I am going to say that the right channel is fully working. I connected a pair of speakers and the amplifier is amplifying using a Discman as source plugged into AUX inputs.
DBT still in use,
Both left and right bias pots now adjust current but the range is narrow. The minimum is about 32mV and maximum is about 36mV.
After the amp idled for an hour or so last night, right channel offset went from 0.1mV to 2.5mV.
DBT still in use,
Both left and right bias pots now adjust current but the range is narrow. The minimum is about 32mV and maximum is about 36mV.
After the amp idled for an hour or so last night, right channel offset went from 0.1mV to 2.5mV.
Hopefully this particular post will not divert attention from the power amp section which this thread is about. This post has nothing to do with that. It concerns a different part of the amplfier. The schematic has the heading Tone Amp and Muting Sw
It is just to satisfy my curiosity.
Something I noticed quite a bit earlier in this thread was that both R501 and R502 (1/2W 820 ohm) resistors had cracked exteriors. I got some replacments but did nothing with them until today. While I was replacing R501/502 I saw that R533 and R536 also had cracked exteriors. These too are 1/2W 820 ohm resistors. When I removed them for replacement, I saw that the PC board underneath was discoloured, obviously subjected to excess heat from those resistors.
Testing the cracked resistors after removal showed that they were still within specc.
I replaced all four of those resistors but was wondering what conditions could have made them get hot in the first place? I suspect it is "old damage" and not something that happened since I've owned this amplifier.
It is just to satisfy my curiosity.
Something I noticed quite a bit earlier in this thread was that both R501 and R502 (1/2W 820 ohm) resistors had cracked exteriors. I got some replacments but did nothing with them until today. While I was replacing R501/502 I saw that R533 and R536 also had cracked exteriors. These too are 1/2W 820 ohm resistors. When I removed them for replacement, I saw that the PC board underneath was discoloured, obviously subjected to excess heat from those resistors.
Testing the cracked resistors after removal showed that they were still within specc.
I replaced all four of those resistors but was wondering what conditions could have made them get hot in the first place? I suspect it is "old damage" and not something that happened since I've owned this amplifier.
The resistors are used to drop the main supply voltage to lower values for the preamp stages. The circuit shows 35 volt as the main supply and the voltage on the other side of the resistor as 17v. So we have 18 volts across the resistor... if the mains voltage is a bit higher than the voltage may be a bit more.
Dissipation in the resistor is (18*18)/820 which is 0.4 watt. So they are run close to their limit, they will run hot and in a case.... its a normal thing to see. So no real issue. I would probably replace with 1 watt types.
So are we saying both channels seem basically OK at this point? If the bias adjusts and offset is low it would seem so.
To get bias lower we need to alter either of these two resistors and make one a bit higher in value. I would try making the 1.8k into a 2.2k or increase the 3.3k to a 3.9k or a 4.7k. There is no way to calculate these, it is trial and error as it depends on the forward B/E volt drop of the driver and output transistors used.
If you make them to high the bias will fall to zero (which is safe), just do not make them lower than they are. Make sure the original values fitted match those on the circuit as these type of parts might be subject to change in production. Always go higher than the value you remove. Also keep using the DBT at this point.
Dissipation in the resistor is (18*18)/820 which is 0.4 watt. So they are run close to their limit, they will run hot and in a case.... its a normal thing to see. So no real issue. I would probably replace with 1 watt types.
So are we saying both channels seem basically OK at this point? If the bias adjusts and offset is low it would seem so.
To get bias lower we need to alter either of these two resistors and make one a bit higher in value. I would try making the 1.8k into a 2.2k or increase the 3.3k to a 3.9k or a 4.7k. There is no way to calculate these, it is trial and error as it depends on the forward B/E volt drop of the driver and output transistors used.
If you make them to high the bias will fall to zero (which is safe), just do not make them lower than they are. Make sure the original values fitted match those on the circuit as these type of parts might be subject to change in production. Always go higher than the value you remove. Also keep using the DBT at this point.
