This topic extends the discussion on unusual Kriesler amplifier design on another thread. Of particular interest is Philco's take on "See-Saw-paraphase" push-pull drivers - a design that first appeared just after WWII.

I found from my console collection a number of Philco model iterations that appeared in the late 1950s and incorporated this design.

The earliest design uses a 12AX7 'seesaw-paraphase' driving 6L6's. In this case, the12AX7 loads & anode-follower resistors are all 100k. A subsequent model that came out a year later has 6BQ5 outputs and changes the anode-follower grid resistors to all 1-megohm while retaining 100k for both plate loads.

Strangely, in an intervening model, Philco went with a quite different 'Williamson'-inspired design. I will have a link for that service folder which, luckily, is online. Later, it appears, models after this reverted to the original paraphase design!

I have a few tech questions regarding these designs that I'll ask later in relation to the schematic.

So first tech question I have is why do they have RF chokes in series on drives to the stator plates of electrostatic panels?

Second engineering question is, could this model benefit from a bypass capacitor at junction of 2.2 megohm & L1 tap connection?

Here is link to the "Williamson" (1960) version of this series that uses the rare 7199 triode/pentodes (DC coupled as in Williamson voltage-gain stage) http://www.audiophool.com/Philco.html (scroll down to model H-1916 to retrieve PDF)

I suggest you download and install LTSpice. It's a free download. Spend a few hours learning how to use it.

Load the .ASC file from my post. That will give you all the models you need to try out that circuit.

I have used LTSpice to simulate lots of analogue designs, in particular a Power Over Ethernet power converter circuit that is very complex.
I have found the LTSpice simulations accurately predict what happens when the circuit is built. Provided your models are accurate of course. You can discover other users' experience of various models and how to modify them on-line.

Re: RF Chokes,

Electrostatic speakers are fantastic at reproducing high frequencies, but absolute slugs at anything approaching low mid-bass.

I am guessing the choke is a cheap way of reducing the "screech" that electrostatic speakers are famous for.

On a slightly different topic, I am a great fan of 6L6's. They are the bees knees, the ants pants, the absolute workhorse of guitar amps.

May I ask if I can adapt this circuit to drive "real speakers".

Yes, I am biased, I hate electrostatic speakers, they should be restricted to mosquito repellent devices.

Hi New Vista,

Re: the addition of a bypass cap from junction 2M2/L1,

This is a very intelligent question. I would hazard to guess it would help the workload of RFC by bypassing any "surplus" highs that may have slipped past.

Ian, could such a modification introduce ultrasonic feedback? I'm out of my depth here.

Additional comment re: RCF and proposed bypass cap---

I must refrain from looking at circuit diagrams after having a few beers (it was New Year's here, yours today, I believe)

The RFC will block unwanted high frequencies.

A bypass capacitor at the junction of 2M2/L1 would decouple audio being fed from the 6L6's via the 0.01 capacitors. Too high a value would reduce drive to the speakers. It would make an interesting experiment in tone control.

I do also wonder if there is a risk, with RF getting into those speakers? Whilst a bit of RF may help a valve run at a steadier current, many audio valves have little problem with amplifying it.

Marc

Good point.

RF making it to those electrostatic speaker plates would see them as a nice antenna, re-radiate all over the place. Any receivers in the area would not be happy.

Reason for the RFCs is to restrict capacitive loading on the amp due to the fact that the capacitive reactance of electrostatic speakers drops with increasing frequency.

In LTSpice, it's very instructive to add capacitance to the output of an amplifier with feedback, it's quite easy to invoke oscillation at supersonic frequencies. Quite tiny amounts of capacitance will quickly degrade the square wave response.

I really recommend you take the time to learn LTSpice. That way, when you build the actual amplifier there is a much better chance it will work the way you want. It's so easy to see the result of small changes to components and how critical or otherwise a design will be when it comes to component tolerances. I rejected a DC coupled valve amp design because of this. It would work but the slightest tolerance shift would degrade it unacceptably.

You can also confirm that all components, valves, transformers etc. will be working inside their ratings.

Ian, A small guitar amp would make a great winter's project.
Push/Pull, common valves, readily available and budget output transformers.
I would be great, and might improve the sound of my much loved "Partscaster".
JJ

I will tell our hobby group about LTSpice (have not checked it out yet)

this quote from Ian on other thread:

QUOTE: The Anode Follower has issues when you start putting respectable amounts of feedback around the amp.

Interestingly, we had trouble with unstable ultrasonics on this 6L6 amp which only disappeared when .0033 cap was removed from feedback line! Though to be fair, the 100k/half-watt resistors need changing as they measure out of spec (most went low to around 84k)(so case not closed on this yet.) Philco's later Anode Follower 6BQ5 amp also has a "large" .0033 cap for feedback compensation while their Williamson version, as you'll note from link, only needs a mere 100pf.

NewVista,

I am intrigued, resistors going low, not high?

Have they been expose to moisture/coastal air??

Old memories of very patient Mr. Wiltshire, our 70's instructor at Sunshine Tech. Vic. telling us resistors always go high, never low.

Anyone else on the forum remember him? Ex RAAF, wonderful instructor.

If they read 84K, I would not replace them, they are probably OK.

Another thought.

Are you measuring them with a DVM perhaps? (very high input resistance)

Old days meter specs usually ran around 20K meter resistance. (I think, been a long time)

Many resistors read low, if measured "in circuit" One should be aware if that, should it occur.

The leaks can be Electrolytic caps and leaking Wax Paper types.

Marc