Floating paraphase stage

[Alan0354]

This is a continuation of this thread. It went totally off subject, it's time to start my own thread. Just read through this and join in:

http://music-electronics-forum.com/t35845-2/

The whole thing started with this post:

What do you think of this PI Alan with its unity gain tube "opamp"

[ATTACH=CONFIG]27813[/ATTACH]

[Dave H]

Seconds out round two!

[Alan0354]

Seconds out round two!

I don't look at it this way, I think it's a good discussion.

[Dave H]

OK then. I was only joking. It's gone midnight in the UK. I'll be back tomorrow.

[Alan0354]

OK then. I was only joking. It's gone midnight in the UK. I'll be back tomorrow.

Great, I know you know your stuff, I like the post you wrote about the distortion etc. I just hope you can stop and see my presentation so we can talk more. Who knows, I could be wrong, but I need to know why as I don't see it.

As I emphasize, this is a discussion, not who's right or wrong.

[Alan0354]

What do you think of this PI Alan with its unity gain tube "opamp"

[ATTACH=CONFIG]27813[/ATTACH]

I did not spend a lot of time in this theoretical design. This is a differential pair, I don't think you can isolate the two triode. This is completely different from the paraphase PI in the OP's circuit where there is a big bypass cap on the cathodes.

The way the different pair works here is in the idle situation, it depends on the matching of Vgs between the two triodes. Assume Vgs characteristics is identical, the tail current developed across the 470 ohm resister is split evenly between the two triodes. The moment the input signal goes a little positive, the cathode of the top triode follow and goes positive, and start pulling the junction between the two cathodes up, plate current increase on the top troide BUT the current DECREASES in the bottom triode. So the plate voltage of the top triode goes down as the plate of the bottom triode goes up the EXACT amount.

This is an explanation of differential amp, they use BJT, but the theory is exactly the same:
Transistor Differential Amplifier Circuit Description, using a single input. Dictionary of Electronic and Engineering Terms
Differential amplifier Circuit Tutorial using BJT and Opamp

The plate voltage of the two triodes depends on the plate resistance. From looking at the schematic, the two 1M is symmetrical to both of the triodes. The bottom triode see 330K +33k to ground and the top triode is only 330K. That is the only difference. The junction between the two 1M resistor has only very little voltage to feedback to the grid of the bottom triode, so there is very little feedback voltage. The less the feedback voltage, the higher the closed gain. This is not a unity gain stage.

This is a strange circuit that I would not use in real life, maybe people using this in a test in a class!!! If you really want unity gain, connect the grid of the bottom tube to the junction between the 330K and 33K. Now you take the same amount of output voltage back to the -ve input ( grid of the bottom triode). This should give you unity gain.

Well, I've tried it in LTSpice. Does that count? It is an old tried and tested circuit. I thought these design concepts came along with opamps but it turns out they are ancient. Studying these old valve circuits is an electronics history lesson. Anyway I've hijacked this thread enough. You can read about it at Merlin's - here.

I went through the article. Actually I was right. If you look at the Gain paragraph. The unity gain refers to V2 is unity gain from V1. BUT the gain of the whole stage is equal to the gain of V1 as common cathode stage!!! The whole stage is not unity gain NFB opamp circuit.

Is this article written before the modern opamp and differential amp days? This article totally disregard the theory of the differential amp. You don't need Rg and Rf to set up anything. Just the nature of the differential amp that V2 has a gain of -1 of the output of V1. In the first figure, the circuit is BALANCED, author declared Rg=Rf, there is no voltage at the grid of V2 as the voltage of the two plates are equal and opposite just by the nature of a differential stage. See the two links I provided in post #40, they explain exactly how the differential long tail pair works. This is the heart of the modern days opamp design. You'll find this circuit works exactly the same after you get rid of Rg and Rf.

Get rid of the Rf and Rf feedback and just ground the grid of V2, you'll get a symmetrical differential amp already with gain of V1. This is because each of the cathode sees ra=1/gm of the other tube. So when you look at the differential output, you get exactly the gain of V1 as common cathode stage!!!

