Single channel AB763 design question?

[markmalin]

Just a quick question.  Trying to track down a buzz in my 1-channel vibrato-less Vibrolux Reverb I was reading on-line and came across this post:

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

If you scroll down a ways one guy drew up some single channel AB763 schematic clips.  My question is, in his discussion he said it's important to include BOTH 220k resistors in the summing circuit that feeds the PI, tying the unused one (from the missing channel) to ground. 

When I built my amp I only had the one 220k (the resistor from the one channel I implemented).  I realize this may have been wrong... or perhaps I should not have even put that one in there?  But in any case, as I'm trying to track down a faint background buzz I stuck the other 220k in there and tied it to ground.  The amp is quieter, but it has less gain now (less punchy) which I'm not sure I like.  It was quite "gainy" before.  For the sake of the hiss and buzz I'm dealing with I'm inclined to leave this other 220k in there, tied to ground, but I want to hear people's opinions/expertice on this. 

Who knows...maybe there are other, better ways to add gain.  Granted, as I said it is quieter now, which I like, but I miss that punch.

Thoughts?
Opinions?

As always, thanks guys!
Mark.

[Fresh_Start]

The amp is quieter, but it has less gain now

Less gain => less noise

We need to know which channel of the AB763 you're building.  If it's the Vibrato channel, I believe you can forget the 220K resistor from the missing Normal channel.  Or, you could duplicate the circuit with a 220K - .047uf - 100K in series going to ground.  (yes, I left out the path to ground through the triode's plate... )  OTOH, if you're trying to duplicate the Normal channel, then yes you do need a 270K resistor going to ground to simulate the Vibrato channel's 220K PLUS the 50K Intensity pot.  (The bigger you make that resistor, the more gain will stay in the circuit.)

Also, if you're building an AB763 with reverb but not roach-type tremolo, you need to include a 50K (or 47K) resistor in the circuit where the Intensity pot would have been.  Look at Hoffman's AB763: http://www.el34world.com/Hoffman/images/AB763Old.gif  Right in the middle, almost straight down from the tremolo footswitch, you'll see a .01uf coupling cap connected to a 47K resistor going to ground.  That's there to replace the Intensity pot because Hoffman uses bias vary tremolo.  If that resistor is left out, you do get too much gain from the Vibrato channel.

Have you specifically located the buzz as being introduced right before the PI?  Chasing noise can drive you crazy.  I highly recommend working up some variation of Hoffman's "listening amp": http://www.el34world.com/Hoffman/tools.htm.  You don't necessarily need a dedicated amp.  I just wired up a small box with an isolation cap and a volume pot which I can plug into a small solid state amp I never use anymore.  It beats the heck out of trying to guess where the noise is coming from!

Hoffman on tracking down "oddball noises":  http://www.el34world.com/charts/fenderservice6.htm

Hope this helps,

Chip

[markmalin]

The amp is quieter, but it has less gain now

Less gain => less noise

We need to know which channel of the AB763 you're building.  If it's the Vibrato channel, I believe you can forget the 220K resistor from the missing Normal channel.  Or, you could duplicate the circuit with a 220K - .047uf - 100K in series going to ground.  (yes, I left out the path to ground through the triode's plate... )  OTOH, if you're trying to duplicate the Normal channel, then yes you do need a 270K resistor going to ground to simulate the Vibrato channel's 220K PLUS the 50K Intensity pot.  (The bigger you make that resistor, the more gain will stay in the circuit.)

Also, if you're building an AB763 with reverb but not roach-type tremolo, you need to include a 50K (or 47K) resistor in the circuit where the Intensity pot would have been.  Look at Hoffman's AB763: http://www.el34world.com/Hoffman/images/AB763Old.gif  Right in the middle, almost straight down from the tremolo footswitch, you'll see a .01uf coupling cap connected to a 47K resistor going to ground.  That's there to replace the Intensity pot because Hoffman uses bias vary tremolo.  If that resistor is left out, you do get too much gain from the Vibrato channel.

Have you specifically located the buzz as being introduced right before the PI?  Chasing noise can drive you crazy.  I highly recommend working up some variation of Hoffman's "listening amp": http://www.el34world.com/Hoffman/tools.htm.  You don't necessarily need a dedicated amp.  I just wired up a small box with an isolation cap and a volume pot which I can plug into a small solid state amp I never use anymore.  It beats the heck out of trying to guess where the noise is coming from!

Hoffman on tracking down "oddball noises":  http://www.el34world.com/charts/fenderservice6.htm

Hope this helps,

Chip

Thanks for the input, Chip.

