Well, often I find emitter resistors aren't matched well. I measure voltage across them if I am suspicious. But the best indicator of matched devices is at idle. The voltage drop across each resistor should agree with it's resistance if the currents are equal, if you want to get real picky. To be honest, I'll check for roughly the same voltage drop across the emitters. The problem is that many emitter resistors have a tempco that I wish was close to zero. "Metal plate" resistors might be the worst.
I only do this with transistors mounted that someone else or the factory built.
Once you begin pounding current through the transistors, the emitter resistors tend to force sharing more effectively.
I match power devices on a heat sink and look at the voltage drops across matched resistors in the collector circuits. Emitters to the common rail and bases fed from a common source through resistors. I let the temperature stabilize. Mostly an issue with mosfets, BJT outputs tend to behave if they are matched in beta at lower currents. Emitter resistors are enough to force shared currents at higher levels.
I only do this with transistors mounted that someone else or the factory built.
Once you begin pounding current through the transistors, the emitter resistors tend to force sharing more effectively.
I match power devices on a heat sink and look at the voltage drops across matched resistors in the collector circuits. Emitters to the common rail and bases fed from a common source through resistors. I let the temperature stabilize. Mostly an issue with mosfets, BJT outputs tend to behave if they are matched in beta at lower currents. Emitter resistors are enough to force shared currents at higher levels.
Because I can connect everything, get my hands out of it, and run the test. You have no idea how many times Ive fried something or gotten shocked moving an alligator clip when things are up and running. A screwdriver in the bias pot is risky enough for me. And I’ll run it that way with music on tube amps, looking a changing operating currents, simultaneously. The cathode resistors are small on fixed bias amps - not enough to force sharing and I want to see how far off they get, as well as relative drifts during a burn in period. I do put a bias port on those - a D-connector so I can used my 8-tube pre-wired monitor on everything. SS amps can be tested at war volume that way too. If there is an issue that could lead to current hogging or run away it will show up more with dynamic conditions than steady state. There was an issue once where one in a parallel set of outputs (QC triple) would spike under clipped music, but match within 5% sine wave and idle was fine. I ignored it, and it fried in the field. No more.
BTW - 3 guesses what amplifier it was.
BTW - 3 guesses what amplifier it was.
Last edited:
Ahhh, a tube amp.
I use cathode current monitor sockets and have insulated screwdrivers for all bias controls. On my amps I have jacks on the top by the tube with a common jack, the pots can be accessed on top also. For others, I will connect hand held DVMs. I look at static, then run test tones so I can see how things go with a stable signal, then drop back to quiesient again to see changes. This is normally more an issue with tube amplifiers and some tubes.
I have had SS amps that spike unexpectedly. Driving those hard did not tease the amp to fail. There are normally two causes, intermittent in the bias control crkt, or a marginally stable amplifier. I did have a couple with flakey (weird as hell) transistors. But those are those things that happen in your career you simply never forget.
The most difficult I had was CJ Premier 1 that would take out the entire set of outputs but ran fine most of the time. That's a whole pile of premium KT-88s every time, plus those 600V fuses (not cheap either). I found out it had been subjected to the line hitting 208 or similar (they weren't sure). Turned out the screen regulator heater bias resistor had gone flakey. Now how is that for a weird fault? That tube then had an intermittent leakage between heater and cathode and that is what would precipitate the failure. I went through three entire sets of KT-88 tubes, matched of course. It would run completely normally, then suddenly BANG! Then it was all over.
I can't guess what amplifier it was. What bit you?
I use cathode current monitor sockets and have insulated screwdrivers for all bias controls. On my amps I have jacks on the top by the tube with a common jack, the pots can be accessed on top also. For others, I will connect hand held DVMs. I look at static, then run test tones so I can see how things go with a stable signal, then drop back to quiesient again to see changes. This is normally more an issue with tube amplifiers and some tubes.
I have had SS amps that spike unexpectedly. Driving those hard did not tease the amp to fail. There are normally two causes, intermittent in the bias control crkt, or a marginally stable amplifier. I did have a couple with flakey (weird as hell) transistors. But those are those things that happen in your career you simply never forget.
The most difficult I had was CJ Premier 1 that would take out the entire set of outputs but ran fine most of the time. That's a whole pile of premium KT-88s every time, plus those 600V fuses (not cheap either). I found out it had been subjected to the line hitting 208 or similar (they weren't sure). Turned out the screen regulator heater bias resistor had gone flakey. Now how is that for a weird fault? That tube then had an intermittent leakage between heater and cathode and that is what would precipitate the failure. I went through three entire sets of KT-88 tubes, matched of course. It would run completely normally, then suddenly BANG! Then it was all over.
I can't guess what amplifier it was. What bit you?
