AB763 Troubles

[RoadShow]

In 2023 I decided to build the Hoffman AB763 single channel.  I ordered parts in May & June 2023 and began the build in July, ordered the faceplates in June, the chassis was up and running in August, and I built the cabinet in September.

The head played fine through November 2023, maybe 100 hours or so.  The power tubes were both fresh and not too far apart on current biasing, set for 60% power.

Early December 2023 guitar sound cut out, so I turned the off song I was playing with, I heard like a sizzle, looked at the amp, and wisps of smoke were coming out.  I hit the power switch, and the transformer was shot as was the JJ GZ34 and the V6 Sylvania 6L6GC. 

January 2024 I replaced the tubes and transformer and brought the unit up on my current limiter, it was fine.  All other tubes were fine, checked on Hickok 6000A.  I checked all the voltages and created a new voltage chart, they had changed a bit but were close.  Again, the power tubes were both fresh and not too far apart on current biasing, set for 60% power.

Left powered up on workbench for 8 hours, all seemed fine.

Started using it again around April 2024, and after an hour or so the output volume started dying so I shut it off, and checking it out found V6 JJ 6L6GC had gone bad.  Again.
I replaced the power tubes and brought the unit up on a current limiter, it was fine.  All other tubes were fine, checked on Hickok 6000A.  I checked all the voltages and created a new voltage chart, they had changed a bit but were close.  Again, the power tubes were both fresh and not too far apart on current biasing, set for 60% power.

Getting tired of pulling the head out of my stack of heads, I just let it sit for several months.  I started using it at the beginning of the year (2025) and after maybe 10 hours the fuse blew.  Back to the bench.  ANOTHER bad power tube, again V6 Penta 6L6GC.  Swap in a new set of power tubes, bring it up on the current limiter, check the voltages.  While measuring the voltages/currents at the 1ohm resistors on the power tubes to set the biasing power at 60%, the V6 1ohm reading would shoot from ~40mV up to as high as 155 mV when probing the 1ohm resistor lead.  It wasn’t the solder joint, it was at the lead going into the resistor.  So, I think ah ha, could this be it?  The actual connection from the cathode to ground intermittently open?  So, I replace *both* 1ohm resistors just to be safe, and all seems well.  I wait for a good time to put it back into the head stack.  So, I did that in April 2025.

May 2025, and guess what, after an hour or so the output started dying so I shut it off, and checking it out found a second V6 Penta 6L6GC had gone bad.  Again.  I decide to check the 1760J OT windings.  Primary Brown goes to V6-3 and reads 89.6ohms to the red CT.  Blue goes to V7-3 and reads 103.8 to the red CT.  The data sheet calls for a reading of 185ohms from Brown to Blue, so effectively I’m getting 193.4ohms, not too far off.

I’ve tried swapping the 12AT7 between the reverb driver circuit and PI, that didn’t help.

Here’s the link to the Hoffman AB763 single channel:
https://el34world.com/Hoffman/files/Hoffman_AB763_1.pdf
I built it with the dual input jack option. 
My most recent voltage chart and picture of the guts are attached. 

Note: In the Hoffman schematic he has the second stage of the tremolo circuit plate connected directly to the B bus.  In the original Fender circuits it goes through the traditional 100k as does the Mojotone schematic.  Is that OK?

Note: I have my music room set up such that I have a rack to stack my heads on, and I run them to a patch panel that goes to my 4x12 speaker cabinet.  Makes it so easy peasy to  change what head I feel like using.  I move one jack and I’m on my way.  Never had a problem with this.

Thanks for looking and hopefully helping me resolve this.

[tubeswell]

Sounds like you didn’t fix the thing that was causing the output tube(s) to go bad.
Take some voltages and report back (Without any output tubes plugged in). What do the voltages look like on the grid pins?

[Rontone]

The GZ34 heaters pins, shouldn't they both be 5V?

[RoadShow]

Sounds like you didn’t fix the thing that was causing the output tube(s) to go bad.
Take some voltages and report back (Without any output tubes plugged in). What do the voltages look like on the grid pins?

Tubeswell,

Grid pin voltages slight change, from -47.2 to -47.5
All bus voltages are naturally higher.
Voltage chart is attached.

Thanks for looking...

[RoadShow]

The GZ34 heaters pins, shouldn't they both be 5V?

Sorry for the confusion,

Measured lead to lead is 5.17VAC.  Those numbers are a little misleading as I was measuring to ground on each rectifier pin.

Thanks for looking.

[shooter]

where does the under-board black wire go??  The cap-can? and....
has it been visually verified as solid??