I did not read pass this point as it really does not make sense. I don't mean I don't understand what the article said, but I don't think it's correct.

Thanks Nick.

Alan, have a look at the schematic below. It operates in exactly the same way as the one I posted from Merlin’s web site but I’ve redrawn it so that it won’t be mistaken for a differential amplifier.

[ATTACH=CONFIG]27860[/ATTACH]

This is a totally different circuit. You have C2 and C4 bypassing the cathodes, the two tubes are independent. The moment you remove the cathode cap with the two cathodes tied together like what you posted, it becomes a Differential amp. No if and buts about it. As the two links provided, it explain how the two devices interact under a differential configuration. This is apple and orange.

You might call floating paraphase, but it is totally different circuit compare to the original paraphase circuit. you cannot talk about them as if they are related. Once you remove the bypass cap, the two stages join together and become one inseparable circuit.

BTW, Your circuit definitely a big improvement over the original paraphase stage. You use the FB to stablize the gain of the second stage so you do get symmetrical output. The two points I criticized the paraphase is the second stage does not have a stable gain and can drift off from the first stage. You fix that with your circuit.

Above is how we started.

[printer2]

I have tried the paraphase circuit but not the floating paraphase yet. I like it better than the cathodyne for sort of clean amplification. I found a bit of imbalance thickens out the sound more. Kind of falls apart when driven harder though. Sort of like grid biased input, not bad at the edge of breakup but not that good when hammered. Need to do the floating paraphase soon.

[Alan0354]

I have tried the paraphase circuit but not the floating paraphase yet. I like it better than the cathodyne for sort of clean amplification. I found a bit of imbalance thickens out the sound more. Kind of falls apart when driven harder though. Sort of like grid biased input, not bad at the edge of breakup but not that good when hammered. Need to do the floating paraphase soon.

My main gripe about paraphase is the way they get the opposite phase is using an inverted stage of unity gain to generate the 180 deg off phase signal. Problem is the extra stage has propagation delay in form of limited frequency response. I tested out a typical common cathode stage with 100K plate resistor. The -3dB corner frequency is about 20KHz. At that frequency, the phase shift is 45 deg.

You can look at phase shift as either phase in deg or propagation delay. If you look backwards to 5KHz or even 3KHz, there will be a build in phase shift cause by the inverting stage. So the supposedly 180 deg differential phase signal might be 190 or 200 deg at 3KHz!!! You don't have exactly a push pull anymore!!!

I am just guessing this might be a lot more noticeable when you start clipping the amp as higher frequency harmonics are generated and the phase error is a lot more obvious with higher frequency. This will be on top of the amplitude drift with the original paraphase stage.

There is a good reason why people mostly go with true differential amp as PI in all the more recent design. Because it is more symmetrical.

[soundmasterg]

The Vox AC100 uses a floating paraphase phase inverter circuit using a 12AU7. It works well in that amp and has fantastic cleans in that application, but then due to using so many 12AU7's that preamp and phase inverter don't have a bunch of gain. I built myself an AC100 clone and the floating paraphase works well....I think the cathodyne has better cleans before distortion with more high frequencies present, and the LTP has a distinctive sound to it that is hard to be without, but the floating paraphase is another good way to get a phase inverter in an amp. In fact it probably has the potential to have more gain than either of the other two designs I mentioned and really the gain depends mostly on the plate voltage...if you make it close to what the power tubes are seeing, even a 12AU7 used as a floating paraphase can push an EL34 quad to overdrive as in the AC100.

Greg

[jazbo8]

Alan0354 wrote: "This is a differential pair, I don't think you can isolate the two triode."

In hindsight, perhaps the above sentence sums up why the discussion went off in the direction that it did... the circuit shown on Merlin's page is a modified form of the RCA's floating paraphase circuit (I mistakenly said Mullard in my earlier post). The RCA FP did not have a shared cathode resistor, so it could not have been construed as a differential amplifier stage at a quick glance. In any case, given the relative small resistance used in Merlin's FP example, it would not make a good differential amplifier as nickb pointed out on this post.

