Interesting Phase Inverter

[andyhardy]

Has anyone built this phase inverter? Seems cathodyne but no grid stopper interesting both triodes linked.

[Andy Le Blanc]

yes,... right out of a hagstrom, impact... maybe even a univox

there's a few variations, direct coupled or as a cathodyne... works with any pentode.
You can simplify it too, as one dual triode and the pentode, the triode covers the the input and invertor and pentode goes between,
I liked it best with 6sl7 and a 6sj7, nice big clean tone... add the extra stage in the pre like the scheme you posted and get enough gain to knock into post distortion.

go with those kt88... kt 100/120, that would rule with a clean jazz tone

[roberto]

I've done something similar with a 12ax7.
A gain stage directly coupled to the concertina PI.
Good as a return fx loop for EL84 amps.

To balance the sound you need a quite high Rp on the first stage (around 270k) and a low Rk (something around 680R IIRC), then the two 47~100k for the concertina. I also used a 470k grid stopper between the two triodes.

[andyhardy]

Its interesting on the schematic there are no big grid stoppers no 1 meg to the top of the cathode.
Plus what is the advantage of running one into the other seems to defeat the purpose of the concertina.

[roberto]

Why?

[andyhardy]

I guess im trying to understand why there are basically 2 phase inverters in parallel instead of 1.
Plus as you mentioned normally there would be a fair amount of signal attenuation with stoppers etc. but in the above circuit there is none.
I thought the output of the ef86 may need some taming as well.
Similar to these.

[teemuk]

I guess im trying to understand why there are basically 2 phase inverters in parallel instead of 1.

I see only one. There's an EF86 tube as voltage amp/driver and a standard cathodyne PI.

The only difference is that there are two tubes in parallel in the cathodyne for more current drive capability (which is needed with those KT88's). Doesn't make it two phase inverters, though.

KT88's probably also explain the EF86 driver: Pentodes have more voltage swing than triodes and KT88's tend to need a lot of drive voltage.

This is basically a classic Dynaco Mk II & Mk III design: pentode driver and cathodyne PI with high current amplification. Even component values are practically same as in Dynaco circuits.

What's actually more interesting is the configuration of that presence control.

[andyhardy]

The only difference is that there are two tubes in parallel in the cathodyne for more current drive capability (which is needed with those KT88's). Doesn't make it two phase inverters, though.

Great! makes sense now.
Thanks

[surfsup]

Its interesting on the schematic there are no big grid stoppers

I believe the big gs is for taming overdriven signals into this type of PI.

What's interesting to me is the 88V on the cathode with 475 VDC supply.

[matt h]

(deleted)

[Andy Le Blanc]

used to be pretty common, amp using a 6an8a, or a 7199... a pentode and triode in one tube...
everything from a Kay to Sunn... some Gibson.

"One of the problems with regular RC coupled stages is that they can cause instability problems at very low frequencies when NFB is used because the loop is not closed down to dc. dc coupling at least reduces and can in some cases completely eliminate these instability problems."

[martin manning]

Here be an olde poste with some background on this amp:
https://tubeamparchive.com/viewtopic.ph ... highlight=

What's the source of that quote Andy?

[teemuk]

Well, it's evident that any RC circuit inside the loop (read: capacitive coupling) will increase phase shifts and therefore decrease margin of stability. Issue? Not likely, but keep in mind that Dynaco circuits tended to use ample amounts of negative feedback.

[surfsup]

Think of the concertina in this application as being fixed bias (hint: see preceding stage's plate voltage).

Yes that was my point. Why wouldn't the designer put more voltage on the pentode's plate, and hence the PI cathode, to increase the headroom if this is supposed to be a clean amp?

[matt h]

(deleted)