AB763 WITH KT66 OT

[ViperDoc]

While waiting for the brown truck, I discovered an AB763 amp that was built for KT66 output. It sounded sensational from what I could tell. My rudimentary assumptions lead me to believe that, aside from a properly built AB763 preamp board, one would need a push-pull OT with an 8k primary, a proper PT/choke, power supply and bias circuit. Sounds simple!

How would this not be the ultimate gigging amp with a normal-lead channel mod? I've only been able to find one out in the open like this. Maybe it's not as good as I think.

[Roe]

8k isn't a great idea. 4k is better. but these draw more heater current than 6l6s. chinese kt66s don't draw as much as the russian and british kt66s though.

[pdf64]

JTM45 type OTs can be modelled after the RS DeLuxe (6k5@16ohms or 5k8@8ohms) or the Drake 784-103 (8k, well 7k8).
I wonder how Fender came up with a 4k primary for p-p 6L6 type power tubes, which has become the norm for guitar amps? 5 - 6k tends to provide a bit more power output, and 5k6 was the suggested value for high power AB1 conditions in the 6L6GC info (not published until after Fender had used 4k).
https://tubedata.altanatubes.com.br/she ... /6L6GC.pdf

[Stevem]

Without question one of the biggest changes in, first off play feel and secondly tone is how light or hard in a class A or A/B1 amp the PI section drives the output tubes!

Also A big thing is to note is that preamp tubes that test out at under there types current output design spec can show a 25% or more slower reaction to the audio signal applied to them, as this can have huge implications if two preamp tubes are that way!

[Roe]

JTM45 type OTs can be modelled after the RS DeLuxe (6k5@16ohms or 5k8@8ohms) o...
I wonder how Fender came up with a 4k primary for p-p 6L6 type power tubes, which has become the norm for guitar amps? 5 - 6k tends to provide a bit more power output, and 5k6 was the suggested value for high power AB1 conditions in the 6L6GC info

Interesting about the 5k6, with 450v plate voltage (at full power - not idle - presumably). any idea what would yield max power with el34s and kt66s (w/o ultralinear)?

the RS dlx seems to be 6k6 (or 8k) at 4/8/16 ohms or?

[martin manning]

JTM45 type OTs can be modelled after the RS DeLuxe (6k5@16ohms or 5k8@8ohms) or the Drake 784-103 (8k, well 7k8).
I wonder how Fender came up with a 4k primary for p-p 6L6 type power tubes, which has become the norm for guitar amps? 5 - 6k tends to provide a bit more power output, and 5k6 was the suggested value for high power AB1 conditions in the 6L6GC info (not published until after Fender had used 4k).
https://tubedata.altanatubes.com.br/she ... /6L6GC.pdf

Fender's choice of 4k goes back at least to 5881 in the Bassman, and maybe to 6L6G in older amps. Taking the Bassman example, where the plate voltage is 450 (per the schematic), and the screen voltage is nearly the same, a 4k load line passes through the knee of the Vg1=0 curve for 5881, 6L6G, GB, or GC, and the power output is maximum. Those tubes are basically the same except for their dissipation ratings, and of course that applies to low and mid frequencies where the speaker load is nominal. In the 6L6GC data sheet linked above, the example shown for 450V plate voltage has 400V screen voltage, and a 5k6 load might be a bit better for that combination.

I don't know why Marshall chose higher primary impedance for the JTM45 (8k was early-on, IIRC). To me, it looks like a mistake that leads to over dissipation of the screen grids.

Interesting about the 5k6, with 450v plate voltage (at full power - not idle - presumably). any idea what would yield max power with el34s and kt66s (w/o ultralinear)? the RS dlx seems to be 6k6 (or 8k) at 4/8/16 ohms or?

4k would be a good load for KT66 or EL34 at 450V Va and ~450V Vg2. At 25W Pd max, they will be dissipation limited if the peak load line dissipation is to be <=200% of Pd max. That load line is also near maximum output power. With a bit of voltage sag in both the plate and screen, it will be very close.

[ViperDoc]

I appreciate this feedback.

Can you explain WHY you would choose i.e. 450V plate voltage with 400V vs. 450V screen voltage in terms of tone and power? How does the OT primary R change that? This is the stuff I want to learn more about. Soldering by the numbers only gets you (me) so far!

[Colossal]

Martin,

A look at the GEC KT66 datasheet calls out a variety of push-pull applications for the KT66, all of which show 7 or 8k for RL. Were The Ancients mistaken in this?

[ViperDoc]

https://frank.pocnet.net/sheets/084/k/KT66_GEC.pdf

[pdf64]

All published typical operating conditions are either triode or UL, both of which take a heap of stress off the g2, compared to regular pentode operation.
And the higher load impedances tend to be more linear, with the KT66 being aimed at the top end audio market.

[martin manning]

A look at the GEC KT66 datasheet calls out a variety of push-pull applications for the KT66, all of which show 7 or 8k for RL. Were The Ancients mistaken in this?

Here's a clip from the MOV KT66 data sheet from '56, pentode with cathode bias. It shows 8k Raa, but note that the screen voltage is well below the plate.

[ViperDoc]

I'm having a difficult time weeding through peanut gallery arguments about Plate voltage vs. Screen voltage. NOT ON THIS FORUM, OF COURSE, NO!

