Need help tracking down a "buzz"

[markmalin]

Hi guys.

This is driving me nuts.  This is my one-channel Vibrolux (with no vibrato) build.  There is a buzz in the amp that goes away when I remove the tube at the input side of the reverb circuit.  It's there with nothing plugged into the amp's input.  It's faint, but just enough to drive the sound guy crazy (and me as well).  It is not affected by the volume control.  The sound guy showed me a spectrum and there is a spike at 60 Hz and 120 Hz.  As I said I've tracked it down to the vicinity of that tube.  I've tried the following:

 - resolder connections
 - swap out 12ax7's

Nothing has changed.  I do notice the following which may be normal.  The wire to input of the second gain section in this tube is very microphonic.  If you tap it, it's a loud thunk out of the speaker.  Also, if I move a chopstick toward it the amp buzzes really loudly.  I shortened it and re-routed it which seemed to help a little, but not much.  Also, the 10pf cap that is on the other end of this connection behaves the same way.  If I touch it with a chopstick I get hum.  The other weird thing is, this amp picks up my cellphone (I think?).  Sort of an occasional "dit..dit..dit...di-i-i-i-i-it" sound.  I assume that will go away when this is solved.

There are a couple of modifications to the circuit that I need to mention.  I added a 10K pot (back) in series with the feedback resistor and have this dialed about 1/4 of a turn.  I also added a Midrange control with a switch that allows me to switch in/out between a normal Vibrolux tone stack (Treb. and Bass) and more of a Super Reverb version (Treb., Mid, Bass).  All it is doing is switching between the MId pot and the resistor to ground at the bass control. 

I'm including a picture of the schematic which shows the vibrato circuitry gone.  The other major mod is that I have added a dumble style OD circuit which is switchable (not shown in the schematic).  The amp buzzes with that circuit switched off (out of the circuit) so I don't think that's an issue.  I also want to mention that it buzzes with the reverb tank connected OR disconnected. 

What else can I say... hm... I'm still also tracking down a hefty hiss, which you will notice I have replaced some resistors with metal film low noise to no avail. 

I'm including some pictures of that area of the amp.

[RicharD]

>There is a buzz in the amp that goes away when I remove the tube at the input side of the reverb circuit.

Possibly a ground loop through the reverb tank.

[HotBluePlates]

I also want to mention that it buzzes with the reverb tank connected OR disconnected.

That would seem to negate it coming from the reverb driver tube, because you would have effectively disconnected that circuit by pulling the tank.

Do you still have the isolating washer for your reverb return RCA jack? This trick is mostly based on anecdotal experience, but try isolating that jack, then running a wire from the ground lug to the ground-side of the reverb return cathode resistor. In your case, this will be the eyelet that the 220k resistor and 820 ohm resistor are in, which then has a black wire to your ground buss.

May or may not be the issue, but takes 1 washer and a hunk of wire to try.

 

[John]

Nothing has changed.  I do notice the following which may be normal.  The wire to input of the second gain section in this tube is very microphonic.  If you tap it, it's a loud thunk out of the speaker.  Also, if I move a chopstick toward it the amp buzzes really loudly.  I shortened it and re-routed it which seemed to help a little, but not much.  Also, the 10pf cap that is on the other end of this connection behaves the same way.  If I touch it with a chopstick I get hum.  The other weird thing is, this amp picks up my cellphone (I think?).  Sort of an occasional "dit..dit..dit...di-i-i-i-i-it" sound.  I assume that will go away when this is solved.

I'm pretty sure I'm  the least knowledgeable guy here, so take this for what it's worth (as in, not much!). But if it was me, I'd replace that piece of wire completely, along with the 10pf cap, and maybe even add an additional layer of shielding; shrink tubing or larger insulation. Only because it seems odd to me that the wire is noisy.

[HotBluePlates]

Well, I've heard noise from a cap before if it is tapped on. That's not really odd; if the wire is solid core (probably is with typical hookup wire, especially cloth-covered) then movement of the wire possibly results in movement of attached components.

But we have too many disparate noise factors to think about. Let's attack and solve one, then see if the others are really a problem.

[The_Gaz]

Shot in the dark, but have you ruled out the pesky heaters? I once chased a buzz near forever before realizing it was the heaters. What solved it was a "humdinger; a 500R to 1k pot that replaces the 100R/100R artificial center tap (or the transformer's heater center tap). The center lug goes to ground, and the outside lugs go to either leg of the heaters (i.e. the right lug goes to pin 2, and left lug goes to pin 7). This can be inserted in the preamp section as well. Just thought I'd mention it because it's very easy to try. You may want to try running the heaters off a lantern battery as well, just to rule them out completely.

