Strange Phase Inverter in Geloso tube amps?

Tube or solid-state, this section goes to eleven!
Post Reply
User avatar
Desdog
Breadboard Brother
Information
Posts: 54
Joined: 25 Feb 2008, 17:18
Has thanked: 62 times
Been thanked: 16 times

Post by Desdog »

I am working on this nice little 60s Italian PA tube amp... It sound very good in stock condition thanks to some smart tricks: shielding the preamp and feeding it negative voltage for the heaters. But the PI is something I have never seen in a guitar tube amp. Both plates of the sides of the PI are connected to the output tubes, but there is an extra connection between the cathode of one of the PI sides to the other output tube.

Can somebody explain why this is done or how it's called? thanks
Attachments
Geloso G215AN
Geloso G215AN

User avatar
marshmellow
Cap Cooler
Information
Posts: 469
Joined: 16 Aug 2007, 07:31
Location: Germany
Has thanked: 20 times
Been thanked: 127 times

Post by marshmellow »

It's a standard AC-coupled cathodyne PI, just stare at it long enough to connect the right dots from the oldschool drawing. See here for reference.

Image

It's the right variant using cathode bias. The additional 10k/500p is a lowpass/treble filter.

User avatar
Desdog
Breadboard Brother
Information
Posts: 54
Joined: 25 Feb 2008, 17:18
Has thanked: 62 times
Been thanked: 16 times

Post by Desdog »

Thanks!

I still don't understand, because in my opinion Cathodyne PI or Split Load or Concertina, only uses 1 triode:
remember, [for a cathodyne PI] we only need one triode, unlike a long-tailed pair. This usually leaves you with an extra triode which can be used as a gain stage
So the part AC coupled seems to be important: you mean that plate1 is connected to grid2 trough a 22nF coupling cap, thereby blocking DC?
If there would be a resistor there, it would be an DC coupled cathodyne?
Geloso 215AN phase inverter detail
Geloso 215AN phase inverter detail
phase inverter.png (111.52 KiB) Viewed 2284 times
It's not like I want to change this setup, I think it sounds good. At least it is the only part of the circuit radically different from any guitar amp I have seen.
The only reason I want to know it that I would want to implement a PPIMV -- post phase inverter master volume... in Marshall style. Is that possible with different resistors on the output tube grids?

User avatar
marshmellow
Cap Cooler
Information
Posts: 469
Joined: 16 Aug 2007, 07:31
Location: Germany
Has thanked: 20 times
Been thanked: 127 times

Post by marshmellow »

The first triode is just a standard gain stage, it is not part of the PI.
So the part AC coupled seems to be important: you mean that plate1 is connected to grid2 trough a 22nF coupling cap, thereby blocking DC?
Yes, that's what it means. A capacitor is used to block DC, hence only AC signals can pass.
It's not like I want to change this setup, I think it sounds good. At least it is the only part of the circuit radically different from any guitar amp I have seen.
This setup has been used in many other amps, like the vintage Orange amps for example. HERE with AC-coupling, HERE with DC-coupling.
The only reason I want to know it that I would want to implement a PPIMV -- post phase inverter master volume... in Marshall style. Is that possible with different resistors on the output tube grids?
The different resistors are rather curious. There is also the 50p cap at the upper EL84, not sure what is the reasoning behind this asymmetrical circuit. Maybe they tried to compensate for any asymmetry stemming from the cathodyne, or some sort of oscillation prevention. In any case, yes you can build a PPIMV into it, the unequal values won't get in the way with that. Keep in mind though, if you reduce the volume low enough, at some point the feedback loop will stop working properly, which does change the sound a bit. But then again, you might not care too much about that at lower volume.

User avatar
Desdog
Breadboard Brother
Information
Posts: 54
Joined: 25 Feb 2008, 17:18
Has thanked: 62 times
Been thanked: 16 times

Post by Desdog »

marshmellow wrote:It's a standard AC-coupled cathodyne PI, just stare at it long enough to connect the right dots from the oldschool drawing. See here for reference.
After three days of staring and thinking it is not the SAME I suddenly saw the light -- thanks. You are absolutely right.
Of course the first part of the second triode is a gain stage: signal from grid to plate -- but still inverts the signal.

