Long Tail vs Paraphase?

[GLPro]

Will the paraphase configuration render me more output swing than the Long Tail cathode coupled PI? Didn't Ampeg use the paraphase PI's also. Using a 6LS7 to drive 2-6v6's PP. Is one necessary over the other? Thanks

[Dave H]

Will the paraphase configuration render me more output swing than the Long Tail cathode coupled PI?

I would think so. The paraphase will only have a volt or two across its cathode resistors compared to tens of volts across the LTP tail resistor which reduces the max output swing of the LTP.

[GLPro]

Yeah, I've seen that Ampeg used the paraphase to drive the same 6v6 pair in their J12 along with 120k load resistors; I've not been able to achieve more than 12.5 W across an 8 ohm load before clipping (driver) with the LT driver. I also didn't want to rebuild the driver without some conformation with anyone who might already know the outcome. Perhaps that's the solution. Thanks

[pdf64]

I've never used a 6SL7 but it would be unusual for a properly designed LTP to have insufficient output voltage swing to be unable to fully drive a pair of 6V6 (in AB1).
What makes you suspect that's the cause of your low output?
Have you tried scoping the waveforms?
Pete

[GLPro]

Yes Pete, the output wave squares when the inverted side (positive going) of the LT starts to round off; Actually, I've seen this do just fine on most Deluxe's; but they are using the 12AT7, puzzled with the 6SL7 PI

[pdf64]

'the output wave squares when the inverted side (positive going) of the LT starts to round off'

That may be caused by the 6V6 hitting grid conduction and diode clamping the wave, rather than the LTP running out of headroom per se.
You could prove that by monitoring the LTP output without the power tubes fitted, or lifting one leg to the coupling caps between the LTP and power tube grids, and seeing how much output signal swing you then get.
Pete

[GLPro]

Well, at 1vac in to the PI, there is 27vac (RMS) swing on the output of the coupling caps; 6V6 plates at 427v with a -37v bias (23mA); perhaps a little high

[Gingertube]

Well, at 1vac in to the PI, there is 27vac (RMS) swing on the output of the coupling caps; 6V6 plates at 427v with a -37v bias (23mA); perhaps a little high

26 VRMS is 37 Volts peak - that is the 6V6 grid (bias at -37V) is trying to go above 0 (positive) when the PI tries to put out more than 37V peak. pdf64's hypothesis is proved. The PI output is being diode clamped by the 6V6 input.

You won't get any more output from teh PI regardless of what splitter you use.

The paraphase splitter in the old AMPEGs actaully works pretty well with 6SL7 - mind you, when I put the 6SL7's from an old EchoTwin I was restoring thru' my AVO MK3 tube tester they measured between 10 and 15% of new (gm) which was why the paraphase splitters in that amp were bad. NOS tubes and it ran beautifully.

Ian's sweeping generalization:
People seem to think they need to change output tubes far more often than they really do and to change preamp tubes far less often than they really do.

Cheers,
Ian

[GLPro]

Thanks guys, appreciate the discourse; will try a couple other changes..

[R.G.]

One thing not to lose sight of is that the split-load phase inverter has no gain, and the long tail does.

The amount of signal an output tube can take is limited. Output tubes are biased somewhere close to cutoff, the voltage that turns the tube completely off. When the grid-cathode voltage gets up to 0V, grid conduction starts. So the peak signal that the tube can amplify cleanly is a bit less than the cutoff voltage. More than that runs it into grid conduction. The grid bias voltage is a bit smaller than the cutoff voltage.

So signal peaks much bigger than the bias voltage are strongly distorted. With the common bias for, say, a 6L6 at -35 to -40V, peak to peak signals bigger than 80-100V are strongly into distortion. You probably don't need drive signals bigger than that.

This is well within the reach of a long tailed pair unless it's running from a quite small power supply voltage, or is itself biased with the cathodes within 100V of its power supply.

So with the LTP, you get almost as much signal swing as the prototypical paraphase as long as you bias it properly, and you get gain before the output tubes.

[printer2]

One thing not to lose sight of is that the split-load phase inverter has no gain, and the long tail does.

But that is only if you assume the split load as one tube rather than the norm which would include the gain triode before it. So with full gain from one triode and unity of the second is not much different than a LTP. If the LTP can use two triodes why not the split phase?

[soundmasterg]

The cathodyne or split phase, when counted with the driver stage usually in front of it, can output as much gain as the LTP but it needs a higher B+ to do so. The LTP being a diff amp has a different tonal character to it when it distorts too that is somewhat iconic. The cathodyne to me sounds really crisp and clear up until when it distorts, and then the character of the distortion isn't very nice without major tweaking...whereas the LTP has a consistent tonal character in all its ranges of operation.

Greg

[ReadyTeddy]

Using a SS current source such as an LM334 instead of a tail resistor in a LTP makes lots of sense. It would allow equal plate resistors, allow greater output swing because you wouldn't have to waste lots of voltage across a tail resistor, would simplify the LTP, and, I think, wouldn't change the sound. Has anyone tried it or heard about someone using this configuration?

[JoeM]

There was an article in Glass Audio years ago where the author used a CCS as the tail in a LTPI. I dont recall than it made a huge difference, I think it might be more of a HiFi type of thing. I'll see if I can find the article.

[R.G.]

But that is only if you assume the split load as one tube rather than the norm which would include the gain triode before it. So with full gain from one triode and unity of the second is not much different than a LTP. If the LTP can use two triodes why not the split phase?

That's true.

ReadyTeddy
Using a SS current source such as an LM334 instead of a tail resistor in a LTP makes lots of sense. It would allow equal plate resistors, allow greater output swing because you wouldn't have to waste lots of voltage across a tail resistor, would simplify the LTP, and, I think, wouldn't change the sound. Has anyone tried it or heard about someone using this configuration?

JoeM
There was an article in Glass Audio years ago where the author used a CCS as the tail in a LTPI. I dont recall than it made a huge difference, I think it might be more of a HiFi type of thing. I'll see if I can find the article.

I get nervous about using ICs in the signal path of a tube amp when they're DC connected. I think the law of diminishing returns applies here - you could get most of the value of a high performance CCS by using a simple CCS instead of a resistance tail.

Something as simple as a MOSFET, a zener, and a few resistors would do a good-enough CCS and be more voltage-tolerant.

It's not hard to get a CCS with more internal impedance than the resistor string of a typical LTP in a tube amp.

Diffamps do have smoother entry into clipping, even with bipolars. It's worth experimenting with.

I'd have to think about what happens to the feedback path on amps using the cathode string as a target for output feedback.