Misterious Cap, noise issue on NOVANEX AUTOMATIC 6 amp

[Rob Strand]

Oh, and all of the pots have same writing on them - RUW100, I'll measure those too.

I'll see if I can find any info.


I noticed the Novanex amp schematics I found use 47k linear pots everywhere.  (The trimpot excluded.)

[Rob Strand]

I had a look for info on the pots.  It seems RUW100 (RUW1D0) is some sort of part model
but value and taper is written elsewhere.

The 18V 12VA transformer seems to line-up quite well with a 6W amplfier into 8 ohms.

I updated the schematic with the new info so you can see what I'm up to.

[schematic obsolete]

[razabri]

Hey Rob, great input once again!

I noticed the Novanex amp schematics I found use 47k linear pots everywhere.  (The trimpot excluded.)

I tried to measure pots and I have 36k at volume, 63k at treble and bass one i a different type - I get 448 Ohm when it's all the way closed, so I guess this one's rev log pot or something, as it works perfectly so it's not due to a fault.

Another thing - I got the wiring of the trim pot wrong - I see now that it's middle lug actually goes to that 4.7uF cap next to the 2n3927 transistor without connecting to other lug(s) - so each goes it's own way.

And the small red cap that is color coded and I marked it as red, black, white - the last color white may not be relevant, as there's only a bit of it on it's lugs...

Anyhow, I'll try to measure those voltages at the first transistor, and the supply - I'll get back as soon as I get those.

Cheers!

[Rob Strand]

I tried to measure pots and I have 36k at volume, 63k at treble and bass one i a different type - I get 448 Ohm when it's all the way closed, so I guess this one's rev log pot or something, as it works perfectly so it's not due to a fault.

Thanks, good info.  So the Volume and Treble are most likely to be 47k and the Bass is probably 470 ohm or 500 ohm.     The low Bass pot value makes more sense.  I only need to add 100 ohm to 150 ohm to knock the gain back on those mods so a 470 ohm to 1k Bass pot is probably going to drop the bass as low as it needs to be.

The tapers are tricky.   To measure the pot values and tapers *in circuit* you need to take a number of measurements.  You also need to set the pot to positions which reduce/remove the effect of the existing circuit.   In some circuits even that's not enough so you are forced to make a number of "wrong" measurements in a way that lets you *calculate* the pot taper (and sometimes even the pot value).

For that circuit, the Treble and Bass pots have joined terminals.   Because of the caps in the circuit your measurements for Bass and Treble are going to be good.   The volume pot can be affect by the 10k power amp resistor so it's best to measure the volume pot with the volume turned all the way down and measure counter-clockwise terminal (pin 1) to clockwise terminal (pin 3).

To get the tapers set the pots mid way.
- Volume: measure wiper (pin 2) to clockwise terminal (pin 3)   ; the top part of the pot is not affected by the rest of the circuit.
- Bass and Treble just measure across the two terminals like you did before for the pot values.

Another thing - I got the wiring of the trim pot wrong - I see now that it's middle lug actually goes to that 4.7uF cap next to the 2n3927 transistor without connecting to other lug(s) - so each goes it's own way.

Has the trimpot or three leads?  To me the trimpot looks like it only has two leads (you can get those).  Perhaps two leads and the wiper side goes to the cap?

For accuracy, perhaps I can draw the trimpot with only two leads connected.

And the small red cap that is color coded and I marked it as red, black, white - the last color white may not be relevant, as there's only a bit of it on it's lugs...

OK, I managed to decipher that one.  The red is two bands merged into one, so red red black white.    The white is the voltage.
The bands read as 220 which is 220pF.    A 220pF cap in that position in the circuit is very common.

Anyhow, I'll try to measure those voltages at the first transistor, and the supply - I'll get back as soon as I get those.

OK great.  Much better than my estimates, they are likely to be close but they will never be 100% correct.

One thing I wasn't sure about is the diode near the trimpot.  Is it possible for you to read the markings?   I assumed it was the same as the other two diodes.   I can read the markings on the other two and they are exactly what I expect.  However the diode near the trimpot seems to be marked differently.

Also the voltage on the 100uF capacitor.  I'm assuming that's 25V but I can't actually read it.

FYI, this is how a three pin trimpot could connect,

[razabri]

Thanks Rob, your guidance is great appreciated. I'm sorry to say that we have a setback as I managed to break something, probably I've handled the board too much getting those measurements, so something got broken and now I don't have sound going through, only the slight buzz when touching the input lug. I've tried to clean the volume pot again before, using the spray, and I may have soaked some other part as well in the process. The pot is audible when moving, so it doesn't seem to be broken...