Had I known those 1/2W cracked resistors were so close to their limit I would have ordered 1W and I may do that yet. FWIW, I mounted the new 1/2W resistors a few millimeters above the board, not flat against it as the originals were..
Regarding the diagram, I will have a look to see what resistors are at hand and try your suggestion.
Assessing what resistors I have would determine which of those two circled resistors I will change.
I will decode the colour bands of R308 and R307 to see what their values actually are.
The service manual cites 25mV as the target bias current for transistors from 1973/74. Considering that the replacement transitors are modern, is 25mV still the number to shoot for?
I am assuming that the objective for upping resistance of R308 or R307 is to achieve near 0.0mV with bias pot at minimum?
Regarding the diagram, I will have a look to see what resistors are at hand and try your suggestion.
Assessing what resistors I have would determine which of those two circled resistors I will change.
I will decode the colour bands of R308 and R307 to see what their values actually are.
The service manual cites 25mV as the target bias current for transistors from 1973/74. Considering that the replacement transitors are modern, is 25mV still the number to shoot for?
I am assuming that the objective for upping resistance of R308 or R307 is to achieve near 0.0mV with bias pot at minimum?
From the list of substitute resistors you provided, the only one I have on hand is 2.2K. So I was about to remove R308 and put a 2.2K resistor there instead. However, R308 is actually a 1.2K not a 1.8K.
The last digit on the schematic is either a misprint or a print distortion.
1.2K is the value of R308 in the parts list and is also what is printed on the component location diagram.
That being the case, what value of resistor I should look for as a substitute for the 1.2K ohm resistor at R308?
The last digit on the schematic is either a misprint or a print distortion.
1.2K is the value of R308 in the parts list and is also what is printed on the component location diagram.
That being the case, what value of resistor I should look for as a substitute for the 1.2K ohm resistor at R308?
It looks like a misprint on the circuit. Parts list actually states 1.2k
I'm going cross eyed looking at this circuit because its not printed very clearly
I've made an error in circling R307 in the diagram earlier. R307 must be left at 3.3k and we can only alter the 1.2k which need to be raised in value. Try your 2k2. If it is to high you would not be able to reach the 25mv bias setting but it would be 'safe'.
I'm going cross eyed looking at this circuit because its not printed very clearly
I've made an error in circling R307 in the diagram earlier. R307 must be left at 3.3k and we can only alter the 1.2k which need to be raised in value. Try your 2k2. If it is to high you would not be able to reach the 25mv bias setting but it would be 'safe'.
Pardon my ignorance but what is a preset in this context?
I did not install a 2.2K resistor because earlier today I discovered something about R308 & R408. The resistors at those positions are not 1.2K resistors, they already are 2.2K resistors. The color bands made me take a closer look. Brn/red/red/ vs red/red/red.
Chosing to stay on the right channel, I removed R408 and it looked like red/red/red to me so I tested it and sure enough it was 2200 ohms.
To see what bias would be, with a 1.2K resistor there, I installed one at R408 and measured bias. The amp had been on for while and my meter displayed 37,3mV with bias pot at minimum. It would adjust higher but not less that 37.3mV.
Minimum left channel bias, with a 2.2K ohm resistor at R308 is 33.3mV.
Does this mean I should be experimenting with resistors of considerably higher values than 2.2K ohms?
I did not install a 2.2K resistor because earlier today I discovered something about R308 & R408. The resistors at those positions are not 1.2K resistors, they already are 2.2K resistors. The color bands made me take a closer look. Brn/red/red/ vs red/red/red.
Chosing to stay on the right channel, I removed R408 and it looked like red/red/red to me so I tested it and sure enough it was 2200 ohms.
To see what bias would be, with a 1.2K resistor there, I installed one at R408 and measured bias. The amp had been on for while and my meter displayed 37,3mV with bias pot at minimum. It would adjust higher but not less that 37.3mV.
