OK will check those tonight.
I'm thinking I should perhaps go back through the amp with a square wave or sine wave and see if I can find where the distortion is starting...
Just thinking; Both channels draw power and, of course, the bias setting current from the same source (FF). The hum starts and rises as you increase bias in this channel only. That is not so easy to understand if both channels operated and delivered audio to speakers at some decent level.
The implication is that the bias generator circuit around the base of Q7 or its series components like CR4B, Q7,Q8 or connections are faulty yet having changed them, nothing is different. Since the amplifier still works, that leaves resistor failures and the need to again compare voltages across the suspects in this amplifier section.
To be certain that this is not arising via the feedback loop, I'd just remove Q10 and test but to be on the safe side, swap Q9 and Q10 and listen again. This is easy enough but I doubt it is the cause since this is bias current related and that means the noise is derived via the current source.
I don't think using a signal generator will track down existing noise. It will add its own but that is not the issue. An instrument that senses and amplifies or displays the existing noise (and I'm suggesting it is hum from the power supply) is what is needed and that means an oscilloscope to see or another amplifier with either a floating supply like a battery or completely capacitive coupling to hear the relative loudness at different points in isolation, as discussed earlier. Gadgets like this were once called signal tracers and in the days that this type circuitry was common, every bench had one or something like it, such as the audio section of an old transistor radio. When cash is tight, this sort of device becomes an attractive, if limited option.
The implication is that the bias generator circuit around the base of Q7 or its series components like CR4B, Q7,Q8 or connections are faulty yet having changed them, nothing is different. Since the amplifier still works, that leaves resistor failures and the need to again compare voltages across the suspects in this amplifier section.
To be certain that this is not arising via the feedback loop, I'd just remove Q10 and test but to be on the safe side, swap Q9 and Q10 and listen again. This is easy enough but I doubt it is the cause since this is bias current related and that means the noise is derived via the current source.
I don't think using a signal generator will track down existing noise. It will add its own but that is not the issue. An instrument that senses and amplifies or displays the existing noise (and I'm suggesting it is hum from the power supply) is what is needed and that means an oscilloscope to see or another amplifier with either a floating supply like a battery or completely capacitive coupling to hear the relative loudness at different points in isolation, as discussed earlier. Gadgets like this were once called signal tracers and in the days that this type circuitry was common, every bench had one or something like it, such as the audio section of an old transistor radio. When cash is tight, this sort of device becomes an attractive, if limited option.
The hum is there all the time, and the ONLY thing that's ever affected it was increasing the bias. It just seemed like increasing the bias was increasing the output level of that section slightly.
I can spend more time comparing voltages at more or less every point in the power amp sections. Something's got to turn up somewhere!
Will try that.
Yeah a signal tracer would be good, the scope works but sometimes it would be useful to hear what the waveform actually sounds like.
What I'm trying to solve most immediately is no longer the hum, which persists, but the new distortion and lower output which is now the bigger problem. Something happened exactly when I was checking the bias voltage across R82 and instantly the output level dropped a notch and substantial distortion is now present in the audio. I was thinking if I sent a square wave through I could see where that distortion was first appearing. Since I have no idea what's causing the distortion, it seems like it might be good to rule out the preamp etc.
Is it likely that the lack of bias voltage would cause distortion? I guess the first order of business should be to restore the correct bias voltage and then see what's what from there, does that make sense? Thanks....
p.s. I did last weekend spend much time looking at the hum on the scope and it did seem strongest at the transformer windings. But like I said 1st I have to find what's causing the lack of bias voltage and this distortion.
OK I think finally a breakthrough. I tried swapping Q's 9 & 10, no change. So I went ahead and swapped Q's 5&6 with Q's 8&7. Guess what. The hum moved to the other side, and so did the distortion.
The question of the moment for me is simply how many transistors should I replace. All of them? 5,6,7 & 8? Do I need a matched quad?
The question of the moment for me is simply how many transistors should I replace. All of them? 5,6,7 & 8? Do I need a matched quad?
