Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: omalikular on September 17, 2025, 11:28:25 pm
-
I purchased a '65 Fender Twin Reverb/AB763 recently as it has long been a dream amp of mine. Previously I opened up and saw it had been recapped at some point and important resistors converted to metal film and was satisfied. After a few months I decided time to roll new tubes and discovered that they previous owner had 6V6 tubes in the power section. I swapped them back to 6L6 and found that I was unable to bias them above 19ma plate current. My bias with the pot was from about -64V to -44V.
At first I figured perhaps it had been modified to better accommodate the 6V6 tubes, however I have been unable to determine why the current is so low. Resistors all measured within spec. No unexpected continuity errors. All values within the circuit appear to be matching to the schematic, including the bias feedback and range resistors. Transformers are correct model numbers. Diodes in rectifier and bias circuit look good. I asked the previous owner and he only ever biased for the -52V bias voltage.
Voltage from wall outlet was 121V AC, Two 8 ohm JBL 120 speakers connected as load
340v tap 349.5V AC
48v tap 50.27V AC
Setting Bias voltage for -52 before bias range resistor change
Bias Voltage -52.04V
B+1 471.5V
B+2 469.5V
Plate voltage 469.5V
Grid -51.0V
Plate current= 8ma
After changing bias range resistor on bias pot from 22k to 10k resistor aiming for 37ma
Bias negative= -36.02V
B+1 455.5V
B+2 455.3V
Plate Voltage 452.2V
Grid -35.2V
Plate current=37ma
All measurements are with 6L6 tubes in circuit. I used a Variac to drop to 117V at one point but it did not impact current enough to account for differences. Bias was measured using a bias one probe from Eurotubes. 6L6 tubes are a matched set.
I was able to get the plate current where I wanted with no red plating by changing the bias range resistor, but I'm wondering if this is an indicator of some other issue that I have failed to determine, and have only bandaged. I'm still new to tube amps and perhaps the voltage values above indicate something that I fail to understand?
Appreciate any help on this!
-
Have you ruled out the 6L6s themselves? New or used? What brand? If they are are very worn out it would account for it.
-
If they had 6V6s in there, they would’ve goosed around with the bias range I’d say. So good on you for changing it back.
I’m assuming all 6L6 tubes in the quad are now within the same tube-current ballpark?
-
Have you ruled out the 6L6s themselves? New or used? What brand? If they are are very worn out it would account for it.
All tubes are new production Tung-Sols and are brand new. They are a matched set and are pulling the same current.
If they had 6V6s in there, they would’ve goosed around with the bias range I’d say. So good on you for changing it back.
I’m assuming all 6L6 tubes in the quad are now within the same tube-current ballpark?
Exactly what I figured; that it was made to accommodate 6V6. However, it wasn't unusual for folks to just run 6V6 in these amps without modifying the circuit from what I've seen, and I have been unable to spot any changes. The part that confuddles me is that I had to move the bias out of spec to get current to not be so cold. -35V on the grid is far outside the -52V approximation these would typically have.
-
I did wonder potentially about maybe the OT being adjusted in some way. I measure the windings for about 38 ohms while doing the voltage drop method. The transformer is fender stamped 125A29A, but these have different load configurations. Is it possible that they swapped this for another to get a load mismatch to account for the 6V6? Would a mismatched load be enough to account for the differences?
-
I did wonder potentially about maybe the OT being adjusted in some way. I measure the windings for about 38 ohms while doing the voltage drop method. The transformer is fender stamped 125A29A, but these have different load configurations. Is it possible that they swapped this for another to get a load mismatch to account for the 6V6? Would a mismatched load be enough to account for the differences?
Cooling off bias is one way of bringing a load line back under the Pmax curve but it’s better to go for optimum load impedance. If it hasn’t got a standard fender twin OT in there, you should do an impedance test - simply measuring DC resistance with an Rmeter won’t tell you anything. You need to determine the Pr:Sec VAC ratio by supplying one of the windings with a known VAC and measuring the other one. (edit: VAC^2 = impedance)
-
What is your screen voltage measured at the socket, not B+2?
-
What is your screen voltage measured at the socket, not B+2?
Latest voltage measurements:
B+1 457V
B+2 454.1V
Screen 452.3V
Plate 453.5V
Grid -35.2V
-
My bias with the pot was from about -64V to -44V.
That's a very good bias voltage range for healthy 6L6s in a TR amp.
-
My bias with the pot was from about -64V to -44V.