The weirdest things I’ve ever seen on tube amps turned out to be h-k leakage related. None were ever fatal, but all of them defied explanation.
Of course you can guess…. The clues were all there. QC triple, marginally stable, does strange things on a variac….. I put in 2SD424’s in them, and add the nested current feedback used in the CS800, move the bootstrap cap from the driver to the output, and never have a bit of trouble out of them again…..
That Nak amp ought to be straightforward by comparison. I’ve rebuilt (and built) amps with up to 10 pairs of outputs per channel - and as long as they were full comp EF3 the only issues ever were due to fake MJ15024’s. Talk about horrible matching…..
Of course you can guess…. The clues were all there. QC triple, marginally stable, does strange things on a variac….. I put in 2SD424’s in them, and add the nested current feedback used in the CS800, move the bootstrap cap from the driver to the output, and never have a bit of trouble out of them again…..
That Nak amp ought to be straightforward by comparison. I’ve rebuilt (and built) amps with up to 10 pairs of outputs per channel - and as long as they were full comp EF3 the only issues ever were due to fake MJ15024’s. Talk about horrible matching…..
Naw, still doesn't come to mind. I've worked on so much over the years, I mentally throw stuff out as "mind garbage". I remember as soon as I see the <whatever>.
If you said what it is I'd probably whack myself in the forehead with my palm and say - of course! I remember a Pioneer that badly misbehaved when you increased voltage slowly, it had to come up fast or it would draw tons of current even if everything was fine.
The PA-7 II are pretty straight forward amps. Matching reduces distortion, but I haven't seen even a mess be unstable. I've had to rebuild units worked on by "plumbers". The bias will run away if you come down from too high, but aside from that.
If you said what it is I'd probably whack myself in the forehead with my palm and say - of course! I remember a Pioneer that badly misbehaved when you increased voltage slowly, it had to come up fast or it would draw tons of current even if everything was fine.
The PA-7 II are pretty straight forward amps. Matching reduces distortion, but I haven't seen even a mess be unstable. I've had to rebuild units worked on by "plumbers". The bias will run away if you come down from too high, but aside from that.
PL400, S2. I was just b*tching about them. They can STAY latched to the rail if brought up slowly. Or snap to the proper operating point once you get above 60VAC - depending on the polarity of the input offset voltage in the op amp front end. Adding local feedback around everything but the op amp prevents this failure mode - always snaps to zero quickly. Also mitigates the local instabilities in the composite fake PNP bank which will sometimes spike the current in one or more of the outputs under high output voltages at low current (No or high z load).
I wonder is Ragnar still on the case. This also happened to me when I was fiddling with my first amp but the sucker didn't discourage me. Back then there was no internet, no proper books nor any competent soul around me to point out my mistakes. Also, no proper parts were available back in the 80's when we were still in Communist Yugoslavia. Shabby 2N3773 made in EI Nish Yugoslavia was the only choice. MJ15003/4 were in my wet dreams only. Eh, bad times, never more !
Hi Chris.. Have you found the value to be adjust on the 2 trim pots version from your notes ? I have a PA-5EII with also two Pots Trim. I adjust it to 3.9VDC and 40mV... What is your notes says ?
Best regards from Denmark
Last edited:
You mean “deleted member 543346”? Perhaps way too thin skinned.
I bet the first amp you were fiddling with wasn’t anything like this Nak 7A. My first “repair job” had PNP Ge’s. And everybody told me I wasn’t qualified to fix it either.
We were all unqualified to fix our first things, and so proud when we managed to do it. But (hopefully) these were our own things. I grabbed things from the garbage to practice.
Those germanium based things really taught you a great deal about being aware of defects and funny issues! Silicon transistors are so much easier to deal with.
I was lucky enough to be mentored and apprenticed, went to school and further trained and mentored after that. That way no one else's equipment was ever at risk. I was never allowed to touch anything more complicated until I was ready, and the PA-7 isn't straight forward. These days this isn't available to techs in training. One massive problem is that the basics aren't being taught. People don't have a feel and assume this is easy to do correctly. It isn't unless you have training and experience, and even then it isn't always straight forward.
Just because something functions does not mean it is repaired correctly. You cannot assume the equipment meets specifications by listening to it no matter how good you feel your hearing is. I get "repaired" stuff in all the time exceeding 1 % THD that was passed by previous "golden eared" techs and "experts".
Those germanium based things really taught you a great deal about being aware of defects and funny issues! Silicon transistors are so much easier to deal with.
I was lucky enough to be mentored and apprenticed, went to school and further trained and mentored after that. That way no one else's equipment was ever at risk. I was never allowed to touch anything more complicated until I was ready, and the PA-7 isn't straight forward. These days this isn't available to techs in training. One massive problem is that the basics aren't being taught. People don't have a feel and assume this is easy to do correctly. It isn't unless you have training and experience, and even then it isn't always straight forward.
