My dear old Creek 5050 has blown its output transistors and its drivers (14 transistors in total). It has taken some resistors out too and one zener is open circuit. I can easily replace the transistors but the resistors are blackened beyond recognition and the zener cannot be identified. The only way to sort this out is to look at the circuit diagram but I cannot find it and Creek are no help. Can anybody help with the power output stage circuit diagram? I have never had an amplifier go into thermal runaway on both channels at the same time. All power supply voltages are correct and there was no overheating before this happened.
Thanks Mooly but there are indeed 7 transistors per channel blown. This is both outputs (TIP35/36), both drivers, both voltage gain amps and one of the transistors setting up the bias (not the same one per channel). This power amp has 6 other transistors that are in the previous stages - it was a sophisticated design for an integrated. I need to find out what caused it to blow - I suspect the thermal paste had dried and cracked - it was running loud when it went. I note that later versions of this amp had fuses in the power supply lines, this one doesn't. Later variations also specified TIP 35c and TIP 36c, mine were B grades which means a lower max voltage. If one goes short there will be 84 volts across the remaining transistor which the B suffix variant won't handle. So it looks like one channel blown will take out the other. The C suffix transistors handle 100V so they should be more reliable.
Both channels would have to be on the verge of failure at overload. If one channel failed, the unregulated power supply would jump to a higher voltage/increased stress which could well have been enough to take out the second channel. Given the OP's "it was running loud when it went" I think that's not unlikely.
Unless it is a conventional design, the problem for anyone trying to assist you is that we won't own one these amplifiers, nor will we have a schematic to base specific suggestions on. You have one now, which is good of Creek and good for you but it does mean the fault finding and solution is all down to you.
Unless it is a conventional design, the problem for anyone trying to assist you is that we won't own one these amplifiers, nor will we have a schematic to base specific suggestions on. You have one now, which is good of Creek and good for you but it does mean the fault finding and solution is all down to you.
Not just if it goes short. The same applies when delivering a normal signal (and even with no speaker attached). As the output swings toward one or the other rail then one of the pair sees its collector-emitter voltage increase.
Thanks for your replies. Ian Finch has answered tick n pop's query. The amplifier did run perfectly (save for new relays) for 26 years so I cannot complain and well done Creek for supplying the circuit. It is a surprisingly sophisticated circuit for an integrated amp. Once I get the components I'll have a go at the repair though one semiconductor is no longer made. Again, this components is rated at a higher voltage in the later versions of this amp. I'll use the advice about connecting a lamp in series with the power rails. The 'zener' I spoke of is actually a 1N4148. The blown one is a small weedy device with no markings, hence the confusion. The other 3 (marked 1N4148 and look like 1N4148s) are all fine. Perhaps this was the culprit.
I've just re-read my replies here and realise that I still don't know why both channels failed. If the output stage of one channel went short that would lower the supply voltages wouldn't it? It is many years since I retired as an AV technician but I remember we never had amps with both channels blown. Both supply lines are 42.5 volts as the circuit suggests. (This gives 85 volts across the power amp stage - 80volt TIP35B transistors is disappointing here). Thanks for your help but I might give up on this one! Mooly I've just re-read your first response - ignore my answer about the lamp.
Transistors fail in milliseconds when provoked 
As one channel went down it would probably do so asymmetrically, first pulling one rail down via the failed transistor and through the speaker initially. The other 'good' channel would be unbalanced with one rail effectively missing and as such probably shove a large current through its speaker momentarily, this causing destruction of its output stage.
An 80 volt combined rail voltage really needs at least 120 volt rated parts to avoid the possibility of secondary breakdown.
https://en.wikipedia.org/wiki/Safe_operating_area#Second_breakdown
As one channel went down it would probably do so asymmetrically, first pulling one rail down via the failed transistor and through the speaker initially. The other 'good' channel would be unbalanced with one rail effectively missing and as such probably shove a large current through its speaker momentarily, this causing destruction of its output stage.
