Hello
I was working on a broken RA 971 amp. It had blown fuses , which despite replacement, did not switch on (on a DBT - 70W halogen bulb).
Here is the schematic : https://www.vintageshifi.com/repertoire-pdf/pdf/telecharge.php?pdf=Rotel-RA-971-Service-Manual.pdf
R
However, I could hear a faint sound from the speakers, so by process of elimination chanced upon the relay RY901 and it associated circuitry. I couldnt find an obvious problem, so reflowed the solders for fuse F901, D902 and couple other passives in that path.
Lo and behold the relay started clicking again upon startup and everything was fine. The amp sounded good in operation, all the while on the DBT. Note that I am not sure what the 'fix' was - a bad solder, or the relay just deciding to come unstuck (it was in storage for a while before that).
The Bulb lights up bright (on switch on) and then doesn't quite go to fully to 0 when it selttles, even when nothing plays.
I have repaired a RA 930 and the bulb goes pretty much 'off' when that amp is in operation, but the 971 having 2x the power i assumed a higher current is normal.
Now when I plug to the socket directly, F902 blows with quite a visible spark and charring inside. For a slow blow 6.3A fuse that does seem like a lot. (note that F901 is associated with the relay)
I am scratching my head about what the issue could be. The caps including the 50v 10mF power caps are all original, and so are other components afaik.
The only other point to note is that where the +/-49V in the schematic shows, I get only about +/-39V (all on the DBT). But then the 50v power caps seem pretty keenly specified at 49v. The RA930 is supposed to have +/-37V and i get slightly less (c34v).
With the power fuses out, and connected directly to mains, the transformer outputs are about 39.5v AC.
The regulated voltages show 16.1v. The regulating transistors q901, q902 do get very warm or even hot at idling.
I did in circuit checks of all the power transistors and the driver/bias transistors that are coupled thermally to the heatsink. I checked all the diodes as well.
I am yet to find an obviously 'bad' component.
Any pointers would be very appreciated.
I was working on a broken RA 971 amp. It had blown fuses , which despite replacement, did not switch on (on a DBT - 70W halogen bulb).
Here is the schematic : https://www.vintageshifi.com/repertoire-pdf/pdf/telecharge.php?pdf=Rotel-RA-971-Service-Manual.pdf
R
However, I could hear a faint sound from the speakers, so by process of elimination chanced upon the relay RY901 and it associated circuitry. I couldnt find an obvious problem, so reflowed the solders for fuse F901, D902 and couple other passives in that path.
Lo and behold the relay started clicking again upon startup and everything was fine. The amp sounded good in operation, all the while on the DBT. Note that I am not sure what the 'fix' was - a bad solder, or the relay just deciding to come unstuck (it was in storage for a while before that).
The Bulb lights up bright (on switch on) and then doesn't quite go to fully to 0 when it selttles, even when nothing plays.
I have repaired a RA 930 and the bulb goes pretty much 'off' when that amp is in operation, but the 971 having 2x the power i assumed a higher current is normal.
Now when I plug to the socket directly, F902 blows with quite a visible spark and charring inside. For a slow blow 6.3A fuse that does seem like a lot. (note that F901 is associated with the relay)
I am scratching my head about what the issue could be. The caps including the 50v 10mF power caps are all original, and so are other components afaik.
The only other point to note is that where the +/-49V in the schematic shows, I get only about +/-39V (all on the DBT). But then the 50v power caps seem pretty keenly specified at 49v. The RA930 is supposed to have +/-37V and i get slightly less (c34v).
With the power fuses out, and connected directly to mains, the transformer outputs are about 39.5v AC.
The regulated voltages show 16.1v. The regulating transistors q901, q902 do get very warm or even hot at idling.
I did in circuit checks of all the power transistors and the driver/bias transistors that are coupled thermally to the heatsink. I checked all the diodes as well.
I am yet to find an obviously 'bad' component.
Any pointers would be very appreciated.
These older rotel amps are not the best design for fusing, with just slow blow/time delay fuses between the transformer and power supply. Quite easy to fry speakers with!
If you have a suitable dual output bench supply you could connect that to the +/-49V DC supplies. Use that to power it at a lower voltage with current limiting to limit any damage.
Make sure you check things like the bridge rectifier D901 doesn't have any shorted diodes. Last Rotel amp I was in had some really weird off brand capacitors, fitted from new, and even the "rubycon" parts were bulging...
If you have a suitable dual output bench supply you could connect that to the +/-49V DC supplies. Use that to power it at a lower voltage with current limiting to limit any damage.