Minimum left channel bias, with a 2.2K ohm resistor at R308 is 33.3mV.
Does this mean I should be experimenting with resistors of considerably higher values than 2.2K ohms?
Yes, R308 needs to go higher. Use what you have that is higher or make a value up of two series resistor. Its not very critical because the preset allows a wide range of adjustment, you just need it a bit higher than it is now.
It means replacing R308 temporarily with a preset (like the bias setting one) but of higher value than R308, say a 4.7k or 5k preset. You use only the wiper pin and one end pin and make sure the preset is turned so it appears as its full value in circuit.
You then set the normal bias preset to the midpoint and this time set the bias with the new preset. Switch off, remove and measure the new preset value and fit the nearest fixed resistor to that value. Now set the main bias preset to minimum and adjust the bias as per the manual and the bias preset should end up around the midpoint.
It means replacing R308 temporarily with a preset (like the bias setting one) but of higher value than R308, say a 4.7k or 5k preset. You use only the wiper pin and one end pin and make sure the preset is turned so it appears as its full value in circuit.
You then set the normal bias preset to the midpoint and this time set the bias with the new preset. Switch off, remove and measure the new preset value and fit the nearest fixed resistor to that value. Now set the main bias preset to minimum and adjust the bias as per the manual and the bias preset should end up around the midpoint.
I have not been able to wrap my head around the "how to" of creating a preset so I tried some higher value resistors at R408.
5.6K ohms produced minimum bias of 32mV
6.8K ohms produced minimum bias of 29mV
9.0K ohms produced minimum bias of 27mV
12K ohms produced minimum bias of 26mV After a short time, this rose to 27mV and brushed 28mV momentarily.
(The 12K was carbon film, the other 3 were metal film)
I was expecting that resistors values so much higher than stock would make a more substantial change to bias current.
I can go higher than 12K. Should I keep going?
If so, I will tack on some pigtails with alligator clips so as to change resistors quickly and save wear and tear to the solder pads
5.6K ohms produced minimum bias of 32mV
6.8K ohms produced minimum bias of 29mV
9.0K ohms produced minimum bias of 27mV
12K ohms produced minimum bias of 26mV After a short time, this rose to 27mV and brushed 28mV momentarily.
(The 12K was carbon film, the other 3 were metal film)
I was expecting that resistors values so much higher than stock would make a more substantial change to bias current.
I can go higher than 12K. Should I keep going?
If so, I will tack on some pigtails with alligator clips so as to change resistors quickly and save wear and tear to the solder pads
You can go as high as you want. Having no resistor fitted should drop the bias to around zero and yes, those values are a bit higher than I would expect but its not an issue of itself. With the amp OFF just make sure that there are no shorts around the driver transistor such as solder blobs between base and emitter as that would reduce the voltage needed by the bias generator (the resistors you are altering) and yet still give the appearance all is OK.
I don't think there will be a problem but check for certainty that all really is OK.
It would be a base/emitter short on the drivers, nothing else would do this... and I'm sure its fine though and just down to the different transistor characteristics.
One other easy check you could do (but be careful) is to set the bias correctly and then measure the voltage between the two 330 ohm resistors that go to the base of the drivers. Red meter lead on one resistor and black lead on the other resistor. Use the end of each resistor that goes to Q304 for the probe position. Ignore the polarity you see.
The correct bias setting needs there to be about 2.4 volts (or a bit more) between those resistors. If that is OK its good to go If it is lower (say 1.8 then we have a simple problem to resolve).
12k is almost there so I would just go next value up (15k). Don't be tempted to set the bias to high, even a very small bias current removes distortion so for reliability you could aim lower the 25mv if you want. 10 to 15mv would be fine.
I don't think there will be a problem but check for certainty that all really is OK.
It would be a base/emitter short on the drivers, nothing else would do this... and I'm sure its fine though and just down to the different transistor characteristics.
