unusual phase invertor?

[tubenit]

Guys,

Has anyone seen this type of phase invertor before?  And if so, what can you tell me about it.  Does this look possible and/or drawn correctly?

With respect, Tubenit

[sluckey]

That's probably the most common phase inverter of all. It's commonly called a paraphase inverter. Look at early Fender or Ampeg amps for examples. There are lots. Here's one...

http://www.el34world.com/charts/Schematics/files/fender/bassman_5b6_schem.pdf

The first gain triode plate directly feeds one output tube. That same plate signal is applied to another gain triode thru a big voltage divider (270K and 12K). This divider knocks the signal down by a factor of 12:270 or 0.04 (that's 1:22.5) and the second gain triode bumps the signal back up to match the signal applied to the top output tube. The signal applied to the lower output tube will be 180 degrees out of phase with the signal applied to the top output tube. The voltage divider and gain of the second triode are responsible for matching the amplitude of the top drive signal.

Valve Wizard has this to say...
http://www.valvewizard.co.uk/paraphase.html

[tubenit]

That's probably the most common phase inverter of all. It's commonly called a paraphase inverter.

 

Yikes ............. I called the most common phase inverter "unusual". 

Sluckey, thanks my friend for the info as always!  You are a wealth of great information and help!

 

With respect, Tubenit

[Mats]

Hi
Can I also ad to take a look at, Supro Thunderbolt S6420 amp.

/Mats

[Willabe]

Yikes ............. I called the most common phase inverter "unusual".  

It's the way it's drawn. I didn't recognize it either.     


                  Brad          

[tubenit]

It's the way it's drawn. I didn't recognize it either.

Excellent!  A perfect excuse for my ignorance!!        

     Thanks!


Exactly,  whoever drew it that way was totally out of line.      

What kind of deeply disturbed individual would ever draw one like that ?   

With deep respect and gratitude,  Tubenit

[kagliostro]

Ciao Tubenit

If I'm not wrong the Floating Paraphase is an evolution with better performances

http://www.valvewizard.co.uk/paraphase.html

K

[rzenc]

what can you tell me about it.

I built a couple of 5E9-A (Fender Tremolux) and it uses this type of phase inverter.

Very touch sensitive. Somewhat hotter sounding than same period tweed 5E3 for instance, which I also have built and still have it on my collection. Both amps sounds very nice but 5E9-A just gets more atitude. I believe it's easier to control between clean and crunch. Thanks to HBP for such tip  .

Probably due to KISS principle it's simply spectacular. Love it, althought I haven't tried it with bigger output bottles but only 6V6.

Here is a Picture of how I did layout.

Hope this helps!

Best Regards

R.

[printer2]

OK, How about this one?

[sluckey]

OK, How about this one?

Google 'self split inverter'.

[bigsbybender]

Quote
That's probably the most common phase inverter of all. It's commonly called a paraphase inverter.

 

Yikes ............. I called the most common phase inverter "unusual". 

We all use the Long Tail Pair so much... everything else is weird.  My latest build is using an Ampeg influenced Paraphase... and it looks weird to me. 

j.

[printer2]

Google 'self split inverter'.

Lot of links for air conditioners. I get the self split on a cathode biased amp, having trouble seeing this guy.


I got it. Missed the cap from the screen to the grid.

[kagliostro]

http://www.ampbooks.com/home/amp-technology/6L6-phase-inverter/

K

[sluckey]

Lot of links for air conditioners.

I totally missed all the A/C links. Google's #1 hit sent me to ax84.com and I never even looked farther. There's a long discussion about self split. This schematic came from that first hit. It gets the inverted signal from the top screen just like your example, but this circuit is cathode biased rather than fixed biased.



There are some other self split PP amps. One gets the inverted signal from the plate of the top output tube and feeds it to the grid of the lower output tube. You can use triodes for the output. Another uses a common unbypassed cathode resistor and a grounded grid. The Firefly amp uses this method. None of these were mainstream circuits. Too many shortcomings for hifi. But the color they add to a guitar amp seems to be popular with some people.

[tubeswell]

And from here: http://jonzeamps.webs.com/apps/documents/

(Don't know whether he actually got this to work)

[sluckey]

(Don't know whether he actually got this to work)

Hmm, I've never actually scoped a screen but I sure thought the signal on the screen would be in phase with the plate signal.

[printer2]

(Don't know whether he actually got this to work)

Hmm, I've never actually scoped a screen but I sure thought the signal on the screen would be in phase with the plate signal.

Same bunch of electrons hitting the screen are hitting the plate (well not the same actual ones) so plate and screen currents should be in phase as far as I can see.

[sluckey]

Same bunch of electrons hitting the screen are hitting the plate (well not the same actual ones) so plate and screen currents should be in phase as far as I can see.