I'm building the Vibrato channel and leaving out the Normal channel.  So you're saying I can either forget about the 220k from the missing Normal Channel, OR try the 220K - 0.047uf - 100K to simulate the missing channel.  Any advantages to either?  I do like the tone better w/out the 220k, but as I said, it's noisy!

So given that I'm building the Vibrato channel WITH reverb but W/OUT vibrato, I do need a 47k resistor to ground to compensate for the Vibrato pot not being there, yes?  I do have a 47K to ground which basically replaced that signal path you see in the schematic leading from before the PI (between the 0.1 and 220k).

As for locating the buzz, all I know at this point is if I pull the reverb return tube it's gone.  But then again, that removes the gain stage before the PI because they're in the same tube.  A listening amp would be a good idea.  I have a friend who has something like this - I'll see if I can get some time with it.

I'll check out "oddball noises".  Thanks for the link!

Mark.

[Geezer]

I do like the tone better w/out the 220k, but as I said, it's noisy!

I would keep the tone you like  & continue to look for the source of the buzzzz.....I think your problem is elsewhere.

[Fresh_Start]

Do you have a footpedal?  If so, use it to ground the grid of the reverb recovery triode.  If not, use a jumper to ground that grid.  If that makes the buzz go away, the problem is in the reverb circuit.

Does it buzz if you don't have the reverb cables connected?  Then it's a problem with the reverb recovery stage.  If not, it's either the tank or some part of the driver stage.

Where do you have the reverb transformer grounded?  I'm pretty sure Hoffman just grounds it to the chassis like Fender did.  I had some sucess in noise reduction by grounding the reverb transformer with the rest of the reverb circuit on the board.

Oh, I also found a big difference in reverb cables.  Vintage style I made myself from Hoffman's cable worked best for me but YMMV.  Do you have the return end of the reverb tank as far away from the PT as possible?  That's really important noise-wise.  Also, have you tried moving the reverb cables around?  I keep them on separate sides of the tank inside the bag and then route them so that they are close to each other (1" or so) but never actually touch.  Finally, have you tried different 12AT7 reverb driver tubes?  My Blackface builds have been picky about 12AT7s for whatever reason.

DIVIDE AND CONQUER!

Sorry for yelling.  I started life as a computer programmer, sometimes in assembly language, and there's no substiture for a methodical, step-by-step approach to debugging anything.

Cheers,

Chip

[markmalin]

Thanks for the input, Chip, I really appreciate it.  At the risk of this becoming a parallel thread to my "buzz" thread I'll answer your questions below ;)  But here's a link to the other thread so you can see what I'm trying to accomplish:

http://www.el34world.com/Forum/index.php?topic=9592.0

Do you have a footpedal?  If so, use it to ground the grid of the reverb recovery triode.  If not, use a jumper to ground that grid.  If that makes the buzz go away, the problem is in the reverb circuit.

I don't have a foot switch, but I grounded the grid on the reverb recovery tube and the buzz is still there.

Does it buzz if you don't have the reverb cables connected?  Then it's a problem with the reverb recovery stage.  If not, it's either the tank or some part of the driver stage.

It actually buzzes with the reverb tank disconnected, so yes I'm pretty sure it's in the reverb recovery stage.  If I pull the rev. recovery tube it stops.

Where do you have the reverb transformer grounded?  I'm pretty sure Hoffman just grounds it to the chassis like Fender did.  I had some sucess in noise reduction by grounding the reverb transformer with the rest of the reverb circuit on the board.

I've modified things since the initial pictures in my original post, so now the amp is laid out as per Hoffman's ground layout for a standard ABxxx chassis configuration.  So according to that I have the reverb transformer grounded to the reverb send jack.  I've tried grounding it on the board where the cathode resistor/cap's are grounded for the reverb recovery tube, but no difference.

So I say it's laid out like the Hoffman doc, but the only difference now is that I don't have my ground bus exactly like he suggests.  I have it floating, then soldered to the grounded lugs on any of the pots, then to the chassis - then I have a separate ground bus from the input jack to the same chassis point.  Tomorrow provided I get the time I'd like to do as he suggests and solder the bus wire to the backs of the pots, then to the input jack shell (ground) lugs.

Oh, I also found a big difference in reverb cables.  Vintage style I made myself from Hoffman's cable worked best for me but YMMV.  Do you have the return end of the reverb tank as far away from the PT as possible?  That's really important noise-wise.  Also, have you tried moving the reverb cables around?  I keep them on separate sides of the tank inside the bag and then route them so that they are close to each other (1" or so) but never actually touch.  Finally, have you tried different 12AT7 reverb driver tubes?  My Blackface builds have been picky about 12AT7s for whatever reason.