Just because something functions does not mean it is repaired correctly. You cannot assume the equipment meets specifications by listening to it no matter how good you feel your hearing is. I get "repaired" stuff in all the time exceeding 1 % THD that was passed by previous "golden eared" techs and "experts".
I don't think so, not yet anyway.
There are times a person knows enough to get so far, then they are at their end. I'm not sure, but maybe this is beyond his comfort level. Only he could say.
Some techs assume they are correct no matter the evidence before them. They'll hold onto whatever was working for them in the past and good luck getting them to do anything different. They get locked in, once that happens they will never solve whatever issue is at hand.
There are times a person knows enough to get so far, then they are at their end. I'm not sure, but maybe this is beyond his comfort level. Only he could say.
Some techs assume they are correct no matter the evidence before them. They'll hold onto whatever was working for them in the past and good luck getting them to do anything different. They get locked in, once that happens they will never solve whatever issue is at hand.
Yes... im sure that the bias must be 40mV but i don't sure about the other adjustment. I have adjust it to 3.9V but i love to hear other what they have or they original that still is factory is reading out on the TP3 that is adjusted by VR12. I have the original service manual for the PA-5EII but its only the one pot adjustment for the bias, so i guess its must be a another service manual somewhere that contain this adjustment info
Hi McJan,
No, there are no other manuals or schematics. I was warranty for them, a major shop. Very few ever required service despite a large number sold in my market area. I did receive all updates and service bulletins and was initially unaware of this version of the circuit. The first one I saw was not that long ago, a couple years maybe. I had to trace out the new circuit to figure out what the pot did. Then I had to measure a few to see what the average value was. There was no reason to measure this value under warranty, so we didn't.
I have run into a few now, and I am very surprised that Nakamichi did not have records for this circuit version. They were normally pretty good for disclosure and not keeping secrets (unlike many Japanese brands).
I have looked everywhere I can think of. What I would suggest is to centre the bias control, or put it where it normally would be (you can feel a small dent in the pot at that positions). Then use a variable resistor instead of R114L (the higher value of the parallel set). Adjust that pot until the measured bias current is close to where it should be. Remember, you have to start off cold with lower bias current than the set point or it may go into thermal runaway. I would suggest starting at 25 mV. After 1/2 hour, bring it up to 35 mV and once that settles, adjust for your target of 40 mV. If it overshoots, reduce. It will drift up and down a bit once you are in the right setting. When you've got this set, measure the voltage across R112 (150R) and let others know. I'll do the same once I get the next one in like that.
Measure the pot and install the closest value fixed resistor, then set your bias again using the bias control. This should get you extremely close, and there is a variance anyway from what I have found.
I hope this helps you. This method will get you to the correct adjustment and reliable repair. No question.
Note this amplifier is sensitive to AC input voltage, so apply 120 VAC or whatever is the specified mains voltage inthe manual for your area. This is for your tests. On final bias setting, adjust your mains to the normal voltages you see in your area.
No, there are no other manuals or schematics. I was warranty for them, a major shop. Very few ever required service despite a large number sold in my market area. I did receive all updates and service bulletins and was initially unaware of this version of the circuit. The first one I saw was not that long ago, a couple years maybe. I had to trace out the new circuit to figure out what the pot did. Then I had to measure a few to see what the average value was. There was no reason to measure this value under warranty, so we didn't.
I have run into a few now, and I am very surprised that Nakamichi did not have records for this circuit version. They were normally pretty good for disclosure and not keeping secrets (unlike many Japanese brands).
I have looked everywhere I can think of. What I would suggest is to centre the bias control, or put it where it normally would be (you can feel a small dent in the pot at that positions). Then use a variable resistor instead of R114L (the higher value of the parallel set). Adjust that pot until the measured bias current is close to where it should be. Remember, you have to start off cold with lower bias current than the set point or it may go into thermal runaway. I would suggest starting at 25 mV. After 1/2 hour, bring it up to 35 mV and once that settles, adjust for your target of 40 mV. If it overshoots, reduce. It will drift up and down a bit once you are in the right setting. When you've got this set, measure the voltage across R112 (150R) and let others know. I'll do the same once I get the next one in like that.
Measure the pot and install the closest value fixed resistor, then set your bias again using the bias control. This should get you extremely close, and there is a variance anyway from what I have found.
I hope this helps you. This method will get you to the correct adjustment and reliable repair. No question.
Note this amplifier is sensitive to AC input voltage, so apply 120 VAC or whatever is the specified mains voltage inthe manual for your area. This is for your tests. On final bias setting, adjust your mains to the normal voltages you see in your area.