For that circuit, the Treble and Bass pots have joined terminals.   Because of the caps in the circuit your measurements for Bass and Treble are going to be good.   The volume pot can be affect by the 10k power amp resistor so it's best to measure the volume pot with the volume turned all the way down and measure counter-clockwise terminal (pin 1) to clockwise terminal (pin 3).

I have strong 60k at volume pot measuring it all the way down on the lugs 1 and 3. Didn't quite realize that rest of the circuit is going to affect the measurements, so thanks for the explanation about that - learned something!

Has the trimpot or three leads?  To me the trimpot looks like it only has two leads (you can get those).  Perhaps two leads and the wiper side goes to the cap?

For accuracy, perhaps I can draw the trimpot with only two leads connected.

Sorry, you're right, it has two leads and they're connected as described earlier - one goes to that diode that looks to me like 1n4148, but I can't really see any print, must be on the other side, and the other one goes to the resistor that measures 150k (by the colors I had 15Gk, must have seen it wrong), and to the 4.7uF cap and the trail then go to the 2n3972 transistor.

I have tried to measure voltages on that transistor, but I only can read 4v at one pin. One is connected to ground. When I touch the pins I have some slight crackling at speaker, that should be normal I guess...

I have managed to measure other voltages - I have 20.4v AC at transformer, 28.4v at one and 27.6v at the other 470uF cap, and I have 19.6v at 100uF cap.

As for troubleshooting the amp, I have no idea about where to start, as I don't really see what could be wrong - no loose wiring, no parts have been burnt, the speaker seems to be working, but audio signal from guitar is not going through... Any guidance on this would be appreciated...

[Rob Strand]

Thanks Rob, your guidance is great appreciated. I'm sorry to say that we have a setback as I managed to break something, probably I've handled the board too much getting those measurements, so something got broken and now I don't have sound going through, only the slight buzz when touching the input lug

That's a bummer.  Unfortunately there's always a risk when poking around.

I have tried to measure voltages on that transistor, but I only can read 4v at one pin. One is connected to ground. When I touch the pins I have some slight crackling at speaker, that should be normal I guess...

So that looks like a problem already.   

If you can measure these points and post the voltages it would help a lot to diagnose the problem,
- Preamp power :   20V  (as you measured already, but recheck in case it has changed)
- Q1 BC549B:   Collector 11V, Base 7.5V, Emitter 6.9V

If those don't check out I suspect the BC549B is fried.  I don't know how it would have got fried since you could short out virtually everything around it with damaging anything at all!

As for troubleshooting the amp, I have no idea about where to start, as I don't really see what could be wrong - no loose wiring, no parts have been burnt, the speaker seems to be working, but audio signal from guitar is not going through... Any guidance on this would be appreciated...

You are lucky nothing happened to the power amp as that more difficult to debug.

If you can measure the DC voltage across the speaker, just to be sure all is good there.

[razabri]

Thanks Rob, I actually got it working again just now, there was a loose connection at 4.7uF cap, I just touched slightly all of the parts and sound came through once I got to this one - resoldered and ready to go.

I've now tried to measure said transistor and I have 11.6v on collector and 6.6v on the emitter, but when touching the base with one probe and ground on the other, I have a bang on a speaker, so I can hold it there long enough to read the value, as I'm afraid to burn something - most probably I'm not doing this right...

If you can measure these points and post the voltages it would help a lot to diagnose the problem,
- Preamp power :   20V  (as you measured already, but recheck in case it has changed)
- Q1 BC549B:   Collector 11V, Base 7.5V, Emitter 6.9V

If those don't check out I suspect the BC549B is fried.  I don't know how it would have got fried since you could short out virtually everything around it with damaging anything at all!

[razabri]

Turned the speaker all the way down so it doesn't crack, and have measured same voltage on the base as on the emitter - 6.6v

[razabri]

Guess I could try that one mod you suggested, with 100 Ohm resistor - I assume I'd need t cut the trace between 10k resistor and 1.5 cap and jump it with 100-150 Ohm resistor to try it out.

Also, I'm not sure what's with that first transistor, 2n3972, should it be checked or maybe switched for different one with same pinout... Should be some biasing be done to it or to the BC549B?

Earlier, I have tried to turn that trim pot and see what happens - that was before your warning that it shouldn't be set to highest value, and nothing really happens, or I haven't noticed much. Only when you turn it all the way counterclockwise the signal gets lost. Perhaps I noticed no difference if it only affects compression and it could be audible only while playing, the way it compresses...