Re the reverb tank and cables, the probalem is the same w/out them connected, so that's where that stainds.  I haven't swapped out any 12AT7's because the issue seems to be in the reverb recovery.  I did swap out a couple 12AX7's I had for the recovery tube, but no difference. Of course, all my tubes could by noisy so that's not saying much.

DIVIDE AND CONQUER!

Sorry for yelling.  I started life as a computer programmer, sometimes in assembly language, and there's no substiture for a methodical, step-by-step approach to debugging anything.

Hey, no problem!  I'm actually a sofware engineer so I can relate to methodical!  And yelling!! ;)

Mark.

[Fresh_Start]

Note:  this started as a PM, then I thought it made sense to share

Sorry, I hadn't noticed your other thread about the BUZZZZZ in your AB763 build.  I didn't have the benefit of your comments, photos, and the other guys' input for my earlier posts.  It makes everybody's life easier if you keep troubleshooting a given build in one thread, or at least include a link to the first thread when you start a second thread. 

No one pointed out that you have your speaker jack impedance selector right next to the reverb recovery triode.  Is the pot next to it variable negative feedback?  I'm asking because you've got BIG current going through that impedance switch!  BIG current close to little tiny reverb tank output signal may equal noise problems.

Can you humor me for an hour or so?  Disconnect all of the OT wires from the impedance switch.  Pick one and connect it to the speaker jack furthest away from the reverb recovery tube.  Disconnect the NFB pot too and put the stock resistor value in place somewhere else.  Is the BUZZZZ still there?  If not, you've found the culprit. 

Do you really need two speaker jacks?  If not, maybe you can squeeze the impedance selector into the speaker jack hole closest to the OT.  Alternatively, pick two impedances and hardwire one to each speaker jack, ditching the impedance switch and the third OT lead.  Maybe just try moving the OT leads so that they run along the "top" edge of the chassis, above the speaker jacks, and come down on the impedance selector instead of curving around underneath and close to the reverb return jack and reverb recovery grid lead.

Try this:  use shielded cable to and from the Reverb pot.  On the "return" cable, use the shield to ground the Reverb pot with the reverb circuit itself instead of grounding the pot on the bus or chassis.  IOW attach one end of the shield to the pot's ground lug and the other end to ground on your board near where the reverb signal is returned to the circuit. 
I can't guarantee this will solve your problem, but it sure helped me.  The grid of the first triode after the reverb tank seems microphonic because it is amplifying a TINY signal.  You can't over-do lead dress in that area - keep that grid lead away from everything else.

Hope this helps.

Chip

[markmalin]

Thanks, Chip.  Sure, I will gladly humor you    Actually, I'll need to do this tomorrow morning.  (I'm at work now and going to be out of town this evening).  Tomorrow I'll hard wire up the 4 Ohm coil and move the others temporarily out of the way.  I hadn't though about the fact that there is a lot of current moving through there.

I actually got rid of the NFB pot and hard wired in a 1.5k feedback resistor, but for the sake of the experiment I'll put back the 820 Ohm just for now. 

Good idea about the impedance switch.  I really don't need 16 Ohms, so I can certainly either wire up 4 Ohm to one jack and 8 to the other, or just ditch the selector idea and hard wire in 4 Ohms.  (I was thinking of building an extension cabinet, though, but for the sake of tracking this down I'm more than willing to simplyfy!).

I'll see how these things go, then work on the idea of shielding to/from the reverb pot.

Thanks again!
Mark.

Note:  this started as a PM, then I thought it made sense to share

Sorry, I hadn't noticed your other thread about the BUZZZZZ in your AB763 build.  I didn't have the benefit of your comments, photos, and the other guys' input for my earlier posts.  It makes everybody's life easier if you keep troubleshooting a given build in one thread, or at least include a link to the first thread when you start a second thread. 

No one pointed out that you have your speaker jack impedance selector right next to the reverb recovery triode.  Is the pot next to it variable negative feedback?  I'm asking because you've got BIG current going through that impedance switch!  BIG current close to little tiny reverb tank output signal may equal noise problems.

Can you humor me for an hour or so?  Disconnect all of the OT wires from the impedance switch.  Pick one and connect it to the speaker jack furthest away from the reverb recovery tube.  Disconnect the NFB pot too and put the stock resistor value in place somewhere else.  Is the BUZZZZ still there?  If not, you've found the culprit. 

