Twin Reverb troublesooting

[Eric_Allard]

Hi,
I'm hoping you guys can help me troubleshoot a few issues with the Fender Twin Reverb I'm restoring. I have very little experience with restoring tube amps and specifically started this project to learn.
A little background information: The amp appears to have been built late in 1969 based on the features and date codes on the components. It was previously modified, poorly, to AB763 specs. The amp was very noisy and picked up every radio station in town.
I have since replaced all of the original electrolytic capacitors, several badly drifted resistors, the coupling caps (a few of them were leaking DC), and the incorrect components used in the original blackface conversion. Also, I have added 1 ohm resistors from pin 8 of the power tubes to ground to aid in biasing the amp. I used 22mfd and 80mfd F&T capacitors in place of the original 20mfd and 70 mfd ones. Bypass caps are Sprague Atoms 25mfd/25v. Bias supply cap is a Sprague Atom 50mfd/50v. Coupling caps are Mallory M150's for which I identified the outside foil and oriented it correctly. Additionally, the power switch and normal channel bright switch were broken and needed replacement.
I put in the new power tubes yesterday and have the following problems:
Biasing: With the bias pot turned fully clockwise (looking from inside the chassis), I am measuring approximately 425V on pin 3, and approximately 20 mA on pin 8, giving me approximately 8W static plate dissipation per tube. What could be the problem? I have double checked that the resistors on the bias pot and the bias supply board are the correct values (27k and 470, respectively). What are my next troubleshooting steps? I have noticed that my multimeter appears to be inaccurate with measurements of small values. For example, the 1 ohm resistors I added to pin 8 of the tubes consistently measure 1.6 ohms, whereas they are 1% tolerance precision resistors. It is highly unlikely they are all incorrect. I had previously assumed that the error should be consistent between readings and cancel out to yield the correct PD value, but now I am reconsidering that assumption.
The second problem appears to be a low frequency oscillation, which, I believe is what is commonly called "motor boating". When hitting a chord, especially at the low end of the neck, as the chord is dying out, i hear a distinct low-frequency oscillation (approximately 0.5-1 Hz). It sounds like a very low speed tremolo. What are the likely candidates here? The lead dress in this amp is terrible compared to my older Bassman. Is that the likely culprit? The amp appeared stable before, but I may have disturbed the lead dress when I was cleaning the chassis (there was a thick coat of cigarette smoke and dust inside the chassis). I haven't yet replaced the preamp tube.
I will post pictures this evening when I am back at home.
Thanks in advance for your help.

[cbass]

I would clean the tube sockets and jacks and pots first then try swapping tubes.
You did replace the elctrolytics in the bias circuit?
What voltage are you getting on pin 5 on your power tubes?

The motorboating is likely due to lead dress or a bad tube.
It might help to post some pics

[Eric_Allard]

Thanks for the reply cbass
I cleaned the pots with DeOxit Fader lube and that eliminated all of the scratchiness.
For the tube sockets, I cleaned them with denatured alcohol and pipe cleaners. I probably should get a can of DeOxit and try again.
I have to double-check the voltage on pin 5, but if I recall correctly, it was around -35V.
I'll post pics of the lead dress tonight. It seems especially bad around the phase inverter and tremolo driver tubes, where several wires are several inches longer than required and repeatedly criss-cross one another. I suspect that may be the problem, but thought it best to seek the advice of a few people with more experience before I start randomly changing the lead dress.

[Milkmansound]

How's the 470R in the bias supply? Drifting high?

You can also change the 27k off the bias pot to a higher or lower value to alter the bias range. 22k might be a good start for you to drop it.

You should get a better meter first though - you are down in milliamps so you don't want too much fluff in the measurements.

Motor boating may simply be an artifact of too much preamp gain - might want to double check that the tubes are in the right sockets.

[Eric_Allard]

I think the amp was originally AA769. It originally had a 3.3k resistor on the bias board and a 15k on the bias pot. Tubes were red-plating when I first got the amp so I replaced these with the 470 and 27k values found on the AB763 schematic. I confirmed their values were correct when I installed them.

Any recommendations for a decent meter? I'm using my friend's "Canadian Tire" brand one, but I don't think it's particularly good (or even decent). Fluke brand gets recommended quite a lot, but I'm not sure I need to go to a $150 MM for as much as I'll be using it. On the other hand, I don't mind paying what it costs to get decent tools if that means less frustration later on.

Thanks

[Milkmansound]

Even a $99 Radio Shack one is pretty good

It's just a re-branded something but it'll get the job done

[Eric_Allard]

I noticed some higher-end multimeters can measure current directly. Does that mean that I could avoid having to add that resistor on pin 8 of the power tubes to ground to take measurements for setting the bias? I found it to be a real PITA to solder the resistor to that braided wire with all the heater wires in the way.

[Teleguy61]

Don't forget that when you are measuring bias CURRENT over 1 ohm resistors, you want to use the MILLIVOLT scale to get the correct reading.
You are not measuring current flow, you are measuring voltage drop, and mathematically converting that measurement to milliamps.
Example: 40mv/1 ohm=40ma.

[cbass]

I noticed some higher-end multimeters can measure current directly. Does that mean that I could avoid having to add that resistor on pin 8 of the power tubes to ground to take measurements for setting the bias? I found it to be a real PITA to solder the resistor to that braided wire with all the heater wires in the way.