Also, I thought I could maybe replace the 470uF caps, if that could change something, as I have a pair of 1000uF axial ones that wouldn't affect it's old school appearance...

[Rob Strand]

Thanks Rob, I actually got it working again just now, there was a loose connection at 4.7uF cap, I just touched slightly all of the parts and sound came through once I got to this one - resoldered and ready to go.

That's a relief.  Good you found it.

I've now tried to measure said transistor and I have 11.6v on collector and 6.6v on the emitter, but when touching the base with one probe and ground on the other, I have a bang on a speaker, so I can hold it there long enough to read the value, as I'm afraid to burn something - most probably I'm not doing this right...

Turned the speaker all the way down so it doesn't crack, and have measured same voltage on the base as on the emitter - 6.6v

Those voltages look good.

The digital multimeter (DMM) looks like a 1M resistor.  When you place the DMM on the base it loads down the voltage.  The change in DC voltage causes a bang.  So yes, you need to turn down the amp.

The DMM loading has little effect on the collector and emitter measurements but it does have a significant effect on the base voltage measurement.   With maths you can show the 1M causes the base voltage to drop by a factor of 0.89.  So the 6.6V base voltage measurement is *really* something like 6.6/0.89 =  7.4 V.    Assuming a base emitter voltage of 0.65V we would expect the base to be at 0.65 + emitter voltage = 0.65 + 6.6 = 7.25V.   So some round errors have occured in the measurements.  Nonetheless it's in the right ball park.   

The measurements (11.4V, ~7.3V, 6.6V) are close to the expected voltages from the simulation (11V, 7.5V, 6.9V).   The lower measured value on the emitter and base are likely to be cause by Q1 having a lower than usual gain.   That could actually make the circuit a little noisier than it could be.

Guess I could try that one mod you suggested, with 100 Ohm resistor - I assume I'd need t cut the trace between 10k resistor and 1.5 cap and jump it with 100-150 Ohm resistor to try it out.

I was thinking something like just lift one end of the cap and put a resistor in series.  You can put heatshrink over the join.

Changing the 330k to 470k will also help prevent clipping.

lso, I'm not sure what's with that first transistor, 2n3972, should it be checked or maybe switched for different one with same pinout 

Earlier, I have tried to turn that trim pot and see what happens - that was before your warning that it shouldn't be set to highest value, and nothing really happens, or I haven't noticed much. Only when you turn it all the way counterclockwise the signal gets lost. Perhaps I noticed no difference if it only affects compression and it could be audible only while playing, the way it compresses...

It's not a good idea to set that pot to the extremes if the power amp has a DC offset it can damage the JFET.

It's difficult to set that without an oscilloscope and a signal generator.   You set the trimpot so the *power amp* never clips.   It's intended to be as subtle as possible.  You don't want clipping but you also don't want the compression to be noticeable.
Without a oscilloscope you might be able to do a good job by using a bass guitar.

Also, I thought I could maybe replace the 470uF caps, if that could change something, as I have a pair of 1000uF axial ones that wouldn't affect it's old school appearance...

These types of mods are a bit of a double edged sword.   With the caps in place you get more power out of the amp  *but*  by doing that you create more stress on the output transistors, transformer and speaker.   Who knows if that will increase the chances of something failing.

The amp is designed not to clip by setting the trimpot correctly.   So if you set the trimpot for 6W output with the mod present you haven't gained anything.  In fact you have lost because now the output transistors will get a little hotter.  The other option is to set the trimpot at the point of no clipping.  With the mod in place the output power will be more then 6W.  So now you have hotter output transistors, more stress on the transformer and on the speaker.

[razabri]

Thanks, Rob, yeah, been lucky with that fault, I'm glad it wasn't anything serious.

It's good to know the voltages look alright to you, referencing the sim and all, and thanks for explaining about the base voltage - haven't realized that factor, and it all makes perfect sense.

Also, your idea on how to approach that slight mod to the preamp is much less invasive than mine - I may as well try doing so once I dig up a resistor with proper value.

As for the clipping trimpot, there's that dab of paint on it that marks default setting, so I tend to keep it there, but I don't really feel that changing it's position does anything to that slight noise issue about which I kind of started this thread - don't know if you had this before, the amp sounds relatively clean, but there's this quiet deteriorating sound that follows a played note and I'm not sure what causes it - idea of changing the power filter caps values came from my (amateur) experience that some guitar (dirt) pedals would be smoother with larger power cap value (this may be complete nonsense), but I understand your explanation on what that may cause and I wouldn't want to stress the components unnecessarily.

Another thing is that I feel that could be done is to verify the schematic you've made - let me know what details I can give to complete it and have correct values all around.

