Well lets focus on the Q116/117 issue for now.
Based on your test of the new A1492 transistor you have in order to begin to bias up the output stage you need somewhere around 550mV drop from B to E on q120 - q133. So you'll need that much drop across R124 & R125 for the output transistors to begin conducting. That comes to about 183mA through Q116 and 183mA through Q117. So if Q116/Q117 is conducting you should measure 183mV or more across the 1ohm emitter resistors R126 & 183mV across R127.
So at around .55V on the base of Q116/Q117 you should begin to see a voltage drop appear across R126 & across R127 if its working properly. The corresponding voltage drop across R124 & R125 will be about 3X higher and may be the better place to measure. You have more than enough base voltage at Q116/Q117 to begin conducting and enough additional voltage range to get the full drop of 183mV across R126 or R127. I would check the mV drop across R126 and R127 and see which of these two is conducting and how much (mV) with the bias pot at max position.
FYI for measuring these voltages you need to measure directly across the resistor just like you're measuring resistance only with your meter set for either DCmV or DCV.
The voltage readings you're getting with your prior method are too high and using all of your available digits up on your meter so you cannot get enough resolution by measuring with one probe on ground and the positive probe on each test point the way you've been doing it. I generally do not worry about the polarity of a voltage drop reading either because I'm just specifically looking to see what the drop is, so neg or pos it doesn't really matter to me unless I suspect something else is going on.
Based on your test of the new A1492 transistor you have in order to begin to bias up the output stage you need somewhere around 550mV drop from B to E on q120 - q133. So you'll need that much drop across R124 & R125 for the output transistors to begin conducting. That comes to about 183mA through Q116 and 183mA through Q117. So if Q116/Q117 is conducting you should measure 183mV or more across the 1ohm emitter resistors R126 & 183mV across R127.
So at around .55V on the base of Q116/Q117 you should begin to see a voltage drop appear across R126 & across R127 if its working properly. The corresponding voltage drop across R124 & R125 will be about 3X higher and may be the better place to measure. You have more than enough base voltage at Q116/Q117 to begin conducting and enough additional voltage range to get the full drop of 183mV across R126 or R127. I would check the mV drop across R126 and R127 and see which of these two is conducting and how much (mV) with the bias pot at max position.
FYI for measuring these voltages you need to measure directly across the resistor just like you're measuring resistance only with your meter set for either DCmV or DCV.
The voltage readings you're getting with your prior method are too high and using all of your available digits up on your meter so you cannot get enough resolution by measuring with one probe on ground and the positive probe on each test point the way you've been doing it. I generally do not worry about the polarity of a voltage drop reading either because I'm just specifically looking to see what the drop is, so neg or pos it doesn't really matter to me unless I suspect something else is going on.
I made a mistake in my earlier statement.
I said "if Q116/Q117 is conducting you should measure 183mV or more across the 1ohm emitter resistors R126 & 183mV across R127."
That was a badly worded statement. I stated it better in the second paragraph. What I should've said in the first paragraph was, If q116/q117 are conducting then you'll begin to show a mV drop across resistors R124,125,126,127. When the voltage drop across R126/R127 gets to approx 180mV then you will have approx the voltage drop you need to begin turning on the output transistors. Thats the point that the DBT will start to glow more brightly as you turn up the bias.
Sorry if I caused any confusion.
I said "if Q116/Q117 is conducting you should measure 183mV or more across the 1ohm emitter resistors R126 & 183mV across R127."
That was a badly worded statement. I stated it better in the second paragraph. What I should've said in the first paragraph was, If q116/q117 are conducting then you'll begin to show a mV drop across resistors R124,125,126,127. When the voltage drop across R126/R127 gets to approx 180mV then you will have approx the voltage drop you need to begin turning on the output transistors. Thats the point that the DBT will start to glow more brightly as you turn up the bias.
Sorry if I caused any confusion.
It's been a busy weekend here, so I haven't had much time to look at the amp, so sorry for the delayed response. Measuring across R126 and R127 and with my DMM in mV mode, I get:
R126, bias pot to the left: 0mV
R126, bias pot to the right: 0mV
R127, bias pot to the left: 0mV
R127, bias pot to the right: 0mV
I think I have one of those super crappy Harbor Freight DMMs somewhere around. My limitation right now is the lack of alligator clips. I need to make up some more.
R126, bias pot to the left: 0mV
R126, bias pot to the right: 0mV
R127, bias pot to the left: 0mV
R127, bias pot to the right: 0mV
I think I have one of those super crappy Harbor Freight DMMs somewhere around. My limitation right now is the lack of alligator clips. I need to make up some more.
