KT88 Class A push-pull

[pdf64]

So 54W class A output from a pair of 6550 is possible, and cathode bias has no downside for class A, it's just that each 6550 would need to idle at nearly 60W

[Colossal]

So 54W class A output from a pair of 6550 is possible, and cathode bias has no downside for class A, it's just that each 6550 would need to idle at nearly 60W

Hahaha, nice, Pete!

[nuke]

It's a guitar amp.

Any references to engineering terms, real or imagined, are purely conjecture and ad-speak.

"It's class-A"

"Wow, is it true class A?"

"yup, truly, the finest sounding amp ever, best in class, so that makes it an A!"

[Stevem]

Personally if your looking for the sweater sounding even order harmonics out of a class A PP stage then your barking up the wrong tree, because you want a SE class A amp in that case.

[Colossal]

It's a guitar amp.

Any references to engineering terms, real or imagined, are purely conjecture and ad-speak.

"It's class-A"

"Wow, is it true class A?"

"yup, truly, the finest sounding amp ever, best in class, so that makes it an A!"

Could their claim just be marketing hype? Possibly. That ridiculous myth about a Vox AC30 being Class A still persists. Komet offers the Vimana 25 which is a 2xEL34 which they say is class A and outputs 25W. That I believe, as I have my own Class A 2xEL34 push-pull amp and it is doing exactly that. So that's why my interest was piqued when I saw the 2xKT88 Vimana which said 54W.

[Colossal]

Personally if your looking for the sweater sounding even order harmonics out of a class A PP stage then your barking up the wrong tree, because you want a SE class A amp in that case.

Steve, my interest in Class A p/p was just to compare the difference in sound, if any, to AB. The Class A (push-pull) 2xEL34 build I've been tinkering with for a while definitely has an amazing presence and clarity. That might be attributed to its topology and output transformer, however. My design goal was touch sensitivity and extreme clarity.

[wpaulvogel]

I don’t think they can do it either. I built a KT88 single ended amp to test the data sheet claim of 19-20 watts. I built my own transformers. The output transformer impedance was something that couldn’t be bought. I don’t remember if it landed on 2.5k, 3k or 3.5k but I couldn’t buy what I figured in the load line so I wound it. The power transformer has two separate windings so I can send the screen a lower voltage than the plate with only a choke filter to the screen. I think that the plate got 420 volts and the screen got 270 volts. Pardon me, I don’t remember everything. I ended up having to use a KT120 because the necessary plate dissipation required to create a nice waveform took something like 49 watts idle dissipation. It’s fixed adjustable bias. Seems like I got 20 watts clean sine.

[Colossal]

I don’t think they can do it either. I built a KT88 single ended amp to test the data sheet claim of 19-20 watts. I built my own transformers. The output transformer impedance was something that couldn’t be bought. I don’t remember if it landed on 2.5k, 3k or 3.5k but I couldn’t buy what I figured in the load line so I wound it. The power transformer has two separate windings so I can send the screen a lower voltage than the plate with only a choke filter to the screen. I think that the plate got 420 volts and the screen got 270 volts. Pardon me, I don’t remember everything. I ended up having to use a KT120 because the necessary plate dissipation required to create a nice waveform took something like 49 watts idle dissipation. It’s fixed adjustable bias. Seems like I got 20 watts clean sine.

Paul,

I remember your build. That was a nice amp. I am wondering if Komet simply copied their copy from the Vimana 25 (class A 2xEL34) specifications and then pasted it for the Vimana (KT88) and changed the power output, but forgot to correct the operating class to AB? 54Wrms at AB would make sense. The Vimana is their production version of the limited Silver Cloud they put out a while back (~5 years ago, I think).

[Helmholtz]

That I believe, as I have my own Class A 2xEL34 push-pull amp and it is doing exactly that. So that's why my interest was piqued when I saw the 2xKT88 Vimana which said 54W.

Clean class A PP output is always less than the plate dissipation per tube at idle.
This follows from the general class A rule I posted earlier.
There's only 2 ways to get 25W clean (or more) from a pair of EL34s:

1) Exceed PDmax (25W) at idle or
2) Enter class (A)B.

You can't fool the laws of physics.

When a cathode biased amp exceeds the class A range, the cathode voltage noticeably increases (e.g. by 20%+ with a classic Vox AC30).
This can be used to detect class AB operation.

[Colossal]

That I believe, as I have my own Class A 2xEL34 push-pull amp and it is doing exactly that. So that's why my interest was piqued when I saw the 2xKT88 Vimana which said 54W.

Clean class A PP output is always less than the plate dissipation per tube at idle.
This follows from the general class A rule I posted earlier.
There's only 2 ways to get 25W clean (or more) from a pair of EL34s:

1) Exceed PDmax (25W) at idle or
2) Enter class (A)B.

You can't fool the laws of physics.

My power tubes are biased to 24.5W at idle. Close enough for government work.

[Helmholtz]

My power tubes are biased to 24.5W at idle. Close enough for government work.

So clean output must be less than 24.5W without exceeding class A.
How did you verify that the amp doesn't leave class A at large outpu?
See my edited post above.

[Colossal]

My power tubes are biased to 24.5W at idle. Close enough for government work.

So clean output must be less than 24.5W without exceeding class A.
How did you verify that the amp doesn't leave class A at large outpu?
See my edited post above.

My load line is 8k, plates at 343V, screens at 320V and cathode at 21.5V (79.6mA cathode current). No NFB. The amp was scoped with a 1kHz signal.

[Helmholtz]

My load line is 8k, plates at 343V, screens at 320V and cathode at 21.5V (79.6mA cathode current). No NFB. The amp was scoped with a 1kHz signal.

Are plate and screen voltage referenced to the cathode?

[Colossal]

Are plate and screen voltage referenced to the cathode (with cathode bias)?
Or do you use fixed bias?

Plates and screens referenced to ground. Both tubes are individually cathode biased. I checked my math and the amp is putting out about 22.8W, so yeah, less than 25W for the pair.