Paraphase Splitter - why is it unpopular?

[ray_moth]

The reason I'm interested to learn more about the (floating) paraphase is that it seems to be a handy and simple way of obtaining good gain, balance and voltage swing.

The LTP splitter, which I use at present and which also gives good gain and voltage swing, is a less simple solution because it needs a CCS in the tail (with attendant negative supply) to achieve good balance.

The paraphase splitter appears to be a thing of the past and I'm wondering why?

[Tubes4e4]

Originally posted by ray_moth
The paraphase splitter appears to be a thing of the past and I'm wondering why?

Comparing to concertina and LTP, the main advantage of the paraphase is that it obviously gives the most voltage gain of that trio.

The drawback is, that it obviously has the highest distortion of that trio. F.e, the concertina is low distortion by design (100% local FB), and in an LTP, much distortion cancels out due to its PP nature.

The paraphase being just two consecutive grounded cathode stages, distortion, especially uneven harmonics, sums up.

Due to the consecutive stages and the coupling caps involved, the phase behaviour is worse than concertina or LTP, too, and obviously it cannot be the same for both "paths" after the splitter. This might add to stability problems with gNFB, which are hard to tame since the phase errors are not symmetrical after the splitter.

Summing up, I guess paraphase was popular mostly because of its high gain and because it doesn´t need additional -ve voltage (or CCS) or elevated heaters, and so on, to get reasonable performance.

Tom

[lndm]

Yes. The paraphase phase splitter does seem easy (and inexpensive) to implement, but the thought of the upper and lower resultant portions of my waves going through different amounts of circuitry doesnt sit well.

The frequency extremes may also exhibit differing levels of phase shift.

[ray_moth]

Thanks for your replies, which make sense.

[audiobot]

I've used it and it seems to work fine for me. Here is a quote from Joseph Diamond, a development engineer for UTC from days goneby, re the circuit as used in his amplifier design,

"The phase inverter used is of the paraphase or anode follower type, and is difficult to surpass. For high frequency applications this circuit has given excellent balance to 1 megacycle."

This refering to a 6SL7 used as input volt amp and paraphase splitter. I will note that my design was a two stage with no feedback and Diamonds was a three stage with local feedback between the power and driver stages but no feedback to the splitter/input.

Easy enough to try, then you can form your own opinion.

Michael

[Tubes4e4]

Hi Michael,

Easy enough to try, then you can form your own opinion.

.

I did

But leaving generalisations aside, I agree. There is no "best splitter" topology anyway, as usual, it just depends on the goals being set to chose the topology that meets the requirements best.

Tom

[Dave Cigna]

... the thought of the upper and lower resultant portions of my waves going through different amounts of circuitry doesnt sit well.

I've heard this argument before, but it seems to me that a LTP does the same thing when only one grid is driven. The 'undriven' side is running as a grounded grid (which is not the same mode of operation as the driven side) and gets its signal only after it passes through the driven side.

I admit that the LTP looks a whole lot more elegant and somehow more balanced than the paraphase on paper. It almost certainly produces more symmetric outputs if a CCS is used as a tail. If that's what you're going for then the LTP is the obvious choice.

Incidentally, I have tried the paraphase and found it sounded surprisingly good despite its apparent warts.

-- Dave

[richwalters]

Originally posted by Dave Cigna


Incidentally, I have tried the paraphase and found it sounded surprisingly good despite its apparent warts.

-- Dave

I agree. The bonus is a 3 stage amp. Most of the para designs worked with an ECC83 which I basically loathe. Short of a complete bench redesign, has anyone tried it with a better tube ?

The 7199 is a good runner for cathodyne but I find pentodes fussy to setup between makes and despite lower gain behave better strapped as triodes.

[lndm]

Originally posted by Dave Cigna

I've heard this argument before, but it seems to me that a LTP does the same thing when only one grid is driven.
-- Dave

Yes, it _seems_ that way. The other triode is driven by its cathode, which is using the same resistor as the first triode. They are intimately connected. I feel this intimate connection is the phenomena that makes local feedback a different proposition to global.

I have used the Williamson phase splitter before. This is a standard split load phase splitter followed by a pair of otherwise independent voltage amps that share a tad of cathode feedback. Elegant, fairly simple in design, even if it takes a couple of extra triodes, and it drives well.

[audiobot]

OK guys, I have to fess up. I was wrong on both points. First, Mr. Diamond did take overall feedback to the cathode of the first section of the 6SL7. Second, my build is actually a sort of funky hybrid. It uses a 6N7, a twin triode with the cathode common to both sections, and the cathode resistor is only partially bypassed. So I guess you could say it is a short tailed paraphase inverter. Try it you'll like it.

Michael