That's a very good bias voltage range for healthy 6L6s in a TR amp.
Correct,however this range would not supply the 6L6 with enough current. I had to increase bias voltage to -34 to get to 70percent dissipation.
-
I think your math is off. You said you measured 38 ohms from CT to primary
I=V/R = 3.5V/38Ω = 92mA
P = IV = 0.092A*453.5V ≈ 42W
That's about 140% dissipation.
-
I think your math is off. You said you measured 38 ohms from CT to primary
I=V/R = 3.5V/38Ω = 92mA
P = IV = 0.092A*453.5V ≈ 42W
That's about 140% dissipation.
I'll double check after work, I know my last check before above measurements was bias probe current at 37. I did OT CT to standby, my measurement could be wrong.
-
I think your math is off. You said you measured 38 ohms from CT to primary
I=V/R = 3.5V/38Ω = 92mA
P = IV = 0.092A*453.5V ≈ 42W
That's about 140% dissipation.
I'll double check after work, I know my last check before above measurements was bias probe current at 37. I did OT CT to standby, my measurement could be wrong.
No idea where the original number came from... I'm getting 97 ohms.
3.5V / 97ohms= 36ma
.036A * 453.5V ≈ 16.8W
-
No idea where the original number came from... I'm getting 97 ohms.
3.5V / 97ohms= 36ma
.036A * 453.5V ≈ 16.8W
I did wonder potentially about maybe the OT being adjusted in some way. I measure the windings for about 38 ohms while doing the voltage drop method.
38 ohms seems high, 97 ohms seems nonsense.
Just to be clear, it's the resistance of the OT primary wires, red to blue, and red to brown, that are needed to derive anode current.
38 ohms seems more like the brown to blue resistance, but even then it's a bit higher than expected.
The relevant readings should be about half that, eg 19 ohms.
https://www.hammfg.com/files/parts/pdf/1750W.pdf?v=1697661946
Fitting 1R cathode current sensing resistors would avoid all this confusion.
The transformer is fender stamped 125A29A, but these have different load configurations. ...
What do you mean by 'different load configurations'?
-
38 ohms seems high, 97 ohms seems nonsense.
Just to be clear, it's the resistance of the OT primary wires, red to blue, and red to brown, that are needed to derive anode current.
38 ohms seems more like the brown to blue resistance, but even then it's a bit higher than expected.
The relevant readings should be about half that, eg 19 ohms.
https://www.hammfg.com/files/parts/pdf/1750W.pdf?v=1697661946
This is on me for my lack of experience, I've been using Rob Robinettes bias guide to determine the windings and not properly understanding the electronics. I had previously gotten 97 ohm measure red to standby. With the transformer still in circuit, I'm getting Red-Blue Red-Brown of 38 ohm.
Fitting 1R cathode current sensing resistors would avoid all this confusion.
The bias probe accomplishes this doesn't it? This is what I'm using to get the 37ma current. I'm just trying to determine why to get this reading I have to increase my negative bias voltage to -34V, out of the typical ~-52V.
What do you mean by 'different load configurations'?
As far as I understand, this transformer is available for 4, 8, 16 ohm loads in mind, looking at after market replacements at least. The only reason I considered that it may be configured for a different load is the fact it has been running 6V6 tubes for who knows how long.
-
This is on me for my lack of experience... With the transformer still in circuit, I'm getting Red-Blue Red-Brown of 38 ohm.
This means the math in my post is valid and you're over dissipating the tubes by biasing them way too hot.
-
This is on me for my lack of experience... With the transformer still in circuit, I'm getting Red-Blue Red-Brown of 38 ohm.
This means the math in my post is valid and you're over dissipating the tubes by biasing them way too hot.
I did consider that previously when I'd done the math, as I had gotten the results similar to you of being hot. But I relegated that to some failure on my part, since the cathode resistor insertion with a bias probe gives .037V, and this is how I've biased other amps. Also the tubes don't look like they are running too hot...
-
Based on the measurements you've provided, the simplest explanation is that you're unintentionally idling your tubes too hot.
If your voltage measurements are correct and complete, then I can only conclude the resistor in your bias probe is significantly off from 1Ω. Try measuring the resistance of the probe with nothing in it.
Failing that, I'd be tempted to do the output transformer shunt method to determine the plate current, but that has fallen out of favor in recent years. So proceed with that at your own risk.