Make sure you check things like the bridge rectifier D901 doesn't have any shorted diodes. Last Rotel amp I was in had some really weird off brand capacitors, fitted from new, and even the "rubycon" parts were bulging...
Last edited:
Hi unfortunately, I do not have a bench PS.
Rectified is good - checked in circuit, and also on DBT the voltages are correct.
As mentioned, on the DBT, the amp actually works and sounds fine functionally.
I checked all transistors in circuit with a DMM but cant see anything obviously wrong.
Would bad power supply filter caps present with this symptom?
Rectified is good - checked in circuit, and also on DBT the voltages are correct.
As mentioned, on the DBT, the amp actually works and sounds fine functionally.
I checked all transistors in circuit with a DMM but cant see anything obviously wrong.
Would bad power supply filter caps present with this symptom?
If the bulb doesn't quite go to dim when it idle and the supply rail voltage drop too much, it seems there is something drawing large current inside the amp.
It maybe the power amp section drawing large current or the power caps problem.
For the power amp, you can check both power amps bias current are normal or not.
For the power cap, if you have any spare power caps of 50V (less than 10mF is OK), try to replacing the original caps to test again.
Would it make sense to connect in parallel a say a pair of new known good 1000uf 63v cap to try augment the existing caps, to see if that helps?
Ref Bias currents - i have set the pot at 6mV (recommended 8 ) but doesnt really change the (low) brightness of the bulb.
Ref Bias currents - i have set the pot at 6mV (recommended 8 ) but doesnt really change the (low) brightness of the bulb.
Try turning the bias to zero while faultfinding. 8mv across 0.22 ohm four times over is nearly 150ma in total... or 15 watts. That would easily increase the bulbs resistance and give low rails.
You say you measure 39.5 volts AC (so you mean 39-0-39) on the transformer windings. That would give nearly 55 volts DC across the reservoir caps at no load.
And they are 50 volt caps?
You say you measure 39.5 volts AC (so you mean 39-0-39) on the transformer windings. That would give nearly 55 volts DC across the reservoir caps at no load.
And they are 50 volt caps?
hey mooly!
Yes they are 50v caps 10mF . And yes they are original and as per schematic as well. So yes .. they do seem not at all over-specced as one might expect.
I turned the bias pots to max resistance. Now, the light is fully off. I get pretty much 0 mV on the test points. But if i connect up speakers, i get very low volume (as one might expect).
I have ordered replacement caps, so waiting for them to arrive. This is more because I am lost for ideas, not since I figured it out.
Just looking about for something that looks 'off' i tried touching some components to find anything too hot.
The regulator transistors Q901 and Q902 are again, quite hot to the touch even at idling. That sure doesn't sound right.
The 10mF caps in the power supply area dont get warm to the touch.
But C909 and C910 which at the 25v 470uF associated with Q901 and Q902 are a bit warm (not as much as the transistors).
The heatsink is now definitely to warming up as much as it was with the bias at more 'normal' levels.
Yes they are 50v caps 10mF . And yes they are original and as per schematic as well. So yes .. they do seem not at all over-specced as one might expect.
I turned the bias pots to max resistance. Now, the light is fully off. I get pretty much 0 mV on the test points. But if i connect up speakers, i get very low volume (as one might expect).
I have ordered replacement caps, so waiting for them to arrive. This is more because I am lost for ideas, not since I figured it out.
Just looking about for something that looks 'off' i tried touching some components to find anything too hot.
The regulator transistors Q901 and Q902 are again, quite hot to the touch even at idling. That sure doesn't sound right.
The 10mF caps in the power supply area dont get warm to the touch.
But C909 and C910 which at the 25v 470uF associated with Q901 and Q902 are a bit warm (not as much as the transistors).
The heatsink is now definitely to warming up as much as it was with the bias at more 'normal' levels.
Low bias should have zero impact on perceived volume. All it will do is slightly increase distortion (and maybe so slightly that you can't even notice under critical listening).
Hot regulator transistors are probably normal and absolutely typical of this kind of design. The Zeners will also be fairly hot and so the caps in that area will heat just by physical proximity. It's not the best design but it is the way it is.
Hot regulator transistors are probably normal and absolutely typical of this kind of design. The Zeners will also be fairly hot and so the caps in that area will heat just by physical proximity. It's not the best design but it is the way it is.
Hmm ... I am not sure what i was missing with the low volume in that case.
Anyways, before I saw your reply, I realised I had 25v 470uF caps lying about. They are 85 degrees, not 105 like the originals, and physically a little smaller. I replaced them.
