I never thought this would happen to me! (PT failure)

[Steve A.]

Can you explain the 27v zener and 470R cathode resistor for the pair of 6V6's?

Thanks!

Steve A.

[Chuck H]

It "fixes" the bias once the cathode voltage rises to 27V with current and prevents further cooling of the bias. I didn't invent it and it was actually suggested to me a long time ago by Ken Gilbert (SN=kg). But I sort of championed it's use with the Paul Ruby mod in a sort of calculated pairing of the two circuits for 2xel84 amps that are prone to buzzing with crossover distortion. Some here and on other forums have taken to calling the Zener across the cathode resistor the "Chuck H" mod anyway. But alas, it can't be found on the Blue Guitar site

[dstrat]

Chuck I have never tried that , how would you describe the sound if different?

[Chuck H]

Well, looked at strictly as a circuit it can be done with different results. You could, for example, measure your cathode voltage and then choose a Zener value at that voltage. This would effectively make the amp fixed bias with the operating point for the power tubes floating at the cathode voltage instead of ground and I would expect it to behave like a fixed bias amp.

The way I do it is to scope the amp to the point of obvious clipping. Not just compressed tops, but visible flats starting to show. I measure the voltage on the cathode at that point and that would be my Zener value. The result is that the amp behaves like a cathode biased amp for clean tones and tightens up for distortion tones rather than continuing to get colder in bias and lending more compression and crossover distortion.

The typical instructions for the Paul Ruby mod (this is the mod where Zeners are placed across the grid leaks) from his site use to suggest measuring cathode voltage and making THAT the Zener voltage. The problem, as I saw it, is with AB amps the cathode voltage rises significantly with current. That should mean that the diode clipping is then in the signal. I didn't want this so I start by choosing the cathode Zener value as described and then choose a PR mod Zener voltage 1 volt higher than the cathode Zener. That way the tubes will always be in cutoff before the PR Zeners clip, effectively keeping the diode clipping out of the signal.

[dstrat]

thanks for that!..

[mikepukmel]

I suspect that this would NOT have happened without crushing the input the way I did. And I don't think I intended the amp to be driven like that. But hey, musicians... what can you do?
But of course 'what I can do' is to understand the design requirements for a robust amp. Over-rate the PT by 30%?

Oh man, when we were kids and got these amps cheap used, we'd crank them up to 9. And this one guy had another amp (some odd name solid state amp, I still remember the name "Mike Matthews Freedom Amp), he'd out the OUTPUT of the freedom amp into the old princeton so he could get that little added extra 'zip'. I do remember he made a few amps go 'pouf'.

[mikepukmel]

Two days ago I was taking this last winter's project out for a test run at full volume, as the missus was not at home. This is a 2x 6V6 amp, kind of a mashup of a BF princeton or Tweed Deluxe power section, with switchable NFB. Had a dirt box in front, driving the amp into heavy grind territory. When suddenly there was a drop in volume for a few seconds, then the mains fuse blew (1.0A time delay). A sniff check seemed OK, and shooting the tubes with an IR thermometer revealed about 200F on the plates in the few minutes that had elapsed since the event.
I had convinced myself that a screen collapse and g2-cathode fault could have drawn enough current to blow the primary fuse. Examination of the circuit board did not show any charred components, and what I tested - screen resistors, cathode resistor, rectifier diodes, power supply caps (testing no short to ground), power supply node resistors - all OK. Power up, all tubes out, on a light bulb limiter, lamp shone brightly. Rats.
measuring PT secondary winding resistance - first in circuit and then out - revealed 43R red-red/yel (CT) on one side, and 163R on the other. Oddly enough, red-red measures about 120R, less than the one side to CT. The nominal winding resistance is about 170R each side.

Hammond 270CX PT spec sheet here

The PT is rated for 550v @ 75ma = 41.25VA
2 x6V6 may pull 12W each diss plus 18W (yeah, I'm optimistic!) audio power for 42W
Seems like the right PT for the application

Is this kind of failure attributable to specific causes? or is this a simple infant mortality? I have contacted Hammond with a nice email, asking what their tech dept thinks. I'm more concerned with spec'ing the right PT for this amp than anything else. Comments?

edit: a quick check of the output tubes (LBL, then a floating guitar cable "noise" test - nice and buzzy, even on the LBL) in a champ suggests no problems with the tubes. Also the OT resistance check seems OK at about 200R each side.

Hammond 1760H OT spec sheet here

So I'm thinking the HV winding short was the problem, not a symptom. Reality check?

Oh boy can't click "like" on this one. About how long were you running the amp loud before the PT went? I haven't worked up the guts to crank deluxe "amp build 0" yet, similar to your setup (at least same output tube set) My brother played it a bit, ran it into clipping, at 8 on vol or so, no overdrive pedals, but only a couple of minutes. its got a 138ma rated PT, very similar to yours, hammond 291bx.

