Super Reverb Hoffman AB763 voltage readings questions

[osing]

Howdy,

I just finished reassembling a Hoffman AB763 Super Reverb that I built several years ago. I have not played much guitar in recent years but wanted to get started again and decided to play around with various switchable mods (normal channel disconnect, tremolo disconnect, cathode caps/resistors, NFB). Anyway, after reassembling, with everything in stock mode, I am seeing voltage readings which I believe are a bit higher than when I originally built the amp (at least for power tubes). I feel like the amp is breaking up earlier than before as well, but I could be imagining that (again, everything is in stock mode). Otherwise, it seems to be working fine. Any thoughts as to why voltages would be higher now than in the past, i.e. any hints as to what I might have connected incorrectly or any components I might have damaged with a soldering iron, etc. that could cause this change yet still allow the amp to function? Are the voltages below actually relatively normal for a Super Reverb?

All tubes are JJ. Power tubes 6L6GC. Preamp tubes are 12AX7 except for V3 &V6 which are 12AT7. Rectifier tube is a GZ34.

I had notes from the original build suggesting a bias of ~45-50ma and plate voltages in the 460V range using the same JJ tubes as I still have in the amp. Below are current readings:

Line voltage at outlet 123.5V

Bias current V7 - 46mA
Bias current V8 - 39mA
Plate voltage V7 - 482V
Plate voltage V8 - 485V
Bias cap @ -50V

V1a - 330V
V1b - 295V
V2a - 271V
V2b - 273V
V3a - 480V
V3b - 480V
V4a - 276V
V4b - 282V
V5a - 230V
V5b - 484V
V6a - 260V
V6b - 251V

Rectifier GZ34  - 486v pins 2 & 8

Thanks!

[eleventeen]

Couple of things:

Yes, your line voltage is on the high end.

Second, does the Hoffman spec a GZ34? Because that will definitely supply more volts (and not just ten volts) more than a 5U4 or a 5Y3. If you are concerned about too-high volts, putting in a direct-heat rectifier is the easiest way to chop 30+ volts out your B+.

Third, V1 volts are really high, like 60 volts high. And V2 volts are maybe 40-50 volts high. Using a Fender AB763 as reference. 

I would measure cathode volts on the various preamp stages. The voltages that appear across the cathode resistors are very sensitive to the current through those little tubes. That measurement may tell you little if the B+ on those tubes is just too high, but it will tell you if proper currents are running through them.

Did you change the resistors between the power supply nodes?

[HotBluePlates]

I just finished ... a Hoffman AB763 Super Reverb ... I am seeing voltage readings which I believe are a bit higher than when I originally built the amp (at least for power tubes). ...

I had notes from the original build suggesting a bias of ~45-50ma and plate voltages in the 460V range ... Below are current readings:

Line voltage at outlet 123.5V

Bias current V7 - 46mA
Bias current V8 - 39mA
Plate voltage V7 - 482V
Plate voltage V8 - 485V
...
Rectifier GZ34  - 486v pins 2 & 8

Okay, an original blackface Super Reverb would have 460vdc on the 6L6 plates back in the 60's. It would be higher now due to higher wall voltages (more like 120v or more instead of 100-115v). The '67 Super Reverb I used to have had higher voltages than indicated on the schematic for this reason.

The AB763 Super Reverb schematic shows a 360-0-360v PT with a 110-115v wall voltage. Doug shows his 022798 transformer for the Super Reverb provides 700v CT with 120v input; call that 350-0-350v with 120v input.

If you had the original Fender transformer and applied 123.5vac, you'd get  (720v/115v) * 123.5v = 773v, or 386-0-386.
If you used Doug's transformer, you'd get (700v/120v) * 123.5v = 720.4v, or 360-0-360.

386v * 1.414 = 545v minus voltage drop in the GZ34.
360v * 1.414 = 509v minus voltage drop in the GZ34.

Either way, and regardless of whose transformer you used, 480vdc seems a likely result. And it's only 480v/460v = ~4% high.

You also are biasing at a lower current than your notes said you used before. Try idling hotter and see if the plate voltage comes down some.