-
For the few cents-worth of 1R1% cathode bias measuring resistors, you can avoid a lot of incorrect measurements
-
For the few cents-worth of 1R1% cathode bias measuring resistors, you can avoid a lot of incorrect measurements
This is the current plan, will update once I get these put in.
-
Just finished installing the 1R1 resistors, getting .04V reading :w2:
-
0.04V = 40mV
I suggest you put your V-meter on the 'mV' setting. Then you will get an even more accurate reading
What Plate voltage have you got?
-
Yeah... Just seems to disagree with bias being too hot. Not sure why I need -34V bias voltage to get this current, but good is good I guess?
-
What Plate voltage have you got?
^ ^ ^ ^ ^
And what is your screen dcv?
-
Screen is 450.5V
Plate is 452.2V
-
I stand by my earlier statement that there is a math/measurement error, and these tubes are being punished.
There should be agreement between the output transformer voltage drop method and the cathode current sense resistor with the cathode resistor giving a slightly higher current number. The difference here is almost 2x. Something else is happening.
-
Assuming you are running 6L6GC rated at 30W Pmax (as you didn't say whether you were running 6L6GC or 5881s)
You measured cathode voltage at 0.04V (If you take the cathode voltage measurement with your V-meter set to 'mV', we possibly could get slightly more accurate cathode current estimation.)
To get a plate current estimate for a 6L6GC from measuring the cathode current, you can assume 5mA screen current, and since cathode current = plate current + screen current, this leaves 35mA plate current.
452.2V x 0.035A = 15.8W
15.8W/30W = 53%
If your measurements are correct and if you are running 6L6GC, you could easily warm them up a bit
-
There should be agreement between the output transformer voltage drop method and the cathode current sense resistor with the cathode resistor giving a slightly higher current number.
Yes there should, assuming the OT voltage drop measurement and the OT primary resistance measurements have been made accurately - and that 2-plates'-worth-of-voltage-drop has been measured across each OT primary half
(However, given the OP has reported a voltage of 0.04V at one of the cathodes - albeit with a 1R 1% resistor - there is still some ambiguity about what the cathode current is. It could be 0.045V or 0.036V, with the '10s-of-volts' setting not able to provide enough decimal places for a decent readout).
-
High precision mode on multimeter says .039V
-
High precision mode on multimeter says .039V
Well then, that’s 39mA cathode current on that tube you measured. Is that the same on each tube? Or ?
The sum of cathode current for each pair on each side of the primary should correspond to the total current measured across each half of the OT primary plus the screen current for each pair, I.e., for each half of the OT primary;
2.Ik = 2(Ip + Ig2)
-
High precision mode on multimeter says .039V
Well then, that’s 39mA cathode current on that tube you measured. Is that the same on each tube? Or ?
The sum of cathode current for each pair on each side of the primary should correspond to the total current measured across each half of the OT primary plus the screen current for each pair, I.e., for each half of the OT primary;
2.Ik = 2(Ip + Ig2)
Then the math error is mine. I forgot this was a 4 x 6L6. That accounts for the 2x discrepancy.
I have no idea what is causing the bias voltage to be so far from the expected range then. Screen voltage lines up.
-
All the tubes are matched, I am getting .037V-.039V range across each one of the resistors individually
-
All the tubes are matched, I am getting .037V-.039V range across each one of the resistors individually
So assume a screen current of 5mA for each 6L6, and subtract that from the cathode current for each tube, and the plate current (for each tube) will be what’s left over.
Then do the plate voltage x plate current equation* for each tube to get each tube’s plate dissipation. A good number for a twin reverb is 60 to 70 percent of 30W. When you get that, button it up and call it done.
*Remember that the units in the equation are Volts and** Amps (so 39mA needs to change to 0.039A in the equation)
**But the equation is Volts x Amps = Watts ;-)
It doesn’t really matter what actual control grid voltage is required to get there - different batches of tubes may naturally run hotter or colder depending on production stats etc, and so the amount of bias voltage required to get to a happy place will vary somewhat between batches.
-
Appreciate all the well written responses. I do feel like I've got what I need for the bias, and yeah bias is adjustable because it the variance of tubes. Ultimately what had me end up here is I would expect the stock bias range to be sufficient for that variance, without needing to change the resistor on the leg of the bias pot. I had figured that this could be an indicator of something wrong, either by my calculations or my amp itself.
The 6V6 tubes also had me wondering for that reason. But if the voltages and currents are getting me dissipation I guess I just role with it :laugh:
-
Well the plot thickens... After deciding to roll with it, I noticed my V7 and V8 tubes are red plating... Measuring .037V and 453V plate on them.