[RoadShow]

where does the under-board black wire go??  The cap-can? and....
has it been visually verified as solid??

That is the OT common.

[AlNewman]

Looks like maybe a dry solder joint to the bias dropping resistor?

[RoadShow]

Looks like maybe a dry solder joint to the bias dropping resistor?

Thx for looking,

No, it's solid.

[shooter]

is it "floating" or chassis grounded?
is it also the ground for the cliff-jacks?

[RoadShow]

is it "floating" or chassis grounded?
is it also the ground for the cliff-jacks?

The wire you asked aobut previously that goes to the OT common comes from the cliff jack ground point.  While not visible in the pic, the cliff jack goes to a chassis ground right below it and between V3 and the chassis wall.

Thanks.

[tubeswell]

You need to sit there in the dark for (maybe quite) a while with the amp idling and multiple V-meters clipped on the power rail filter cap nodes and the output tube plates and screens (and especially on the output tube grids) as well as the LTP plates and the bias supply.


You possibly may have a bad resistor or bias supply cap in the bias supply that goes wonky as the amp heats up, causing the bias voltage to die off.

[dogburn]

I'm far from an expert, but is it possible that there is arcing between a couple of the joints on the pins on the tube socket? Some of those pins and joints look rather close to each other, and maybe when you are playing through the amp, it's enough to cause arcing and that's why readings look good when idling, but after playing a few hours you burn out the tube? This would be another thing you could see by looking at it in the dark (but would have to be running a signal through or playing guitar at your typical volume).
 

[RoadShow]

You need to sit there in the dark for (maybe quite) a while with the amp idling and multiple V-meters clipped on the power rail filter cap nodes and the output tube plates and screens (and especially on the output tube grids) as well as the LTP plates and the bias supply.


You possibly may have a bad resistor or bias supply cap in the bias supply that goes wonky as the amp heats up, causing the bias voltage to die off.

My thoughts were headed in that direction, that it had to be an issue somewhere in the power system, the supply or biasing.   I have a signal generator that I can use to at least run a sine wave in or an mp3 song set to guitar input level as a stimulus.  I have 3 DVMs I can hang along with a scope.

Looking over the schematic this AM, I also started to wonder about the tremolo system.  That pulls on the bias supply, so I could at least disconnect it.  And that thought brings me back to a statement I made yesterday in the original post,  the tremolo output stage has the B bus supply run directly to the anode instead of through a 100k like the original Fender AB763 designs, Mojotone, and other schematics.  Yes, I realize Hoffman used the 6G16 design, and that's how it's set up.
 Also, the original Fender designs do not have it integrated into any of it's biasing.  Even in the footswitch disabled condition I have run it so far, could an issue in the tremolo section affect the bias supply in a bad way?  I think I'm going to pull the connection for now, I'm just not trusting it...

[RoadShow]

Well,

No luck watching for any out of the ordinary voltages or burning or unlit tubes.

Clearly, this only rears itself during real action.

My next step is to disconnect the tremolo (that I don't use anyway), replace the PS electrolytics, and replace everything from the bias supply to the power tubes.  Yeah, it's shotgunning, but is faster and cheaper than looking for a ghost.

Any other ideas, LMK...

Thx.

[stratomaster]

Since it's always the same tube that fails I'd be inclined to replace the PI coupler to the grid of V6 and nothing else.

[RoadShow]

Since it's always the same tube that fails I'd be inclined to replace the PI coupler to the grid of V6 and nothing else.

You think it's breaking down and letting a smack of DC through?  Worst case it going from -47 to 0 would let uncontrolled current through, right?  Then bye bye tube...

[tubeswell]

Another item that fails on all silverface amps is the white mallory bias supply cap(s) - so replace as well


Oops, why did I say ^that^, already knowing it’s a home built amp. (Must be all the late night action.)

[p2pAmps]

Quick note,  I notice you have the center tap on the heaters grounded and also the 100Ω artificial center taps on the board.  Decide which you want but not both.  I doubt this is the issue but something I noticed...

~m

[tubeswell]

Did you add any customisation to the build when you built it (like stuff that’s not normally in an AB763 output stage)?

[RoadShow]

Quick note,  I notice you have the center tap on the heaters grounded and also the 100Ω artificial center taps on the board.  Decide which you want but not both.  I doubt this is the issue but something I noticed...

~m

OK,

Thanks.

[RoadShow]

Did you add any customisation to the build when you built it (like stuff that’s not normally in an AB763 output stage)?