Do you really need two speaker jacks?  If not, maybe you can squeeze the impedance selector into the speaker jack hole closest to the OT.  Alternatively, pick two impedances and hardwire one to each speaker jack, ditching the impedance switch and the third OT lead.  Maybe just try moving the OT leads so that they run along the "top" edge of the chassis, above the speaker jacks, and come down on the impedance selector instead of curving around underneath and close to the reverb return jack and reverb recovery grid lead.

Try this:  use shielded cable to and from the Reverb pot.  On the "return" cable, use the shield to ground the Reverb pot with the reverb circuit itself instead of grounding the pot on the bus or chassis.  IOW attach one end of the shield to the pot's ground lug and the other end to ground on your board near where the reverb signal is returned to the circuit. 
I can't guarantee this will solve your problem, but it sure helped me.  The grid of the first triode after the reverb tank seems microphonic because it is amplifying a TINY signal.  You can't over-do lead dress in that area - keep that grid lead away from everything else.

Hope this helps.

Chip

[markmalin]

Hi Chip

No one pointed out that you have your speaker jack impedance selector right next to the reverb recovery triode.  Is the pot next to it variable negative feedback?  I'm asking because you've got BIG current going through that impedance switch!  BIG current close to little tiny reverb tank output signal may equal noise problems.

Can you humor me for an hour or so?  Disconnect all of the OT wires from the impedance switch.  Pick one and connect it to the speaker jack furthest away from the reverb recovery tube.  Disconnect the NFB pot too and put the stock resistor value in place somewhere else.  Is the BUZZZZ still there?  If not, you've found the culprit. 

I did as you said, disconnecting the other 2 OT tap wires and actually pulled them back through the chassis temporarily.  Then I disconnected the impedance jack and hard-wired the jack up to the 4 Ohm tap only, with the NFB resistor (hard wired) direct to that point (i.e. no adjustable NFB pot). 

Here's what I came up with.  The amp still buzzed faintly (as it's been doing since reducing the gain by adding that 220k I mentioned above), but I discovered something else.  It is only buzzing when the volume is turned off.  Then...I had noticed if I passed my hand over the chassis it would quiet down, so I turned off the overhead florescent lignt near my workbench.  No buzz.  So the SOURCE of the buzz is the light, but the amp is picking it up like an antenna when the volume is set at ground (turned all the way off).  I wound up shortening some of the grid wires some more - I'm wondering if at this point it's an issue anymore?  When I play live, it is going to pick up buzz from stage lighting, but only if I had the volume pot turned off (which would be never).  Your thoughts???

All that said, I can certainly try shielding the reverb signal to/from the pot.  One thing I did while I was at it was to complete the Hoffman style ground bus layout, AND while I was at that, I shortened the reverb pot wires.  I think all these little things are helping.  I would like to try to go back to the way I had the PI input before, which was without that second 220k resistor in the old "summing" circuit used for 2 channel Fenders.  I liked the gain/punch then.  I suppose it's going to make the buz and background hissing more pronounced, though :(

Do you really need two speaker jacks?  If not, maybe you can squeeze the impedance selector into the speaker jack hole closest to the OT.  Alternatively, pick two impedances and hardwire one to each speaker jack, ditching the impedance switch and the third OT lead.  Maybe just try moving the OT leads so that they run along the "top" edge of the chassis, above the speaker jacks, and come down on the impedance selector instead of curving around underneath and close to the reverb return jack and reverb recovery grid lead.

I may try using 2 jacks, one for 4 Ohms and one for 8.  However, if I do this, I would need to buy another shorting tip jack, wouldn't I?  I've never seen an amp with individual jacks for different impedance.  Any thoughts on how to do that?

Thanks again for the input
Mark

[Fresh_Start]

OK!  You found the source of the noise.

If you flip the chassis over so that the insides are shielded from the flourescent light, does the buzz go away?  If so, you may need the shield the top of your cab, assuming it's a combo.  I have a sheet of steel in my workshop to put over an open chassis specifically because I've got a flourescent light too. 

BTW only buzzing when the volume pot is turned all the way down may be a clue that you've got a bad ground on that pot.

Also, please take the extra 220K resistor out of the circuit and listen to it.  The schematic you posted in the other thread looks right - that other forum's thread would only be correct if you were copying the Normal channel. 

