When you try this take the precaution of fitting a 100R wire-wound resistor rated at 5W in the positive supply rail to limit the current drawn.
The trimmer should be adjusted for maximum resistance. For safety you should use the 10k one before going to 1k or 2k in value.
To decide which of these two you could solder a 10k resistor in parallel with the 10k trimmer to bring the combined resistance closer to the target.
At turn on keep one hand on the on-off button and moisten a finger and put it say 5mm above the wire-wound to detect any heat. Turn off immediately if this gets hot.
If not measure the voltage drop across it and use Ohm's law to deduce the current. If this can be adjusted to close to zero then progress will have been achieved.
This circuit is similar to one shown in Motorola application note AN484A although that is rated at 3-5Watts.
Above that level driver transistors are necessary. Both output transistors have 0R56 emitter resistors to stabilize output currents. In that regard what you have is an economy that may not have paid off.
In all amplifiers in this paper there is an output zobel stability network of 10R and 100nF. The other stability measure of increasing the compensation capacitor from 22pF that I mentioned has not been implemented - options to consider if luck is not on your side.
The trimmer should be adjusted for maximum resistance. For safety you should use the 10k one before going to 1k or 2k in value.
To decide which of these two you could solder a 10k resistor in parallel with the 10k trimmer to bring the combined resistance closer to the target.
At turn on keep one hand on the on-off button and moisten a finger and put it say 5mm above the wire-wound to detect any heat. Turn off immediately if this gets hot.
If not measure the voltage drop across it and use Ohm's law to deduce the current. If this can be adjusted to close to zero then progress will have been achieved.
This circuit is similar to one shown in Motorola application note AN484A although that is rated at 3-5Watts.
Above that level driver transistors are necessary. Both output transistors have 0R56 emitter resistors to stabilize output currents. In that regard what you have is an economy that may not have paid off.
In all amplifiers in this paper there is an output zobel stability network of 10R and 100nF. The other stability measure of increasing the compensation capacitor from 22pF that I mentioned has not been implemented - options to consider if luck is not on your side.
In class B, everything works as intended by the manufacturer. The product is 1976, and the scheme is 1966 (Amperex) with simplifications 
But the owner does not like the original sound
I liked the sound with a high quiescent current (class A), but the heatsink and the power supply are not designed for this.
There is no temperature compensation in the AB, there is no convenient adjustment of the quiescent current, and thermal runaway occurs.
I guess there was never a magic sound here
I would suggest using a PCB with TDA2030 (LM1875) with a unipolar power supply (32v will do).
Caution! Online treatment. No one is responsible for the patient's health.
But the owner does not like the original sound I liked the sound with a high quiescent current (class A), but the heatsink and the power supply are not designed for this.
There is no temperature compensation in the AB, there is no convenient adjustment of the quiescent current, and thermal runaway occurs.
I guess there was never a magic sound here
I would suggest using a PCB with TDA2030 (LM1875) with a unipolar power supply (32v will do).
Caution! Online treatment. No one is responsible for the patient's health.
It's working The 1k resistor did the trick. The useful range is between 1k - 2k. With the 10k pot the adjustment is not easy. So far it's around 1k6 and the distortion has gone away. But 100r less and the voltage starts to drop.
By the way, I've been taking a look at the old output caps I replaced time ago and they are rated at only 16v. I know for sure they were the originals. How that makes any sense?
The 1k resistor did the trick. The useful range is between 1k - 2k. With the 10k pot the adjustment is not easy. So far it's around 1k6 and the distortion has gone away. But 100r less and the voltage starts to drop.By the way, I've been taking a look at the old output caps I replaced time ago and they are rated at only 16v. I know for sure they were the originals. How that makes any sense?
Last edited:
Economy class. The manufacturer decided that 16v (1/2 U) would be enough at the midpoint. Capacitors could BOOM !, but did not explode - good quality The design is optimized in its own way, so it is difficult to make changes that require other changes ...
You could only use only 1k pot (no R1 resistor). The only problem is the reliability of the middle contact.
The design is optimized in its own way, so it is difficult to make changes that require other changes ...You could only use only 1k pot (no R1 resistor). The only problem is the reliability of the middle contact.
Last edited:
Yeah, but .. um .. when the wiper goes a little intermittent, bias goes up. (Perhaps catastrophically !)
I would place a fixed resistor Base-Collector, then connect the trim-pot Base-Emitter. That way any indignities the wiper gives will REDUCE bias. Of course it won't sound very good, but it usually won't instantly eat the outputs.
nonost, Glad you noticed that 16V output coupling capacitor -- would be a good candidate for replacement. But the circuit tolerates it because in that circuit, the outputs always come 'up' from ground, rather than 'down' from supply-max.