On switching back on, you are right, the volume is 'normal'. And I cannot discern serious distortion, but I am on my cheapo test speakers.
The Bulb is (still) not glowing at all now. So I am left a bit puzzled as to what is going on.
I tested the old 25v 470uF caps on my lcr-tc1 component tester, and it does find them pretty close to spec, but unless the new ones are totally inappropriate for the function i will leave them in.
So to summarise--
After turning the bias to 0, the bulb is off, and the heatsink doesn't get anywhere as warm as it did before.
Would you expect that to make sense, or did some thing get magically fixed...
Anyways, before I saw your reply, I realised I had 25v 470uF caps lying about. They are 85 degrees, not 105 like the originals, and physically a little smaller. I replaced them.
On switching back on, you are right, the volume is 'normal'. And I cannot discern serious distortion, but I am on my cheapo test speakers.
The Bulb is (still) not glowing at all now. So I am left a bit puzzled as to what is going on.
I tested the old 25v 470uF caps on my lcr-tc1 component tester, and it does find them pretty close to spec, but unless the new ones are totally inappropriate for the function i will leave them in.
So to summarise--
After turning the bias to 0, the bulb is off, and the heatsink doesn't get anywhere as warm as it did before.
Would you expect that to make sense, or did some thing get magically fixed...
Well, I decided to connect up to mains to see what happens, since I had 5A fuses in there as I didnt have 6.3 to hand. And... it doesn't blow.
I had the DMM hooked up to see the bias voltage, whereas it operated at about 5mV on the DBT, it went up all the way to about 14.5 and settled at 11.5 ish currently.
The heat sink is quite warm too.
I will let it sing, and maybe even cook itself for a while, unless any one has any further inputs?
EDIT: Forgot to add the rail voltages, when not on the DBT are +/- 49 volts; which is also as per the schematic. The collector voltages on the output transistors are also the same. So that does seem like a very tightly specified reservoir caps.
I had the DMM hooked up to see the bias voltage, whereas it operated at about 5mV on the DBT, it went up all the way to about 14.5 and settled at 11.5 ish currently.
The heat sink is quite warm too.
I will let it sing, and maybe even cook itself for a while, unless any one has any further inputs?
EDIT: Forgot to add the rail voltages, when not on the DBT are +/- 49 volts; which is also as per the schematic. The collector voltages on the output transistors are also the same. So that does seem like a very tightly specified reservoir caps.
Last edited:
So is it working now?
If so then you need to be 100% sure those 6.3A fuses were up to the job. This might be a case of trying to find out how the fuses were specified. Were they super slow with a TT marking as opposed to just T.
If so then you need to be 100% sure those 6.3A fuses were up to the job. This might be a case of trying to find out how the fuses were specified. Were they super slow with a TT marking as opposed to just T.
Might be oscillating - give the amp full rail voltage and oscillation becomes really bad and pops fuse? Just a thought.
A variac would be very useful here to monitor the current's dependence on voltage.
A variac would be very useful here to monitor the current's dependence on voltage.
Hey mooly
just recapping my whole thread:
The very original issue was that it wasn't playing any audio. I tracked it down to the relay. But I am not sure how i fixed it, i reflowed a few components in that circuit. Maybe it was that , maybe it was just the relay coming unstuck after a few power cycles. After that , the amp was always 'working' on the DBT. When i went to the mains, it blew the fuses.
Now, it is working on the mains on 5a fuses, but i still feel the heatsink seems much hotter than it ought to be even when idling with the bias set as per manual to 8mV. I mean, i cannot keep my hand on it for more than a few seconds.
So I still cannot shake off a feeling this is not 'normal'.
So i dont know if there is a latent issue yet.
just recapping my whole thread:
The very original issue was that it wasn't playing any audio. I tracked it down to the relay. But I am not sure how i fixed it, i reflowed a few components in that circuit. Maybe it was that , maybe it was just the relay coming unstuck after a few power cycles. After that , the amp was always 'working' on the DBT. When i went to the mains, it blew the fuses.
Now, it is working on the mains on 5a fuses, but i still feel the heatsink seems much hotter than it ought to be even when idling with the bias set as per manual to 8mV. I mean, i cannot keep my hand on it for more than a few seconds.
So I still cannot shake off a feeling this is not 'normal'.
So i dont know if there is a latent issue yet.
Well it shouldn't be so hot you can't touch it for more than a few seconds.