There are some other things I see in your gut shot which aren't in the schematic.  Can you show us an as-built schematic with that switch in it?  What's that trim pot near V1? (gain for the boost stage, eh?)   The cap on V1?  Or is that actually V2?  Is the long blue/white lead from the switch a sensitive signal?  If so, it's too close to the board and your B+ "rail".  Maybe shielding that would help too.  You can run leads so you get vertical separation too, IOW 3/4" above the board is relatively far away from everything even if it's "up". 

I may try using 2 jacks, one for 4 Ohms and one for 8.  However, if I do this, I would need to buy another shorting tip jack, wouldn't I?  I've never seen an amp with individual jacks for different impedance.  Any thoughts on how to do that?

If moving the OT leads away from the reverb recovery tube didn't make any difference in the noise level, put it back the way it was (or maybe with the OT leads tucked under the "upper" lip of the chassis on their way to your impedance switch). 

I've seen plenty of schematics and home brew amps with multiple jacks on them for different impedances instead of a switch.  Cheaper I guess, and there's nothing wrong with that!  Just checked the AX84 P1 Extreme documents.  It shows a 220K-5 watt resistor permanently between the 8-ohm OT lead and ground.  Now I'm puzzled - we usually see a 470K "safety resistor" and it's switched on the jack.  I don't know what effect having that resistor in the circuit permanently would have.  My guess is that it's close to zero.  220K in parallel with 8-ohms barely affects the net resistance mathematically.  However, my gut doesn't like the idea.  Frankly, my Super Reverb build just had two non-switching jacks (even with an impedance selector switch).  IIRC that's how Leo did it... you just have to make sure you've always got a load attached to one of the speaker jacks!

Please post your as-built schematic.

Cheers,

Chip

[markmalin]

OK!  You found the source of the noise.

If you flip the chassis over so that the insides are shielded from the flourescent light, does the buzz go away?  If so, you may need the shield the top of your cab, assuming it's a combo.  I have a sheet of steel in my workshop to put over an open chassis specifically because I've got a flourescent light too. 

I have a spare piece of sheet metal and if I put it over the top of the chassis, the buzzing (at volume off) is gone.  So I have a piece of sheet metal mounted to the inside of the cabinet (combo) so I'm good to go there.

Also, please take the extra 220K resistor out of the circuit and listen to it.  The schematic you posted in the other thread looks right - that other forum's thread would only be correct if you were copying the Normal channel. 

Done.  Sounds much nicer to my taste, but much more hiss (another topic).  But no buzz and lots of punch.

There are some other things I see in your gut shot which aren't in the schematic.  Can you show us an as-built schematic with that switch in it?  What's that trim pot near V1? (gain for the boost stage, eh?)   The cap on V1?  Or is that actually V2?  Is the long blue/white lead from the switch a sensitive signal?  If so, it's too close to the board and your B+ "rail".  Maybe shielding that would help too.  You can run leads so you get vertical separation too, IOW 3/4" above the board is relatively far away from everything even if it's "up". 

It's going to take me a while to draw up the schematic as the current circuit is laid out.  I'll try to post something by tomorrow evening. 

Re the trim pot and V2 and all...this is a Dumble Overdrive circuit like you'd see on TheAmpGarage.com. I have a clip of that part of the schematic in my other thread, but I'll try to get a full schematic drawn up.  The trim pot is the entrance of the Dumble OD's first gain stage. It regulates the level of the signal in.  It's kind of a hack, I know... I mean putting the D OD into a Fender pre-amp, but it turned out OK. The switch on the front of the amp takes the output of V1 of the Fender pre-amp channel and re-routes it through the D'style OD, then back in. It's inserted in the secion of the Fender layout labeled [z] (that's the white/blue wire with the pink shrink tubing on it that you mentioned above).  This is not tuned (yet another thread topic some day!!), just hacked in, so it's way too high a level and it's hard to get any good smooth overdrive from it - it's just too loud.  When the switch is off, the signal is routed per Leo's original desing and the input to the D' OD stage is grounded. 

Please post your as-built schematic.

WIll do... It's just going to take me a while!

Thanks again, Chip.

Mark.

[markmalin]

Here's a schematic.  Sorry for the poor quality - it's cut/pasted from a Fender schem. and a Brownnote schem., and some hand drawn stuff in Paint.  The only things I don't show here are the multi-tap OT and switch, and that I added a Mid pot per the Super Reverb, but have a switch to go between the using the Mid pot, and taking it out of the circuit and just using a 6.8K resistor as per the standard Vibrolux.  (I may remove that because it's kind of pointless, and was more of an experiement than anything...)