KT88 Push-Pull

[denny9167]

I have a schematic using a pair of 6SN7's to drive a pair of KT88's in a Push Pull config. it's rated at 30 watts, is it possible to add an additional pair in parallel to increase the power output?

[tubewade]

It would be nice to have a look at your schematic to say for sure. Is it the classic GEC 30 W KT88 schematic? At any rate, you can probably add a second pair of valves for more power out. It will consume twice the current and require half the load impedance if you do that. There may be other things to consider such as bias, too, but it can probably be done.

Another thing to consider is that doing so will increase the cost tremendously, i.e. four valves, bigger output transformer, bigger power transformer, maybe more rectifiers, but will only increase your output by about 3 decibels. Is it worth the cost?

Wade

[TubeHead Johnny]

IMHO, I think 4 KT88's are overkill, unless you're building a seriously loud bass amp, or P.A. With the right PT (and OT), you can get some incredible power with just a pair in push-pull, with fixed bias.

[denny9167]

Originally posted by tubewade
It would be nice to have a look at your schematic to say for sure. Is it the classic GEC 30 W KT88 schematic? At any rate, you can probably add a second pair of valves for more power out. It will consume twice the current and require half the load impedance if you do that. There may be other things to consider such as bias, too, but it can probably be done.

Another thing to consider is that doing so will increase the cost tremendously, i.e. four valves, bigger output transformer, bigger power transformer, maybe more rectifiers, but will only increase your output by about 3 decibels. Is it worth the cost?

Wade

You probably have a point,I thought what I might do is use the original design,and just bi-amp my system by using the tube amps on the mids and highs,and use the pair of SS monoblock amps I have for the low end.

Thanks for the advice though!

[Johan Potgieter]

Denny9167,

It is not clear whether you already have some parts e.g. the output transformer, power transformer, etc. As Tubewade said you will have to reckon with cost.

But only 30W for a pair of KT88s? Is this triode connected? I don't know over there, but in RSA the KT88 is quite expensive, and it will pay to get as much as is safely possible from a pair. As pentodes they can give up to 100W, but I would prefer UL. There they can give an easy 60W with 500V h.t. and fixed bias.

I like 6L6s at half the heater current of the KT88. A twin pair (i.e. 4 in push-pull parallel) of 6L6GCs can give a comfortable 100W with fixed bias in UL. The list is endless, but just to position you within the scope of your question.

[tubewade]

Denny,

If you use an active filtering setup, the P/P KT88s at 30 watts on the medium and high range drivers should certainly be more than enough power, even if the speakers are fairly inefficient.


Johan,

I believe the reason for the relative low power from this circuit is to promote extremely long life of the valves. Running them at the operating points required for 30 watts should make the valves last practically forever. I agree that it would also be very easy to get 50 or 60 watts from the same set of valves and still not be pushing them very hard.

[TubeHead Johnny]

The 30 watt rating has to do with cathode biasing, no?

[navin]

Originally posted by denny9167

You probably have a point,I thought what I might do is use the original design,and just bi-amp my system by using the tube amps on the mids and highs,and use the pair of SS monoblock amps I have for the low end.

This makes the most sense. There a circuit posted by Wayne on a thread called EL34 schematic confusion optimised for just such a configuration. It needs 1 6SL7, 2 6SN7 and 4 power tubes like the KT88. The original has been designed for the EL34 but I think the 6SN7 can drive KT88s as well.

Originally posted by tubewade

If you use an active filtering setup, the P/P KT88s at 30 watts on the medium and high range drivers should certainly be more than enough power, even if the speakers are fairly inefficient..

if the speakers are 90db/1m/1m and 84-85db/1w/3m you will get 94-95db at 10W and 100db at 40W. More than enough for most applications.

[tubewade]

Originally posted by navin


if the speakers are 90db/1m/1m and 84-85db/1w/3m you will get 94-95db at 10W and 100db at 40W. More than enough for most applications.

This is true. Also consider that this is for only one channel. If there are two channels, there will an additional 3 decibels due to the additional electrical power, and an additional 3 decibels due to doubling the radiation area, for a total gain of +6dB. Considering that power distribution in music is about 50 percent below 350 cps and the remaining 50 percent above 350 cps, if filtering is done in this range or above the amplifier will sound MUCH larger than it is. Dispensing with power wasting passive filter and level matching networks will make things even louder. I agree with you that it will be more than enough.

[richwalters]

Originally posted by denny9167
I have a schematic using a pair of 6SN7's to drive a pair of KT88's in a Push Pull config. it's rated at 30 watts, is it possible to add an additional pair in parallel to increase the power output?

So long one stays in class A (which you are implying) the load on the drivers is bearable and isn't suprising to find in some designs an ECC83 driving a KT88. The troubles really start when amp tips class B or specifically true power triodes with high Cag Miller cap.

Cribbing the literature for the GEC 400Wattt amp, it quotes << with four o/p valves the grid resistors could be increased to 100K and the cathode followers dispensed with.>>

This is a fixed bias set-up. With my sim designs I use separate grid coupling caps for each tube and bias presets, (a bit more complexity) but an advantage if unmatched tubes are used with differing neg bias volts ..


richj