The symptoms you've now clarified are really those of having a transistor fail - either one of the outputs (Q1,Q2) or drivers (Q7, Q8). When you probe about a lot, the odds of a short occurring increase as I can attest, having had the same thing happen more times than I like to admit. Still, there's no firm evidence that's the case yet. Transistors most often fail with an internal short, where you may still have 0.65V across the B-E junction but a shorted C-E. It should be easy enough to read any short as a low resistance with a multimeter, since you can pull the drivers and that should also allow testing the output transitors without having to remove or otherwise disconnect them. I assume you know that power should be off for this kind of testing where you rely on the meter battery for test current.
The key symptoms for a blown output stage are low volume, distorted sound and inability to set bias. Bias setting, unfortunately, is the cause of many accidents when eager beavers slip with probes or misunderstand just how rapidly things go pear-shaped when they set it to a extreme position because the setting does not settle fully until there is thermal equilibrium, as when the cover is on. Realistically, with today's consistent performance components, it would be better to have fixed resistors with no adjustment possible.
Anyways, to be sure you are looking at a short, reverse the meter probes and it will still be there.
The key symptoms for a blown output stage are low volume, distorted sound and inability to set bias. Bias setting, unfortunately, is the cause of many accidents when eager beavers slip with probes or misunderstand just how rapidly things go pear-shaped when they set it to a extreme position because the setting does not settle fully until there is thermal equilibrium, as when the cover is on. Realistically, with today's consistent performance components, it would be better to have fixed resistors with no adjustment possible.
Anyways, to be sure you are looking at a short, reverse the meter probes and it will still be there.
I'm sorry I missed this post for some reason ...it would have saved a bit bandwidth but yes, that's the matter and RCA 36643 is long obsolete but there may be some stock around in the US, at least. The standard replacement type is 2N3053 but there were many suppliers of that part and I couldn't say that performance will always be the same using it.
It would be better to have both drivers in each channel matched to some degree such as within 10% for Hfe, or beta. However, you would go through a lot of transistors to only a small benefit to get this and I certainly wouldn't try this in an old amplifier that almost certainly didn't have transistor matching. Still, using parts from the same batch or container often is a good sorting method so taking some care otherwise would be fair enough.
In short, use the same types in the faulty channel and listen to results before doing the same substitution in the other. Price is going to have a lot to do with this anyway.
From #66, I deduced the problem was most likely drivers, not output transistors and I was aware you had gone through cleaning and remounting the outputs earlier.
For the present, getting the right parts is all that can be done to get operational but likely the hum will return and as you confirm, this varies with the bias current setting.
That means that the more current drawn through the output stage, the more hum and that hum is not signal from the input stage, which is unchanged, but from the power supplies to the output stage, where current does change.
We'll have to look at that again when you have the parts or you could work on just this channel.
For the present, getting the right parts is all that can be done to get operational but likely the hum will return and as you confirm, this varies with the bias current setting.
That means that the more current drawn through the output stage, the more hum and that hum is not signal from the input stage, which is unchanged, but from the power supplies to the output stage, where current does change.
We'll have to look at that again when you have the parts or you could work on just this channel.
I did not run the amp for very long with the driver transistors swapped, but from what I could tell, the hum switched channels with the transistors. Hoping that's the case. Can a transistor cause a hum or pick up power supply hum in any way? Maybe wishful thinking
Thanks Ian.p.s. it may be a week or two before I get the Q's so if you don't hear from me in this thread for a bit that will be why...
No problem - it's the way of our modern globalised supply world. Huge market and huge delays. Funny, it takes only 4 days for Mouser, Texas to deliver a carton with 40 items to me in a rural district 350 miles from a major city airport and 10,000 miles distant, but 2 -5 weeks to airmail a couple of tiny signal transistors from Hong Kong 
We were looking at this hum issue before the incident with the driver and if you just consider this section of the amplifier, the causes can only be due to the supply and a simple list of parts and connections. If, when you replace the drivers, this small residual hum doesn't transfer with a driver swap, you can be fairly sure you have a resistor issue and one is open.
The lesser possibilies of a fault in the previous stage could be eliminated by pulling both input stage transistors (Q9,Q11 or Q10,Q12) and rechecking hum. Note,that R58 should also be lifted or the connection GG lifted too, to ensure no possible supply hum from the small remaining current through the transformer primary there.