[markmalin]

I also want to mention that it buzzes with the reverb tank connected OR disconnected.

That would seem to negate it coming from the reverb driver tube, because you would have effectively disconnected that circuit by pulling the tank.

Do you still have the isolating washer for your reverb return RCA jack? This trick is mostly based on anecdotal experience, but try isolating that jack, then running a wire from the ground lug to the ground-side of the reverb return cathode resistor. In your case, this will be the eyelet that the 220k resistor and 820 ohm resistor are in, which then has a black wire to your ground buss.

May or may not be the issue, but takes 1 washer and a hunk of wire to try.
 

>There is a buzz in the amp that goes away when I remove the tube at the input side of the reverb circuit.

I wanted to clarify something.  The tube that I pulled which quieted the circuit down was the guy on the return side of the reverb tank.  So in the picture this is the tube to the right of the reverb transformer wires.  Then, though I have that white isolation washer on the reverb send jack (the one on the left) the face of that jack is still in contact with the chassis, so the "ground" side of the reverb jack is connected to chassis ground.  Then, on the reverb return jack, that guy doesn't have an isolating washer and is also in contact with the chassis - but I also ran the black wire from the reverb tank to that jack's ground lug as well as one side of the 220k resistor on that jack.  Is this not good.

Anyhow, given that should I still try isolating things like you mentioned?

Thanks!
Mark.

[HotBluePlates]

Fender got away with it, but sometimes Fender reverbs hum and buzz.

There is a 220k resistor on the reverb return jack. It is the grid reference resistor for the reverb recovery tube. The problem is, it finds ground at the jack itself, not at the same point that the tube's cathode resistor finds ground.

Any currents flowing through "ground" between that return jack ground and the cathode resistor ground might end up mixed with your reverb signal. That includes any noisy power supply currents.

Best plan is completely isolate the return jack from the chassis, and run the wire. The send jack is not critical, and can be grounded to the chassis. After all, it's isolated by the reverb transformer (and maybe at the tank too; would have to check Accutronics to know for sure).

[PRR]

Why is there no connection to the shell of the TO Tank jack?

Why is the black which which should be there tied to the FROM Tank jack?

I don't see any actual ground here.

[markmalin]

Here's what I did (and PRR, I may have something messed up - there was no connection on the TO tank jack's shell. 

Ok - please tell me if I did this wrong:

- I removed both reverb jacks (TO and FROM)
- I put an insulating washer on the FROM tank (return) jack and NO insulating washer on the TO tank (send) jack.
- Then, on the TO tank (send) jack I re-connected the green wire
- and since the Send jack is no longer isolated from the chassis, I used its "ground" or "shell" lug to connect the reverb transformer's black wire to chassis ground.
- On the now isolated Return jack I connected the tube's grid wire and the 220k resistor, then ran a black wire from the resistor to the cathode ground of the reverb return tube.

Here's where I got confused.  I left the shell connection on reverb tank Return jack unconnected.  The reverb didn't work and I got lots of hissing.  So, I then took a black clip-lead wire and tied that shell connection to chassis ground.  Reverb worked now, but it still buzzed.  Then, I tied the clip lead from the shell connection on the return jack to the same cathode ground.  Everything's working, but I have to say there is still a faint buzz there.  I don't believe it has changed anything.

So - I appreciate all your input.  Sorry to sound like an idiot here.  On top of helping with the buzz, can someone please answer these question?:
- Is it OK to have the reverb tank return side jack isolated and the send side jack not-isolated from the chassis? 
- what am I supposed to connect the shell lug on both the send and return jacks to? 
- Finally, where is the best place to tie the black reverb transformer connection to chassis ground?

Thanks for your patience with me, guys.
Mark.

Fender got away with it, but sometimes Fender reverbs hum and buzz.

There is a 220k resistor on the reverb return jack. It is the grid reference resistor for the reverb recovery tube. The problem is, it finds ground at the jack itself, not at the same point that the tube's cathode resistor finds ground.

Any currents flowing through "ground" between that return jack ground and the cathode resistor ground might end up mixed with your reverb signal. That includes any noisy power supply currents.

