Hi, I hope posted the thread in right section.
I have some driver and power transistors dead and I want find a good replacement with specs as close possible to the originals.
Appreciate any helps!
Thanks
Here the parts:
2SC897 (SCHEMATIC SHOW 2SD217)
2SC959
2SA606
2SA762
2SD427
2SB557
Thanks all
S.
I have some driver and power transistors dead and I want find a good replacement with specs as close possible to the originals.
Appreciate any helps!
Thanks
Here the parts:
2SC897 (SCHEMATIC SHOW 2SD217)
2SC959
2SA606
2SA762
2SD427
2SB557
Thanks all
S.
Those look and sound like old devices to me.
If you are thinking of doing a big swap of semiconductors in an old amp then be aware that modern parts made with modern processes are fundamentally 'different' in several areas. They will be generally faster which can lead to stability issues, and also (and possible of great importance in driver and output stages) have slightly different forward bias voltages due to modern doping processes in manufacture.
This can cause issues with for example bias setting by putting it out of the range of adjustment, and while all these issues are normally easily worked around by an experienced tech, to just fit replacements and hope it works may lead to disappointment.
Don't want to sound all doom and gloom but these are things you do need to be aware of 🙂
If you are thinking of doing a big swap of semiconductors in an old amp then be aware that modern parts made with modern processes are fundamentally 'different' in several areas. They will be generally faster which can lead to stability issues, and also (and possible of great importance in driver and output stages) have slightly different forward bias voltages due to modern doping processes in manufacture.
This can cause issues with for example bias setting by putting it out of the range of adjustment, and while all these issues are normally easily worked around by an experienced tech, to just fit replacements and hope it works may lead to disappointment.
Don't want to sound all doom and gloom but these are things you do need to be aware of 🙂
Hi, Thanks for your advice, but unfortunately these original parts are no longer found, and I don't want to buy Chinese parts that are just fakes.
S.
S.
I'd guessed as much 🙂
What I'm saying is that you need to look more closely at the circuit and see whether it might lend itself to modern current production replacements and/or if not what possible measures you might take to get such devices to work.
For that we would need to know the unit in question... and even then it may not be possible to give definitive answers.
What I'm saying is that you need to look more closely at the circuit and see whether it might lend itself to modern current production replacements and/or if not what possible measures you might take to get such devices to work.
For that we would need to know the unit in question... and even then it may not be possible to give definitive answers.
We worry and go silly over straws and carrier bags.. and yet a transistor can mean the difference between the dump and another few years of happy use... just seems insane that discrete parts come and go so quickly.
I just googled the 1060. What a beauty.
Have you seen this:
Marantz 1060 Repair/Recap | Audiokarma Home Audio Stereo Discussion Forums
Last edited:
I see the Marantz uses the once popular double diode package for thermal compensation. These have a reputation for failing, possibly intermittently and the effect is a massive increase in bias and ultimately failure of the output stage.
If I was faced with this I would look to fit modern parts such as 2N3773 for the outputs. The drivers I think are round T05 packages and so modern equivalents are out in that style. BD139/140's could be workable... if the drivers have clip on heatsinks then you would have to fit some suitable ones to the flatpak drivers.
The diode pack is more problematic and I would probably ditch that and use a vbe multiplier in its place. That's easy to do with a single transistor and preset.
You also need to use a bulb tester to initially test the amp and use a scope to make sure stability is OK. If not then you need to tweak the compensation to suit.
If I was faced with this I would look to fit modern parts such as 2N3773 for the outputs. The drivers I think are round T05 packages and so modern equivalents are out in that style. BD139/140's could be workable... if the drivers have clip on heatsinks then you would have to fit some suitable ones to the flatpak drivers.
The diode pack is more problematic and I would probably ditch that and use a vbe multiplier in its place. That's easy to do with a single transistor and preset.
You also need to use a bulb tester to initially test the amp and use a scope to make sure stability is OK. If not then you need to tweak the compensation to suit.
I just looked up dim bulb tester. Couldn't you just use an RCD?
Edited to add.. oh I see.. the idea is to slowly feed power to the amplifier.. an RCD doesn't do that.
Edited to add again: er.. regular bulbs only go up to 100W.. an integrated amp might consume anything from 100W to 1000W?!?
Edited to add.. oh I see.. the idea is to slowly feed power to the amplifier.. an RCD doesn't do that.
Edited to add again: er.. regular bulbs only go up to 100W.. an integrated amp might consume anything from 100W to 1000W?!?
Last edited:
An RCD is a totally different thing.
The bulb tester uses the positive temperature coefficient of a filament to safely limit current in the event of excess current being drawn. If the amp draws little current then the filament is cool and presents as a low resistance thus allowing almost full mains to be delivered. If current increases, the filament resistance rises sharply automatically limiting the current the amp can draw. Result... no blown output transistors while you are working on the unit.
The RCD is purely safety related and won't help save an amp from destroying itself. The RCD compares live and neutral currents and should a difference exist then the circuit is tripped.