The transistors don't generate or receive hum at this stage of the amplifier where voltage gain is only 1, so as I'm saying, the influence is from supply current flowing where it shouldn't, as when parts have shorted or gone open circuit. The rectifier switching noise that is always on the power supply rail, depends on caps and resistors to attenuate and keep it out of the input stage where voltage gain raises it to audible. When the current exceeds the design level, the caps are less effective at their filtering job and hum rises, as you hear. You've replaced the caps, but that may have just exposed another fault that has been there for years. The obvious place to start is at C16 by simply paralleling another large electro. of sufficient voltage rating across it. Then you know where the problem is - or not.
If it were my amp, I would pull all the old resistors in the output stage, bar the 0.5 ohm output emitter ones and replace with cheap and cheerful half or 1W carbon film types from your local store. They will be more reliable, at least. That'a long and somewhat expensive process though, if you have to buy packs of more than 2 parts from Radio Shack etc.
Re: matching - #69 already suggests that there will be little benefit and you can't match parts without a large stock to sort from. With these parts the only matching would be the process wafer quality and packaged lots, which are still quite good matching methods. From just a few loose parts stock? Nope.
We were looking at this hum issue before the incident with the driver and if you just consider this section of the amplifier, the causes can only be due to the supply and a simple list of parts and connections. If, when you replace the drivers, this small residual hum doesn't transfer with a driver swap, you can be fairly sure you have a resistor issue and one is open.
The lesser possibilies of a fault in the previous stage could be eliminated by pulling both input stage transistors (Q9,Q11 or Q10,Q12) and rechecking hum. Note,that R58 should also be lifted or the connection GG lifted too, to ensure no possible supply hum from the small remaining current through the transformer primary there.
The transistors don't generate or receive hum at this stage of the amplifier where voltage gain is only 1, so as I'm saying, the influence is from supply current flowing where it shouldn't, as when parts have shorted or gone open circuit. The rectifier switching noise that is always on the power supply rail, depends on caps and resistors to attenuate and keep it out of the input stage where voltage gain raises it to audible. When the current exceeds the design level, the caps are less effective at their filtering job and hum rises, as you hear. You've replaced the caps, but that may have just exposed another fault that has been there for years. The obvious place to start is at C16 by simply paralleling another large electro. of sufficient voltage rating across it. Then you know where the problem is - or not.
If it were my amp, I would pull all the old resistors in the output stage, bar the 0.5 ohm output emitter ones and replace with cheap and cheerful half or 1W carbon film types from your local store. They will be more reliable, at least. That'a long and somewhat expensive process though, if you have to buy packs of more than 2 parts from Radio Shack etc.
Re: matching - #69 already suggests that there will be little benefit and you can't match parts without a large stock to sort from. With these parts the only matching would be the process wafer quality and packaged lots, which are still quite good matching methods. From just a few loose parts stock? Nope.
Hi, It wouldn't be the first time hopes are raised by some false listing on the net.
Sorry about that but I think NTE 128 will work OK in most conditions though I ran a short thread in the Parts Forum to ask about these RCA 3xxxx codes which don't appear in contemporary RCA general catalogues but their later replacement SK 3024/128, which lasted until the 90's apparently, did. http://www.diyaudio.com/forums/parts/232080-rca-transistor-vintage-search.html
I prefer 2N3019 as a heavier duty model than NTE128 which would suit 4 ohm speakers and today's bass levels better with its higher current rating. 'Just my view, mind you.
Sorry about that but I think NTE 128 will work OK in most conditions though I ran a short thread in the Parts Forum to ask about these RCA 3xxxx codes which don't appear in contemporary RCA general catalogues but their later replacement SK 3024/128, which lasted until the 90's apparently, did. http://www.diyaudio.com/forums/parts/232080-rca-transistor-vintage-search.html
I prefer 2N3019 as a heavier duty model than NTE128 which would suit 4 ohm speakers and today's bass levels better with its higher current rating. 'Just my view, mind you.
OK Thank you SO much for researching this for me! I'll probably go ahead and find some 2N3019's and order them next time I'm ordering parts. The NTE's will arrive sooner and I can at least use them to work on the hum. I don't really want to run the amp at all the way it is right now. Thanks again.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.