No problem. I've been busy here too up until today anyway. Lots of rain here in DFW today.
So either there's an open circuit somewhere or one or both of the transistors Q116/117 are bad.
If you measure the resistance from the + DC rail to the collector of Q116 and you get approx 3ohms. Then measure the resistance from the Emitter of Q116 to the Emitter of Q117 and get approx 2 ohms. Then from the collector of Q117 to the - DC rail measures another 3 ohms then the circuit pathway is good.
If you're getting +/- 550mV or more to the bases of Q116/Q117 and that circuit pathway you just measured is good but you're stilll not showing any current flow through R124/125/126/127 then one or both of those transistors is bad. If either transistor doesn't conduct then no current will flow in that current path.
I'm kind of thinking that the Pos polarity transistor Q116 is working because you're getting a positive voltage response on TP12 when you adjust the bias potentiometer.
So either there's an open circuit somewhere or one or both of the transistors Q116/117 are bad.
If you measure the resistance from the + DC rail to the collector of Q116 and you get approx 3ohms. Then measure the resistance from the Emitter of Q116 to the Emitter of Q117 and get approx 2 ohms. Then from the collector of Q117 to the - DC rail measures another 3 ohms then the circuit pathway is good.
If you're getting +/- 550mV or more to the bases of Q116/Q117 and that circuit pathway you just measured is good but you're stilll not showing any current flow through R124/125/126/127 then one or both of those transistors is bad. If either transistor doesn't conduct then no current will flow in that current path.
I'm kind of thinking that the Pos polarity transistor Q116 is working because you're getting a positive voltage response on TP12 when you adjust the bias potentiometer.
+ rail to Q116 collector: 3.2 ohms
Q116 emitter to Q117 emitter: 2.4 ohms
Q117 collector to - rail: 3.5 ohms
I removed Q116 and Q117 and measured:
Q116:
BJT-NPN
hFE=55
Ie=6.3mA
Vbe=517mV
Q117:
BJT-PNP
hFE=112
Ie=6.3mA
Vbe=543mV
I don't seem to have any other known good uninstalled C3856s to compare to Q116, which seems to have an hFE that is a bit low.
Q116 emitter to Q117 emitter: 2.4 ohms
Q117 collector to - rail: 3.5 ohms
I removed Q116 and Q117 and measured:
Q116:
BJT-NPN
hFE=55
Ie=6.3mA
Vbe=517mV
Q117:
BJT-PNP
hFE=112
Ie=6.3mA
Vbe=543mV
I don't seem to have any other known good uninstalled C3856s to compare to Q116, which seems to have an hFE that is a bit low.
I thought of a test you could do to test them and that is to reinstall Q116/117 and then hook a resistor up across the speaker out for the channel you're testing. This test assumes that the speaker protection relay is engaging. Use something in the 10 - 20 ohm range, 10 ohm would be better. The voltage you're going to be putting on this resistor is low so more than 20 ohms will drop the current down to a really low level. The resistor will give a current path back to ground and allow one of the transistors (if working) to conduct. You will then see a voltage drop across the 3ohm resistor for the device that's working. This would be a good test to have two meters connected one on each of the 3 ohm resistors monitoring DC mV. Start with the bias pot at its minimum position and slowly increase while watching the meter(s). If one starts to show a voltage drop across the 3 ohm resistor and it increases as you turn up the bias pot then its working. The most you're going to see is maybe 45 - 50mV across that 3 ohm resistor but that really depends on how high of a resistance value you put across the speaker channel binding posts. If you find a bad one then maybe pull one of the output transistors and put it into the driver Q116/117 location to see if it'll work in the same test.
I would verify that you have the full +/- DC rail voltage on the emitter leg of each transistor before beginning the test. You should have the full voltage as your continuity test says you have continuity, but I'd check it anyway.
If you do not have a resistor and would have to order one then it would be better to get a few new output transistors to use for testing. If you had two new ones of each polarity you could install one each in the driver location q116/117 and one into each polarity of the output stage Q126/127 and you should would be able to see if it would bias up using only 4 transistors. Like I was saying early on in the first page or two of the thread, make sure you use the Q126 position if you only put one output transistor in. This is because that transistor Q126 is the output stage bias measurement leg.
There's really no reason I can see that the output stage should not be biasing up properly. You have continuity everywhere that you have checked and you have the proper bias voltage coming into the base of the driver transistors q116/117 to turn them on. That really only leaves you with a bad transistor or two. Unfortunately the drivers and output stage were hit pretty hard and a few of them failed shorted which also means the others in parallel were also hit hard, so some of them may be damaged as well.
