Harman Kardon Citation II PI question

[Printer2]

I am trying to understand the PI circuit in this amp. I am a little confused about the output to the power tubes, one side is taken from the plate which has a 12k resistor on it while the other side is taken on the other side of a 4.7k resistor in series with a 8.2k resistor. Are they using these two resistors as a split load in order to balance the two outputs to the power tubes?

Also the two screens on the on the 12BY7A's are tied together and go through a 22k resistor to the supply. Are they running in triode mode? Or is there something funny that goes on when they are run through a common resistor? I do not know of any other amps that use pentodes in the PI section so I have not come across an explanation on this type of circuit before. Thanks in advance for your help.

[SGregory]

Your first question is answered by your comment. The difference in resistor is to balance the output. This appears to be a hybrid of the LTP phase splitter.

I can't answer your second question as easily. I am not familiar enough with this circuit. It looks like it is just applying a screen voltage and is still a pentode connection.

[Miles Prower]

Originally Posted by Printer2

I am trying to understand the PI circuit in this amp. I am a little confused about the output to the power tubes, one side is taken from the plate which has a 12k resistor on it while the other side is taken on the other side of a 4.7k resistor in series with a 8.2k resistor. Are they using these two resistors as a split load in order to balance the two outputs to the power tubes?

Yup, that's what they're doing. The phase splitter is an LTP with a passive tail load, and it looks to be a fairly small one at that. Given that VTs don't have the gain capability that transistors have, you end up with an inherent phase imbalance since the driven side sees: Vgk= Vi - Vtail

Whereas the passive side simply sees Vgk= -Vtail

If you don't have an active tail load, you won't see phase-to-phase balance, so you need to force it with unequal plate load resistors.

Also the two screens on the on the 12BY7A's are tied together and go through a 22k resistor to the supply. Are they running in triode mode?

No, pure pentode operation here.

Or is there something funny that goes on when they are run through a common resistor?

Produces nearly constant current through the screen dropping resistor, since as one side is pulling a heavier current, the other side is pulling less current. Since it's almost pure DC, you don't need to bypass the screens since the AC cancels out.

I do not know of any other amps that use pentodes in the PI section so I have not come across an explanation on this type of circuit before.

This comes close: Le Renard Main Schemo

Cascoding is quite similar to using pentodes in some respects.

[Printer2]

Got to love this place, a lot of smart cookies.

Produces nearly constant current through the screen dropping resistor, since as one side is pulling a heavier current, the other side is pulling less current. Since it's almost pure DC, you don't need to bypass the screens since the AC cancels out.

I kind of thought the currents would sum together but did not know if that would be a good or bad thing. I wonder why everyone uses separate screen resistors in p-p amps when this would seem to be an advantage. Thanks for your answers.

[smoking-amp]

If the splitter output signal gets big enough to drop the plate voltages below the screen voltage, it will start producing odd harmonic distortion big time (signal compression). For low plate voltage swings though, the un-bypassed common screen resistor probably acts similar to additional tail resistance. The screens will intercept a near constant fraction of the cathode currents up till the plate voltages drop down near the screen voltage. Any tendency to draw more than constant plate current (the sum) will drop the screen voltage, turning off the tubes somewhat, just like the tail resistor voltage rises to turn them off. Would help some to DC auto-bias the splitter tubes too, improving the constant current regulation.

[anatech]

Hi Miles,
Really good explanation, thank you. There is one thing I'd like to bring up though ...

Since it's almost pure DC, you don't need to bypass the screens since the AC cancels out.

... as long as the tubes are matched.

-Chris

[Printer2]

Originally Posted by anatech

Hi Miles,
Really good explanation, thank you. There is one thing I'd like to bring up though ...

... as long as the tubes are matched.

-Chris

Matched tubes? At my place?

Not too concerned about distortion, actually part of the reason I am looking into this. Just for the heck of it I would like to do an all pentode amp.

[Eli Duttman]

Not too concerned about distortion, actually part of the reason I am looking into this. Just for the heck of it I would like to do an all pentode amp.

There are 6 adjustments on a "Deuce": a bias trim pot. for each "final" and an AC balance trim pot. for each channel. Perfectly matched 12BY7s in the splitter positions are nice, but not essential.

It's very hard to better Stu Hegeman's effort in an all pentode amp, other than adding Jim McShane's modernizations. Jim takes advantage of parts Stu did not have access to.

Cloning the Cit. 2 is no easy matter. The O/P trafos are among the BEST ever made. Bandwidth is phenomenal. Edcor and Hammond stuff are dead out. The only off the shelf item that is remotely suitable is Plitron's 4004-01, which is somewhat shy in the power handling dept. Be sure to protect against infrasonic noise induced core saturation.

[Printer2]

Originally Posted by Eli Duttman

There are 6 adjustments on a "Deuce": a bias trim pot. for each "final" and an AC balance trim pot. for each channel. Perfectly matched 12BY7s in the splitter positions are nice, but not essential.

It's very hard to better Stu Hegeman's effort in an all pentode amp, other than adding Jim McShane's modernizations. Jim takes advantage of parts Stu did not have access to.

Cloning the Cit. 2 is no easy matter. The O/P trafos are among the BEST ever made. Bandwidth is phenomenal. Edcor and Hammond stuff are dead out. The only off the shelf item that is remotely suitable is Plitron's 4004-01, which is somewhat shy in the power handling dept. Be sure to protect against infrasonic noise induced core saturation.

Sorry, left out one important word. All pentode guitar amp. Bandwidth, what is that?

[anatech]

Hi Printer2,
Well, given that it's an HK tube amplifier, it would seem logical to do a reasonably good job on it. I have one here I'll be rebuilding for a very good friend. These are worth more nicely restored than as a different rebuild.

There is no shortage of amplifiers using real pentodes in the output stage. The less expensive units that use 6BQ5As would be a nice proof of concept type deal without ruining a larger amplifier. Not to mention that matching a new set of tubes is only a dollar or two (Canadian) extra for matching. May as well if you are replacing outputs.

-Chris