Some pics I took last weekend, should have all electrolytics replaced by this weekend:
Fisher 440T
Fisher 440T
Nice, interesting pics too! You sure did the right thing in replacing those electrolytics. I note you didn't find out
much by disconnecting the input to one channel and this is reasonable because you can't compare it with the
other any more for signal differences at least. What to do is earth both or fit equal resistors of around half the
volume pot. value. I'd be more specific but the manual you posted has disappeared from your website.
You could link it here if you don't want to post it any more.
The images remind me just how old those carbon composition resistors are and how far they drift (go high in value)
from the colour coded values. It's easy to say 'check and replace' but some will read as a combination with other
parts in the circuit and confuse the measurement. It means you may have to lift one end to be sure what you have.
We are certainly in the right area if you can kill all hum by grounding the input before the first capacitor.
(I'm sure there's 0.22uF there) You likely have the fault in the cap or first couple of transistors and parts before
the transformer. Let's assume there is no fault with the transformer like shorted or open windings. You can test
those when all else fails but it's not something to look forward to.
much by disconnecting the input to one channel and this is reasonable because you can't compare it with the
other any more for signal differences at least. What to do is earth both or fit equal resistors of around half the
volume pot. value. I'd be more specific but the manual you posted has disappeared from your website.
You could link it here if you don't want to post it any more.
The images remind me just how old those carbon composition resistors are and how far they drift (go high in value)
from the colour coded values. It's easy to say 'check and replace' but some will read as a combination with other
parts in the circuit and confuse the measurement. It means you may have to lift one end to be sure what you have.
We are certainly in the right area if you can kill all hum by grounding the input before the first capacitor.
(I'm sure there's 0.22uF there) You likely have the fault in the cap or first couple of transistors and parts before
the transformer. Let's assume there is no fault with the transformer like shorted or open windings. You can test
those when all else fails but it's not something to look forward to.
Last edited:
Fixed the link, had accidentally blown it away when I put the pics in the same place. I woke up thinking about a 47uF (was 50) cap I replaced that physically is located in about the right area of the quiet channel - I haven't traced it's connections, just replaced it along with everything else as I was going through it. If it's part of the left channel circuitry then it's missing on the other side. But it might just be a power filter for one of the boards; there was only one of them. Anxious to get back to trace that. Visible in the next to last pic, upper right, black/gold...
C19,20,23,24 form high pass filters and look suspicious, if you haven't already given them the heave-ho. Just recheck the soldering there, though I'm sure you've been careful.
The amplifier negative feedback goes from the output connections at C29,30 to C27/R87 and C28/R88 then back to the emitters of Q11,12 respectively. Those are quite sensitive runs and the dress or positioning of any wires there might cause trouble with hum. 'Perhaps worth a look and prod if it is the case.
The amplifier negative feedback goes from the output connections at C29,30 to C27/R87 and C28/R88 then back to the emitters of Q11,12 respectively. Those are quite sensitive runs and the dress or positioning of any wires there might cause trouble with hum. 'Perhaps worth a look and prod if it is the case.
19,20,23,24 long gone; 19 & 20 are blue Nichicons and 23&24 are little Silmics. You can see them on the left side of the last photo, for the right (noisy) channel. I'll poke around there and maybe resolder all of those just in case. Likewise with the coupling cap and surrounding connections... So C27 & 28 are part of the feedback circuit? Need to learn how that works... Is that the whole function of Q's 11 & 12?
Thanks a bunch, helps a lot to have this support.
-Mark
OK I believe I've tracked the hum down and I do believe it's one of the power amp's input transformers. The noise is relatively loud and clear on the scope, on the positive side of C24. One component away in any direction, it's weaker. Done deal, yes?
I guess the other thing I should do is compare the signal at that point to the other channel at the same point. That will be for Saturday...
I guess the other thing I should do is compare the signal at that point to the other channel at the same point. That will be for Saturday...