I would verify that you have the full +/- DC rail voltage on the emitter leg of each transistor before beginning the test. You should have the full voltage as your continuity test says you have continuity, but I'd check it anyway.
If you do not have a resistor and would have to order one then it would be better to get a few new output transistors to use for testing. If you had two new ones of each polarity you could install one each in the driver location q116/117 and one into each polarity of the output stage Q126/127 and you should would be able to see if it would bias up using only 4 transistors. Like I was saying early on in the first page or two of the thread, make sure you use the Q126 position if you only put one output transistor in. This is because that transistor Q126 is the output stage bias measurement leg.
There's really no reason I can see that the output stage should not be biasing up properly. You have continuity everywhere that you have checked and you have the proper bias voltage coming into the base of the driver transistors q116/117 to turn them on. That really only leaves you with a bad transistor or two. Unfortunately the drivers and output stage were hit pretty hard and a few of them failed shorted which also means the others in parallel were also hit hard, so some of them may be damaged as well.
I reinstalled Q116 and Q117. I powered on and checked the emitters:
Q116L emitter: 0.1v
Q117L emitter: -0.42v
However, I think you meant to measure the rails on the collectors, which are attached to the rails via the 3 ohm resistors R124 and R125:
Q116L collector: 78.0v (this seems to vary in voltage, dropping down to 70v and then back up)
Q117L collector: -78.0v
Then I powered off and attached a 10 ohm resistor to the speaker terminals on the left channel. Measuring across the 3 ohm resistors, I get the following voltages after a power cycle for each:
R124, bias pot to the left: 186.0mV, then drops to 36.5mV, then pops up to 66mV
R124, bias pot to the right: 132.8mV, no fluctuations
R125, bias pot to the left: 0mV
R125, bias pot to the right: 0mV
So I guess this implies that Q117 is toast? Q119 is still out of the circuit, but since they're effectively working in parallel, that shouldn't matter, right? The behavior of Q116 is strange to me too, maybe it needs Q117 to be working for it to work properly.
Q116L emitter: 0.1v
Q117L emitter: -0.42v
However, I think you meant to measure the rails on the collectors, which are attached to the rails via the 3 ohm resistors R124 and R125:
Q116L collector: 78.0v (this seems to vary in voltage, dropping down to 70v and then back up)
Q117L collector: -78.0v
Then I powered off and attached a 10 ohm resistor to the speaker terminals on the left channel. Measuring across the 3 ohm resistors, I get the following voltages after a power cycle for each:
R124, bias pot to the left: 186.0mV, then drops to 36.5mV, then pops up to 66mV
R124, bias pot to the right: 132.8mV, no fluctuations
R125, bias pot to the left: 0mV
R125, bias pot to the right: 0mV
So I guess this implies that Q117 is toast? Q119 is still out of the circuit, but since they're effectively working in parallel, that shouldn't matter, right? The behavior of Q116 is strange to me too, maybe it needs Q117 to be working for it to work properly.
Yes you're correct I meant the collector leg of the transistor.
Do you see the same drop from 78V to 70V on the + DC rail too or is it only at the Q116 Collector?
Yes it does look like Q117 isn't conducting so its toast.
Q116 may need a bit to settle after adjustment. The bias pot could be a little flaky too. You put the old ones back in right?
Do you see the same drop from 78V to 70V on the + DC rail too or is it only at the Q116 Collector?
Yes it does look like Q117 isn't conducting so its toast.
Q116 may need a bit to settle after adjustment. The bias pot could be a little flaky too. You put the old ones back in right?
Yep, I've got the old bias pot in the left channel. The right channel is still not hooked up, but has the newer one.
When I power up the amp, the DBT dims completely after the relays click on. But once I probe the DC rail at C114+, the DBT comes back on dimly and the rail voltage drops to something like 40-50v. This is with the 10 ohm resistor on the speaker terminal in place FYI. Very odd behavior.
When I power up the amp, the DBT dims completely after the relays click on. But once I probe the DC rail at C114+, the DBT comes back on dimly and the rail voltage drops to something like 40-50v. This is with the 10 ohm resistor on the speaker terminal in place FYI. Very odd behavior.
I think you're still battling bad output transistors.
The only thing that I can think of is that one or more of the + polarity outputs is coming on and conducting which is dropping the rail volts. There could be a faulty transistor in the output positive polarity that once it gets a signal of any type its turning on for some odd reason. Once you touch your probe to the C114+ there may be a small spike of noise that goes through the circuit and it could be enough to possibly cause that faulty output to come on partially.