The only deviation is I used a dual 32uF can for the A & B bus and a dual 16uF can for the C & D bus.

[RoadShow]

OK folks,

As discussed, I basically replace all the components on the back end, with the exception of the transformer.  See the pic with all the components circled.  I also set the tubes up for individual biasing which is noted in the voltage chart attached.  I disconnected the tremolo circuit to eliminate that, just in case.

I let it run on my bench for 4 or 5 hours idling and there were no problems.  So, I put it back in the head cabinet and started using it from June 20 until Wednesday 7/09.  During that time, I used it maybe 8-10 times for 15-20 hours.  While playing Wednesday, suddenly there was a loud hum so I instantly shut it down while I pulled it out and pulled the back off so I could directly see the tubes without looking through the mesh.  At that point I turned it back on and V6-6L6GC lit up with a couple of bright blue flashes and was toast.

I get the same readings from the OT as before: V6-3 to Red (A) is ~90ohm and V7-3 to Red (A) is ~104.  Could this OT be breaking down during use?  What else could be the problem?  I’ve replaced everything else.  I’ll replace it, but if this happens again, I’m lost and I’m done with it.  It’s evil.

Help!!

Thanks in advance...

[stratomaster]

Do you have a negative feedback resistor or was that just a schematic mistake/omission?

[pdf64]

Has V6 socket been replaced yet?

[RoadShow]

Do you have a negative feedback resistor or was that just a schematic mistake/omission?

Sorry,

The 820ohm per the Hoffman schematic is present, neglected to put it into the regenerated schematic without tremolo and dual bias.

Updated the schematic, thanks for noticing.

Bill

[RoadShow]

Has V6 socket been replaced yet?

No it hasn't.  I just tested each socket jack with a 3/32 pin (same size as the 6L6GC pin) and each one is snug and make good contact.

What do you suppose could be wrong there?

[stratomaster]

Has V6 socket been replaced yet?

Along these lines, I was looking more closely at the photos. There is a lot of questionable workmanship in this amp.  There is bird's-nesting happening with stranded wires (see 5v heaters on rectifier socket, hard to see if it's on the power tubes as well) that could make an unintended connection. There are solder starved joints, there are tack solder connections at the can cap with no wrapping, a lot of stray burns, and deviations from standard/established/best practices.

I'm leaning towards an unintended short on that V6 socket, be it from workmanship or a prior arcing event.

[RoadShow]

Along these lines, I was looking more closely at the photos. There is a lot of questionable workmanship in this amp.  There is bird's-nesting happening with stranded wires (see 5v heaters on rectifier socket, hard to see if it's on the power tubes as well) that could make an unintended connection. There are solder starved joints, there are tack solder connections at the can cap with no wrapping, a lot of stray burns, and deviations from standard/established/best practices.

I'm leaning towards an unintended short on that V6 socket, be it from workmanship or a prior arcing event.

That pic was from the original build early last year and was cleaned up long ago.  Picture from this AM attached.

I'll swap that socket, and give it a whirl.

[stratomaster]

That stacked trim pot arrangement looks tenuous at best.  If you're losing just one side you could have an intermittent wiper on that trim pot.

Try using a standard bias arrangement and see if the problem vanishes.

[RoadShow]

That stacked trim pot arrangement looks tenuous at best.  If you're losing just one side you could have an intermittent wiper on that trim pot.

Try using a standard bias arrangement and see if the problem vanishes.

I actually started with a standard bias arrangement per the Hoffman AB763.  I decided to set it up with dual bias as a test in this last go around to get the same biasing on both tubes while dealing with this whole issue.  They are epoxied together and it is actually pretty stable and secure.  If you look at the wipers on them you'll notice the different settings and if you then compare to the voltage sheet attached they are biased almost identical.

[dogburn]

The light in the photo is not very good, but it looks like the end of the heater wire on the tab for pin 7 is wrapped around so it is rather close to the tab 6 - could be arcing there. You can bend the tabs back to get more separation but may be better to rewire.

[Willabe]

Left powered up on workbench for 8 hours, all seemed fine.

This really doesn't test much, as the power tubes are not under the stress of playing through the amp. They have to work much harder when played through.

Note: I have my music room set up such that I have a rack to stack my heads on, and I run them to a patch panel that goes to my 4x12 speaker cabinet.  Makes it so easy peasy to  change what head I feel like using.  I move one jack and I’m on my way.  Never had a problem with this.

Has this amp head ever lost the V6 tube while NOT plugged into that patch panel?