Perhaps that slight noise issue is something comming from one or both of the output transistors, if that make any sense - checking their voltages may diagnose it, so perhaps if you have a full sim of the amp I could measure those. Or maybe there's something in the compression state that's off the track - there are 6 transistors there, as I understand, and those may be worth checking too...

[Rob Strand]

Also, your idea on how to approach that slight mod to the preamp is much less invasive than mine - I may as well try doing so once I dig up a resistor with proper value.

I always go for least invasive - well unless I'm convinced what's there is a design disaster .

As for the clipping trimpot, there's that dab of paint on it that marks default setting, so I tend to keep it there, but I don't really feel that changing it's position does anything to that slight noise issue about which I kind of started this thread

Well the intent of that part of the circuit is that it does nothing for unless the amp is pushed.  In reality it could be doing something detrimental.  I feel the only way to test that theory would be to remove the JFET.  It's either going to do nothing or be the cause.

- don't know if you had this before, the amp sounds relatively clean, but there's this quiet deteriorating sound that follows a played note and I'm not sure what causes it -

Unfortuately, there's so many causes.   The only way to find the problem is to eliminate the "possibles" from the list.

Bad/Dirty input sockets:   Just bad connection and/or vibrations from the amp cause a subtle fizz.   I guess try the other input socket.

Bad connection or solder joints on the PCB.   You can go over the connections with a magnifying glass.  Some people will "touch up" (resolder) any connections that look suspicious.

Faulty or Failed components.      No rules here you really need to listen to various point in the circuit to narrow it down.   Some guys just replace stuff until the problem is gone but you can end-up replacing a lot of parts with no result other than a damaged PCB.

Faulty pots can do weird stuff not unlike faulty sockets.

I can see the one of the tropical fish caps has some cracks on the leads.  Maybe an issue there.

Electrolytic caps dry/faulty.   This can cause all sorts of weird behaviours.   Even oscillations.

Faulty speaker.

Design problems:

- It is very possible the power amp has some bad crossover distortion.  The whole design is set-up so the power transistors don't have much bias current.   There's no way to increase the bias current reliably it affects the design too much: no emitter resistors on the output transistors, no thermal feedback on the bias (only R19 330R), output transistors on different heatsinks.   It would be a headache to fix properly and anything less could end up in the amp melting down because of "thermal runaway".

- oscillation.    No Zobel network across the speaker.    You could add one.     If the amp is marginal in design the network could cause issues.

PCB Layout issue:   There's a few points where tracks are close together.  It may be causing problems it might not.

idea of changing the power filter caps values came from my (amateur) experience that some guitar (dirt) pedals would be smoother with larger power cap value (this may be complete nonsense), but I understand your explanation on what that may cause and I wouldn't want to stress the components unnecessarily.

It is very possible that the supply caps are old and are the cause.    You can try replacing them.   For a short term test using 1000uF would be fine.  For long term there's a small risk (for reasons mentioned before).

Perhaps that slight noise issue is something comming from one or both of the output transistors, if that make any sense - checking their voltages may diagnose it, so perhaps if you have a full sim of the amp I could measure those. Or maybe there's something in the compression state that's off the track - there are 6 transistors there, as I understand, and those may be worth checking too...

If the DC voltage at the output of the amp, ie. across the speaker, is near 0V then it's highly likely the transistors in the amp are OK.   You might check the screws on the power transistors as a bad connection there could do something weird.

*** I think the biggest step forward would be to isolate the problem to the preamp stage or the power amp stage. ***

One way would be to build an audio probe and listen to points throughout the preamp to see if the weird sound is present or not.  I don't recommend using an audio probe on the power amp, especially probing internal points like the small transistors as this can cause the power amp to oscillate which could end-up burning out the speaker.

Another thing is that I feel that could be done is to verify the schematic you've made - let me know what details I can give to complete it and have correct values all around.

I think most of it is done.  I fixed up some resistor and capacitor values already from your overlay markup.   Also one of the JFET pins connected to the wrong point.   I'm used to fixing this type of stuff as I go.   Overall I'm happy with it.  There was only a handful of smaller points,

- pot tapers
- the diode which is hard to read.   we can only assume it is the same as the others.
- the voltage rating on the 100uF cap

I'll release another schematic.  (not sure if I've changed anything since last time anyway.)

[Rob Strand]

Well, kind of as expected the power amp has an enormous crossover distortion.
Getting rid if it will certainly risk thermally runaway.

It would be good if you could use the audio probe to confirm the preamp isn't the cause.

I'll check over the schematic later this afternoon.  I did find an error.

