Fender Bassman 70 buzz

[JAelec]

I have said this before here, but Uncle Doug is not a reliable source of information. Some times he is correct, but other times he is very wrong. It is a bit disheartening that he has become such a widely attributed source. (ex. https://music-electronics-forum.com/...544#post951544)
He did go back and note his error in the video description: "My comments regarding "hot bias = lots of headroom" are wrong. Hot bias = Early Breakup. Cold bias = Lots of Headroom"

I stand corrected, along with Uncle Doug... Too bad...I kind of liked his teaching style but I guess he's not always correct. Good to know.

He's simply wrong. Doesn't even explain where the borderlines come from and why they would change.
In reality the borderlines are given by supply voltage which at large output doesn't depend on bias.

Actually the increase in "headroom" with cold bias is rather small, maybe 20% (or 1.6dB).
As a player I mainly notice a lack of "headroom" with amps not providing enough power/SPL for the band context (drummer, room size, open air, etc.)
But power output doesn't change with bias.

Nevertheless a little hotter bias can sound better with some amps/tubes due to increased compression and other involved effects.

As far as bias effecting headroom, it's B.S as has been stated. You need to understand what tube bias accomplishes. Bias sets up the idling condition of the tubes and decides how much signal is needed to get the tubes into a certain area of conduction. If tubes are biased at cutoff, it will take more signal to get them conducting than it will if they are already biased on. If biased too low, you will get crossover distortion because tubes are off near the crossover point. If biased too hot, tube life is unnecessarily shortened. None of this has anything to do with maximum power output or "clean headroom". As Helmholtz stated, it is primarily supply voltage that decides max output or "headroom".

So if I were to sum up what you are saying here:

- a cold bias can only increase clean headroom a tiny bit (20% or 1.6dB)
- headroom is mainly set by the supply voltage. I'm guessing that higher supply voltage = more headroom
- a hotter bias can increase compression and other effects which may or may not be desirable
- but the main purpose of biasing is to avoid the extremes of cutoff distortion on the cold end and lower tube life on the hot end. After that to set the area of conduction to taste.

The Bassman 70 is often referred to as sounding cold or sterile. Perhaps that is why the bias has been altered in this amp to the hotter end.

As far as the buzzing, all adjustments should be made once the amp is fully warmed up. The matching may not be valid as the tubes warm up, but should get there fairly soon. If there is horrible buzzing while the tubes warm up, there is probably some other fault. Maybe a slightly leaky coupling cap from the PI could do that.

Would that be the 0.1uF between the plates if the PI and the 1.5k grid resistors?

Are there any other caps (coupling or other) that could cause this kind of buzz?

I read something about the bassman 70 having "Shunt capacitors" to adress oscillations in the amp caused by sloppy wiring...
https://www.tdpri.com/threads/ul-bassman-mods.644789/

Also I watched a video on some of these 70s amps conducting DC through the wax on the fibreboard causing noise...

The buzzing could also be a defective filter cap that starts working as it slowly charges. Try temporarily tacking a cap across each filter cap to see if one might be causing the hum.

I'll try this once again and see if I effects the buzz during warmup.

[JAelec]

The buzz during tube warmup has gone down in volume level and is now at an acceptable level though I'm not sure why... which bugs me.

I tightened the socket pins and scraped a bit inside them. This did nothing.
I also swapped between different combinations of tubes from my quad matched set to see if there was a winning combination. No big difference.

I was not sure how to check for leakage on the coupling caps (as suggested by g1).
I measured the plate voltages of the PI (Pin1: 268V, pin6: 262V) and the other side of the coupling caps connected to the PI plates ( -42V and -42V), to see if there was DC leakage. But the other side has the -42V from the negative bias supply so not sure how to tell if there is leakage from the plate side?

The multimeter test probes between ground and the coupling caps (grid side) made the noise increase a lot while measuring. Why is this?

If anyone knows the proper way to test for leakage here I would appreciate your thoughts!

The coupling caps from the PI plates are new and have been already been replaced by someone. So I wouldn't expect them to leak.

I also strapped 100uF across the filter caps one at a time... I checked many times: restarting the amp with or without the extra 100uF.
At some point I noticed the noise was not as loud as before, no matter if the 100uF was there or not.

So not sure what has happened here and if the positive change in buzz level is reliable.