-
Swapping tubes the red plating followed, but I do see very slight red plating in the dark on the others that weren't full on red plating noticibly
-
The 6V6 tubes also had me wondering for that reason.
You say above 6V6. ^ ^ ^ ^
Which do you have in the amp, 6V6 or 6L6? :dontknow:
Big difference in those 2 types of tubes.
-
The original post started with me rolling tubes and discovering it had 6V6 tubes. I swapped to 6L6 tubes, that is what I'm currently working with.
-
Well the plot thickens... After deciding to roll with it, I noticed my V7 and V8 tubes are red plating... Measuring .037V and 453V plate on them.
This could be because these two tubes either aren’t bonafide 6L6s, or they could be 6L6s that have already been run badly previously (or are old worn out ‘NOS’ tubes) and have deformed grids etc that mean that they should be thrown away.
If your measurements are correct, these tubes are running at 16.7W, which is cold for 6L6s, but too hot for 6V6s
Can I ask, what makes you think these are redplating? You don’t necessarily need to be in the dark to see red plating (but you might not see normal red heater filament glow in broad daylight).
-
Attached an image. Potato phone camera quality! These are new Tung Sol 6L6 STR. Bias probe is in but it red plates without it.
-
Okay that is red-plating.
If your voltage and current measurements are correct, these must be very poor quality tubes.
Have you got the same negative DC voltage on every 6L6 grid pin*? (pin 5)
*never mind what the RI schematic test point says, you need to measure the actual tube socket pin voltage (in case there is a fault somewhere on the PCB)
-
Okay that is red-plating.
If your voltage and current measurements are correct, these must be very poor quality tubes.
Have you got the same negative DC voltage on every 6L6 grid pin*? (pin 5)
*never mind what the RI schematic test point says, you need to measure the actual tube socket pin voltage (in case there is a fault somewhere on the PCB)
-34.89V - red plating
-34.89V - slight red plating
-34.92V - slight red plating
-34.91V - red plating
-
Okay, next thing to check is how fresh is the battery in your meter?
And confirm you have 1R 1%tolerance resistors installed between each 6L6 socket cathode pin and the ground return. (And not 1.1R or ‘1R1’ as that would make a 10% error in the calculation)
-
Hold your horses!
Are you making these measurements through your bias probe? Because that could be giving you a false indication if that has a 1R resistor installed between the cathode return wire and the ground if you have also installed 1R resistors at each tube socket! (Depending on how the probe is wired, you might be getting 1R||1R , which would be half an ohm - over which you are measuring 0.037V , which would be 74mA!
-
I do see what you are saying but no that is not the case :D I had removed the 1R resistor for this tube exactly for that reason. This probe does have the 1R resistor setup, I had realized that earlier. I did it as more of a sanity test, to confirm all my readings with the bias are accurate. I'm soldering the 1R back in on this tube. All four have the 1R.
-
When you use your R-meter to ohm-out the 1R on the tube socket, do you get 1 ohm (after the DC-resistance of your probes is deducted)?
-
I am wondering if this amp is oscillating ultrasonically and that is why the grid bias voltage is bizarrely low when the cathode current appears to read normal?
-
I am wondering if this amp is oscillating ultrasonically and that is why the grid bias voltage is bizarrely low when the cathode current appears to read normal?
Checking the bias with no tubes should answer that.
-
Pull the PI and check the bias, unless the oscillation is in the power amp.
-
When you use your R-meter to ohm-out the 1R on the tube socket, do you get 1 ohm (after the DC-resistance of your probes is deducted)?
.99-1.01 when subtracting the .15 internal resistance of the probe.
-
Pull the PI and check the bias, unless the oscillation is in the power amp.
To confirm: leave in the power tubes, pull V6 phase inverter tube, and check bias voltage or current passing through the cathode?
-
I have
.038V across 1R
-35.30V Screen
-36.20V Negative bias at the 220k junction
With PI pulled
-
I will say now that I'm in high precision mode I see a .03 fluctuation on these negative bias voltage measurement...is that normal or an indicator of an oscillation?
-
It's time to bite the bullet and put a fresh set of 6L6s in the amp.