Went looking further on the web and found some local forums discussing the assymetrical output stage: however the unit discussed here is the Geloso 1040, with pushpull EL34's. But it has a similar phase inverter
The output impedance on the anode side is much higher than on the cathode side. For an ECC81 in this scheme roughly 40k and 2k. You can 'equalize' this by setting a higher resistance in series with the cathode signal output than in series with the anode signal output. Normally that would not matter much, but something else will be added: if the lower output tube grid current were to start pulling, that would cause enormous distortion especially at the anode output of the ECC81.
In order to prevent worse, the lower output tube apparently gets the 47k grid 'stopper'. The upper one also gets 1k; it is customary to equip end tubes with grid stoppers (resistors) of, for example, 1k, in order to nip VHF oscillations in the bud.
They say it's called the KANGOROO PHASE INVERTER, I found this explanation dated 1947
Kangoroo phase inverter
Kangoroo phase inverter
The reasons it is not popular in guitar amps is proably the gain factor being 0.9 as mentioned here.


User avatar
alexradium
Resistor Ronker
Information
Posts: 313
Joined: 31 Jan 2010, 13:31
Has thanked: 64 times
Been thanked: 68 times

Post by alexradium »

its not popular as a guitar amp because at the time these were conceived and built there were only generic mono amps for microphones,accordions,phonographs etc.
the whole guitar thing happened later with proper choice of eq section,gain,LTPI.

User avatar
teemuk
Breadboard Brother
Information
Posts: 147
Joined: 17 Aug 2009, 19:54
Has thanked: 3 times
Been thanked: 55 times

Post by teemuk »

Yes, nothing but a generic "cathodyne" phase inverter there. ...And overall the setup is pretty generic too: Since cathodyne does not have voltage gain (only insertion losses) there is a common cathode amp input stage for voltage amplification. The same stage also receives negative voltage feedback from the output. Feedback goes to the cathode, as is the most generic setup, and the feedback ratio is defined by the feedback and cathode shunt resistors, which effectively form a voltage divider. Very typical arrangement to couple with cathodyne indeed.

So I beg to disagree about CC+cathodyne input stage not being "popular" in guitar amps. No, maybe not in Fender or Marshall amplifiers (and all those amplifiers that were derived from those designs) but yes, still very typical and popular choice in many other guitar amps, such as the aforementioned Orange, Sunn, Ampeg, Magnatone, Gibson, etc. Overall I would think this is the second most popular phase inverter scheme after long tail pair.

The reason why this type of PI is avoided in some amps is that it clips very nastily. It clips HARD and also produces ton of other "not-so-nice" overdrive artifacts that create a lot of high order harmonics. If overdriven hard it has very asymmetric output due to different output impedances of plate and cathode. Basically it doesn't have the smooth "soft" clipping overdrive characteristic of a differential-type phase inverter (e.g. LTP) and that will be reflected in overall tone of the overdriven power amp.

User avatar
Desdog
Breadboard Brother
Information
Posts: 54
Joined: 25 Feb 2008, 17:18
Has thanked: 62 times
Been thanked: 16 times

Post by Desdog »

teemuk wrote:The reason why this type of PI is avoided in some amps is that it clips very nastily. It clips HARD and also produces ton of other "not-so-nice" overdrive artifacts that create a lot of high order harmonics. If overdriven hard it has very asymmetric output due to different output impedances of plate and cathode. Basically it doesn't have the smooth "soft" clipping overdrive characteristic of a differential-type phase inverter (e.g. LTP) and that will be reflected in overall tone of the overdriven power amp.
Hey thanks for the comment! This is absolutely the case... the distortion was pretty nasty. I will eventually change to LTP, it's unavoidable if this amp is to have a future for guitar. For now I put a 1M resistor before the grid of the PI, which fixed part of the nastiness.

Post Reply