[g1]

The easiest way to check the coupling caps is to disconnect one end of them. Then see if the buzzing has changed at all while they are disconnected.
As the -42V seems to be good at all the grids, it's not likely that either of them are leaking bad, but I'm wondering if one might be a little wonky during charge up.

[Helmholtz]

I might have missed it, but did you check/replace the bias filter caps?

[JAelec]

I might have missed it, but did you check/replace the bias filter caps?

I have neither checked nor replaced them, though they have been replaced by a previous technician.

As far as I understand the bias supply is half wave rectified, so ripple should be at 50hz not 100hz so that shouldn’t be my problem…But I’ll test them for sure.

Can I test them the same way as the filter caps, only remembering the reverse orientation of the negative bias caps and be sure to have the same orientation on the paralleled cap, correct?

[Helmholtz]

As far as I understand the bias supply is half wave rectified, so ripple should be at 50hz not 100hz so that shouldn’t be my problem…But I’ll test them for sure.

So you scoped and measured the frequency of your "buzz"?
If so, please post a scope pic.

Can I test them the same way as the filter caps, only remembering the reverse orientation of the negative bias caps and be sure to have the same orientation on the paralleled cap, correct?

Yes.

[JAelec]

So you scoped and measured the frequency of your "buzz"?
If so, please post a scope pic.

I did at in the beginnings of this thread… but I havn’t scoped specifically this shorter buzz but I might as well do that to be on the safe side. Though I’m quite confident I can hear the difference between 50 and 100 hz by now.

[Helmholtz]

Though I’m quite confident I can hear the difference between 50 and 100 hz by now.

That's not so easy to tell, as the 50Hz bias ripple is distorted and contains 100Hz as second harmonic.
Naturally the guitar speaker emphasizes the higher frequency.
I've often seen people being wrong about hum frequency.

[JAelec]

Here's the scope image of when the noise is at its loudest...

5ms/div
10mV/div
= 100hz

Paralleling a 100uF cap to each of the negative bias caps one at a time does nothing, so they seem to not be the issue.

Desoldering one leg of each of the coupling caps does nothing either.

I feel the amp has been long enough on the bench... and the noise is at an acceptable level since it is only during startup... so I've put the amp back in the chassis.

Before putting it back I replaced the 22k resistor on the tube matching pot with a 33k and re-checked the bias:

V5:
Resistance: 44.4 ohms
Vdrop: 1.11 V
Plate voltage: 456.3 V

Plate current: I = V / R = 1.11/44.4 = 0,025 A
Plate dissipation: 0,025*456.3 = 11,4 W
% of max dissipation = 0,38 = 38%

V6:
Resistance: 53.3 ohms
Vdrop: 1.37 V
Plate voltage: 465.1V

Plate current: 0,026 A
Plate dissipation: 0,026*465.1 = 12.0 W
% of max dissipation = 0,40 = 40%


So now the bias is more conservative and tubes will hopefully last longer.

[Helmholtz]

No scope pic showing.

Of course your buzz during startup could be caused by power tubes having different heat-up times. I.e. if we're speaking of seconds.

[JAelec]

No scope pic showing.

Of course your buzz during startup could be caused by power tubes having different heat-up times. I.e. if we're speaking of seconds.

Yes perhaps it is different heat up times as you say... but it is sounds about the same even when switching between different tubes.

Another possible cause could perhaps be the wiring? The internals don't look super tidy and since it has been modified with the several extra fuses, perhaps some wire somewhere has been moved to close to a HT wire...

Strange about the scope pic... it is showing in my browser, but I'll post it again:

[Helmholtz]

With 5ms/div the scope pics shows a fundamental frequency of 100Hz.
But the additional ~ 600Hz oscillation is not normal.

[JAelec]

With 5ms/div the scope pics shows a fundamental frequency of 100Hz.
But the additional ~ 600Hz oscillation is not normal.

Thanks, yes there seems to be another tone there... It is still there after the 100hz has died away. However it is not that noticeable so I will have to leave it that way for now and perhaps in the future do a more complete overhaul of the amp. Perhaps a full recap and possibly some of the mods proposed in this thread: https://www.tdpri.com/threads/ul-bassman-mods.644789/ that should deal with some of the design flaws in this amp including oscillations.

Thanks for your kind assistance and answers to my questions especially Helmholtz and g1, plus other voices who joined the discussion.