-
I'd be concerned about repeating the experience with yet another new set of tubes. It seems something fishy is going on with the bias. I'm wondering whether it could be something dodgy to do with the PCB traces - given it's one of those modern reissue jobs, and the amp has obviously been tinkered and meddled with before (with parts being pulled and substituted). I've seen several re-issue fenders with cracked/bad PCB traces that have needing jumpering to get the things to work. Hairline cracks in traces can cause oscillation. Given the seeming insolvability of this, I'm thinking if it were my amp, whether it might be more satisfactory to completely gut the innards and install traditional eyelet board setup etc complete with new pots and jacks and so on. $1,000 bucks worth of parts and work but at least you could exert complete control over the quality.
-
I'd be concerned about repeating the experience with yet another new set of tubes. It seems something fishy is going on with the bias. I'm wondering whether it could be something dodgy to do with the PCB traces - given it's one of those modern reissue jobs, and the amp has obviously been tinkered and meddled with before (with parts being pulled and substituted). I've seen several re-issue fenders with cracked/bad PCB traces that have needing jumpering to get the things to work. Hairline cracks in traces can cause oscillation. Given the seeming insolvability of this, I'm thinking if it were my amp, whether it might be more satisfactory to completely gut the innards and install traditional eyelet board setup etc complete with new pots and jacks and so on. $1,000 bucks worth of parts and work but at least you could exert complete control over the quality.
Oh this is not a reissue, this is from '65
-
Were the bias capacitor(s) and bias diode updated?
What about setting the bias at 30ma per tube.
-
Were the bias capacitor(s) and bias diode updated?
What about setting the bias at 30ma per tube.
All the original capacitors in the amp were replaced with orange drops and modern electrolytics, though an unknown time ago. At 30ma at each tube it still red plates. I get in to the low 20s before it really stops red plating. Unsure of the diode, I did a test with the multimeter and it drops the expected voltage and drops .58V
-
Oh this is not a reissue, this is from '65
Sorry I misread the '65 Fender Twin Reverb in your first post - my bad.
Is there anything else we don't know about your amp? (Like - has anyone added g-k caps to the output put tube sockets? or other possible modifications that could explain the bias conundrum?) Can you upload some hi-res gut shots?
-
Traveling for work for a couple days. I grabbed some old pics from the listing, the only thing different is the caps on the normal channel were replaced and the 1 ohm resistors added.
-
More closeups
-
Have to tested the 2 x 0.1uF couplings caps going to the 6L6 grids for DC-leakage?
-
Have to tested the 2 x 0.1uF couplings caps going to the 6L6 grids for DC-leakage?
His grid voltage is the same as his bias supply voltage, so the coupling caps are fine. Anyway that wouldn't explain why his tubes are red plating with only ~30mA plate current.
-
Maybe something is wired incorrectly? How else can there be red plating with only 37mA cathode current? How can current get to the plate other than through one of the other electrodes (cathode, screen, suppressor or grid)?
-
haven't followed close so;
does this amp have flyback diodes on the plate?
has the tubes been verified good in another amp?
The bias was verified at the grid, was it verified at the the tube side of sockets?
guessing this is all being troubleshot at DC level...No AC signal, has the OT primary been ohmed for partial shorts?
-
Appears to be missing the artificial center tap as there is no filament center tap to ground. Confirm this with measuring the resistance of one leg to ground when powered off though. Some of these early 60s PTs have internally grounded filament CTs. Haven't seen it on a Twin Reverb though--but most of my experience with TRs is post '68...aka What my clients can afford.
The bias filter cap looks like it's several decades old. Its worth investigating. Put your meter in Vac and check if you have significant ripple riding on the bias voltage. You may have a 60Hz AC spike to low voltage that the DC measurements aren't picking up. (Grasping at straws with that one.)
The OT is definitely aftermarket, so that carries the risk of it being improper spec and installed incorrectly. You can try disconnecting the NFB at the speaker jack and rechecking bias to see if it was oscillating above audible frequencies which is causing the red plating. If this causes the bias to behave and the tubes to stop red plating, then reverse the secondaries and reconnect the NFB.
Finally, I always replace tubes that have a history of red plating. I would want a fresh set in there after you've fixed the problems, regardless. So perhaps have a set on order. This isn't as thoughtless a suggestion as it was even 10 years ago with the pricing being what it is now vs then. But in my experience, once they pop the fun don't stop--so a clean slate is warranted after the servicing is complete.
-
You can try disconnecting the NFB at the speaker jack and rechecking bias to see if it was oscillating above audible frequencies which is causing the red plating.
That would answer my earlier question thanks. UHF oscillation from an incorrectly wired NFB loop can cause thermal runaway in the output tubes