The bulb tester uses the positive temperature coefficient of a filament to safely limit current in the event of excess current being drawn. If the amp draws little current then the filament is cool and presents as a low resistance thus allowing almost full mains to be delivered. If current increases, the filament resistance rises sharply automatically limiting the current the amp can draw. Result... no blown output transistors while you are working on the unit.
The RCD is purely safety related and won't help save an amp from destroying itself. The RCD compares live and neutral currents and should a difference exist then the circuit is tripped.
100W incandescent bulb will not pass 100W in series with load, but will allow you to power up a CUT (circuit under test) while at idle or low output for testing and manipulation. If a short is created under this circumstance, it will prevent catastrophic failure.
The benefit of this simple tool is also great when working with off-line SMPS.🙂
The benefit of this simple tool is also great when working with off-line SMPS.🙂
Another technique that can be even more protective than a dim-bulb tester is to add temporary series resistors between the filter caps and the amplifier rails. This prevents the filter caps discharging high currents through the amp if something is wrong.
With something like 220 ohm 2W resistors on each rail you are unlikely to burn out drivers or output devices, and there is enough current to bias up to realistic operating point for a class B. You lose some supply volts, but its better than losing half the transistors!
When testing a new amp circuit on a breadboard I normally progress from +/-25V supply and 500 ohm series resistors, to 100 ohm resistors, then something like 10 ohms and progress to full rail voltage (I have homebuilt supply with two transformers that I can switch from +/-25V to +/-50V or so). I start with something like 50 or 100 ohm resistor in series with a small loudspeaker (prevents deafening sound levels if there is oscillation, and reduces current requirements for testing). 8 ohm dummy follows after everything tests out, and if that's stable at low levels I remove the current limiting resistors.
Replacing a lot of semiconductors is probably treated in a similar manner, ie like building a new amp circuit, given the scope for circuit error, oscillation, overheated parts, wild bias levels, etc. Limit the current and most problems will be diagnosable without smoke. If you have access to a thermal camera that's a great aid too (but a cautious finger works for detecting hot spots too).
With something like 220 ohm 2W resistors on each rail you are unlikely to burn out drivers or output devices, and there is enough current to bias up to realistic operating point for a class B. You lose some supply volts, but its better than losing half the transistors!
When testing a new amp circuit on a breadboard I normally progress from +/-25V supply and 500 ohm series resistors, to 100 ohm resistors, then something like 10 ohms and progress to full rail voltage (I have homebuilt supply with two transformers that I can switch from +/-25V to +/-50V or so). I start with something like 50 or 100 ohm resistor in series with a small loudspeaker (prevents deafening sound levels if there is oscillation, and reduces current requirements for testing). 8 ohm dummy follows after everything tests out, and if that's stable at low levels I remove the current limiting resistors.
Replacing a lot of semiconductors is probably treated in a similar manner, ie like building a new amp circuit, given the scope for circuit error, oscillation, overheated parts, wild bias levels, etc. Limit the current and most problems will be diagnosable without smoke. If you have access to a thermal camera that's a great aid too (but a cautious finger works for detecting hot spots too).
If I go to a *serious* lamp supplier, I can still easily get 250 Watt incandescents. They are useful in boathouses and such, where large light is needed for short hours (efficiency not as important as lamp cost).
That's in Edison screw sockets (in the US, above 250W we must use the larger Mogul socket, and that is getting scarce).
But "tubular halogen" is again an incandescent lamp, the T3 size available in 250 Watts and 500 Watts. You need a proper fixture but they have always been cheap. (Can be free today: I have thrown-out several for LED replacements.)
And very few transistor amps suck large power when idle, or playing very soft. A "200W+200W" amp may idle under 50 Watts. And the cold filament resistance is low, so very little voltage is dropped in the lamp *unless* the amplifier sucks BIG current.
Attachments
Please, be aware if your amplifier self oscillates without power limiting the power stage will be destroyed. I used the series incandescent lamp limiter and it works. Without it I would not have been able to test the amplifier.
If your amplifier does not self oscillate and your bias is correct, the filament should remain red hot or dimmer. If it turns yellow or light orange, re-inspect the amplifier.
When you are changing transistors, you are also changing the collector-base capacitance. The latter causes a phase difference between the collector current and collector voltage. As your amplifier has been designed with different transistors, it means, the new base-collector capacitance was NOT accounted for in the original circuit. This is why you may get free oscillations which often are at several hundreds of kilohertz. The original circuit did not account for these unwanted high frequency oscillations, and they are, very demanding on output transistors. The output filters, will conduct an unusually high current if frequencies like those are generated.
If your amplifier does not self oscillate and your bias is correct, the filament should remain red hot or dimmer. If it turns yellow or light orange, re-inspect the amplifier.
When you are changing transistors, you are also changing the collector-base capacitance. The latter causes a phase difference between the collector current and collector voltage. As your amplifier has been designed with different transistors, it means, the new base-collector capacitance was NOT accounted for in the original circuit. This is why you may get free oscillations which often are at several hundreds of kilohertz. The original circuit did not account for these unwanted high frequency oscillations, and they are, very demanding on output transistors. The output filters, will conduct an unusually high current if frequencies like those are generated.
- Status
- Not open for further replies.