One other easy check you could do (but be careful) is to set the bias correctly and then measure the voltage between the two 330 ohm resistors that go to the base of the drivers. Red meter lead on one resistor and black lead on the other resistor. Use the end of each resistor that goes to Q304 for the probe position. Ignore the polarity you see.
The correct bias setting needs there to be about 2.4 volts (or a bit more) between those resistors. If that is OK its good to go
If it is lower (say 1.8 then we have a simple problem to resolve).12k is almost there so I would just go next value up (15k). Don't be tempted to set the bias to high, even a very small bias current removes distortion so for reliability you could aim lower the 25mv if you want. 10 to 15mv would be fine.
Well, I installed a 15K ohm resistor and if anything, minimum bias is a little higher. It was fluttering between 29mV and 30mV and would not increase more than about 0.1mV.
To accomplish the above inspection, I removed the board from the amp to examine with better lighting and better magnification. Base and emitter of both driver transistors free from shorts.
I removed the 15K ohm resistor. With no resistor in place, bias was mid 20mV range not zero.
I am confined to the right channel so 4** numbers.
Bias cannot be adjusted but since it is is near to 25mV with that resistor removed I went ahead with the above measurment. The 330 ohm resistors I located are R413 and R414. With DBT still in place the voltage I measured between them is 0.76V.
I am keen to learn what the simple problem to solve might be.
To accomplish the above inspection, I removed the board from the amp to examine with better lighting and better magnification. Base and emitter of both driver transistors free from shorts.
I removed the 15K ohm resistor. With no resistor in place, bias was mid 20mV range not zero.
One other easy check you could do (but be careful) is to set the bias correctly and then measure the voltage between the two 330 ohm resistors that go to the base of the drivers. Red meter lead on one resistor and black lead on the other resistor. Use the end of each resistor that goes to Q304 for the probe position. Ignore the polarity you see.
The correct bias setting needs there to be about 2.4 volts (or a bit more) between those resistors. If that is OK its good to go
I am confined to the right channel so 4** numbers.
Bias cannot be adjusted but since it is is near to 25mV with that resistor removed I went ahead with the above measurment. The 330 ohm resistors I located are R413 and R414. With DBT still in place the voltage I measured between them is 0.76V.
I am keen to learn what the simple problem to solve might be.
We need to investigate this. It should fall to zero.
I wonder if this is a sign of very high frequency instability (oscillation) caused by modern parts being so much faster than the old originals. It is something that happens.
I'll have to think what to try next but my first thought would be to place a very small cap (just a few pF) across the feedback resistor R320 (or R321 as they are in parallel) and see if it resolves the problem. The size of cap would be small, perhaps in the range 10pF to 47pF but no higher.
Let me think today. I might throw the design into a simulation and look at the stability margins but that would be later today/tonight.
This is where an oscilloscope comes into its own and it would show anything like this instantly
Please know that I am thankful for your expertise and very grateful for you dilligence and patience.
I put a 47pF capacitor in parallel with R321 as suggested and bias remains at 27mV.
This is where an oscilloscope comes into its own and it would show anything like this instantly
You have planted an oscilloscpe seed. But there have been none offered for sale even remotely closeby.
An acquaintance of mine has a well equipped test bench. But since my move 4 years ago, he is now 700km distant. Unlikely he will have a spare oscilloscope to sell but he may know of one or may be willing to browse online ads in his area.
I would trust his judgment but he has champagne taste and I have a
As useful and educational as an oscilloscope can be, you really need someone with some audio/video or even automotive diagnostic experience to show you how to identify and get meaningful results. Otherwise, you may be just looking at squiggles on a screen, since there's no indication of what they represent.