Also, the GZ34 voltage at pins 2&8 doesn't tell anyone much that isn't already said by the output tube plate voltages. Instead, the a.c. volts at pins 4&6 would be more useful to know what the transformer is doing.

I would measure cathode volts on the various preamp stages.

Agreed, because just knowing the plate volts are somewhat high doesn't tell much. Knowing the cathode voltage might help.

... decided to play around with various switchable mods (normal channel disconnect, tremolo disconnect, cathode caps/resistors, NFB). ...

Tremolo disconnect won't apply to the Hoffman board because it doesn't use an optoisolator for tremolo like the original circuit.

Normal channel disconnect will be (in my opinion based on owning a number of Fender amps) underwhelming at best. I didn't find it to do much for gain or distortion that turning the volume knob a hair higher doesn't already do.

[osing]

Thanks guys for your input. I made some cathode voltage measurements:

V1a - 2.84V (I forgot to mention earlier that I at some point in the past had done an "British Mod" on the normal channel which I since mostly removed, but I left the 2.7k/0.68uF resistor/cap, so this voltage might be expected to be fairly high given this?)
V1b - 2.05V (this seems in line with the Super Reverb schematic which shows 2.1V)
V2a - 2.25V (higher than the 2.1V found in the Super Reverb schematic)
V2b - 2.23V (higher than the 2.1V found in the Super Reverb schematic)
V3a/b - 9.12V
V4a - 2.27V
V4b - 2.18V
V5a - 2.47V
V5b - 236V
V6a/b - 105V

I am a bit puzzled by the V2a and V2b cathode voltages above, but I just had a thought. I am using 2.7k and 3.3k resistors in parallel to achieve a ~1.5k equivalent resistance when in stock mode (I have this as a switchable mod right now to go from ~1.5k to 2.7k). If the resistors are a bit off their target values (i.e. at the 5% extreme of their tolerances), perhaps that could be causing the slightly higher cathode voltage?

I noticed that the old Super Reverb AB763 schematic in the link from HotBluePlates actually has 270V at V1a, V1b, V2a and V2b. Further, it shows 410V before the 100k plate resistors, so there appears to be a 140V drop consistently across each 100k plate resistor for V1a, V1b, V2a, and V2b. My voltage measures 431V before the plate resistors, and as you can see below I am not getting the same consistent voltage drop at V1b, V2a and V2b (I am assuming that V1a is higher due to the 2.7k cathode resistor, but should the others not be equal, or is this caused by the cathode voltages being different between V1b and V2a/V2b?).

Another observation - in the Fender schematic, V1a and V2a both have independent 1.5k resistors, whereas V1b and V2b share an 820 Ohm resistor. Since 820 Ohm is about 10% more than half (750 Ohm) of 1.5k, why would Fender be getting the exact same 270V at each plate of V1a, V1b, V2a, and V2b?

To HotBluePlates' point, maybe my power tube plate voltages are reasonable at ~20V more than the original AB763 Super Reverb, and thus perhaps the 431V I am measuring before the preamp plate resistors is reasonable at ~20V higher than the 410V found in the Fender schematic. Does this then mean that my 295V V1b plate voltage is reasonable? Why would V2a and V2b be lower at 271V & 273V? Is it because of the cathode voltages being different, or does that not make sense? Shouldn't a higher cathode voltage lead to a higher plate voltage (as is the case with my V1a)?

The Hoffman AB763 as well as the original AB763 Super Reverb call for the GZ34 rectifier tube, if I am not mistaken. I am glad to hear that ~480V on the plates is not unreasonable, but I am confused as to why it would be higher now than when I originally did the build. When I was biasing it today, I did not notice a huge difference in plate voltage with a higher bias current, but I can take another look tomorrow. I will also measure AC voltage at pins 4 & 6 of the rectifier.

I suspect you are right about the Normal disconnect mod. It seems to basically allow me to do the same as if just turning up the volume knob, but I'll play around with it some more to see if there might be some subtle differences. I did find a post from Sluckey that had some info on how to do a tremolo disconnect on a Hoffman AB 763. Basically, you just cut the circuit after the 47k resistor prior to V5. It might end up having the same effect as just increasing the volume, though. I need to listen some more first. Both mods were definitely noticeable (easy to tell as I have them wired to switches), but again, it might not be any different then just turning up the volume.