There should'nt be enough current draw in the front end of the circuit to cause a high wattage DBT to glow. The R160 10r 1/2w CRC resistor or another resistor in the front end would go up in smoke in a hurry if that were the case. Only the drivers or output transistors could cause that DBT to glow.
Do you see any kind of shift in the Q116/117 base voltage after measuring the C114+ pin?
The only thing that I can think of is that one or more of the + polarity outputs is coming on and conducting which is dropping the rail volts. There could be a faulty transistor in the output positive polarity that once it gets a signal of any type its turning on for some odd reason. Once you touch your probe to the C114+ there may be a small spike of noise that goes through the circuit and it could be enough to possibly cause that faulty output to come on partially.
There should'nt be enough current draw in the front end of the circuit to cause a high wattage DBT to glow. The R160 10r 1/2w CRC resistor or another resistor in the front end would go up in smoke in a hurry if that were the case. Only the drivers or output transistors could cause that DBT to glow.
Do you see any kind of shift in the Q116/117 base voltage after measuring the C114+ pin?
You know now that I think about it a bit more. If the neg rail isn't dropping off in voltage and only the positive rail is then this drop is not due to the DBT. A drop due to a glowing DBT should show up on both rails.
You should measure the DC voltage at the primary caps to see if it's showing the drop there too. C001/C002 that is. Also measure AC volts across the primary caps with the problem present.
I would also verify that you have good solid connections all the way back to those primary filtering capacitors. It could also be that possibly one of the power supply diodes feeding those large caps has opened up or that the main filtering cap itself is going bad and has developed a high current draw across it. Normally those diodes and capacitors fail shorted, but there's always a possibility that one of the diodes is open or a cap has partially shorted.
You should measure the DC voltage at the primary caps to see if it's showing the drop there too. C001/C002 that is. Also measure AC volts across the primary caps with the problem present.
I would also verify that you have good solid connections all the way back to those primary filtering capacitors. It could also be that possibly one of the power supply diodes feeding those large caps has opened up or that the main filtering cap itself is going bad and has developed a high current draw across it. Normally those diodes and capacitors fail shorted, but there's always a possibility that one of the diodes is open or a cap has partially shorted.
It's a bit hard to measure at the primary caps due to the way the left channel is connected to the central chassis. But the big cables supplied the rail voltages are easy enough to probe.
-rail: -76.0vdc after about 30s. When I remove my probe, the DBT starts to glow. Then when I remeasure, I get -55vdc and it starts to charge up again to the full rail and the DBT dims as it does this. AC measurements starts at about 2v and drops down to 0v.
+rail: +76.0vdc after about 30s, same effect as above, but I see it get down to about +50vdc, then it charges back up again to the full rail. AC measurement starts at about 1v and drops down to 0v.
To do the diode measurements and voltage measurements on the caps themselves, I'd need to disconnect the left channel and drain the charge. I should be able to do that a bit later.
-rail: -76.0vdc after about 30s. When I remove my probe, the DBT starts to glow. Then when I remeasure, I get -55vdc and it starts to charge up again to the full rail and the DBT dims as it does this. AC measurements starts at about 2v and drops down to 0v.
+rail: +76.0vdc after about 30s, same effect as above, but I see it get down to about +50vdc, then it charges back up again to the full rail. AC measurement starts at about 1v and drops down to 0v.
To do the diode measurements and voltage measurements on the caps themselves, I'd need to disconnect the left channel and drain the charge. I should be able to do that a bit later.
Measuring the diodes on the power supply PCB is pretty difficult. In part because the PCB is not removable without unscrewing a bunch of wires and removing the cross-straps on the - side of the big caps.
But I can get my probes in there and make some measurements which I think are accurate. Nothing that the 401 and 402 parts are different polarities, and using the diode setting on my Fluke, I see the following
D401L: both diodes measure about 1v
D401R: both diodes measure about 1v
D402L: both diodes measure about 1v
D402R: both diodes measure about 1v
Powering on an measuring the caps themselves, I see:
L+ 81.8v
L- -81.8v
R+ 81.8v
R- -81.8v
The DBT never dims when I measure them.
But I can get my probes in there and make some measurements which I think are accurate. Nothing that the 401 and 402 parts are different polarities, and using the diode setting on my Fluke, I see the following
D401L: both diodes measure about 1v
D401R: both diodes measure about 1v
D402L: both diodes measure about 1v
D402R: both diodes measure about 1v
Powering on an measuring the caps themselves, I see:
L+ 81.8v
L- -81.8v
R+ 81.8v
R- -81.8v
The DBT never dims when I measure them.
