PT and tube limitations

[Randall]

I have my sights set on building a Garnet Herzog, which is basically a 5F1 Champ with a fixed resistor instead of a speaker. I would like to use a salvaged PT I have that puts out 342-0-342 unloaded, and 480 B+ with just the rectifier diodes and first cap connected. I'm worried this may be to much for the tubes, but on the other hand it is a tortured sounding device I am going for. Rob Robinette recommends on his Champ page anywhere from a 190-0-190 to 300-0-300 PT going as high as 430v for B+, my rig is already running about 20% higher than that. I really don't read tube data sheets very well, especially the JJ 6V6S I am looking at. I think it has some language barrier misprints, like commas for periods, and I think the Letter U for the letter V, very confusing. I know the JJ 6V6 is pretty robust, so I ask, I am tempted to try this PT and torture the JJ 6V6, especially since it is just an effect and likely won't see a lot of stage time relative to what an amp would.

Yay, or Nay?

[eschertron]

Use a tube recto, or sag resistor, or both... JJ6V6's are fine, cathode biased, at 360vdc B+ IME, not sure they'd handle 400vdc plus.
Or use any of the tricks to reduce B+ mentioned across these boards. As you've indicated, this isn't rocket science.

edit: looking at the JJ spec sheet, it says you're good to 500vdc B+. So go ahead and torture away!

[Chuck H]

+1 to what eschertron said. The caveat being that such tubes may not always be available for the amp.?. But even then we can refer back to his post and qualify the suggestion of a dropping resistor. I mean, the idea of the Herzog is to accomplish class A tube chaos and channel it. Ideally, in class A the current won't sag significantly through the tube so the dropping resistor dissipates watts without much consequence to the overall tone. Therefor, if you later find that additional dropping resistance is necessary to accommodate another tube the adjustment should be easy I've said this about el84's in the past, but it seems to apply here WRT the 6V6 in this design. Smoke 'em if you got 'em.

[Randall]

Could someone enlighten me about the dropping resistor? I have no objection to trying that.

[eschertron]

Could someone enlighten me about the dropping resistor? I have no objection to trying that.

A 5F1 amp has a 10k resistor between the plates and screens of the power tubes. This drops the voltage a little for the plates, as well as (the primary purpose) reduces ripple. You could put a resistor in front of the plate node, the resistor simply "eats" some power and the plates see a lower voltage. This allows you to torture a PSU resistor while en route to torturing the signal.

Alternately, build an entire pi filter in front of the plate node. This will actually benefit the sound by reducing ripple that doesn't get cancelled in a SE design. While I'm thinking out loud... Make the load resistor on the herzog device greater than the typical speaker impedance. This adds a double benefit of reducing total power consumption from the PSU (current-limiting the PA, I think?) thereby forcing the amp into clipping earlier. Both desirable traits for a sonic-madness-inducing effect unit.

[Chuck H]

Literally just a high wattage resistor (15 watts should do) in line with the B+ or the HV center tap if a full wave recto is used. Value will probably be between 100R and 1k depending on the voltage drop required. Remember it's going to get hot and mount it accordingly and away from electrolytic caps.

[mozz]

If you have a unused 5v winding, that can be used as a bucking winding to reduce the output HV. Also a small choke input power supply. I'm gonna say any dropping resistor should be about 25 watts, they get hot, you will want to dump at least 50v.

[pdf64]

..While I'm thinking out loud... Make the load resistor on the herzog device greater than the typical speaker impedance. This adds a double benefit of reducing total power consumption from the PSU (current-limiting the PA, I think?) thereby forcing the amp into clipping earlier. Both desirable traits for a sonic-madness-inducing effect unit.

I think that would work for an AB amp (which draws more current under high signal conditions), but not for class A, where the average HT draw shouldn't change hardly, from idle to full load.

[pdf64]

...you will want to dump at least 50v.

We don't know what the loaded HT is going to be yet, do we?
We've got PT and rectifier sag under load to factor in yet, which may be sufficient.

[mozz]

He can put a 4k-8k resistor across the setup he has and get an idea of how much B+ he will have. He is using diodes currently. Duncan PSU software will also put him very close. Just from experience i would use at least 3x power resistor wattage.

[Dave H]

My 'Champ' has a pi filter using 2 x 47u caps and a 1k1 resistor (2 x 2k2 3W in parallel was all I had). It drops 58V which is 3W in a 1k1 resistor. An added bonus is the amp is dead quiet (no hum)

[eschertron]

While I'm thinking out loud... Make the load resistor on the herzog device greater than the typical speaker impedance. This adds a double benefit of reducing total power consumption from the PSU (current-limiting the PA, I think?) thereby forcing the amp into clipping earlier. Both desirable traits for a sonic-madness-inducing effect unit.

I think that would work for an AB amp (which draws more current under high signal conditions), but not for class A, where the average HT draw shouldn't change hardly, from idle to full load.

Maybe I should have been more clear. This resistor is the substitute for the speaker load. Increasing OT secondary impedance will 'flatten' the transfer curve and get to clipping earlier. Also reduce total power output (relative to the ideal output impedance bridge) when massively driving the PA.

[Chuck H]

Another "benefit" to using a higher than rated load resistor value is that speakers don't have a flat impedance. In fact, most of the tonally preferred resistive attenuators employ a higher than rated load. The idea being that you want to average out the work load across the frequencies to something more like what an actual speaker might present. I'm just speculating about this though.

[Randall]

As it happens, this PT does have an unused 5v winding. I did not realize it could be used as a bucking transformer. Not really sure how to wire that up.

[pdf64]

Given the issue, why not use it for a tube rectifier?
I'm not sure how a bucker could be arranged on the bi-phase winding?