Best plan is completely isolate the return jack from the chassis, and run the wire. The send jack is not critical, and can be grounded to the chassis. After all, it's isolated by the reverb transformer (and maybe at the tank too; would have to check Accutronics to know for sure).

[markmalin]

Why is there no connection to the shell of the TO Tank jack?

Why is the black which which should be there tied to the FROM Tank jack?

I don't see any actual ground here.

please see the post above.  I admit, I'm not quite sure how it's supposed to be connected up. 
Mark.

[HotBluePlates]

I count maybe 13 hours between your first and last post on this topic. Did ya just try running the wire from the reverb return jack ground lug to the reverb recovery stage ground (with the jack not-connected/isolated from the chassis)?

It would have taken minutes to see if this solves the problem.  If it didn't, then it's a clue.

Why is there no connection to the shell of the TO Tank jack?

Why is the black which which should be there tied to the FROM Tank jack?

I don't see any actual ground here.

please see the post above.  I admit, I'm not quite sure how it's supposed to be connected up. 
Mark.

You also said that the insulating washer was only on the inside of the chassis, which means there's still a possible ground connection from the jack to the outside of the chassis; so maybe it's grounded.

Speculating takes time; trying takes very little.

[markmalin]

I count maybe 13 hours between your first and last post on this topic. Did ya just try running the wire from the reverb return jack ground lug to the reverb recovery stage ground (with the jack not-connected/isolated from the chassis)?

It would have taken minutes to see if this solves the problem.  If it didn't, then it's a clue.

Now I'm confused...    Your original post said:

...snip...
Do you still have the isolating washer for your reverb return RCA jack? This trick is mostly based on anecdotal experience, but try isolating that jack, then running a wire from the ground lug to the ground-side of the reverb return cathode resistor. In your case, this will be the eyelet that the 220k resistor and 820 ohm resistor are in, which then has a black wire to your ground buss.

May or may not be the issue, but takes 1 washer and a hunk of wire to try.

but now you're asking if I just tried

running the wire from the reverb return jack ground lug to the reverb recovery stage ground (with the jack not-connected/isolated from the chassis)?

What I did (or thought I was doing) was following your first post.  Isolating the return jack.  I didn't have another isolating washer for the return jack but only had the outer plastic washer, so since you said to isolate the return jack and in a later post, that it didn't matter about the send jack, I just removed the washer from the send and put them both on the return.  I had to disconnect/desolder them both to do this.  I guess I got confused because I had the black wire from the reverb tank connected to the send ground jack and it was no longer at chassis ground because I had run a separate wire from the 220k resistor - it seemd to make more sense.  Argh!  I'll change things back to better reflect what your original post said.

(the 13 hours was mostly time with the grand-kids ;)

Thanks for your patience with me.

Mark.

[markmalin]

I count maybe 13 hours between your first and last post on this topic. Did ya just try running the wire from the reverb return jack ground lug to the reverb recovery stage ground (with the jack not-connected/isolated from the chassis)?

It would have taken minutes to see if this solves the problem.  If it didn't, then it's a clue.

Speculating takes time; trying takes very little.

Ok.  The wiring is now as you origionally suggested.  Basically a wire from the reverb return jack groudn lug to the rever recovery stage ground (with the jack isolated from the chassis).  Still have the buzz.

One additional note, if I touch the back of the chassis with my finger near the power and standby switch, the buzz gets very quite (almost not there).


Mark.

[PRR]

> if I touch the back of the chassis with my finger near the power and standby switch

How odd!! What about when you touch 1/2" to the left of the Treb pot?

Do this logically, experimentally.

Take both jacks OFF the chassis. Balance them on a newspaper so they don't short to anything.

The reverb trans green AND black go to the Send jack. It does not actually matter which is which, but do it like Fender did it. It is NOT essential that one side of the reverb trans output be grounded. (Slightly improves "safety" in some very unlikely situation.)

At the Return jack: The center goes to reverb recovery tube grid. The shell goes to bottom of reverb recovery cathode network (a ground, but THE ground for THIS stage).

That should work.

What is that pot and rotary switch near reverb recovery jack? Should they be there? I'm thinking this is Speaker Output Stuff. That really should not be adjacent to Reverb Recovery Stuff.

If it works balanced on newspaper, get Return jack over the chassis and gently slide the newspaper out, touch the Recovery jack shell to chassis. Is buzz more, less, or the same? If not "more", then mount Return jack bare to the chassis; if "more" then mount jack with insulating washers. Verify still quiet. Now repeat the trial with the Send jack. Again, if direct connection is not "more", then mount bare; else mount insulated.