NFB
Q11 & 12 with Q9 & 10 buffer the input signal and provide some gain as well as a summing point for global feedback with the input signal. When connected to the emitter, feedback is in anti-phase with the base input and output at the collector, hence it becomes negative feedback and the amplifier is then fed a differential signal that incorporates instantaneous correction of errors between the input and output. That also means it fixes gain by the ratio of its level to that of the input with the effect of scaling resistors R87,R56 etc. taken into account. In this case, the voltage ratio of the transformer windings provides most of the amplifier voltage gain but this is reduced to (R88+R56)/R56 or x 28 when the feedback loop is added (closed)
That's quite simplistic because the whole story of negative feedback quickly becomes a massive mathematical nightmare that causes eyes to glaze over and gives most people a strong desire to think of something else.
Possibly the transformer is a problem but yes, compare signal and DC voltage at C24+, C23+ or Q10, Q9 emitters. Then power-off and measure R50 resistance if possible. Recheck connection polarity of C24, C20. Also check resistance from Q10, Q9 emitters to transformer frame. If it's not the same, try a fly clip grounding the transformer frame to C20- terminal. It seems unlikely that it would be poorly grounded and I don't really know how sensitive this condition would be. For a longshot, you could also swap Q9 & 10 and then Q11 & 12 and power up again.
Looking at your last pic, I see that some caps are joined before connecting them to grounding lugs. Judging by the empty lugs alongside, these may have been separate runs. It may make no odds but if they were separately grounded, try to keep it that way for hum and/or circuit interaction reasons.
Q11 & 12 with Q9 & 10 buffer the input signal and provide some gain as well as a summing point for global feedback with the input signal. When connected to the emitter, feedback is in anti-phase with the base input and output at the collector, hence it becomes negative feedback and the amplifier is then fed a differential signal that incorporates instantaneous correction of errors between the input and output. That also means it fixes gain by the ratio of its level to that of the input with the effect of scaling resistors R87,R56 etc. taken into account. In this case, the voltage ratio of the transformer windings provides most of the amplifier voltage gain but this is reduced to (R88+R56)/R56 or x 28 when the feedback loop is added (closed)
That's quite simplistic because the whole story of negative feedback quickly becomes a massive mathematical nightmare that causes eyes to glaze over and gives most people a strong desire to think of something else.
Possibly the transformer is a problem but yes, compare signal and DC voltage at C24+, C23+ or Q10, Q9 emitters. Then power-off and measure R50 resistance if possible. Recheck connection polarity of C24, C20. Also check resistance from Q10, Q9 emitters to transformer frame. If it's not the same, try a fly clip grounding the transformer frame to C20- terminal. It seems unlikely that it would be poorly grounded and I don't really know how sensitive this condition would be. For a longshot, you could also swap Q9 & 10 and then Q11 & 12 and power up again.
Looking at your last pic, I see that some caps are joined before connecting them to grounding lugs. Judging by the empty lugs alongside, these may have been separate runs. It may make no odds but if they were separately grounded, try to keep it that way for hum and/or circuit interaction reasons.
Last edited:
Just reading that paragraph before coffee this morning almost sent me back to bedQ11 & 12 with Q9 & 10 buffer the input signal and provide some gain as well as a summing point for global feedback with the input signal. When connected to the emitter, feedback is in anti-phase with the base input and output at the collector, hence it becomes negative feedback and the amplifier is then fed a differential signal that incorporates instantaneous correction of errors between the input and output. That also means it fixes gain by the ratio of its level to that of the input with the effect of scaling resistors R87,R56 etc. taken into account. In this case, the voltage ratio of the transformer windings provides most of the amplifier voltage gain but this is reduced to (R88+R56)/R56 or x 28 when the feedback loop is added (closed)
That's quite simplistic because the whole story of negative feedback quickly becomes a massive mathematical nightmare that causes eyes to glaze over and gives most people a strong desire to think of something else.
Will do. Polarities of C24,C20 are good (last picture, bottom left). The bottom of this amp is really clean; the top side had some moisture exposure and there's a bit of mold on the outer covering of the transformers. Not impossible there'd be some corrosion between the xformer frames and chassis...
I guessing - tell me what you think - I'm OK with these. Two of those are the pair on either side of the FM Stereo lamp, C's 10&11, and the other pair are C's 14&15, on either side of the 24V zener regulator. These were originally larger axials the spanned that entire distance (see 1st pic). This topic has always confounded me a bit, how could there be a difference between an inch of copper wire connecting two caps vs an inch of aluminum. In any case, I do get that there are some aspects of audio electronics that approach voodoo in their mysteriousness and etherial nuance...