What kind of jacks and wire gauge are in the patch panel? What kind of speaker cables/jacks are you using to go from the head's speaker out to the patch panel, then from the patch panel to the speaker cab? What wire gauge?

What about the heads output speaker jack? V6 tube sockets connection to the OT fly lead? What OT are you using? Who made it? Is it for a 2x6L6GC amp? Are you using the correct ohm speaker load? Have you tried the amp with different speakers? 

Do you have another OT you could swap in to test with? 

I would not plug the amp into that panel until I got it to stop burning up V6.

I know V6 is before the OT output, so the problem should be before the OT output, but since you've changed all the parts from the PI to the OT.......

[Willabe]

And your impedance selector switch, that could be bad.

[Willabe]

Your dual -bias has no safety R across the pot, so if the wiper looses contact with the R tract the tube will have no -bias.

I know you had single -bias at 1st, but that was the same type of pot and was wired up the same.

Those pots might be bad? I had a small pot like that for -bias adjust and as I would turn it it would ark and loose the -bias.

Look at the -bias drawings attached. In both of those, if the pot's wiper looses connection, full -bias is applied. (Those are from a Kevin O'Connor TUT book.)

[RoadShow]

Left powered up on workbench for 8 hours, all seemed fine.

This really doesn't test much, as the power tubes are not under the stress of playing through the amp. They have to work much harder when played through.

Thanks for all the brainstorming, I appreciate the time and thought.

Understood, but after making any changes I reset bias and remeasure all the voltages.  Then, I just let it sit and idle with a signal generator going into the amp and feeding a speaker.

Note: I have my music room set up such that I have a rack to stack my heads on, and I run them to a patch panel that goes to my 4x12 speaker cabinet.  Makes it so easy peasy to  change what head I feel like using.  I move one jack and I’m on my way.  Never had a problem with this.

Has this amp head ever lost the V6 tube while NOT plugged into that patch panel?

Yes, once.  I've used this patch panel for years and the dual JTM45/2203 I build a couple of years ago, the JTM45 and 18W I built several years ago all use this panel, SV20H, DSL20, never a problem with any head with this patch panel.  During this trouble I did try a known good jack.

What kind of jacks and wire gauge are in the patch panel? What kind of speaker cables/jacks are you using to go from the head's speaker out to the patch panel, then from the patch panel to the speaker cab? What wire gauge?

All Switchcraft jacks & plugs, 16 gauge wire.

What about the heads output speaker jack? V6 tube sockets connection to the OT fly lead? What OT are you using? Who made it? Is it for a 2x6L6GC amp? Are you using the correct ohm speaker load? Have you tried the amp with different speakers? 

Head output speaker jack is a Cliff jack.  Not sure what you mean by "V6 tube sockets connection to the OT fly lead?" Do you mean is it a good connection? Yes.  The OT is Hammond 1760J, recommended for the Bandmaster which this essentially is.  Yes, the speaker load setting is correct.  No, I haven't tried it with different speakers, although I could.  The 4x12 set up is what I drive with the aforementioned heads.

Do you have another OT you could swap in to test with? 

No, but per my post a couple of days ago I was headed in that direction as I'm running out of possibilities.  $65 + ship from HawkUSA, whatever.  Cheap compared to the number of 6L6GCs this amp has eaten!! 

I would not plug the amp into that panel until I got it to stop burning up V6.

Per your earlier question, I think I'll run it through a 2x10 set up I have.  I swapped the tube socket yesterday per pdf64's thought and played through it for 2-1/2 hours earlier today, so far so good.

I know V6 is before the OT output, so the problem should be before the OT output, but since you've changed all the parts from the PI to the OT.......

Right...

[RoadShow]

And your impedance selector switch, that could be bad.

Anything's possible, if I fry another V6 I'll swap the OT and that's a good time to swap the selector.

[RoadShow]

Your dual -bias has no safety R across the pot, so if the wiper looses contact with the R tract the tube will have no -bias.

I know you had single -bias at 1st, but that was the same type of pot and was wired up the same.

Those pots might be bad? I had a small pot like that for -bias adjust and as I would turn it it would ark and loose the -bias.

Look at the -bias drawings attached. In both of those, if the pot's wiper looses connection, full -bias is applied. (Those are from a Kevin O'Connor TUT book.)

Per the safeguard theory, yes.  Considering that I'm using (2) 25k now, at full swing to get 25k per pot I would need a 50k pot and 47k resistor... Resulting equivalent resistance is 12.5k in both cases.  Easy enough to do.

[Willabe]

Then, I just let it sit and idle with a signal generator going into the amp and feeding a speaker.