If -both- jacks prefer to be insulated, then add 1,000-5,000pFd (any type) cap from Return shell to near-by chassis, preferably a dedicated screw. You don't need this until you play under a radio transmitter; then you may NEED it.

The Reverb pot "should" kill all these buzzes.... does it??

How silent is reverb switch in Kill mode?

Why does there seem to be a resistor in your reverb return ground lead?

[markmalin]

Thanks for the input, PRR.  I'll try these steps later today.  I wanted to answer a couple of your questions first, though.

What is that pot and rotary switch near reverb recovery jack? Should they be there? I'm thinking this is Speaker Output Stuff. That really should not be adjacent to Reverb Recovery Stuff.

The pot is a negative feedback resistance adjuster.  It's a 10k pot in series with the feedback resistor - I was trying to get more "punch" out of the amp.  I like where it is set now so I had considered measuring the resistance and just changing out the NFB resistor to a fixed value.  I can certainly take that out, or move the pot to the front of the amp - would you recommend I do that before experimenting?  Or maybe first??

The rotary switch is impedance select.

Why does there seem to be a resistor in your reverb return ground lead?

That is a 220k that would normally be off the reverb footswitch (which I don't have).  I thought if I didn't have a reverb on/off footswitch like Fender did, I should set the circuit up as though the footswitch were disconnected.  Did I get this wrong??

Again, many thanks.
Mark.

[markmalin]

> if I touch the back of the chassis with my finger near the power and standby switch

How odd!! What about when you touch 1/2" to the left of the Treb pot?

when I touch 1/2" left of the Treb pot, the buzz does not go away.  Only if I touch the actual toggle on the Standby or power switch does it go completely away. If I touch the back of the chassis all around the side that has the power transformer, it goes down, but not completely away.

Here are the results of my experimentation:
(... please note, I wasn't sure whether you meant that I should have the reverb tank connected, but given that the issue happens with no tank, I left it disconnected for the steps below...)

Do this logically, experimentally.

Take both jacks OFF the chassis. Balance them on a newspaper so they don't short to anything.

The reverb trans green AND black go to the Send jack. It does not actually matter which is which, but do it like Fender did it. It is NOT essential that one side of the reverb trans output be grounded. (Slightly improves "safety" in some very unlikely situation.)

At the Return jack: The center goes to reverb recovery tube grid. The shell goes to bottom of reverb recovery cathode network (a ground, but THE ground for THIS stage).

That should work.

I connected everything as you said above, with the jacks isolated.  The thing is, I added that 200k resistor to ground.  This is the resistor from the recovery grid to ground which you see in Fender's schematic.  This worked, but now I noticed the buzz would increase if you increase the reverb pot, even with the volume off.

What is that pot and rotary switch near reverb recovery jack? Should they be there? I'm thinking this is Speaker Output Stuff. That really should not be adjacent to Reverb Recovery Stuff.

Just a note here - I removed this NFB adjustment pot and just changed the NFB resistor to a 1.5k, which is the total series value that the regular 820 Ohm NFB resistor and the pot (where that pot was set) equalled.

If it works balanced on newspaper, get Return jack over the chassis and gently slide the newspaper out, touch the Recovery jack shell to chassis. Is buzz more, less, or the same? If not "more", then mount Return jack bare to the chassis; if "more" then mount jack with insulating washers. Verify still quiet. Now repeat the trial with the Send jack. Again, if direct connection is not "more", then mount bare; else mount insulated.

I did this as you stated, first with the send jack.  No difference.  The result was not "more" buzz, so I connected the send jack shell to the chassis.  I repeated the situation with the return jack.  Again, no difference.  The buzz is still there.  So it seems neither the return or send jacks care whether they are insulated or not.  (again, the conditions here for my tests were that the reverb tank was not in the circuit and the 220k resistor from the grid to chassis ground that you asked about above was in place.)

If -both- jacks prefer to be insulated, then add 1,000-5,000pFd (any type) cap from Return shell to near-by chassis, preferably a dedicated screw. You don't need this until you play under a radio transmitter; then you may NEED it.

The Reverb pot "should" kill all these buzzes.... does it??

There is still a faint buzz, the one I am tracking down, even with the reverb pot off.  Granted, if I turn the reverb pot on, there is a much louder buzz that increases as you turn the pot.

How silent is reverb switch in Kill mode?