Last edited:
grounding
DC supplies, lamps or other devices aren't fussed if you connect return leads together before grounding or not, if that's all that we are concerned with. With audio signals using a steel or aluminium chassis, or a large PCB, consequences become blurred because the chassis and wiring have some measurable resistance which enables currents to sum with others that can be profoundly dirty like AC wiring, rectifier, regulator and filter cap currents or the amplified audio itself. If there were zero ground resistance and very short leads throughout, this couldn't happen.
The effects may only be tiny, but because of the amplification, we start to hear them along with any distortion products. So, if you are choosy about really clean audio, then it's a matter of concern to make sure these are inaudible.
When we look at quality commercial products, we also see techniques borrowed from high-end ones but only to the extent that significant costs are still avoided. That's probably the lowest expected level of attention we should pay to these nuances of design but unfortunately, it means we have to learn and carefully juggle these grounding principles, construction guidelines etc. which, as you suggest, seem to be "out there" in comparison to the basic circuits, parts and wiring that we may be more familiar with.
For a concept that helps me wrestle with grounding, I think of it as a mass of copper and the connections to it are best if leads are only mm rather than inches long. The bigger the circuit, the harder to do all those connections and group them around the copper mass, so the circuits are broken up into isolated sections that are then connected to a single point or "star" ground. Then the rules of priorities in what should have shortest leads, be routed here or there etc. start to come into play.
You get the feeling it's not such an ideal world in electronics, after all
.
- rant over.
DC supplies, lamps or other devices aren't fussed if you connect return leads together before grounding or not, if that's all that we are concerned with. With audio signals using a steel or aluminium chassis, or a large PCB, consequences become blurred because the chassis and wiring have some measurable resistance which enables currents to sum with others that can be profoundly dirty like AC wiring, rectifier, regulator and filter cap currents or the amplified audio itself. If there were zero ground resistance and very short leads throughout, this couldn't happen.
The effects may only be tiny, but because of the amplification, we start to hear them along with any distortion products. So, if you are choosy about really clean audio, then it's a matter of concern to make sure these are inaudible.
When we look at quality commercial products, we also see techniques borrowed from high-end ones but only to the extent that significant costs are still avoided. That's probably the lowest expected level of attention we should pay to these nuances of design but unfortunately, it means we have to learn and carefully juggle these grounding principles, construction guidelines etc. which, as you suggest, seem to be "out there" in comparison to the basic circuits, parts and wiring that we may be more familiar with.
For a concept that helps me wrestle with grounding, I think of it as a mass of copper and the connections to it are best if leads are only mm rather than inches long. The bigger the circuit, the harder to do all those connections and group them around the copper mass, so the circuits are broken up into isolated sections that are then connected to a single point or "star" ground. Then the rules of priorities in what should have shortest leads, be routed here or there etc. start to come into play.
You get the feeling it's not such an ideal world in electronics, after all
. - rant over.
Is this to say that since these particular caps are related to DC supplies and lamps, it's not so critical?
If I were going to redo these caps, should I mount new terminals closer to the caps? Else I'll be in the inches of lead length.
I still have a hard time understanding how a potential can exist between two points on an aluminum chassis, and at the same time how having two caps ground to different points on a chassis is better than grounding to the same point. Arcane stuff...
OK Now I've done it. The good news is the hum is gone in the right channel. The bad news is the bias on Q1 is also gone and I have distortion in the right channel instead of hum. This happened as I was doing a last bias check before calling it done and putting the cover back on.
I'd just finished probing around for an hour becoming more convinced the hum is associated with the right input transformer. Not 100% but as best I can figure.
It's possible I shorted the base to the emitter of Q1 as I was checking the voltage across the resistor (R82). Now my new problem is figuring out what happened and whether it can be fixed. The other possibility is a problem with the socket for Q1. I just realized that while I'd cleaned all the small transistor sockets, it hadn't really registered that the T03's are socketed too. Any bets or guesses on what I did? Argh.
I'd just finished probing around for an hour becoming more convinced the hum is associated with the right input transformer. Not 100% but as best I can figure.
It's possible I shorted the base to the emitter of Q1 as I was checking the voltage across the resistor (R82). Now my new problem is figuring out what happened and whether it can be fixed. The other possibility is a problem with the socket for Q1. I just realized that while I'd cleaned all the small transistor sockets, it hadn't really registered that the T03's are socketed too. Any bets or guesses on what I did? Argh.
