45v on the variac.
I have a feeling I will just have to 'shotgun' and replace pretty much all transistors on the right channel. Now I will have to go on a search to find suitable replacements if that's even possible
In the meantime, I will compare the voltages at safe voltages to see if there are discrepancies as you suggest, and gradually increase power to see what happens to those values.
I can confirm that the voltage rails seem to be acting properly at their proper load and no load specifications (at reduced power and comparing both channels). I think you may be on to something in checking the transistors, despite them showing consistent readings in circuit using my multimeter transistor-test setting. I am in belief now that when power is increased on the variac to a certain point, the increase in current causes the faulty transistor(s) to increase load through a (partial) short. This may account for the steady voltage readings on the output transistors, but beyond a certain power level, the partial short causes large current draw causing the overheating, to finally going into a full shorting of the circuit which is causing the unstable/erratic readings at power levels beyond 45v on the variac.
I have a feeling I will just have to 'shotgun' and replace pretty much all transistors on the right channel. Now I will have to go on a search to find suitable replacements if that's even possible
In the meantime, I will compare the voltages at safe voltages to see if there are discrepancies as you suggest, and gradually increase power to see what happens to those values.
Progress report: Good news!! I think I may have fixed the problem on the right channel but further testing is needed
I decided to replace TR1 and TR2 (MPSA93) and TR4 and TR5 (MPSA42) where we believed the problem to be stemming from but I also replaced TR3 (MPSA93) and TR6 (MJE350) along with TR7 (MJE340) while I was in there to be sure. Voltages are now stable at full power and I do not have massive heat loading at low voltages!
Since TR1 and TR2 seemed to be crucial (MF even makes a side note that they must be matched), lo and behold I measured out of circuit and the hFe values were not matched and out of spec. Everything else seemed a little inconsistent but nothing too noticeable.
It should be noted I am only running two output transistors (SAP15) at the moment but I decided to bias the amp with no load to get a proper baseline while the amp warmed up for about 10 min.
I ended up with these values and holding steady at full 120v power:
(Bias readings)
8.82mV on PNP
8.55mv on NPN
-I also measured around 2.2mV DC on speaker output terminals and 6.7mV AC (if that's of any use) at idle with no load.
My next step is obviously install the remainder output transistors and re-test and re-bias if needed. What else would you guys recommend? If those values look good - I suppose hooking up a speaker is next!
I decided to replace TR1 and TR2 (MPSA93) and TR4 and TR5 (MPSA42) where we believed the problem to be stemming from but I also replaced TR3 (MPSA93) and TR6 (MJE350) along with TR7 (MJE340) while I was in there to be sure. Voltages are now stable at full power and I do not have massive heat loading at low voltages!
Since TR1 and TR2 seemed to be crucial (MF even makes a side note that they must be matched), lo and behold I measured out of circuit and the hFe values were not matched and out of spec. Everything else seemed a little inconsistent but nothing too noticeable.
It should be noted I am only running two output transistors (SAP15) at the moment but I decided to bias the amp with no load to get a proper baseline while the amp warmed up for about 10 min.
I ended up with these values and holding steady at full 120v power:
(Bias readings)
8.82mV on PNP
8.55mv on NPN
-I also measured around 2.2mV DC on speaker output terminals and 6.7mV AC (if that's of any use) at idle with no load.
My next step is obviously install the remainder output transistors and re-test and re-bias if needed. What else would you guys recommend? If those values look good - I suppose hooking up a speaker is next!
Glad to hear it! Bias and DC levels are good.
I expect it's in your plan, but check once again once all the Darlingtons are back in circuit.
For the final bias check I'd play the amp for ~2hrs to allow it to fully warm up, then measure.
At some point it will be worth getting matched pairs of MPSA93s - let me know if you have a good source as I'll probably complete the same exercise on my A308.
Hopefully some very enjoyable listening time coming soon!
I expect it's in your plan, but check once again once all the Darlingtons are back in circuit.
For the final bias check I'd play the amp for ~2hrs to allow it to fully warm up, then measure.
At some point it will be worth getting matched pairs of MPSA93s - let me know if you have a good source as I'll probably complete the same exercise on my A308.
Hopefully some very enjoyable listening time coming soon!