I don't have a reverb kill switch in my amp, but if I manually ground the reverb recovery grid (in the same way the kill switch would have), that faint buzz is still there...but if I turn the reverb pot up the louder buzz that increases with the Rev pot is gone.

Why does there seem to be a resistor in your reverb return ground lead?

Just for completeness, as I mentioned above, this is the 220k from reverb recovery grid to ground that you see in the Fender schematic. 

Any thoughs?

Mark.

[PRR]

> If I touch the back of the chassis all around the side that has the power transformer, it goes down, but not completely away.

The "half-inch left" was joking..... but you DO have an incredibly odd symptom. This is a metal chassis, correct? Any metal conducts lots better than your finger. It should NOT matter "where" you touch it.

What grounding scheme? What deviations?

[markmalin]

> If I touch the back of the chassis all around the side that has the power transformer, it goes down, but not completely away.

The "half-inch left" was joking..... but you DO have an incredibly odd symptom. This is a metal chassis, correct? Any metal conducts lots better than your finger. It should NOT matter "where" you touch it.

What grounding scheme? What deviations?

Yes, the chassis is metal.  

As for grounding, I tried to follow the AB964 schematic/layout for the Fender Vibrolux.  (I guess it is considered a bus ground scheme after further reading).  Diviations include the following:

- I didn't have that angled metal plate that Fender amps use where you solder the ground wires to, so I ran a solid coper wire as a ground bus.  This attaches to the chassis in one spot, in front of the volume pot (then runs past the pots to the Reverb pot.  If you look at the Fender Vibrolux layout, wherever it showw a ground leaving the board, I ran a wire to the ground bus near that point.  This amp has only one channel and no Vibrato, so some of those wires would not be there.  (it might be easiest to look at the pictures to see)
- For the input jacks, their ground lugs are tied together then to chassis at the same location as the ground bus (but are a separate copper wire).
- My circuit inserts a Dumble style OD gain stage in-between the first pre-amp tube and the send tube for the reverb circuit.  So that section's ground runs to the bus at that location.
- As for power supply, I did it as per the Vibrolux and attached these grounds directly to the chassis as the Vibrolux did.
- One additional deviation: I was having a hiss problem so I added a shielded cable for the input jacks into the first gain stage. For this, I tied the shield on the jack end to the bus wire for the input jacks' ground.

Does that help?

(PS - yah... I actually touched the thing 1/2" left of the Treble pot.  What can I say?  D'oh!).

[HotBluePlates]

Now I'm confused...   embarrassed Your original post said:...

My mistake. I couldn't see your pictures from the computer I was on, so I couldn't see that you had already tried my suggestion.

[markmalin]

In case anyone's still listening <laughs>... 

Last night I re-routed the grounds according to the article on grounds in the Hoffman Amps library.  The only thing I didn't do was re-do my bus to be soldered to the back of all the pots.  It's suspended over the pots connected to any of the grounded lugs on any of the pots, then obviously all the pre-amp grounds are connected to it.  In any case, re-routing things to be like the Hoffman menthod didn't change anything.

I think next I'll try running the ground bus soldered to the backs of all the pots and terminated at the ground lug on the input jack like Hoffman suggests.

One thing, I'm not 100% certain the power supply board on my amp is laid out like what is discribed in the picture in the Hoffman doc.  There are 3 caps with a common ground that goes to the bus, and 2 caps with a common ground that goes to the chassis (at the main transformer ground point).  But I still have to convince myself these 2 and 3 caps are the correct ones.  I laid it out like Fender did on the 67 Vibrolux, so I assume that's correct.

Mark.

[John]

My name is not Frazier, but I'm listening. :)

[markmalin]

(thanks for listening, John... )

Last night after doing some reading I found a thread on thegearpage where someone was discussing the necessity of having both 220k resistors that sum the channels to feed the PI even on a 1 channel ABxxx design.  Apparently you tie the second one to ground:

http://www.thegearpage.net/board/showthread.php?t=601134&page=2

I did that and it really quieted the amp down...but I've lost the "gainyness" of the amp now, which I'm not sure I like.  Really, all I've done is reduced the overall gain, so yah, the buzz is quieter (and the hiss is quieter!), but it's still there...just a lower level.

Anyhow, the next thing I think I'll try is using a ground bus that's physically soldered to the back of the pots.  I'm not sure that will have any affect, but I figure sticking closest to Hoffman's ground deisgn is best in any case.

Mark.