Hmm....just as you were beginning to have some fun.
There isn't much to consider here - only 2 groups of 2 transistors and a diode each are in this isolated part of the circuit. Check the marked 0.65 and 1.2V voltages are close at the bases of the transistors and there is 0.65V across each diode. That should tell you if you still have healthy semis.
I mentioned early in the thread that there is likely weakness in the pots used to adjust bias. If you adjust these, and I suggest the less the better, then check that the small voltage across it and R72 varies in concert. If all this checks out, we have to look at the transformer secondary windings. These will have some DC flowing but initially, check for continuity of each winding with the power off and supply voltages down to zero. They should have similar resistance in both channels if all is well there. There are also the resistors to consider and any composition resistor that old is suspect but checking some of them won't be all that simple - voltage comparison checks between the channels are a better method at this stage.
Take care with the meter probes, it's all too easy to smoke parts with them when you are probing about. I have a small bag full of reminders why I should not hover about over a powered circuit.
Gajanan & Ian, thanks, I'll check all these things to the best of my ability.
The only thing I can add is I was just getting the multimeter probes situated on the sides of the resistor when this happened, and had not touched the ADJ pot yet. They have both worked very smoothly so far, and the voltage across the resistor always responded directly. Now the voltage across that resistor has dropped to maybe 5mV (too low to really trust the multimeter) and turning the ADJ pot has no effect.
Thanks much
-Mark
The only thing I can add is I was just getting the multimeter probes situated on the sides of the resistor when this happened, and had not touched the ADJ pot yet. They have both worked very smoothly so far, and the voltage across the resistor always responded directly. Now the voltage across that resistor has dropped to maybe 5mV (too low to really trust the multimeter) and turning the ADJ pot has no effect.
Thanks much
-Mark
Apologies, was refering to R82, which is where I'm measuring for 40-60 mV when setting bias.
I've never been so frustrated not to find anything wrong Just spent an hour or two checking things. Removed each of the power transistors, cleaned them and their sockets, measured resistances on them, and gave them very thin films of silicone thermal compound on each side of their mica gaskets. Q1 measured just the same as the others, which is to say 140-150 Ohms between the pins, same between one pin and the case, and open or perhaps very high resistance between the other pin and the case. Each 140-150 Ohm measurement was open with +- leads swapped.
The transformers both had similar measurements which appeared to be about 15 Ohms through each of the windings. I did not disconnect them so I don't know if any other components would have affected these.
Voltages were good and same or close for both channels everywhere except for across R82 where I had very little to measure. I checked the resistance of R82 vs R81 and they were close.
Also I think the hum is still there, just the output level of the right channel has dropped a bit.
Is checking resistances across the transistors as I did an adequate test? Should I similarly test Q's 5-8? Or do some further testing of Q1?
Oh, the 4 diodes. Got .7V across each, measured resistance and each had similar high/low resistance depending on lead swapping.
Also found similar resistance across the BAL pots. Also just realized I've been measuring bias only for Q1 and Q4, should go back and check Q's 2&3...
Just spent an hour or two checking things. Removed each of the power transistors, cleaned them and their sockets, measured resistances on them, and gave them very thin films of silicone thermal compound on each side of their mica gaskets. Q1 measured just the same as the others, which is to say 140-150 Ohms between the pins, same between one pin and the case, and open or perhaps very high resistance between the other pin and the case. Each 140-150 Ohm measurement was open with +- leads swapped.The transformers both had similar measurements which appeared to be about 15 Ohms through each of the windings. I did not disconnect them so I don't know if any other components would have affected these.
Voltages were good and same or close for both channels everywhere except for across R82 where I had very little to measure. I checked the resistance of R82 vs R81 and they were close.
Also I think the hum is still there, just the output level of the right channel has dropped a bit.
Is checking resistances across the transistors as I did an adequate test? Should I similarly test Q's 5-8? Or do some further testing of Q1?
Oh, the 4 diodes. Got .7V across each, measured resistance and each had similar high/low resistance depending on lead swapping.
Also found similar resistance across the BAL pots. Also just realized I've been measuring bias only for Q1 and Q4, should go back and check Q's 2&3...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.