Blue Junior Hum even with all controls at zero - help with PD, please?

[nickb]

Thanks again nickb.

OK - so some useful results from today's latest test

I did as you suggested, and connected the 100nF capacitor from the PI tube (v3, not v4) pin 7 to ground and ... no buzz! However, the amp also lost a huge amount of volume and lots of treble - I could wind up both Volume and Master to full and strum the guitar without having the neighbours call the police.

Doing the same thing with pin 2 had no effect. So - an interesting result, but not sure what that means for a solution to the problem...

The loss of volume is to be expected. Adding the cap shunts all the signal and hum to ground and that was the purpose of the test. Well now we know the hum is coming in via the grid pin 7. I should have said to keep the master at zero otherwise it complicates things.

Since you get hum with the master at zero it narrows the problem down to the trace from R22 thru W10 and then C11 to pin 7. Try putting your 100nF cap from the end of R22 that connects to C11 to ground and see if the hum goes away. If that does then then try from the other end of R22 ( since the master is at zero.

[LondonSi72]

The loss of volume is to be expected. Adding the cap shunts all the signal and hum to ground and that was the purpose of the test. Well now we know the hum is coming in via the grid pin 7. I should have said to keep the master at zero otherwise it complicates things.

Since you get hum with the master at zero it narrows the problem down to the trace from R22 thru W10 and then C11 to pin 7. Try putting your 100nF cap from the end of R22 that connects to C11 to ground and see if the hum goes away. If that does then then try from the other end of R22 ( since the master is at zero.

OK - I see what you were testing for now. In that case I need to explain exactly how I connected up the 100nF capacitor, as I haven't fully removed the PCB from the amp - I'm working on it whilst it's fitted in the chassis. This limits my access to only the front of the board where the components are mounted, and I can't access the solder joints on the trace side of the board. To test the 100nF cap I attached it to the grounded leg of resistor R26 and the tube side pin 7 leg of resistor R22. There is no leg available this side of the board from the capacitor C13 to which I could attach the 100nF. I didn't solder them as I'm just testing, and the Blue Junior is known for having sensitive traces, so I don't want to risk lifting the trace by soldering until I have to. I made sure they had good mechanical contact, though...

The second test I did was with the capacitor still attached to R26 grounded side (which is the top when you look at the board) and the bottom of R23, which connects to pin 2 of the tube.


Can I just clarify your next test - is my connecting the 100nF to the top of R26 as the ground connection valid for your test? If it is, then I have already done the test where the other end of the cap is connected to the bottom of R22 (the leg of the resistor that then connects to the PCB and via the ribbon cable to pin 7 of the PI tube). The second test you suggest is connecting the other end of the cap to the junction of R22 and C11. Would connecting the cap to the leg of R22 be a valid test for this? If not I'll need to remove the PCB from the chassis so I can access the trace at the back and try running these tests from the solder joints.


Thanks for all your help with this - it really is appreciated, and apologies if I'm asking daft questions.

[nickb]

The way you did it is just fine. I'd do it the same way. There are no daft questions

You can continue to use the same ground at R26. I think if you can just get to the C11 side of R20 that will do. When following traces you can't assume that everything shown on the schematic is joined together - bad joint will change the picture. That is the reason for checking both ends, if possible. With the board in place it would be impossible so we'll skip it, for now anyway.

My suspicion is that the trace from R20 to pin 7, being quite high impedance, is picking up the buzz. I've also fear that it may be inherent. To make any difference would require bypassing the track and rewiring with a screened lead.

On my to-do list for tomorrow is a FBJ. I'll put a scope on the output and see how much buzz/hum I get and we can compare notes.

[LondonSi72]

The way you did it is just fine. I'd do it the same way. There are no daft questions

You can continue to use the same ground at R26. I think if you can just get to the C11 side of R20 that will do. When following traces you can't assume that everything shown on the schematic is joined together - bad joint will change the picture. That is the reason for checking both ends, if possible. With the board in place it would be impossible so we'll skip it, for now anyway.

My suspicion is that the trace from R20 to pin 7, being quite high impedance, is picking up the buzz. I've also fear that it may be inherent. To make any difference would require bypassing the track and rewiring with a screened lead.

On my to-do list for tomorrow is a FBJ. I'll put a scope on the output and see how much buzz/hum I get and we can compare notes.

I'd be very interested in your findings on a second FBJ. It seems some are noisier than others with no obvious reason. Aside from poor dressing of the wires, poor solder joints is the next biggest culprit I think. Some have said the export models with their more complex PT are noisy too (I adjusted mine for 240v - they ship configured for 230v)

One point I should make, is my FBJ is a 'cream board' version. It's not the first green board model and it's not the latest FBJ III. I mention this because you said 'the C11 side of R20'. The cream board has another capacitor between these - C10. In the BJ III it was removed and a wire jumper put in place. Part of the 'sparkle mod' apparently.