I think those meters only measure AC current Im using a $40 dollar sears MM It works fine measures capacitance too

[martin manning]

I noticed some higher-end multimeters can measure current directly. Does that mean that I could avoid having to add that resistor on pin 8 of the power tubes to ground to take measurements for setting the bias? I found it to be a real PITA to solder the resistor to that braided wire with all the heater wires in the way.

Yes, by using the somewhat-advanced OT primary shunt technique, where you put the meter probes across each side of the OT, from plate lead to CT lead, shunting virtually all of the current through the meter. It's nice to have a "hold" feature for this since you don't want to be looking away from what you are doing with the probes. If you slip you can easily short the high voltage to ground or something else since the probes are essentially connected together by a wire in this mode.

Another way to get idle bias current is to first measure the DC resistance of each half of the OT primary with power off, measure voltage with power on (which is much safer than the shunt technique), and then calculate current using Ohm's law.

[Eric_Allard]

Thank you all for your replies.
As promised, here are some photos of the chassis:
https://plus.google.com/photos/11359400 ... 8097595905
I apologize for the quality; they were taken with my Nexus 4, which has a rather poor camera.
I suspect the lead dress near the phase inverter and tremolo driver tubes is responsible for the oscillation, but would also like to know if you guys can spot any other areas of concern.
I also noticed that the heater wiring is not very tightly twisted as I often see on some amps, especially near the pilot light. Is that a concern?
I also picked up a can of deoxit on my way home after work. I'll try cleaning and retensioning the tube sockets tonight.
Regarding the accuracy of the multimeter I'm using, one of my friends pointed out that there's likely 0.6 ohms in the test leads alone. I tested it by setting the meter to measure resistance and touched the probes together and sure enough, there is 0.6 ohms on there. What does this mean for my readings? Is it a concern? Can I compensate for it or should getting a better multimeter still be top my priority? Is it possible to "zero" the meter?
Thanks again for the help.

[Eric_Allard]

Yes, by using the somewhat-advanced OT primary shunt technique, where you put the meter probes across each side of the OT, from plate lead to CT lead, shunting virtually all of the current through the meter. It's nice to have a "hold" feature for this since you don't want to be looking away from what you are doing with the probes. If you slip you can easily short the high voltage to ground or something else since the probes are essentially connected together by a wire in this mode.

Another way to get idle bias current is to first measure the DC resistance of each half of the OT primary with power off, measure voltage with power on (which is much safer than the shunt technique), and then calculate current using Ohm's law.

Thanks for the reply. I read about these methods, but also read that they're not recommended for a novice, nor for people without a good multimeter.

[cbass]

opps guess iwas wrong I wouldnt recomend the shunt method anyway
measuring voltage drop across the primaries is a much safer method IMO
Once you measure the primaries resistance you can calculate what voltage drop you want to see and just adjust the bias pot till you get the correct voltage drop

1ohm resistors are the safest way to measure but not the most accurate but close enough

[martin manning]

Don't worry about the accuracy of the meter at 1 ohm. Even a top-shelf DMM will not measure accurately in that range. In measuring bias current you are measuring the voltage across that 1 ohm, and the lead resistance becomes inconsequential.

Here's a test you can do: Connect a 4k7 resistor in series with a 1ohm and put that assembly across a 9V battery. Measure the voltage across the battery and you should get about 9.2V. Then measure the voltage across the 1ohm resistor and you should see something very close to 2mV, indicating current of 2mA. That should agree with battery voltage divided by 4k7 (9.2V/4700R = 1.96mA).

[Eric_Allard]

I got some work done last night and this morning:

Cleaned the tube sockets and jacks with DeOxit.

Cleaned up the lead dress around the phase inverter and tremolo driver tubes. I used the AB763 layout as a reference.

I listened to that oscillation more closely so I could hopefully describe it in more detail. It sounds kind of like a buzzy, distorted lower harmonic of the chord root that fades in and out slowly. I think it's always there, but it becomes far more noticeable when a chord is sustaining out. I'll try replacing the preamp tubes and report back.

I got quite a few voltage measurements done so I could compare if they're in the same ballpark as the original schematic. I found that many voltages are down 5-10% compared with the published values. I didn't expect it to match perfectly, but I expected that the discrepancy would be in the opposite direction. I assume that these values were calculated assuming a line voltage of 110 or 115 volts, whereas I measured 122 volts at the outlet this morning. Here is a summary of a few measurements (M=measured, P=published):
Transformer secondary (red wires): P=340VAC, M=320VAC
Bias supply (red-blue wire): P=48VAC, M=53VAC
Bias pot (turned fully clockwise): P=-52V, M=-46.8V
Power tube pin 4: P=458V, M=421V
Power tube pin 3: P=460V, M=424V
Phase inverter pin 6: P=235V, M=219V
Phase inverter pin 1: P= 245V, M= 227V

Does this help identify what might be wrong? Could the resistors under the cap pan affect these readings? I measured them and they appeared to be within spec, but maybe I should check again. It doesn't indicate a problem with the power transformer, does it? It certainly looks like it was exposed to some pretty intense heat at some point.

Bias measurements are as follows (with the pot fully clockwise):
Pin 3: 424V, Pin 8: 21.3 mA, which yields a static plate dissipation of about 9 watts per tube. Also, I confirmed the pot is functioning properly. I read about 37k resistance with the pot fully clockwise and 27k with it fully counter-clockwise. Given the above measurements, should I change the resistor to 15k or 22k to get the current up or is there another underlying problem?

There are some updated pictures that can be seen at the same link I shared earlier.