The first mention of the "self-balancing phase-inverter", aka the "floating paraphase" that I can find is in RDH4's footnote on phase-inverters and it dates back to 1938! If anyone has a copy of the original RCA application note, please share it!

For detailed descriptions on how the circuit works, we can refer to two patent filings by RCA (way after the application note came out);
I think this is the actual patent: RCA patent 1
One kinda related: RCA patent 2

Here is an article from 1940 showing the RCA circuit:


A geek with nothing better to do...
Jaz

[Alan0354]

Alan0354 wrote: "This is a differential pair, I don't think you can isolate the two triode."

In hindsight, perhaps the above sentence sums up why the discussion went off in the direction that it did... the circuit shown on Merlin's page is a modified form of the RCA's floating paraphase circuit (I mistakenly said Mullard in my earlier post). The RCA FP did not have a shared cathode resistor, so it could not have been construed as a differential amplifier stage at a quick glance. In any case, given the relative small resistance used in Merlin's FP example, it would not make a good differential amplifier as nickb pointed out on this post.

The first mention of the "self-balancing phase-inverter", aka the "floating paraphase" that I can find is in RDH4's footnote on phase-inverters and it dates back to 1938! If anyone has a copy of the original RCA application note, please share it!

For detailed descriptions on how the circuit works, we can refer to two patent filings by RCA (way after the application note came out);
I think this is the actual patent: RCA patent 1
One kinda related: RCA patent 2

Here is an article from 1940 showing the RCA circuit:
[ATTACH=CONFIG]27889[/ATTACH]

A geek with nothing better to do...
Jaz

I really have a hard time reading the article as it is blur. But if the original RCA FP does not share a common cathode resistor, it will work as intended. In fact I did reply to Dave that his second drawing works a lot better than the original paraphase because now, the gain of the second stage is controlled by NFB and it will not drift with time and temperature. I have no issue with that.

Regarding to Merlin. It uses a 470 ohm. I just did a quick calculation using 12AT7 with gm=5000, so the cathode output impedance is approx 1/gm= 1/0.005=200 ohm. So the cathode impedance of V2 is 200 ohm. The impedance the cathode of V1 sees is 470//200. It is still strongly affected by V2. There is still enough coupling even with resistor as low as 470 ohm.

Another way to look at it is if the grid of V1 goes positive, the cathode of V1 goes up following the grid. The cathodes of V1 connects to the cathode of V2. So the cathode of V2 get dragged up when the grid of V1 goes up. This, will cause the grid to cathode voltage of V2 become more negative and decrease the plate current. So when grid of V1 goes up, the plate current of V2 goes down while the plate current increase in V1. This is the very simple way of looking at the differential pair in Merlin even when the resistor is only 470 ohm.

Let me try to read the article. I just want to reply first as this change everything. I would agree with the RCA.

[jazbo8]

No worries, you are correct that even with a small cathode resistor, the stage still acts like a differential pair (Merlin also mentioned it was used to improve the output balance) - I guess I am just trying to point out that the diff amp aspect is not the "main feature" of the FP design. I think the design boils down to the plate-to-grid feedback for V2, which is the self-balancing part patented by RCA. But as you rightfully pointed out, the FP is far from ideal, so it is no wonder that it is seldom used in hi-fi or guitar amp designs (post 1950), although it still has its own following, because it sounds/reacts differently to the split load and the LTP phase inverters, anyway, just another tool at our disposal.

[Alan0354]

I printed out the RCA, it is still very blur. But I can clearly see Tube A and Tube B has it's own cathode resistor and capacitor. They a absolutely independent stages and it will work. This I totally agree.

I bet the drawing in Merlin has a typo that missed out the cathode bypass cap. Because the circuit will work as he explained in the book with the bypass cap. Maybe that's the reason I was puzzled why he talk as if there is no differential stage effect.

[printer2]

I wouldn't mind one.

[ReadyTeddy]

The 5E9-A schematic shows a poorly designed PI for all the reasons previously stated. I'm surprised ole Leo didn't know better at the time.