[eschertron]

Most 6v6-gt / gta tubes seem to have max signal plate of 92ma for 2 tubes (ab1) and only give 285vdc for plate and screens in the datasheets.
the screens at max signal is 13.5ma.
what it seems to me 105.5ma at max signal. I am not sure how that would differ for cathode bias?
classictone has a 50s deluxe (5e3) at 100ma and the BF, SF deluxe at 120ma.
knew I had seen that but had to look again.

So where are you getting your current consumption numbers from? I need to get a hold of that...

[dstrat]

they can be found online Tube Data Sheet Locator
or just google 6v6-gt datasheet,or what ever tube.
some have charts some do not. but general class a1 , ab1 , ect is usually there.
jj has some also, have not seen any for the Russian reissues.

note: I use tube datasheets as general data, some guitar amps do vastly exceed max values. ( I know,..go figure )
but they give an idea of what to expect.

[g1]

So where are you getting your current consumption numbers from? I need to get a hold of that...

I think he is getting them from the tube data sheets, like attached. Suggest getting as many as you can for each tube type, as they often give different examples as far as plate voltages used, etc. If you look at enough, you will probably find an example that is somewhat close to your circuit voltages/values. (or at least you can guess-strapolate )


edit: what he said ^^^^

[dstrat]

woops spoke too soon.... Russian Tungsol :|

Tung-Sol.com:: 6V6GT Vacuum Tube page

oh boy!

[eschertron]

OK. Curious. The two spec sheets agree that the "maximum signal plate current" is 92mA, for whatever B+ (285vdc is mentioned in one document). I wouldn't have taken it as a guideline, but on re-reading it I put it together now as
"Radiotron 6V6-GT maximum ratings are design center values Push-pull AB1 max sig plate current 92mA".
I had previously thought it was a limiting value, so specing a PT that could provide less than that number would be a good idea.

[eschertron]

OK, since there's a consensus about plate draw at max, I'm gonna assume that the question of how much (or where to find how much) has been answered. Appears the Old-timey engineers actually put good info on their data sheets. Who knew?

Can I assume that the failure was thermal? I don't see a way that HV spikes at the plates could make their way back to the PT, could there?

[pdf64]

OK. Curious. The two spec sheets agree that the "maximum signal plate current" is 92mA, for whatever B+ (285vdc is mentioned in one document). I wouldn't have taken it as a guideline, but on re-reading it I put it together now as
"Radiotron 6V6-GT maximum ratings are design center values Push-pull AB1 max sig plate current 92mA".
I had previously thought it was a limiting value, so specing a PT that could provide less than that number would be a good idea.

It's the designer's job to ensure that the circuit's design is such that the tube's limiting values are not exceeded; the tube won't somehow self limit.
Unless some sort of current limiting circuit was included (RG's MOSFET HT regulator) the max current that the power amp can draw will mainly be limited by the loaded VHT / OT primary impedance / 4.
So even with that info sheet, if the HT is higher (as is likely with your amp) more current will be drawn.
And the PT will supply as much current as the circuit draws, albeit with a slightly lower winding voltage.

In regard of circuit limits on the peak HT current that can be drawn, your amp is somewhat less restricted than a tweed deluxe, eg silicon rather than 5Y3 rectifier, fixed rather than cathode bias, 1k rather than 5k g2 HT dropper (plate current will be lowered as Vg2 is reduced).

Hence the potential for your amp's PT to become overloaded by excessive HT current demand, whilst it may be have been ok with a circuit that had the same idle conditions but a saggier HT and cathode bias.

[eschertron]

It's the designer's job to ensure that the circuit's design is such that the tube's limiting values are not exceeded; the tube won't somehow self limit.
Unless some sort of current limiting circuit was included (RG's MOSFET HT regulator) the max current that the power amp can draw will mainly be limited by the loaded VHT / OT primary impedance / 4.
So even with that info sheet, if the HT is higher (as is likely with your amp) more current will be drawn.
And the PT will supply as much current as the circuit draws, albeit with a slightly lower winding voltage.

From above, current draw in the 6V6's could approach 330vdc/6.6k/4 = 200mA? Yikes!

In regard of circuit limits on the peak HT current that can be drawn, your amp is somewhat less restricted than a tweed deluxe, eg silicon rather than 5Y3 rectifier, fixed rather than cathode bias, 1k rather than 5k g2 HT dropper (plate current will be lowered as Vg2 is reduced).

I sometimes fail to consider that once the zener voltage is exceeded, the current through the cathode resistor is only a fraction of total cathode current.

Hence the potential for your amp's PT to become overloaded by excessive HT current demand, whilst it may be have been ok with a circuit that had the same idle conditions but a saggier HT and cathode bias.

The original design had a HUGE screen resistor and a cap to ground (4.7k/10W and 8uF, I think), which gave me the kind of degeneration that i wanted with an SE version of the circuit. I didn't like the way it sounded here, so I removed it and went with individual 100R stops. 470R or 1k stops are probably in this amp's future, then.