If both DC rails are dropping and rising similarly then it is probably the DBT that is dropping the voltage.
I was under the impression earlier that only one polarity was dropping.
Its odd that touching the probe to the front end DC rail is having that effect but it may be an artifact of the amp being powered through DBT and not only that the fact that the output stage is still having issues. I think you need to just get four good devices in the back end of the amp, two drivers and two output transistors and get it biased and idling properly and remove the DBT to see what happens.
Once the amp is properly biased and the rails are loaded and the DBT is out of circuit, then its liable to act totally different.
I was under the impression earlier that only one polarity was dropping.
Its odd that touching the probe to the front end DC rail is having that effect but it may be an artifact of the amp being powered through DBT and not only that the fact that the output stage is still having issues. I think you need to just get four good devices in the back end of the amp, two drivers and two output transistors and get it biased and idling properly and remove the DBT to see what happens.
Once the amp is properly biased and the rails are loaded and the DBT is out of circuit, then its liable to act totally different.
I reinstalled R163 to its nominal 1k value, reinstalled R128, reinstalled Q119, replaced Q117 with my last new A1492.
I first powered up through the DBT to make sure there were no hard shorts. Then, with my probes measuring bias, I powered up without the DBT, and adjusted the pot to 40mV successfully! It's pretty touchy (it's the old, reinstalled pot), and I should probably replace it with a new part when I get a chance. But I can get it within 1mV or so. It seemed to walk around slightly for the first 10 minutes, requiring readjustments to 40mV. But after that point, it seemed to settle in place pretty well. I switched it off after 20 minutes.
I guess the next step might be to reinstalled the 4 transistors in the right channel, attach it, and power it back up and check bias there. What do you think?
I first powered up through the DBT to make sure there were no hard shorts. Then, with my probes measuring bias, I powered up without the DBT, and adjusted the pot to 40mV successfully! It's pretty touchy (it's the old, reinstalled pot), and I should probably replace it with a new part when I get a chance. But I can get it within 1mV or so. It seemed to walk around slightly for the first 10 minutes, requiring readjustments to 40mV. But after that point, it seemed to settle in place pretty well. I switched it off after 20 minutes.
I guess the next step might be to reinstalled the 4 transistors in the right channel, attach it, and power it back up and check bias there. What do you think?
That sounds like a good plan. Be sure to bring up the right channel on the DBT first, just in case something is wrong...
Single turn pots are notoriously touchy especially for bias and usually you have to let the bias settle after adjustment then readjust again. So all that you're seeing there sounds normal. Installing a multi-turn pot usually helps with much of that.
Single turn pots are notoriously touchy especially for bias and usually you have to let the bias settle after adjustment then readjust again. So all that you're seeing there sounds normal. Installing a multi-turn pot usually helps with much of that.
I reassembled the right channel, powered on with the DBT with success, then without, at which point I got the fixed 2mV bias issue we saw previously. I had checked the transistors before I reinstalled them and got:
C3856#1
BJT-NPN
hFE=34
Ie=6.2mA
Vbe=532mV
C3856#2
BJT-NPN
hFE=31
Ie=6.2mA
Vbe=541mV
A1492#1
BJT-PNP
hFE=115
Ie=6.3mA
Vbe=554mV
A1492#2
BJT-PNP
hFE=112
Ie=6.3mA
Vbe=556mV
Checking voltage diff across R124R and R125R, the same method as before, I see:
R124R, bias left: 3.8mV
R124R, bias right: 3.8mV
R125R, bias left: 0mV
R125R, bias right: 0mV
Based on that, it sure seems that all of Q116-119R might be blown. What do you think?
C3856#1
BJT-NPN
hFE=34
Ie=6.2mA
Vbe=532mV
C3856#2
BJT-NPN
hFE=31
Ie=6.2mA
Vbe=541mV
A1492#1
BJT-PNP
hFE=115
Ie=6.3mA
Vbe=554mV
A1492#2
BJT-PNP
hFE=112
Ie=6.3mA
Vbe=556mV
Checking voltage diff across R124R and R125R, the same method as before, I see:
R124R, bias left: 3.8mV
R124R, bias right: 3.8mV
R125R, bias left: 0mV
R125R, bias right: 0mV
Based on that, it sure seems that all of Q116-119R might be blown. What do you think?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.