Plexi KT88?

[martin manning]

Is the screen voltage up near 400V too? If so, then a 2k Zpri 100W OT would be a better match, or you could bring the screen voltage down to 300 or so. As it is now, the screens will be dissipating a lot of power at high volume.

[jazbo8]

Is the screen voltage up near 400V too? If so, then a 2k Zpri 100W OT would be a better match, or you could bring the screen voltage down to 300 or so. As it is now, the screens will be dissipating a lot of power at high volume.

Agreed, 2k would be a better match, but in practice the screen grid dissipation isn't that bad even with 3.3k, since most modern KT88's have relatively high saturation voltages - much higher than what the old datasheets might suggest, so the screen current stays low even during high volume, also the screen grid voltage sags a bit with the screen resistor (assuming the typeical 1k is fitted)...

[martin manning]

Is the screen voltage up near 400V too? If so, then a 2k Zpri 100W OT would be a better match, or you could bring the screen voltage down to 300 or so. As it is now, the screens will be dissipating a lot of power at high volume.

Agreed, 2k would be a better match, but in practice the screen grid dissipation isn't that bad even with 3.3k, since most modern KT88's have relatively high saturation voltages - much higher than what the old datasheets might suggest, so the screen current stays low even during high volume, also the screen grid voltage sags a bit with the screen resistor (assuming the typeical 1k is fitted)...

This trace matches the parameters on the Genalex KT88 data sheet for 200V Vg2 (1974). Note that this modern EH KT88 has higher screen current.

[jazbo8]

This trace matches the parameters on the Genalex KT88 data sheet for 200V Vg2 (1974). Note that this modern EH KT88 has higher screen current.

It would be great if you can run the trace at higher Eg2, say 300V to see if the diode line slope leans more to the right, anyway, using the figure provided, it appears that the diode line has a slope of ~200R, so if you extend it to the proposed operating voltages, the saturation voltage would still be relatively high (I am guessing quite a bit above 100V), so the load line intersects the Eg0 curve well away from the high current portion of Ig2 curve.

[martin manning]

I can try to get some data at higher Vg2. I don't think that slope will change though, it doesn't on the old data sheets. Regardless, look at the g2 current shown in the 200V Vg2 trace above: at 25V Va-k it's already 110 mA, where the Genalex plot shows only 70. The g2 current of the EH is much higher than the old data sheet on the flat part of the Ig2 curve at high Va-k too.

[jazbo8]

I can try to get some data at higher Vg2. I don't think that slope will change though, it doesn't on the old data sheets. Regardless, look at the g2 current shown in the 200V Vg2 trace above: at 25V Va-k it's already 110 mA, where the Genalex plot shows only 70. The g2 current of the EH is much higher than the old data sheet on the flat part of the Ig2 curve at high Va-k too.

No worries, don't waste your time with the higher Vg2 if the slope does not change. I will try and do some calculations to see how the screen grid does with the 200R slope.

[MHProd]

Is the screen voltage up near 400V too? If so, then a 2k Zpri 100W OT would be a better match, or you could bring the screen voltage down to 300 or so. As it is now, the screens will be dissipating a lot of power at high volume.

How would I go about that? Right now, the B+ line dropping resistor for the screens is 18K.

And yes, there's a 1K on the tubes sockets.

[jazbo8]

Right now, the B+ line dropping resistor for the screens is 18K. And yes, there's a 1K on the tubes sockets.

With the dropper resistor, and the screen grid resistor, the screen voltage is only around 280V, so not much to worry about...

[martin manning]

Jazbo8, heres a little more data with an expanded scale. The diode line slope is constant with Vg2, but it looks more like 160 ohms. Can you share or point a link to your max screen current estimation method?

[jazbo8]

Jazbo8, heres a little more data with an expanded scale. The diode line slope is constant with Vg2, but it looks more like 160 ohms. Can you share or point a link to your max screen current estimation method?

160R is a bit steeper, so the screen dissipation would be higher, in any case, I use the method outlined in RDH4 to estiamte the Pdg2, which requires Ig2_max, Ig20, and Vg2 as the inputs. Ig2_max is taken at the saturation voltage where the load line intersects the Vg1=0V line.

[martin manning]

Ok great I'll have a look. Got a page number for that article by chance?

[jazbo8]

RDH4, P. 562, equation (10).

[martin manning]

Thanks! Looks like RDH4 p. 562 (10) is to be applied in cases where the load line passes through the knee, as shown in figure 13.21 below it.

As you say, the screen voltage will be reduced by screen grid resistors and by power supply dropping resistors, and if there is enough voltage drop the screen dissipation will be controlled. If the static load line passes much below the knee, I would want to make some measurements with the power stage grids driven to zero volts, and assume that the screens would have to stand up to that level of dissipation with a 50% duty cycle.

[jazbo8]

I would want to make some measurements with the power stage grids driven to zero volts, and assume that the screens would have to stand up to that level of dissipation with a 50% duty cycle.

You are welcome, I think there is a worksheet done by "es345" floating around that does exactly that, it uses a sine wave signal over one cycle and records the maximum and the average anode and screen dissipations. But I wonder how much more dissipation occurs when the waveform is closer to a square wave instead of a sine wave, which is typically the case for guitar amps?