Are you certain your meter is accurate on a low voltage range. If both channels are behaving in the same way then it sounds more like something you are doing rather than a real problem tbh 🙂
If you have a scope then a quick check as Mark suggests is always useful.
Are you certain your meter is accurate on a low voltage range. If both channels are behaving in the same way then it sounds more like something you are doing rather than a real problem tbh 🙂
If you have a scope then a quick check as Mark suggests is always useful.
Setting the bias to very low (close to 0) , and running off the mains. No blowouts. And the heatsink is just at room temperature. I cannot honestly say I feel any difference in the sound quality. So what does the bias actually affecT?
The bias current minimises crossover distortion by evening out the non linearity in the transfer curve of the NPN/PNP output pairs.
Even a very small current is enough to reduce the distortion to very low levels.
Why not set the bias for something really low like a couple of millivolts and see how hot it runs at that. Without further testing that would also include checking the calibration of your meter it is impossible to say why it feels so hot at the recommended setting.
Even a very small current is enough to reduce the distortion to very low levels.
Why not set the bias for something really low like a couple of millivolts and see how hot it runs at that. Without further testing that would also include checking the calibration of your meter it is impossible to say why it feels so hot at the recommended setting.
Hi mooly
aha, i was doing exactly that in the meantime. I have set it to 3mV and the heatsink is now only slightly warm, about what I see on my Ra 931.
Re the meter - I have two. A Fluke 73iii and another cheaper one with a few more functions. I have been using both to check the TP's simultaneously. I swapped them around, keeping the same test leads as well and they agree to within 0.1 mV , so the meters are 'OK'
Intuitively i guess the bias shouldn't have such a disproportionate impact on the total current draw then.
If i were to pinpoint the hottest places on board, apart from the regulators discussed above, they would be Q624/5/6/7 .
These are lifted off the board and screwed on to 4 of the power amp transistors so they are thermally coupled. These smaller transistors are almost as hot to the touch as the regulators. The power transistors which do not have these little ones screwed to them are less hot, so id be willing to bet its these smaller ones causing the most heat on to the heatsink.
aha, i was doing exactly that in the meantime. I have set it to 3mV and the heatsink is now only slightly warm, about what I see on my Ra 931.
Re the meter - I have two. A Fluke 73iii and another cheaper one with a few more functions. I have been using both to check the TP's simultaneously. I swapped them around, keeping the same test leads as well and they agree to within 0.1 mV , so the meters are 'OK'
Intuitively i guess the bias shouldn't have such a disproportionate impact on the total current draw then.
If i were to pinpoint the hottest places on board, apart from the regulators discussed above, they would be Q624/5/6/7 .
These are lifted off the board and screwed on to 4 of the power amp transistors so they are thermally coupled. These smaller transistors are almost as hot to the touch as the regulators. The power transistors which do not have these little ones screwed to them are less hot, so id be willing to bet its these smaller ones causing the most heat on to the heatsink.
Those four transistors are the drivers and they should not be hot at all in their own right (when it is just idling).
They supply the base current for the output pairs and so their current flow is much much lower and so less heat generated by them.
The bias is the current draw in the output transistors and the voltage across the 0.22 just a convenient way to deduce it (ohms law). Providing the amp isn't unstable which is something Mark suggests you check (you need a scope) then the heat vs bias current will be normal. There is no other contributing factor meaning it is just the way it is.
If the meter is accurate and the 0.22 ohm is what it says it is then ohms law says we have I=V/R which is 0.008/0.22 giving 36 milliamps per transistor.
Each output transistor has 50 volts across it and so 0.036 * 50 gives 1.8 watts dissipation per device.
Just a thought... the test only checks the current in one pair of devices. Try measuring the voltage across the 0.22 ohm belonging to the other pair and see if it is similar. If not then use the highest reading pair to set the voltage.
They supply the base current for the output pairs and so their current flow is much much lower and so less heat generated by them.
The bias is the current draw in the output transistors and the voltage across the 0.22 just a convenient way to deduce it (ohms law). Providing the amp isn't unstable which is something Mark suggests you check (you need a scope) then the heat vs bias current will be normal. There is no other contributing factor meaning it is just the way it is.
If the meter is accurate and the 0.22 ohm is what it says it is then ohms law says we have I=V/R which is 0.008/0.22 giving 36 milliamps per transistor.
Each output transistor has 50 volts across it and so 0.036 * 50 gives 1.8 watts dissipation per device.
Just a thought... the test only checks the current in one pair of devices. Try measuring the voltage across the 0.22 ohm belonging to the other pair and see if it is similar. If not then use the highest reading pair to set the voltage.