kt66 screen voltage (jtm45/100)

[Roe]

What screen voltage would you recommend for kt66s?

The power transformer allow plate voltages as high as 570v at idle but I am aiming at 545v max at idle with new tubes. At max power, the plate voltage sags around 10%.

Z=4k (quad)
Individual screen resistors are 1k.
The screens overdissipates greatly with high screen voltages

[pdf64]

Is this the schematic? https://el34world.com/charts/Schematics ... remelo.gif

[Roe]

Is this the schematic? https://el34world.com/charts/Schematics ... remelo.gif

yes, but I am considering a dual rail supply to reduce screen voltage

[Helmholtz]

With the original power supply wiring the screen supply tracks the power tube supply and the preamp tubes track the screen supply.
With a separate screen supply at least some of this effect will be lost - resulting in less dynamic compression.

Apart from that, a significantly lower screen voltage will lower power tube gain and available output.
Are you using a 20H/690R choke?

A less invasive method to lower the screen voltage would be using a zener in series with the choke.

[_Steve]

When I built mine I seem to recall being advised everywhere to keep the screens (and plate) sub-500V for new production KT66s. My PT had two taps for 560 and 500V so i made the entire B+ switchable. I dont really hear much difference at 560V so I just run at 500V, but I know others that swear the higher voltage sounds better.

As per the poster above I would advise using zeners, but exclusively on the screens feed, to keep the preamp voltages unchanged. I've done this on later 1959s and it sounds great, but havent tried on the 45/100. It changes the balance to distort slightly later in the PI and earlier in the KT66s.

[Roe]

With the original power supply wiring the screen supply tracks the power tube supply and the preamp tubes track the screen supply.
With a separate screen supply at least some of this effect will be lost - resulting in less dynamic compression.

Apart from that, a significantly lower screen voltage will lower power tube gain and available output.
Are you using a 20H/690R choke?

A less invasive method to lower the screen voltage would be using a zener in series with the choke.

Yes, I am using the 20h/690R choke on one amp with around 525v on the plates.
On another with 510v I am using a 3h/110 ohms choke instead. Both amps are overdissipating the screens.

Also, I've built a 32ohms loadbox that can be added in parallel with the 16 ohms cab, for a 2k67 reflected primary

the power transformers of these amps have 220/230/240v primaries and two secondaries (roughly 405-365-0-365-405v).
I was considering using the 240v primary with a termistor and the 405v taps for the plate (aiming at 545vdc). For the screens I was considering the 365v tap into a CLC filter. To adjust the voltage, I was planning a small cap before the choke. I wanted to try 400+vdc on the screens

[Helmholtz]

Also, I've built a 32ohms loadbox that can be added in parallel with the 16 ohms cab, for a 2k67 reflected primary

What is the idea behind?
While this will lower screen dissipation, it will increase plate dissipation, DC supply current and HT winding current.

[martin manning]

I would consider a VVR feeding the screens, with its input voltage coming from the screen node (just after the choke). I convinced TAG member bcmatt to try it with a pair of KT88 and it worked fine. It will track the voltage sag and keep the screens at a more or less fixed fraction of the plate voltage. With your target B+ and 4k load the max dissipation is ok, just dial the screens down to ~60% of the plate voltage, say 325V.

[Roe]

Also, I've built a 32ohms loadbox that can be added in parallel with the 16 ohms cab, for a 2k67 reflected primary

What is the idea behind?
While this will lower screen dissipation, it will increase plate dissipation, DC supply current and HT winding current.

yes, the idea is to lower screen dissipation and experiment with the sound. around 500+v 2k67 Z seems ok for kt66s. the power transformer is doing fine. the same transformer was used on the early el34 amps, with a 1k75 Z. It the latter scenario it sags more due to increased current

[Helmholtz]

I forgot to mention that a lower Zaa also increases output power and OT primary and secondary currents.

[nworbetan]

I would consider a VVR feeding the screens, with its input voltage coming from the screen node (just after the choke). I convinced TAG member bcmatt to try it with a pair of KT88 and it worked fine. It will track the voltage sag and keep the screens at a more or less fixed fraction of the plate voltage. With your target B+ and 4k load the max dissipation is ok, just dial the screens down to ~60% of the plate voltage, say 325V.

I found your post https://ampgarage.com/forum/viewtopic.p ... 14#p440814 by searching, but using a vvr to reduce screen voltage seems like such a widely useful idea that it might deserve its own thread?

[Roe]

I would consider a VVR feeding the screens, with its input voltage coming from the screen node (just after the choke). I convinced TAG member bcmatt to try it with a pair of KT88 and it worked fine. It will track the voltage sag and keep the screens at a more or less fixed fraction of the plate voltage. With your target B+ and 4k load the max dissipation is ok, just dial the screens down to ~60% of the plate voltage, say 325V.

would 400v be too high with 1k screen resistors and 520ish volts on the anode with 4k Z for a quad of kt66s?

[Roe]

BTW my two amps have never blown a tube. But a friend's 45/100 has blown chinese kt66s and russian kt88s (it has around 545v at idle, 4k Z, 110ohms choke, 1k individual screen resistors)

[martin manning]

Here is the difference in Vg1=0 knee location. 325V looks better, but 400V is for sure much safer than running the screens at Va. As you reduce the screen voltage the power stage will have more headroom. Larger screen resistors help by dropping screen voltage dynamically and throttling screen current.

[Helmholtz]

PDmax is an average limit.
Exceeding PDmax in class B operation by up to a factor of 2 isn't a problem, because the plate can cool down in the other half-cycle.
Problem with a high plate impedance and high screen voltage is screen dissipation.