That's not very good either. Not really putting much stress on the amp, only 1 note. No harmony/multiple notes at once, no volume dynamics changes, no low bottom end, etc. Much better to run a CD player with the CD set to loop. Then you get many different notes high to low and all kinds of sounds, drums, etc. 

And you can burn up speakers like that, the voice coil will get hot. Guys would burn up speakers with single note feed back held too long. The speakers coil can't take.

Yes, once.  I've used this patch panel for years and the dual JTM45/2203 I build a couple of years ago, the JTM45 and 18W I built several years ago all use this panel, SV20H, DSL20, never a problem with any head with this patch panel.  During this trouble I did try a known good jack.

I'd try switching jacks on the panel that you hook the amp up to with 1 of the other heads jacks on that panel. Just swap the amps where you plug them in the panel with each others jack connections. That should fully rule out the panel if it still burns up V6. 

I swapped the tube socket yesterday per pdf64's thought and played through it for 2-1/2 hours earlier today, so far so good.

I agree with pdf64 on that.

[Willabe]

Per the safeguard theory, yes.  Considering that I'm using (2) 25k now, at full swing to get 25k per pot I would need a 50k pot and 47k resistor... Resulting equivalent resistance is 12.5k in both cases.

Not sure what you mean? Are you saying your at the pots end of wiper swing to get enough -bias dcv? So you need to use a larger value bias pot to get more -dcv range?

You don't need to increase the -bias pot values to 50K, just increase the 15K pot tail range R. Just use alligator clip leads to swap in different value R's until you the -bias pot's wiper to ~center on the pot when biased up where you want it.

The values in those 2 drawings have plenty of bias -dcv swing. Just might have to adjust the raw -dcv.

[tubenit]

I would be tempted to change it to cathode biased to see if that made a difference?


With respect, Tubenit

[RoadShow]

Yes, once.  I've used this patch panel for years and the dual JTM45/2203 I build a couple of years ago, the JTM45 and 18W I built several years ago all use this panel, SV20H, DSL20, never a problem with any head with this patch panel.  During this trouble I did try a known good jack.

I'd try switching jacks on the panel that you hook the amp up to with 1 of the other heads jacks on that panel. Just swap the amps where you plug them in the panel with each others jack connections. That should fully rule out the panel if it still burns up V6. 

That's what I was saying when I said "During this trouble I did try a known good jack."
I used jacks assigned to a different head and it happened.

[RoadShow]

Per the safeguard theory, yes.  Considering that I'm using (2) 25k now, at full swing to get 25k per pot I would need a 50k pot and 47k resistor... Resulting equivalent resistance is 12.5k in both cases.

Not sure what you mean? Are you saying your at the pots end of wiper swing to get enough -bias dcv? So you need to use a larger value bias pot to get more -dcv range?

You don't need to increase the -bias pot values to 50K, just increase the 15K pot tail range R. Just use alligator clip leads to swap in different value R's until you the -bias pot's wiper to ~center on the pot when biased up where you want it.

The values in those 2 drawings have plenty of bias -dcv swing. Just might have to adjust the raw -dcv.

What I'm saying is, with the values as is, right now the pot sweep is pretty much centered for 60% bias.  I like that.  Yes, I can change the fixed resistors after adding the parallel resistors to the pot.  Or, maintain the same equivalent resistance of the parallel pots by increasing those to 50k and adding the 47k parallel resistors.

However, to your point, changing the fixed resistors would be easier.

[Willabe]

If you add those safety R's across the bias pots, I would take the power tubes out, and with the amp on, measure the -dcv on both pots with the the pots turned full clockwise and then full counter clockwise. Write those numbers down.

Then add the safety R's, then measure those -dcv's again. Now you know what you need to get to as far as -dcv swing on the pots to get back where you were.

Then if you need to, increase/decrease the pots tail R as needed. 

[Willabe]

Then if you need to, increase/decrease the pots tail R as needed.

You might be able to use a 25K pot to dial in the new R value.

[RoadShow]

So,

I was thinking this week, the Dual 2203/JTM45 and JTM45 I build have essentially the same power requirements as they have a single KT66/6L6GC PP tube output and I have never had a problem with those amps, so what do they have for OT, see the attachment.

Well, today V6 6L6GC went again.  Only the OT is really left.  I can't get the Classictone anymore, but I can sure put in the Hammond 1750Q.  Maybe, even though the 1760J is what was recommended by Hammond and technically should work, when I change OT I should swap in the 1750Q.

Thoughts??