Regards
edit: Oh, and forgot -- your question in post 105, the answer is *yest*, that'll give better resolution than the 10k pot. And congrats on finding a satisfactory solution.
!)I would place a fixed resistor Base-Collector, then connect the trim-pot Base-Emitter. That way any indignities the wiper gives will REDUCE bias. Of course it won't sound very good, but it usually won't instantly eat the outputs.
nonost, Glad you noticed that 16V output coupling capacitor -- would be a good candidate for replacement. But the circuit tolerates it because in that circuit, the outputs always come 'up' from ground, rather than 'down' from supply-max.
Regards
edit: Oh, and forgot -- your question in post 105, the answer is *yest*, that'll give better resolution than the 10k pot. And congrats on finding a satisfactory solution.
Last edited:
Hi Rick, thanks for passing by.
Yep, I still have to do more tests: with the turntable running and stuff, but it looks good. I run the amp with the vol pot at 12:00 for like 30 minutes and the voltages were ok and not going down. In the next days I will do the fine adjustment with the 1k pot and measuring current as well. With the 10k pot the sweet spot is very very narrow.
But yeah, I can't hear any distortion I don't have golden ears but this amp doesn't sound like 10% distortion at all.
I will let you know how it goes in a few days.
pd: that electro was replaced time ago. I just noticed today while I was messing around with the old components I swapped
Yep, I still have to do more tests: with the turntable running and stuff, but it looks good. I run the amp with the vol pot at 12:00 for like 30 minutes and the voltages were ok and not going down. In the next days I will do the fine adjustment with the 1k pot and measuring current as well. With the 10k pot the sweet spot is very very narrow.
But yeah, I can't hear any distortion
I don't have golden ears but this amp doesn't sound like 10% distortion at all. I will let you know how it goes in a few days.
pd: that electro was replaced time ago. I just noticed today while I was messing around with the old components I swapped
It's working great
These trimmers are the key. I've set the output transistors at 35mA (0.5R) for both channels. I've gone as far as 100mA, but the heatsink gets hotter. What's the right mA for this thing? At 35mA and after an hour of loud playback, the heatsinks right above the transistors are warm-hot, but you can put your finger on and stay there forever, no pain. However, at 100mA it can hurt after 20 mins or so, it's stable but hot.
Well, I'm quite happy. Actually I can't believe it. I thought it was a never ending thing this amp restoration.
I'm attaching a couple pics. You can see the heatsink for each channel: the bd139 is sandwiched between the bd131 & bd132 output transistors.
Thank you all!
These trimmers are the key. I've set the output transistors at 35mA (0.5R) for both channels. I've gone as far as 100mA, but the heatsink gets hotter. What's the right mA for this thing? At 35mA and after an hour of loud playback, the heatsinks right above the transistors are warm-hot, but you can put your finger on and stay there forever, no pain. However, at 100mA it can hurt after 20 mins or so, it's stable but hot.
Well, I'm quite happy. Actually I can't believe it. I thought it was a never ending thing this amp restoration.
I'm attaching a couple pics. You can see the heatsink for each channel: the bd139 is sandwiched between the bd131 & bd132 output transistors.
Thank you all!
Attachments
The heat sink appears to be part of the chassis which looks to be an alloy of a kind rather less efficient in removing heat than aluminium. You have to de-rate the power of transistors with increases in heat.
You could improve this by fitting a bar of aluminium of 5mm thickness as long as the ledge these are sitting on and wide enough to cover the plastic casing of the power transistors.
With this the devices should run relatively cool with a standing current of 35 m.A.
With Class B operation increasing the standing current too much forces the crossover between outputs to occur and a higher voltage setting where the gm slope is steeper where the base-emitter voltage has greater leverage in increasing current than it does near the turn on region where the slope is milder.
The bias adjustment scheme you have is basic and it could handle mild slopes such as with 35 m.A. I see 100 m.A. standing current as being in another league.
You could improve this by fitting a bar of aluminium of 5mm thickness as long as the ledge these are sitting on and wide enough to cover the plastic casing of the power transistors.
With this the devices should run relatively cool with a standing current of 35 m.A.
With Class B operation increasing the standing current too much forces the crossover between outputs to occur and a higher voltage setting where the gm slope is steeper where the base-emitter voltage has greater leverage in increasing current than it does near the turn on region where the slope is milder.
The bias adjustment scheme you have is basic and it could handle mild slopes such as with 35 m.A. I see 100 m.A. standing current as being in another league.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.