Exactly. That self split circuit you posted (or the one I posted from ax84) would not work if there was not an inversion from the grid to screen.

[printer2]

I am not sure of phasing on that 6sj7, but G2 has a higher voltage than the anode, so the potential to catch electrons would be controlled by G1.  I believe you are looking at this portion of the pentode acting as a triode, G2 is acting as an anode.  G3 has a lower potential, so it may be acting as a quasi cathode, with some ac  and therefore the anode g3 pair may be acting as a diode.

Gurus tell me that I'm chasing wild electrons. 

Could somebody spice this design?

G1 controls the amount of electrons that flow by it. If it is positive a lot of electrons go by, if it is negative fewer go by. Once they are by G1 the electrons pretty much hit the rest of the tube at the same time. G2 acts to give a uniform field to the electrons, the voltage on the plate can go up and down depending on how much voltage is developed across the load. G2 isolates the plate from G1 so G1 sees a uniform field. G3 catches the backsplatter of electrons that hit the plate so they do not go back and much up the screen.

If G1 allows more electrons to pass the other elements see an increase in electron flow because there are more electrons stopping by for dinner. Whether one element or the other has more voltage on it or not they all see more so they all act in phase.

[PRR]

> Could somebody spice this design?

Be very wary. Most SPICE models do a pee-poor job of modeling G2 current in unusual circuits.

> plate and screen currents should be in phase as far as I can see.

That's my quick-guess too.

Also agrees with how the screen-drop system works. As top-tube plate pulls down, bottom-tube G1 goes negative, implying that top-tube G2 current increases.

There's odd G2 effects. Especially in true tetrodes. Or at violent voltages. But most are so odd they could not give "linear" output with useful level.

I know there are unusual tubes which can give non-inverting gain, but 6J7 isn't the type.

The 22K cathode resistor suggests he thought he was building a constant-current source; but the 4.7u cap bypasses that idea.

[printer2]

This one was bugging me some, so I looked into it. It has a brief description in the Radiotron Designers Handbook 1954, fig 12-29C.

Seems my memory is going and I have seen this circuit before.

http://74.220.215.222/~neampfes/ampgarage/forum/viewtopic.php?p=226140&sid=b2671392f7302694f25c1d7a82bf4025

[tubeswell]

Yes TFT printer2 - knew I'd seen that somewhere before. Guess its one of those 'always puzzled me' things.

My first line of thought when i came across this was that it somehow works like an unbypassed screen (because hey! that's what it is) and the screen current and plate current are out of phase (because pentodes are constant current devices, are they not?). Discussion welcome on this BTW.

[printer2]

I've got no time to muss over this anymore. As they say, one test is worth ten theories. Hope you do not mind me taking liberties of using a 6AU6 rather than a 6SJ7 and not driving some output tubes. Also used 0.022uF caps and 270k resistors rather than 0.047 and 220k. Already had them in my breadboard. So is it the same? Close enough for me.

So to cut to the chase, I got 35V P-P on the plate capacitor, and 10V on the screen cap. Running with a 200V supply. Not a good balance between phases don't you think? Mind you, give you a big fat sound. Well that is if there actually was a phase shift. Good old ALT function. Two traces, no difference. Well one is bigger than the other, but they are both in phase.

Do you think I screwed up the phase shift not using 0.047uF and 220k? Not going to try it so I may never know.

[sluckey]

Good old ALT function. Two traces, no difference. Well one is bigger than the other, but they are both in phase.

But how were you triggering the scope? If trigger was set to internal, then as the vertical input is alternated, so is the trigger and the channels will appear to be in phase even if not. But if you were triggering externally or say on channel 1 or channel two, the phase difference will be valid.

This is the most unusual phase inverter I have ever seen. I don't have any memory of ever seeing it before. I'm inclined to store it in the round file along with a grid leak biased preamp circuit. 

[printer2]

Good old ALT function. Two traces, no difference. Well one is bigger than the other, but they are both in phase.

But how were you triggering the scope? If trigger was set to internal, then as the vertical input is alternated, so is the trigger and the channels will appear to be in phase even if not. But if you were triggering externally or say on channel 1 or channel two, the phase difference will be valid.

This is the most unusual phase inverter I have ever seen. I don't have any memory of ever seeing it before. I'm inclined to store it in the round file along with a grid leak biased preamp circuit. 

Triggering off Ch 1.

[Mats]

Hi
Pentode, 6AC7, pin 3 is connected to "earth" and not to Cathode is that something to do with this kind of PI ?
/Mats

[printer2]

Could the circuit just have been a mistake and nobody checked it out? Not like they describe how it works. Now if the plate resistor was 22k rather than 220k and the lower cap was connected to the cathode resistor this would be a simple cathodyne. Oh right, the cap that bypasses the lower resistors. Still think it is wrong and upping the capacitance will not change anything.

