Loadlines, Plate Dissipation, and Commercial Amps

[El Jefe]

I'm trying to build a Class AB amp using a pair of KT-88s without melting them down. I would like to push them as far as they can go, but no farther.

My "worst case" is a square wave (representing a heavily clipped signal) at B+/2. I understand that the conventional wisdom is that at B+/2, the plate dissipation should be no higher than 200% of max rating.

But consider a commercial amp that uses KT-88s, the Hiwatt 405. The output transformer has winding for 2 secondaries, which are stacked (thus the "DR" in the model number, meaning "dual rail). One is at 288 VAC and the other at 178 VAC. Both feed full wave rectifiers per the schematic, so the lower rail should be at 407 VDC and the upper rail stacked on top of it should be at 659 VDC (the schematic claims 650 on the B+).

3 push-pull pairs feed the output transformer, with Za-a at 1k45. Multiply by 3, each pair sees 4k35. At cutoff, each tube sees 1/4 of that, or 1k0875. At 329.5 VDC (B+/2), the current should be 329.5/1087.5 = 303 mA. And that means plate dissipation is 329.5 X 0.303 = 99.8 W, which is 238% of the KT-88's rating of 42 W.

I understand that 200% is a "soft" limit, but somewhere beyond it is a hard limit. Does anyone know of other amps that run a loadline as far to the right as that Hiwatt without having a reputation for spontaneously combusting in the middle of a gig? My thoughts are that with good design and build practices, I should be able to go as far as the pros, but probably not any farther than that.

I am also aware that screen survival at full output is also a problem, but want to see what I can and cannot do with plate dissipation before tackling that.

[Stevem]

First off why do want to push them so hard, and second how many watts are you looking to durive out of them?
These Are not cheap tubes to ware out sooner then need be!
If you drive them in class AB2 you can yank far more then 42 watts out of each of them if need be!

[martin manning]

...plate dissipation is 329.5 X 0.303 = 99.8 W, which is 238% of the KT-88's rating of 42 W. I understand that 200% is a "soft" limit, but somewhere beyond it is a hard limit.

To get that condition, where you are doing max load line (if you assume a line) dissipation 50% of the time you need a continuous square wave of a particular amplitude, which won't occur often or for very long, and in addition there might be some B+ sag that will reduce the max load line dissipation.

Does anyone know of other amps that run a loadline as far to the right as that Hiwatt without having a reputation for spontaneously combusting in the middle of a gig? My thoughts are that with good design and build practices, I should be able to go as far as the pros, but probably not any farther than that.

As I recall some 100W Marshalls ran plate voltage and load impedance that implied excessive dissipation, at or above the 238% calculated for the Hiwatt, but thanks to significant voltage sag at high current draw, they were reasonably reliable. Cloners who spec'd transformers with better regulation had issues, though.

[El Jefe]

First off why do want to push them so hard, and second how many watts are you looking to durive out of them?

At this point, I'm just trying to understand what's possible and what isn't. I've seen circuits that pull 150 watts from a pair, but they aren't suitable for guitar use (Class B, and far too fragile).

[El Jefe]

As I recall some 100W Marshalls ran plate voltage and load impedance that implied excessive dissipation, at or above the 238% calculated for the Hiwatt, but thanks to significant voltage sag at high current draw, they were reasonably reliable. Cloners who spec'd transformers with better regulation had issues, though.

I think that's just the answer I was looking for. My interpretation?

0-200%: Absolutely safe.
200-220%: Safe for 99% of users.
220-240%: Kinda iffy, some blowups may happen.
>240%: Here there be dragons.

[martin manning]

But, as in the above example, the robustness of the power supply can be a significant factor.

[wpaulvogel]

“DR” in Hiwatt Amps stands for Dave Reeves. Not dual rail.

[BobSimpson]

+1 on Dave Reeves

[M Fowler]

All I learned from building 200w bass heads was to either switch to 6550's or lower the screen voltage for the KT88's because they could not handle the higher screen voltage.

I just custom ordered some Heyboer SVT PT and OT for a 300w bass build. That has the lower screen voltage taps needed but I'm going to run six 6550's so not really a test for KT88's.

Mark

[gingertube]

M Fowler,
Can you confirm the above. KT88 has higher screen voltage rating than 6550 and I had always ASSumed that switching from 6550 to KT88 was the safer option.
MY KT88 experience has only been with a HiWatt 400 and a pair of Jadis JA80 HiFi Monoblocks, no experience at all with 6550, although I have 3 matched quads of old Sovtek 6550C on my shelf looking for a project.

Datasheets show operating points for either Pentode or Ultralinear at B+ of 550V to give 100W per pair. Kevin O'Connor covered this in TUT3 (I think it was TUT3) and showed that you could get significantly more power that that without impacting tube reliability.

Cheers,
Ian

[jbrew73]

I’m trying to follow along and learn so forgive me if my question seems trivial. Where does the B+/2 come from in the equations above? I understand most of the other numbers but this one has me lost.

[martin manning]

Running screen voltage below plate voltage is about optimizing the position of the grid curves with respect to the load line for a given tube type. I think typically you will see Vs at anywhere from 1/2 to 2/3 of Va for the big beam power tubes.

[M Fowler]

The only problem I have with data sheets vs real time is those KT88's of various brands bought by me and the owners did not fair well and they were not exceeding any voltage or current limits. These players were previously using old Ampeg SVT which they spent so much money on tubes and repairs that going with less power but locally built amp was a better choice.

No builder error found I decided to go with 6550's and the amp has not been back since and is gigged every weekend, death metal stuff high volume low bass melt your ears music.

I would follow the advice of Martin and others they know their stuff, I just go for it and see what happens. I still think one should have the windings for HT and Screen separate or create a lower screen voltage of around 350vdc. I did use a filament transformer to handle the preamp tubes and the PT handled the power tube filaments.

[wpaulvogel]

I also believe that is the reason for the harshness people find when using 6550 in an amp designed for EL34. The 6550 wants to pass so much current with the screen voltage so high that the tube operates so far below the knee on the diode line that it sounds “solid state “. I’ve had no experience with the 6550 but want to build an amp with them and I was going to set the screen voltage at about 2/3 of plate voltage just to see how it performed. That’s what I find while working out the load lines.