Anyway, good to know I'm on the right track. I'm not averse to putting in a shielded bypass if needed, but hope that would be a last resort.

Thanks again

Addendum: I just tested with the 100nF cap on R26 for ground and the other end of R22 (the circuit side rather than the tube pin 7 side, if you know what I mean). The buzz was still there.

Also tried connecting to R20 at the junction of R20 and C10, and the buzz again disappeared. But that's not surprising I guess as that's shunting all the incoming signal from the master to ground.

Speaking of which - the trace from the Master to R20 takes a very long and tortuous route in this PCB, so you could be spot on with this picking up the buzz.

Looking again at the signal at the Green terminal on the speaker with the scope, there is a definite if small spike in the waveform. seems to be a 50Hz signal though, so not sure what's causing it (would be 100hz if it were the rectifiers, wouldn't it?)

[LondonSi72]

Yep, the buzz is certainly excessive when you are playing. One question that you've probably already answered: does the buzz get louder when you turn up the amp volume?

This is not meant as an insult but sometimes it is worth it to pay a good amp tech to diagnose the problem. I have a hunch that once he finds the problem the fix will be pretty easy (like a poor wire connection.) I assume that you have tried unplugging all of the connectors and plugging them back in. I would also set my DMM to check continuity between chassis ground and the various grounds you'd expect to find after studying the schematic and layout drawing.

Steve Ahola

Hi Steve. Yes I've done all the basics. Buzz doesn't noticeably change with any of the controls, though hiss increases when volume and master wound up, which makes the buzz sound slightly different, but not louder.

I took the amp to a tech and he resoldered a couple of dodgy looking connections in the power amp tube sockets but couldn't figure out where the buzz was coming from. Don't think he went through it with a scope, though...

[nickb]

OK. The FBJ III I have here also with cream board gives 10.2mV hum + noise on the output. Looking at the spectrum I see that 50Hz is strongest but there is plenty of energy on all the harmonics. That corresponds with what I hear, a mixture of hum and buzz. By the way it's pretty low level and I would not consider it a problem. If you touch (CAREFULLY!!!) the ribbon that goes to V3 you hear the noise rise dramatically. This tells us what we already suspect - the noise is being picked up in the ribbon and possibly the PCB trace to the grid.


I took a small length of sticky-back aluminium foil and wrapped it around the cable, grounded it and it cut the noise by about 2-3mV. Something, but not as much I hoped. The conclusion is that the PCB trace is also involved.

[LondonSi72]

Interesting - so you get similar results on the FBJ III. It's the buzz more than the hum that's the problem. I'm probably more aware of it because I play in my garage conversion at home - quite a small space, and I don't play particularly loudly so I hear the buzz a lot of the time. It is an issue when I try to record though, as you could hear in my uploaded sample.

I know the v3 ribbon cable is very sensitive to noise - one of the suggestions on the billmaudio.com website is to adjust the position of this cable to try and reduce noise (and in particular high frequency oscillation).

But now we've got this far ... and you seemed to be able to reduce noise by shielding the ribbon cable, I'm interested in knowing a little more about how you did that, and wondering if it would be possible to bypass the PCB trace that connects the output from the master pot to R20 with shielded cable instead.

Where did you ground the aluminium foil you wrapped the ribbon cable? Would grounding to R22, like I did in the tests be a good position? And if I were to try running shielded cable from the master to R20, what kind of cable should I use (what brand/make do you use?), and should I ground at the master pot end or the R20 end? Assuming you'd want to avoid grounding at both ends, like with the guitar input jack?

[nickb]

I used a wire with crocodile clips between the foil and the chassis.

I think I suggested using a screened wire to feed pin 7 to bypass as much as possible of that signal a little ways back.


One other thing that occurred to me is that the source impedance on that signal could be made much lower. Right now it's ~220k set by R20. If you were to simply bypass R20 then it could significantly cut the noise pickup, especially at the low volumes you use. It's worth a try. I'm really not sure why they elected to have that resistor at all but it may be to do with blocking distortion. In that case it would only apply to a heavily driven power amp situation.

[LondonSi72]

Hi. Yes you did suggest bypassing the trace - that's why I was thinking it might be time to try it, if I want to reduce the noise further

I guess I could jumper the 220k resistor with another of similar value to halve the impedance and see what effect that has before I start slicing the trace or removing the master pot

I had a quick look at shielded/screened cable and wasn't sure what type is best for this kind of use. Any suggestions?

One thing I didn't understand though, the noise was removed when I connected the cap from ground to R22 (pin7) but not when I connected to R23 (pin2) or the other side of R22. Wouldn't that suggest the noise is coming in after the R20 220k, C10 and C11 and between C11 and pin 7 itself?