[tubeswell]

Me wonders whether the screen compression in an unbypassed screen is the thing that's 'in phase' with the plate. Is it not that case that even with an unbypassed screen, the screen current is in opposing phase to the plate current (cos that how pentodes work right)? Because the screen is still at some +ve potential pulling electrons away from the cathode, even when the plate voltage swings 'down' right? (They are constant current devices are they not?) The only way that a screen in a pentode/tetrode gets in phase with the plate is if its (triode) connected to the plate. Surely the internal screen resistance and the screen resistor has got something to do with the relative amount of screen current as well?

[printer2]

Anybody have a spare 6SJ7 lying around? Should be pretty easy for someone in this crowd to try out?

[PRR]

> a 6CA7, read el34, as the pentode

Again read.

6AC7, not 6CA7.

http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6AC7.pdf

[DummyLoad]

did anyone notice the nice near-flat transconductance curve of the 6AC7 when the screen is run un-bypassed, or that the spec GE spec sheet tells us there is also a nice extended g1 cut-off characteristic when the screen voltage applied is not fixed?

perhaps this characteristic is why the 6AC7 tube is specified for the application?

do most sharp-cutoff pentodes exhibit similar behavior when biased likewise?

http://tubedata.milbert.com/sheets/093/6/6AC7.pdf

--pete

[tubeswell]

Found this 1948 article by Peter Sulzer (See pages 10-11) http://www.americanradiohistory.com/Archive-Communications-Magazine/Communications%201948%2008%20August.pdf

{updated URL: https://www.americanradiohistory.com/Archive-Communications-Magazine/1940s/Communications-1948-08.pdf --- PRR, 2019}

[jazbo8]

Just posted over at MEF as well, but I could not get the phase splitter to work with EF86 (could be my model as well), perhaps there is something special about the 6AC7, which I am not able to locate a SPICE model with a separate suppressor grid yet, if anyone has one, please post a copy.

[tubeswell]

If nobody as a 6AC7, national valve museum lists 6AJ7, CV660, CV747, CV846 and CV849 as 6AC7 equivalents.

[printer2]

So if the circuit only works with a specific tube I would guess that is why we have not seen it in much use. How about different part values/voltages for other tubes? Anybody with free time? The article gives a good indication of what you are looking for. I have too much on my plate for now to play.

[jazbo8]

So if the circuit only works with a specific tube I would guess that is why we have not seen it in much use. How about different part values/voltages for other tubes? Anybody with free time? The article gives a good indication of what you are looking for. I have too much on my plate for now to play.

Tried to simulate with 6SJ7/6J7 with what I had on hand (suppressor tied to cathode), and could not make it split... I am sure the circuit works as claimed, just not with the SPICE models I have at the moment, more work required... Will post some results if something useful come up.

[DummyLoad]

IN THE NORMAL METHOD of operation a single-stage amplifier will produce a phase reversal between the input and output voltages. Pentodes when operated in this way have a screen-grid voltage supplied by a source with a very low internal impedance. It is known that the screen-grid supply impedance will produce a degenerative effect, decreasing the gain of the amplifier.'  Further study has indicated that this effect might be increased to the point where the amplifier gain would he zero, and that a still greater increase in screen-supply impedance would cause the amplifier gain to increase from zero, but with opposite phase.

so what mr. sulzer is saying is: if you move the supressor grid and connect it to ground, add enough impedance to the screen supply, you've manipulated the behavior of the space charge between the elements so that the plate phase will follow the phase of the control-grid and the screen grid will then act similar to a conventional amplifier and invert phase. so now one has a single tube phase-inverter with gain...OK, so what's so wrong with a CC stage coupled to a concertina? it's a single tube PI with gain... in fact, more gain than the pentode PI mr. sulzer presents.    

SPICE model with a separate screen grid yet, if anyone has one, please post a copy.

you'll probably need to write your own. perhaps someone at duncan amp pages or prodigy pro has something you can tweak? the suppressor grid in this bias scheme is another element that needs to be considered along with it's manipulation of space charge the author mentions as why this bias scheme works; in short, wouldn't this SPICE model need to consider the effects of the suppressor grid as well?

i'm no engineer, but my thoughts are this is something one should probably build and experiment with in the meat-world, acquire as much understating of the physics, and then build the SPICE model. another thought is that this should work with any pentode; just a guess, this may work better with high-slope sharp-cutoff pentodes. lastly, i wonder how a high-slope pentode such as 6CL6 or perhaps even the smaller high-slope power pentodes such as the EL84 would perform...  

