Thumping Tremolo

[tubeswell]

With a bit of spare time on my hands I'm twiddling around today with my 5G9 and with everything dimed (including tremolo) there is a horridable chopper-like thumping.

When the inst channel vol on max, speed on max, the depth can only go to about 2/3 before it starts making the helicopter-like knocking sound at max vol (with the tone on 1/2) The speaker (G12M) gets kind of a knocking sound and you can visibly see the cone bobbing back and forth.

With the depth on max the speed can only go to about 3/4 before you get the knocking chopper-like sound.

When the speed and depth on max, the inst vol can only go to about 3/4 before you get the knocking sound

Interestingly on both settings the vol pot is has a scratchy pulse when you turn it with the trem on. This doesn't occur with the trem switched off. I've tried swapping V3 (trem tube) with several 12AX7s but to no avail.

I have a 12AY7 in V1 and a 12AX7 in V2 and in the V3 (trem position)

The voltages are like this:

V1 (VA stage)

Inst channel Plate 174 Mic channel plate 172, Common cathode 2.67V

V2 (LTP)

Plates 199 and 203, Cathode 27.9V, Grid resistor/tail junction ~27V

V3
The Trem LFO triode's Plate voltage is 232 at idle. The (schematic says 270). The cathode for the same stage is 1.94V (compared to the schem's 1.7). I've tried swapping in several other 12AX7s and they didn't seem to fare much differently (- a volt or two on the plate)

The LFO CF cathode is 246 (the schem says 260). The CF plate is 378 (which is the same node as the screen supply node)

Cap and resistor values in the oscillator are mostly as per the 5G9 schematic. i.e.; .03uF off the plate going to a 2M5 RA pot tapered with a 120k resistor, then a .01uF with a 1M going to the LFO cathode and the footswitch, and then .01uF with a 1M to ground


The output tubes are JJ6V6s -

Plate is 377 (vs about 369-370 on the schematic)

Screen on the same tube is 377. Scr resistor is 3W 467R - voltage drop is .46V (therefore screen current is about 1mA)

Tube current is 27.5mA (taking away 1mA for screen) = 26.5mA Plate current - dissipation is 10W (71%)

OT is a 5k Pr Z

Is it the speaker (maybe?)? Is it the LFO tube? (don't think it can be), or is it something else??

[loudthud]

The tremolo LFO signal cancels in theory at the output transformer. That theory depends on matched power tubes. Tried other tubes? The current modulation of the output stage puts quite a strain on the power supply and some of the LFO signal gets onto the preamp's B+ and into the signal path. Bigger filter caps will reduce this as will a little colder bias of the power tubes. A stiffer rectifier might help. Bias tremolo requires some compromises, limiting the depth is one of those. Note the 1meg in series with the depth control.

[tubeswell]

Originally Posted by loudthud

The tremolo LFO signal cancels in theory at the output transformer.

Hi loudthud

Does that mean the more evenly matched each side of the OT primary is, the better it will be at cancelling out the LFO signal?

The OT in the amp has a slightly unbalanced lengths of winding either side of the CT;- one side of it measures about 47R, the other about 61R from memory - due to one side being wound on top of the other. This makes even matched tubes have differing idle current draws (by about couple of mA)

Originally Posted by loudthud

That theory depends on matched power tubes. Tried other tubes?

Yep tried 'Tung-sol' RIs and now the JJs. They both do the same thing.

Originally Posted by loudthud

The current modulation of the output stage puts quite a strain on the power supply and some of the LFO signal gets onto the preamp's B+ and into the signal path. Bigger filter caps will reduce this as will a little colder bias of the power tubes. A stiffer rectifier might help. Bias tremolo requires some compromises, limiting the depth is one of those. Note the 1meg in series with the depth control.

The reservoir is 2 x Sprague Atom 500V 20uF in parallel, and I didn't want to go higher because of the rating on the 5U4GB chart

Similarly there is one of the same type of cap at the screen node and one at the pre-amp node.

I have tried higher and lower bias. How low do you reckon I should try and go?

What other ways can you limit the depth besides the 1M series resistor?

Cheers

[txstrat]

Hi TW,
don't know if it helps you but my 6G3 also makes a chopper like sound when the tremolo is dimed. You can see the power tubes blue glow go up and down with the thumping (sort of a soft thumping as if turning the vol pot up and down very quickly). This seems normal to me, since the tremolo is actually changing the bias to the hotter side when engaged. Haven't tried this with the whole ampd dimed, though. Don't know for sure whether the speaker cone would visibly move, but I believe it would (I'm gonna try it if you want me to).

[tubeswell]

Yep the cone sure does move. Being a little 20Wer, I wondered how arduous it was for the G12M - supposedly they are rated to take a bit more, but I don't know about this sort of treatment.

Edit

I rebiased it to idle at about 22mA plate current. The whumping has decreased a bit. Incientally, I noticed with the trem on max intensity that the tube current did rise quite a bit.

[txstrat]

Well, actually, if you hit the strings really hard with a power chord or somethin', I guess it would be even harder for the speaker to eat that. I wouldn't expect the speaker to get damaged as long as it's fed by "normal" signal strength. (depending on the wattage of course)

[txstrat]

TW
One minute ago I tried the amp with all six knobs dimed (remember 6G3 Deluxe). The chopper effect is quite frightening
The effect is the same whether a guitar cable is plugged in the normal channel or not. With a cord in the bright channel the effect is even worse.
The tubes glow nice blue (more or less with the tremolo).
I think this is quite normal with a bias vary tremolo.

[loudthud]

The DC resistance of the OT shouldn't matter, it's the turns ratio that makes it cancel. Pull the PI tube to judge the OT's rejection (see last paragraph). Swapping tubes side to side may make a slight difference.

The amount getting to the preamp's B+ can be judged by how much LFO comes thru the volume pots. You can take this down with a bigger cap and/or resistor on last filtering node. Also, the .1 caps in the preamp and the .02 at the input to the PI can be reduced to limit LF response.

You may also notice that faster trem frequencies get thru a little more than the lower ones. A cap to ground between the 1Meg and the depth pot will knock the higher ones down a little. Experiment with the value, something between .01 to .1 or so.

Look at what the LFO signal sees where it hits the grid resistors of the power amp. There is a voltage divider made up of the 220Ks, the coupling caps, and the plate resistors of the PI (and the PI tube's plate resistance). Note the slight imbalance caused by the 82K verses 100K. To correct this imbalance, reduce the 220K on the 82K side to 200K or 180K. This is going to mess with the balance of the PI so you may not like what this does to the non-trem sound. Everything is going to interact so you may have to compromise the non-trem sound for the trem sound.

[tubeswell]

Thanks loudthud, I'll try some of your suggestions with coupling caps first. The LFO already has a .03uF and 2M5 RA (tapered with a 120k from wiper to ground) at the first filter node.

[tubeswell]

Well I have done something of a fix. I put in a switchable cathode bypass cap in V1 - 2.2uF, which can be paralleled with 22uF.

When it is on 2.2uF, the thumping disappears/ is a lot more reduced (and also the amp is not as hard on the little G12M20 on the bottom string).

[tubeswell]

Found the problem

I had the speaker on the wrong OT tap . (I was getting a reflected load of 10k instead of 5k). I swapped it to what it should be and the thumping was gone.