[Rob Strand]

Here's the updated schematic with all new info and the fixes,



The voltages are slightly different to your measurements.  I just edged them a bit to make them more consistent.

[razabri]

Alright, Rob! Great input and thanks for updating/debugging the schematic.

Agreed on focusing to find whether it's the preamp or not - I've made an audio probe and have poked around a bit, as this is the first time I'm using it I'd appreciate if you could lead me on which points are important to check. I have disconnected the speaker and have another, clean amp on the probe, probe's ground connected to the Novanex's and with the guitar plugged in - I can hear the output well, louder on some points and so far without that fizzle.

Also, I did tried out a different speaker with no change, so amp's original speaker should be fine. Tried both inputs, but no change, since they actually both go to the same input points.

[razabri]

Oh, and one more thing, I don't know if this shows well on the drawing I made, but the collector pins of both 2n5492 transistors are cut off, only connections are at the base, emitter and it's heat-sink.

[razabri]

Also, 100uF cap is rated at 25v.

[PRR]

Well, kind of as expected the power amp has an enormous crossover distortion....

If D2 D3 were on the power stage heatsink (either one, they should hot-up the same), it would be thermally over-compensated. But at ~~2.2Vbe and we want 3Vbe. But if we get too close without emitter resistors, you are right, it will try to fry. Since this will never be the loudest amp in town, I would think 1Ω emitter resistors as safest. That is if they can be hacked into the PCB (lift the E legs?).

Then the "330Ω" bias resistor computes like 465Ω, with large uncertainty. And if too large, disaster. I'm thinking parallel 1K fixed and 1k trimmer, start with trim half-way . There may be better schemes.

[Rob Strand]

If D2 D3 were on the power stage heatsink (either one, they should hot-up the same), it would be thermally over-compensated. But at ~~2.2Vbe and we want 3Vbe. But if we get too close without emitter resistors, you are right, it will try to fry. Since this will never be the loudest amp in town, I would think 1Ω emitter resistors as safest. That is if they can be hacked into the PCB (lift the E legs?).

Then the "330Ω" bias resistor computes like 465Ω, with large uncertainty. And if too large, disaster. I'm thinking parallel 1K fixed and 1k trimmer, start with trim half-way . There may be better schemes.

Adding emitter resistors is always going to be the safest bet.  With the emitter resistors that scheme would probably work.  And yes, I realized those diodes could be use for sensing.  I've been pondering what to do as well.

It's pretty common to use two diodes and a series resistors.   With the emitter resistors present you can get these to work without the diodes on the heatsink provided you aren't too greedy with your bias setting.  I think both of these amps have smaller driver transistors *which aren't on the heatsink* and bias diodes which aren't on the heatsink, just in the general area of the o/p stage.

https://www.ozvalveamps.org/coronet/scan0003.jpg
https://i.imgur.com/U3Lia47.gif

The separate heatsinks, small drivers off the heatsinks is going to make the thermal situation less than ideal.  So the simple schemes relying on low bias, emitter resistors and no diodes on the heatsink is probably going to be more practical.  Maybe one diode on the heatsink would buy some insurance.  I have seen amps work fine on power-up without a sense diode but when you run the amp hot it upsets the bias and gets stuck with a very high bias.  Loosely coupling the diode to the heatsink fixes it.

I actually did some simulations  playing around with the minimal bias on output transistors to at least reduce the crossover distortion.   I was getting around 1.75V between the bases of the driver transistors (with no emitter  resistors).   The output transistors were between 10uA and 500uA bias.  However, my experience is in order to remove audible crossover distortion you really need to be *very* close to the point where output transistors are biased.   That's exactly where you start getting problems if you don't have emitter resistors.

[Rob Strand]

Agreed on focusing to find whether it's the preamp or not - I've made an audio probe and have poked around a bit, as this is the first time I'm using it I'd appreciate if you could lead me on which points are important to check. I have disconnected the speaker and have another, clean amp on the probe, probe's ground connected to the Novanex's and with the guitar plugged in - I can hear the output well, louder on some points and so far without that fizzle.

Also, I did tried out a different speaker with no change, so amp's original speaker should be fine. Tried both inputs, but no change, since they actually both go to the same input points.

The simplest test is to listen to the output of the Volume control.   If the signal gets that far and is still free of the fizzy sound it means nothing in the preamp is creating the problem.  It shifts the blame to the power amp in one step.  If you do hear the problem in the preamp then check the signal at the JFET.  That will tell you if the fizz is at input of the preamp.   At that point it gets trickier to narrow down but it might be worth removing the JFET.