Digital 'scopes now, are relatively cheap and easy to use but I think you're very likely to need some 1 on 1 type instruction to ensure you're on the track to meaningful results. In my experience with living in isolated communities, there has usually been someone around who was a capable tech. or at least able to say something is wrong and how and where to find out what it is, if not repair it. These useful guys may often have little spare time to show you how but you'll need their kind of help with instrument use, because internet articles often don't squarely meet your needs. In fact, most of the articles on using a 'scope that I've read on the net, seem to skim over the steps to learning and just waste space on talking up the benefits - or simply flashing product images. Many articles seem to ignore the basics of use and focus on specific applications that are probably too advanced or irrelevant anyway.
Digital 'scopes now, are relatively cheap and easy to use but I think you're very likely to need some 1 on 1 type instruction to ensure you're on the track to meaningful results. In my experience with living in isolated communities, there has usually been someone around who was a capable tech. or at least able to say something is wrong and how and where to find out what it is, if not repair it. These useful guys may often have little spare time to show you how but you'll need their kind of help with instrument use, because internet articles often don't squarely meet your needs. In fact, most of the articles on using a 'scope that I've read on the net, seem to skim over the steps to learning and just waste space on talking up the benefits - or simply flashing product images. Many articles seem to ignore the basics of use and focus on specific applications that are probably too advanced or irrelevant anyway.
You're very welcome
although it has become a bit of a puzzler.If the cap made no difference at all then we have stop and think and not go down any rabbit holes.
Can we just confirm that when you removed the resistor and still measured 25mv that you are in fact effectively measuring across one of the 0.47 ohm resistors. The procedure in the manual is correct but also unusual in its method of where you connect the meter. If you were measuring from ground or from the speaker negative terminal the reading would not be accurate. We just need to be sure on that
There must also be NO load (speaker) connected because any DC offset will skew the result as an offset would cause current to flow in the load adding or subtracting from the bias measurement depending on the offset polarity.
Assuming that is all OK we are back to this:
The theory is sound here, if you have 0.76 volts between the two 330 ohm resistors then that alone is nowhere near enough to turn the driver and output stage on. We need approx 0.6 volts (the typical forward base/emitter volt drop) per transistor so the two drivers plus Q311 and the output pair (the parallel outputs make no difference) mean we need about 0.6 times 5 or 3 volts. So 0.76 is way below that and there should be zero bias current flowing.
So as a simulation we can see it agrees with the theory. The offset of this simulated amp is virtually zero (I trimmed the 51k at the input to achieve that) and the DC voltage on 330 ohm resistors is +1.2 and -1.8 volt giving us that 3 volt figure between the resistors. The bias is set to approx the recommended value.
(For your amp, measuring between the resistors means any DC offset in your amp is accounted for as both values would shift up or down accordingly. The same applies to the 25mv on the 0.47 ohm resistors)
This is the same set up but with a small DC offset of +85mv. Although the voltages as measured from ground look very different, in reality little has changed. There is still the same voltage across each 0.47 ohm and still the same voltage between the two 330 ohms.
This is why the manual tells you to measure the 25mv in the way it does and why I say to check between two parts rather than measuring from ground. It removes the errors caused by any offset.
One final image with 0.76 volts between the 330 ohms. R308 is now removed and it gives around 856 mv between the two 330 ohm resistors. That's fine and agrees closely with your measurement
The bias current is zero with -19mv (offset) on all nodes. There is no voltage across the 0.47 ohm and no bias current flowing. This should be what you see. There should be zero voltage across each 0.47 ohm resistor even if that 0.47 ohm resistor actually has a small voltage on it relative to ground.
Did we replace Q311???
I'll post the sim in case anyone wants a play but it will need your own models substituting.
And thinking aloud while looking at the circuit. Is R328 OK which is the 2.7k between base and emitter of Q306 and making connection OK, no breaks around it.
Q306 and Q311 form a super high gain Darlington configuration and even a slight leakage in Q306 could cause an issue as could R328 being open as it would leave the base of Q311 'floating' and that could cause unwanted conduction.
Q306 and Q311 form a super high gain Darlington configuration and even a slight leakage in Q306 could cause an issue as could R328 being open as it would leave the base of Q311 'floating' and that could cause unwanted conduction.