Eleventeen - I did not change any resistors between power supply nodes, but I did disconnect and reconnect the filter caps as I removed the entire turret board when drilling some holes in the chassis, so I could have potentially damaged the adjoining resistors with excessive heat upon reconnecting. What would the impact of a damaged resistor be - lower resistance, perhaps?

Thanks!

[osing]

I just measured pins 4&6 of the GZ34 rectifier tube and got 760VAC. I guess it is safe to conclude that I should expect to see the high plate voltages at the power tubes since the voltage coming in is fairly high.

I also measured the cathode resistor values (without lifting any legs, which I realize is not always good practice for getting reliable results) and got:

V1a - 3k (per my second post, I have a 2.7k resistor here rather than the typical 1.5k)
V1b - 1.58k (standard 1.5k resistor here)
V2a - 1.52k (this is a 2.7k/3.3k parallel resistor pair)
V2b - 1.52k (this is a 2.7k/3.3k parallel resistor pair)

I don't believe this explains why I am seeing higher voltage drops across the cathode resistors of V2a (2.25V) and V2b (2.23V) compared to V1b (2.05V) and why the plate voltage at V1b is a good amount higher at 295V compared to the 271V/273V at V2a/V2b. Any insights would be appreciated!

[sluckey]

I don't believe this explains why I am seeing higher voltage drops across the cathode resistors of V2a (2.25V) and V2b (2.23V) compared to V1b (2.05V) and why the plate voltage at V1b is a good amount higher at 295V compared to the 271V/273V at V2a/V2b. Any insights would be appreciated!

Those voltages are only slightly higher. No need to be alarmed. It's all because of the 760VAC input to the rectifier.

[osing]

Thanks Sluckey, that is good to know then. I tried swapping V1 and V2, and the V1 plate voltages went down a bit and the V2 plate voltages went up quite a bit to around 290V or so. It looks like the tubes might be the main reasons for the differences among the voltage readings on the V1b and V2a/b plates.

Theoretically speaking, if I increased the 4.7k resistor located immediately before the filter cap closest to the preamp, I would be able to lower the preamp plate voltages, right? What would be the expected result from a sound perspective?

[HotBluePlates]

V1a - 2.84V (I forgot to mention earlier that I at some point in the past had done an "British Mod" on the normal channel which I since mostly removed, but I left the 2.7k/0.68uF resistor/cap, so this voltage might be expected to be fairly high given this?)

If you've changed the cathode resistor value, all bets are off for cathode and plate voltages.

When you made V1a's resistor bigger than the stock 1.5kΩ, the same amount of tube current would cause a bigger voltage drop across the cathode resistor. The bigger drop (2.84v vs. 2.1v) is a large bias voltage, which tends to tun the tube more-off and reduce tube current. These are contradictory, but in the end the tube settles at a lower current than with the smaller cathode resistor.

Less current through the plate load resistor yields less voltage drop across the plate load, and less reduction of the plate voltage from the supply voltage. So plate volts goes up. This accounts for the rise in measured plate voltage, along with the fact voltages are high throughout because of increased wall voltage.

...
V1b - 2.05V (this seems in line with the Super Reverb schematic which shows 2.1V)
V2a - 2.25V (higher than the 2.1V found in the Super Reverb schematic)
V2b - 2.23V (higher than the 2.1V found in the Super Reverb schematic)
...

You measured voltages are -2.4%, +7% and +6.2% compared to the schematic values and are smaller than the tolerances of the resistors themselves and the tolerance note on the schematic itself.

... I am not getting the same consistent voltage drop at V1b, V2a and V2b (I am assuming that V1a is higher due to the 2.7k cathode resistor, but should the others not be equal, or is this caused by the cathode voltages being different between V1b and V2a/V2b?). ...

Yep, with different cathode volts (because of different cathode R's) you have different tube currents, so different voltage drops across the plate load resistors.

... Another observation - in the Fender schematic, V1a and V2a both have independent 1.5k resistors, whereas V1b and V2b share an 820 Ohm resistor. Since 820 Ohm is about 10% more than half (750 Ohm) of 1.5k, why would Fender be getting the exact same 270V at each plate of V1a, V1b, V2a, and V2b? ...