--pete

[PRR]

> if you move the supressor grid and connect it to ground

Not just "ground". *Strongly Negative* of cathode and G1.

You need the G3 connection; otherwise I suspect the tube type is not that critical.

You may have real trouble modeling it. Tubes are just a few bucks: build it!

Agree that the performance is not terrific, bias may be fussy, there's other one-bottle drivers which cost very little more.

[printer2]

OK, it is still in my breadboard. Supply 200V 220k plate, 120k screen 1.5k cathode, 0.022u caps, 4.7u bypass. 6AU6

22k 'tail' resistor
Anode....152V
Screen....188V
22k...........6V
Anode...35V p-p
Screen....10V p-p


51k tail...6.7V, smaller AC but no inversion, reduced same amount.

100k tail...7V, smaller AC yet but no inversion, reduced same amount.

500k tail....7.8V, smaller yet

1.2M tail.....9V, no inversion
Plate 194V
Screen 197V


Maybe the 6AU6 needs a different cathode resistor? Didn't even to think to do any calculations, looking at graphs, parts are beside me I am sticking them in, just want to see a change.

Cathode to 220 ohms, max voltage at tail with 500k (pot was a 500k pot) 7.8V

Cathode to 10k, max voltage with 500k, 7.6V

Under no condition has the signals been out of phase.



Take it for what it is worth.

[tubeswell]

6AU6 - sharp cutoff pentode, high transconductance - sounds promising. Did you try biasing the suppressor grid?

[sluckey]

Take it for what it is worth.

Exactly!

  If this circuit had any real world value, the market would have been flooded with it 60 years ago.

[printer2]

6AU6 - sharp cutoff pentode, high transconductance - sounds promising. Did you try biasing the suppressor grid?

The suppressor to ground as in the schematic. I thought I would get more voltage with the higher resistance values but not much luck as shown in the data. I am busy working on an amp right now so I didn't even think of how it should work or what I should adjust to make for a relevant test. I just plugged in parts, took down readings, and then got it off my workbench. I need to finish the current amp I am working on by tomorrow, I really should not have bothered with this as it was not a priority.

I was thinking of hooking up a separate power supply to the G3 and driving it negative but that will be for another day I guess.

[tubeswell]

Take it for what it is worth.

Exactly!

  If this circuit had any real world value, the market would have been flooded with it 60 years ago.

But that's defeats the point of endless and protracted discussion involving obscure and pointless fascination with obsolete technology! ;-)

[sluckey]

But that's defeats the point of endless and protracted discussion involving obscure and pointless fascination with obsolete technology! ;-)

I totally get that side of it. And I am entertained. Just ain't gonna smoke any rosin. 

[printer2]

And I am entertained. Just ain't gonna smoke any rosin.  

Yeah I miss those days.

[tubeswell]

And I am entertained. Just ain't gonna smoke any rosin.  

Yeah I miss those days.

But the point about mid-life is to nostalgically re-live those days all over!

[sluckey]

But the point about mid-life is to nostalgically re-live those days all over!

Another good point. Since I retired there are no more random drug tests. Hmm. Let me re-evaluate! 

[PRR]

Current is generally too low. We need to jam the G2, so it against G3 diverts current away from plate.

Take the 1.5K grid-cathode bias resistor way down. 470, 220, like that.

You want 40V-60V at cathode (i.e. G3 is ~~50v negative of cathode).

> market would have been flooded with it 60 years ago.

Agree.

[printer2]

Current is generally too low. We need to jam the G2, so it against G3 diverts current away from plate.

Take the 1.5K grid-cathode bias resistor way down. 470, 220, like that.

You want 40V-60V at cathode (i.e. G3 is ~~50v negative of cathode).

> market would have been flooded with it 60 years ago.

Agree.

And that is what I thought. I put in a 220 ohm and I did not get the current and the voltage drop i was looking for on the tail (Hope you don't mind me calling it that). That was when I started doubting myself and threw in a 10k just because I was pissed. I'll have to look at the breadboard to make sure it is all hooked up right, stranger things have happened than pushing a wire in the wrong hole. Pretty sure it was all right.

[jazbo8]

Nothing to report on the Sulzer PI yet, but I did manage to verify my pentode model using another interesting suppressor driven pentode PI found here. I also found the original posts by David Jones on his "Angie" amp, but it is not clear if he ever got it working...

Here is the simulation result with Darius' PI with the 6AC7, which seems to work just fine, suggestions on how to get the Sulzer circuit to work is appreciated. I must admit that sometimes it is more interesting to work on a quirky circuit than the try-and-true ones - file it under the "useless pet tricks department"...

pentode: v(7), screen: v(8)