Two things:

1) Can I take the "pentode operation" on a KT66 datasheet to simply mean "not triode", being as it is a KINKLESS TETRODE?

2) It seems that when both plate voltage and screen voltage are high, you get more output power. When screen voltage is low relative to plate voltage, or when ripple is on the screens, you get sag. I could be wrong. Is there any application for putting a variable dual-ganged screen resistor between B+2 and the screens to dial that in? Perhaps that would be more investigative for finding the desired fixed value.



I think I would not be wanting for power in a dual KT66 build, I'd be more interested in nice touch sensitivity and smooth break-up with sag on the lower end. Thinking about that in terms of voltage is enigmatic. Perplexing, even!

[martin manning]

1) Can I take the "pentode operation" on a KT66 datasheet to simply mean "not triode", being as it is a KINKLESS TETRODE? Or more importantly, that we're talking about Class AB operation? I suppose you could run triode mode in Class AB as well.

The key is that there is a functioning screen grid, the voltage on which can be varied to expand or compress the grid curves. If you tie the screen to the plate, that function is eliminated and it operates as a triode. A KT is a beam power tube, just like 6V6, or 6L6. They have five electrodes, so they are pentodes, but the beam forming plates take the place of the wire suppressor grid seen in "true" pentodes like EL84 and EL34..

2) It seems that when both plate voltage and screen voltage are high, you get more output power....

I suggest a read here: http://www.valvewizard.co.uk/pp.html

[ViperDoc]

Thanks Martin!

[pdf64]

Without question one of the biggest changes in, first off play feel and secondly tone is how light or hard in a class A or A/B1 amp the PI section drives the output tubes!...

I don't know what the above actually means, eg what criteria are used to quantify the lightness or hardness of PI drive

...A big thing is to note is that preamp tubes that test out at under there types current output design spec can show a 25% or more slower reaction to the audio signal applied to them, as this can have huge implications if two preamp tubes are that way!

Great, I'll be able to use weak tubes to transpose stuff down then, so I'll finally be able to sing 'stairway' without changing guitar
Not sure if it's meant to be ironic, if not where's this stuff's from? Again, I can't make sense out of it

...Interesting about the 5k6, with 450v plate voltage (at full power - not idle - presumably). any idea what would yield max power with el34s and kt66s (w/o ultralinear)?
the RS dlx seems to be 6k6 (or 8k) at 4/8/16 ohms or?

Yes, I assume that the GE 6L6GC conditions are with a stabilised lab type power supply.
A few years back, I incremented the secondary load resistance value of a JTM45 type amp, noted down the max rms Vac at the onset of clipping and calculated the reflected primary loading. It was a pretty tedious process, which yielded pretty much the same results for 6L6GC, KT66 and EL34, in that they all exhibited max power output when loaded at about 5 - 6k.
Bear in mind the unusual screen grid supply for the JTM45, ie a choke dropper to the g2 HT node, then a 1k shared followed by 470ohm individual current limiting resistors. So saggier g2 than typical.

The Philips and Mullard EL34 info shows various typical conditions, though I'm not sure how applicable the class B conditions in the Philips document are for our application. I assume that Marshall saw the Mullard 400V Vb fixed bias conditions (I assume Vb means the HT supply voltage ), and then ramped up the HT voltage over the years but without adjusting the load impedance. I think that a JTM45/50 will typically sag its HT down to about 400V or less at full load, especially with EL34 and the 3k4 load.
https://tubedata.altanatubes.com.br/she ... e/EL34.pdf
https://tubedata.altanatubes.com.br/she ... e/EL34.pdf

Regarding the RS De Luxe, see their info https://valveampstuff.blogspot.com/2020 ... utput.html
With respect to its '6k6' primary (1 - 9), my RS De Luxe's secondary ratios are 40.11 (A - F), 80.22 (C - B) and 80.22 (E - D).
Wired all in series for a '15' ohm secondary gives a total effective ratio of 20.15, hence a 16 ohm load reflects 6k5 back to the primary (a 15ohm load would reflect 6k1).
Wired for '7 / 8' ohms, the two 80.22 windings are in parallel, with that parallel pair being in series with the 40.11 winding. Hence the total effective ratio is 26.74, such that an 8ohm load reflects 5k7 to the primary.
Wired for '3' ohms, the two 80.22 windings are in series, with that series pair being arranged in parallel with the 40.11 winding. Hence the total effective ratio is 40.11, such that a 4ohm load reflects 6k4 to the primary, and the noted 3 ohm load only reflecting 4k8.
So there's a fair degree of variation on the primary loading, according to the particular arrangements on the secondary.
In regard of winding polarity, a positive going half wave at the '1' primary terminal results in positive going half waves at the 'A', 'C' and 'E' secondary terminals.

...Fender's choice of 4k goes back at least to 5881 in the Bassman, and maybe to 6L6G in older amps...

My impression is that that up to the 5Fx Tweeds, various loads / operating points were used by Fender for a 6L6 p-p pair. From the 5Fx series onwards, they seemed to move to the long term standard of 4k at 450V +/-about 30V, so I guess it's the 5F6a Bassman being referred to above
http://www.classictone.net/40-18088.html
https://www.mercurymagnetics.com/product/fttwin-o/
https://ampgarage.com/forum/viewtopic.p ... 38#p320938