(1640Ω - 1500Ω)/1500Ω = 9%    Fender didn't worry about the slight difference, as it's within the marked tolerance of the schematic (and the 10% resistors themselves).

Novice amp tinkerers are stress about tight tolerancing, when experience amp repairmen know 20% changes aren't that big a deal. When they look for 2v, and see 0.5v or 0v or 45v or 150v, then they know there is a problem to be fixed. The tubes themselves will likely have tolerances greater than 20%.

[HotBluePlates]

... I am confused as to why it would be higher now than when I originally did the build. ...

Your meter will have some range of tolerance (good meter, maybe 1-2% or maybe higher). I already explained 480v is only 4% high of 460v. I find it unlikely that you had a measured 460v before unless your wall voltage was low, as no modern amp build seems to hit the voltages of the vintage amps they copy (almost universally high).

And your wall voltage may have (and very likely has) varied. I've measured wall voltage as high as 127-128v at times. The few times I've measured, I rarely see 120v even or anything below that.

... I did find a post from Sluckey that had some info on how to do a tremolo disconnect on a Hoffman AB 763. ...

What exactly are you hoping to accomplish with this mod?

The blackface Fender tremolo in the Super Reverb used an optoisolator to ground/unground the signal to get trem. Even with trem off, there was some signal loss through the circuit. Some folks mod these amps by disconnecting the trem so there's no associated signal loss.

The Hoffman AB763 does not use an optoisolator; the trem signal varies the bias of the output tubes. There is no associated signal loss, and nothing to be gained by disconnecting it. You will lose your trem just like with the other circuit.

Don't mod just to mod. You should have a goal to be gained by performing the mod.

Most mods I've seen are about more distortion, but I'll tell you from my experience they don't add much (if anything), do not have the impact that simply adding a pedal has, nor add as much gain as having another triode tube or using a different circuit. Most of these mods (at least the ones you've mentioned) ask you to eliminate a feature to gain the effect of the mod (kill your Normal channel, kill the tremolo).

If the amp were mine, I'd just let it be a blackface Fender instead of trying to make it into something it's not. You can build a different gain machine if that's what you need.

Anyway, I think you're worrying too much about this amp. Just play and enjoy!

[sluckey]

The Hoffman AB763 does not use an optoisolator; the trem signal varies the bias of the output tubes. There is no associated signal loss, and nothing to be gained by disconnecting it. You will lose your trem just like with the other circuit.

That's true. But Doug put a 47KΩ resistor in the circuit where the 50K VIB INTENSITY pot was. That 47KΩ loads the signal path down just like the INT pot does on an original AB763. If you put that 47K on a switch and remove it from the circuit you will get a very significant increase in signal level. Switching it does nothing to the tremolo on the Hoffman circuit.

I think switching that resistor is what osing is referring to. Here's the schematic. The 47K is shown on page 1 and is on the output of V4-B, just before the 220K mixing resistors.

     http://home.comcast.net/~seluckey/amps/misc/Hoffman_AB763.pdf

[osing]

The trem disconnect mod Sluckey references is indeed one of the mods I implemented, and it has a noticeable effect. I am doing these mods as a learning experience. The web is flooded with information (and probably misinformation) about how much better a certain mod sounds than the stock design, and I figured I might as well try some out and learn something in the process. I suspect a lot of people implement a mod and then because they want it to sound great, they convince themselves that it made a huge difference, even though they are comparing the new sound to what they remember the old sound being. In my case, I am making all the mods switchable so I can tell for myself in a real time A/B test how they sound. My amp is a gutted Reissue Super Reverb with little value and I am using it as an ongoing project amp as I try to learn as much as possible.

At the end of the day, I might end up concluding that most of these mods, even though they have noticeable effects when switched on, sound about the same as if I had just turned up the volume knob on the original design 

Thanks for all of your input!

[HotBluePlates]

... But Doug put a 47KΩ resistor in the circuit where the 50K VIB INTENSITY pot was. That 47KΩ loads the signal path down just like the INT pot does on an original AB763. If you put that 47K on a switch and remove it from the circuit you will get a very significant increase in signal level. ...

I see. Sorry that I overlooked that on the Hoffman board.