I'm out for a couple of days so not going to be able to test anything but will have a go late Sunday or Monday if I get a chance.

Thanks...

[nickb]

Hi. Yes you did suggest bypassing the trace - that's why I was thinking it might be time to try it, if I want to reduce the noise further

I guess I could jumper the 220k resistor with another of similar value to halve the impedance and see what effect that has before I start slicing the trace or removing the master pot

I had a quick look at shielded/screened cable and wasn't sure what type is best for this kind of use. Any suggestions?

One thing I didn't understand though, the noise was removed when I connected the cap from ground to R22 (pin7) but not when I connected to R23 (pin2) or the other side of R22. Wouldn't that suggest the noise is coming in after the R20 220k, C10 and C11 and between C11 and pin 7 itself?


I'm out for a couple of days so not going to be able to test anything but will have a go late Sunday or Monday if I get a chance.

Thanks...

I'd simply short out R20. There is an excellent chance of great results for minimal effort. OTOH it might make it more oscillation prone as R20 does provide some attenuation of higher frequencies and that might be the reason for it. It is does then see how low a value you can get away with.

BTW: Any miniature shielded cable will do. Preferably choose something with a stranded core but the choice in small quantities is limited. RG178B will do if you can get a small length but it's a little stiff. Do you have any old RCA/Phone audio cables? Strip it down a use a bit of that.

[LondonSi72]

Problem solved!

I have to admit I'm not entirely sure why this works, but as I saw this suggestion on another forum and had already ordered the parts I thought I'd try it out and it has completely removed the buzz. There is now just a very quiet hum. I converted the tube heater supply to DC using a bridge rectifier and a 4700uF electrolytic capacitor and it has reduced the signal at the loudspeaker to 1.8mVrms and, as far as I can tell with such a small signal, eliminated the spikes.

The bridge rectifier does get very hot, so will screw it to the chassis and attach a heat sink to the other side to help manage the heat. The DC voltage is also a little on the low side at 6.2v with quite a large ripple, but it doesn't seem to be effecting the amp and the test point voltages (from the Fender schematic) are all still well within the tolerances.

Thanks to everyone for their suggestions and help - especially nickb and Steve Aloha. I'm one happy bunny

[LondonSi72]

Problem Solved!

Problem solved!

I have to admit I'm not entirely sure why this works, but as I saw this suggestion on another forum and had already ordered the parts I thought I'd try it out and it has completely removed the buzz. There is now just a very quiet hum. I converted the tube heater supply to DC using a bridge rectifier and a 4700uF electrolytic capacitor and it has reduced the signal at the loudspeaker to 1.8mVrms and, as far as I can tell with such a small signal, eliminated the spikes.

The bridge rectifier does get very hot, so will screw it to the chassis and attach a heat sink to the other side to help manage the heat. The DC voltage is also a little on the low side at 6.2v with quite a large ripple, but it doesn't seem to be effecting the amp and the test point voltages (from the Fender schematic) are all still well within the tolerances.

Thanks to everyone for their suggestions and help - especially nickb and Steve Aloha. I'm one happy bunny

[Tom Phillips]

You previously said that the bad buzz stopped when V3 was pulled. Therefore we can conclude that the power tubes are not contributing to the noise so you could continue to run the power tube heaters on AC. The resulting reduced current load on your DC heater supply will reduce the temperature of your new bridge rectifier and reduce the ripple on the DC heater supply. The overall circuit will be more reliable and the final hum level may be even lower. Reference the DC heater discussion at www.valvewizard.co.uk/heater.html , particularly the information about grounding.

[LondonSi72]

You previously said that the bad buzz stopped when V3 was pulled. Therefore we can conclude that the power tubes are not contributing to the noise so you could continue to run the power tube heaters on AC. The resulting reduced current load on your DC heater supply will reduce the temperature of your new bridge rectifier and reduce the ripple on the DC heater supply. The overall circuit will be more reliable and the final hum level may be even lower. Reference the DC heater discussion at www.valvewizard.co.uk/heater.html , particularly the information about grounding.

Thanks Tom. I read through that article, and it does look like a better way to arrange the heater wiring, but as my setup required no permanent changes to the PCB (I have just used the heater wire plugs and cables and inserted the rectifier in series) and it seems to be working well at the moment, I'll stay with this for now. It does get hot, and I'll have to keep on an eye on that when I play for longer, to make sure it
S not getting too hot and damaging any of the components, but this amp ALWAYS got hot, so I'm not sure how much is due to the changes I've made, In any case, if I do decide to cut the wiring in the PCB and add the rectifier after the power tubes section, I'll update this post with the results.

Thanks again to all those who contributed.