PI for lotsa drive, little supply voltage

[imaradiostar]

Hi everyone, just finished a cross-country move and I'm building an amp for a friend and I need advice. I'm using some parts I've had laying around so I've ended up with a 400VCT PT and a cathode biased 6L6 output section. I'm running the screens and preamp from the center tap of the PT. I figure I'm going to need a lot of grid drive to get full power out of the output tubes.

I'm happy with the operating point of the output tubes- cathode is at 20.5 volts, screens are at 268, B+ supply is 536. 268 doesn't seem like very much for the phase inverter. I have a split load PI wired up using a 6u8 triode with the pentode driving it- it distorts pretty nicely but it isn't very clean.

The amp is largely a "power amp" for an effect processor so I want it to distort in a pleasant way but have plenty of headroom.

thoughts?

jamie

[imaradiostar]

Oh yeah...forgot to mention, it's cathode biased with separate 500 ohm resistors per cathode. Yeah, that relaxes drive requirements a bit.

jamie

[imaradiostar]

I meant increases...oops.

[Steve Conner]

Because the screen voltage is low, it'll need less grid drive for full power than a regular 6L6 design, not more. You said you have 20V across your cathode resistors, so it'll only need a bit more than 40V p-p of drive per grid. (The "bit more" is to account for the transition into Class-AB.)

Even so, the split-load PI is hardly famous for its drive capability. If you want more drive, you should change to a LTPI, or add two triode stages after the PI a la Williamson amp.

And/or power the PI off the 536V node.

[imaradiostar]

originally I was using a "fixed bias" split load PI with large value resistors and a pentode driving a volume control before it- not because I wanted to but because it was already in the chassis and I wanted to make sound. I clipped a few wires and added some jumpers so it's direct coupled now and the triode is sitting at about 1/4 of b+. It's quite a bit louder than it was and of course has more drive.

The amp in question will probably end up with a LTPI and I'll just use a dropping resistor off of the B+. If I miss the sound of the pentode I'll go back to it. Now that I'm driving the output tubes harder I think I'm going to add some zeners to clamp the cathode voltage rise under heavy load. I feel like it may be squishing a little bit too much.

On that note- the 6u8 pentode directly coupled to the triode really does sound pretty great. I'm surprised it isn't used in more guitar circuits. Other than Sunn I can't think of any guitar amps with this kind of design. I read somewhere that the Sunn amps where one of the motivating factors behind some of the Dr Z amps- but the pentode/triode didn't make it to the Dr. Z amps so there must be a reason. I wonder why- availability of current production tubes maybe?

should be asleep but can't...

jamie

[MWJB]

500ohms per cathode sounds like a small value to me, given the B+? 1K-1.5K would have been the area I'd have expected, of course this will push up B+ a little.

[Steve Conner]

Well, remember his screens are only 250-something volts, and it's the screen voltage that determines how much grid voltage you need for a given plate current.

[MWJB]

But Jamie also says that he feels it's squishing too much, my first stop would be to back off plate current a little and/or go bigger on the cathode bypass caps.

[imaradiostar]

The output tubes are idling at exactly 70%...so I don't see a reason to change the cathode resistors. I got a good deal on the resistors anyway.

I want the amp to be cathode biased because I don't want the guy I'm building the amp for to have to worry about adjusting the bias. I like the idea of a zener to clamp the cathode voltage because I like the compression to an extent- I haven't tried this method yet and I've been wanting to for a while to see how it sounds.

I'm using an antek inc power transformer and it sags very little under load and doesn't really even get warm even though I'm exceeding it's rated abilities under load. I've seen a few other tube amps using them recently and my other projects with them work very well. I'm excited to build something with some giant WWII potted output TF's and a huge antek inc PT in the future...I'm thinking a very high power tube amp for PA use.

jamie

[Albert Kreuzer]

What about a paraphase ("see-saw") inverter?
Works great in a Leslie 125 with a pair of cathode biased 6L6's.
Best B+/output ratio IMO.

Cheers,
Albert

[imaradiostar]

A little update, I seem to have arrived it fairly conventional ideas at this point. So much for the thread topic! I'm really killing two birds with one stone- I've got the PT for this customer project wired into the chassis from another project so that I can test a few things before drilling holes in another chassis and starting a new turret board. Meanwhile I've managed to iron out a few bugs on the test chassis which is a prototype for another friend.

The screens are still connected to the center tap of the pt and the junction of the first filter caps which are two 160uf 300v caps in series. I'm dropping the high B+ voltage through a large value resistor tuned for about 400 volts on the phase inverter. It's plenty clean for what I'm doing and I've gone back to a volume control between the pentode and the phase inverter. There is no way to inject nfb but I'm not worried about it- I like the way it sounds now and I have plenty of headroom.

I'm still torn on the PI design- I think the customer's amp is going to end up with a LTPI and a similar voltage drop arrangement and a little nfb. I'm tempted to build an amp in the immediate future with several different phase inverters on board to allow me to select them on the fly to hear the differences. I like what pentodes do to the sound and up till the point of distortion the split load PI is very linear- I like that. I think that it distorts prettier now than when it was direct coupled.

jamie