Parallel Single-Ended Questions

[Sherman]

I'm once again planning a winter project. This time around I'm thinking of doing a parallel single-ended amp using KT88's. My questions concern stages. All of the PSE designs I've come across (I've just started so I'm sure I've not scratched the surface) so far use a 2 stage input/driver.

Can I do this with only a single stage using a relatively high mu tube such as the 6N1P? It appears I can get a gain of 22+ with a 6N1P, meaning with a 1.5V source I have plenty of gain to drive a KT88, but what about driving 2 of them?

Playing with SE Amp CAD it appears that >10W is possible with this config. (Actually SEACAD says 15+.)

Does this look like it will make sound and not explode ?

[Fuling]

Looks perfectly fine to me, assuming that the input stage is capable of driving the output stage to full power past 20kHz.
The quick´ndirty way to ensure that would be to insert a cathode (or source) follower directly after the driver.

[dhaen]

Depending on the quality of your KT88's and the dissapation, you may have to lower the 200K shared grid leak resistor. That will unfortunately increase driver distortion. Best to try and see.

[Sherman]

Originally posted by dhaen
Depending on the quality of your KT88's and the dissapation, you may have to lower the 200K shared grid leak resistor...

Well, I'm planning on using some relatively low-quality Valve Arts to start. (They are already on my shelf!)

I was wondering about that resistor, but I didn't see any references about how to determine the optimum value when paralleling tubes. Maps I've seen don't seem to be using 1/2 the recommended value but something like 2/3rds. Maybe 120K in this case? I think I'll take your advice and try it first before making any decision about it.

I'm encouraged that so far no one seems to think this design will let the magic smoke out!

[tubelab.com]

I think that your output circuit looks pretty straight forward. The grid resistor is one concern. If you see any tendancy toward runaway or current hogging, you may need to lower its value. I am not familiar with the 6N1P but any high gain tube will have a tough time driving the miller capacitance of two KT-88's. I (a definitely biased opinion) might be tempted to try a PowerDrive circuit here. The CCS would allow the triode to achieve maximum gain while the mosfet would isolate it from the KT-88 load.

[Sherman]

Originally posted by tubelab.com
... The grid resistor is one concern. ... you may need to lower its value.

...any high gain tube will have a tough time driving the miller capacitance of two KT-88's....

Thanks for the reply!

Now I'm starting to get it. I honestly hadn't considered that the Miller capacitance would double with the two output tubes in parallel. I have used the 6N1P as the input/driver stage for a single-ended KT88 and it works beautifully, but with twice the capacitive load to drive...

I might be better off using a lower mu tube in two stages. I guess that is the reason all the schems I've found do it that way. However, being a minimalist sort of guy I'll probably just give the 6N1P a shot first.

[Joel]

I don't think your 6N1P is going to enjoy a 2.2k bias resistor, or having 1.5V of signal shoved into it.

Personally, if I was doing PSE I would transformer couple the driver stage. A Lundahl LL1660 is only going to set you back $95US.

Joel

[Sherman]

Originally posted by Joel
I don't think your 6N1P is going to enjoy a 2.2k bias resistor, or having 1.5V of signal shoved into it.

Personally, if I was doing PSE I would transformer couple the driver stage. A Lundahl LL1660 is only going to set you back $95US.

Joel

Joel,

Thanks for your feedback!

I'm using the 6N1P currently in a pair of KT88 SE monoblocks. My CD player's signal is close to 2V so I'm not too worried about that end of things.

I am concerned about the 2.2K bias resistor. I am using 1K in my monoblocks and will probably end up with a similar value here. (Meaning also changing the plate resistor and decoupling network but that's life. )

I was hoping to avoid multiple stages on the preamp side and things like transformer coupling. I have no bias against those things it is just that I'm sort of trying to follow a simple and somewhat linear path to increasing my knowledge. I do not expect that this amp will be the end-all SEP or that it will even become my main system amp. It is a learning thing.

At the end of the day I might have to go with at least a two-stage pre section. The transformer coupling might be a great thing to try after that. I should be able to re-use virtually everything else.

[porkchop61]

Hey Sherman-
Did you ever get any results out of this test?
I'm looking to do the same thing to my SE KT88 which a lot of us have built on this forum (including you I believe).
I'd be interested in adding another KT88 to my setup for a little more power.

Glenn

[Sherman]

Glenn,

I only built one channel and that was just breadboarded. However... the 6N1P was in fact able to drive both KT88s. I ended up using 1.2K bias resistor on the 6N1P, everything else was pretty much per the schematic. Measured output was pretty good. My monoblocks measure better than 6.5 watts, this channel measured about 11.

I don't have the equipment or software to measure distortion but it sounded pretty good to me. There was some hum but I'm convinced it was because of the breadboard layout and with more careful layout and dressing it should be extremely quiet.

For the record I also tried a 6922 as the driver and felt it lost something with the change (I didn't change any values so it might have been better with the values tweaked for that tube.) It was also suggested that I try a 6DJ8 in that position but I never got around to it.

Why didn't I build it? After putting together one channel I felt I had learned what I wanted to. The sound wasn't better than my monoblocks (but it could play louder) so I just didn't pursue it.