An 80 volt combined rail voltage really needs at least 120 volt rated parts to avoid the possibility of secondary breakdown.
https://en.wikipedia.org/wiki/Safe_operating_area#Second_breakdown
Thanks Mooly. The left channel failed first as I recall and the right channel was still playing when I switched off the amp. The relays disconnect the speakers but leave the power supply to the amps on. The second channel must have blown at this time. The amp always did run hot (especially the drivers) so it probably didn't need a lot of provoking. Curiously, I've pressed into service an old Leak stereo 30 that I had given years ago. Simple old fashioned design that runs quite cold, even at high volumes. The bulk carbon resistors have been replaced - common to go high in value as I recall but otherwise as it was built in the 70's. Wonder why modern electronics does run so hot - how many devices now have fans?
Ah, a Leak 30. Classic. You might actually find you like the sound better than the Creek
Modern components are specified to operate at high temperatures, often way over 100C but that doesn't always mean its good practice to do so. PA and stage gear often has fans but not so much for home audio. To noisy.
Modern components are specified to operate at high temperatures, often way over 100C but that doesn't always mean its good practice to do so. PA and stage gear often has fans but not so much for home audio. To noisy.
I've got this up and running now. There seems to have been no 'cause' for the blown output transistors though the heat sink now gets quite warm. I did think it odd when I replaced the relays that the heatsink seemed quite cold even though the drivers were hot. The amp is quite crowded so it wasn't possible to see how hot the actual outputs were previously. Now the drivers are only luke warm. Perhaps the thermal paste had failed over the years and caused the output transistors to overheat and it was just a matter of time....
I notice quite a lot of people have viewed this thread. Those looking for info about it might note that it is an excellent amplifier with a really expensive toroidal transformer. It has excellent clarity and timing which is missing from many modern amps though it is a bit sharp. Best go for a later one (mine was 1989 according to a label inside) if you are looking to buy. I agree with Mooly though - the Leak is a very pleasant sound and I might keep it on a while though I will need to bypass the equalisation feedback to convert the disc inputs to line level so that I can attach more devices!
One final question - I couldn't find gain matched TIP35/36 transistors so I bought a few to try to match the nearest. In fact all of the TIP 35s were about the same, as were the TIP 36's but not the same as each other. To set the quiescent current Creek advise setting 6mV across the emitter resistors. Mine are slightly unbalanced at 5mV and 6mV. Is this significant? It sounds OK.
I notice quite a lot of people have viewed this thread. Those looking for info about it might note that it is an excellent amplifier with a really expensive toroidal transformer. It has excellent clarity and timing which is missing from many modern amps though it is a bit sharp. Best go for a later one (mine was 1989 according to a label inside) if you are looking to buy. I agree with Mooly though - the Leak is a very pleasant sound and I might keep it on a while though I will need to bypass the equalisation feedback to convert the disc inputs to line level so that I can attach more devices!
One final question - I couldn't find gain matched TIP35/36 transistors so I bought a few to try to match the nearest. In fact all of the TIP 35s were about the same, as were the TIP 36's but not the same as each other. To set the quiescent current Creek advise setting 6mV across the emitter resistors. Mine are slightly unbalanced at 5mV and 6mV. Is this significant? It sounds OK.
That's great that you have it fixed
The difference in bias is nothing.... do you mean you have 5 mv across one emitter resistor of the pair and 6 across the other ? That kind of imbalance can be down to two things.
1/ You are measuring the voltage with speakers connected, and any slight DC offset pulls current through the speaker and skews the result. Always set the bias with no load attached.
2/ Depending on the circuit design, the drivers may be pulling slightly different currents through one or both of the resistors. Nothing you can do about that.
The difference in bias is nothing.... do you mean you have 5 mv across one emitter resistor of the pair and 6 across the other ? That kind of imbalance can be down to two things.
1/ You are measuring the voltage with speakers connected, and any slight DC offset pulls current through the speaker and skews the result. Always set the bias with no load attached.
2/ Depending on the circuit design, the drivers may be pulling slightly different currents through one or both of the resistors. Nothing you can do about that.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.