6S19P Push-Pull Amplifier

[sorenj07]

I've been making noises about a 6S19P amplifier and have worked up some more schematics. I'd considered direct-coupling the SRPP's to a 5687 gainstage, but I'm not quite sure that's needed. Besides, the 5687 is a current hog. The setup as it is takes a bit under 6A at 6.3V.

Iron is a pair of 10K transformers taken from a Hammond PR40 organ. They were used with 6BQ5's so I assume they can take this role.

A note on the LTP supply: I want to drop A LOT of voltage in the tail resistor of the LTP (down to 400V, as seen in the Dissident Audio 6AS7 design) so that it acts very much like a constant-current source without actually being one. That's why I have a line of 650V or so. Or is this impractical? Should I put a big RC filter after the rectifier or something?

Finally, I should give credit to the DiyTube Eiclone for the bias adjustment scheme. How much current should I assume? All the resistor values and wattages kind of hinge on this, which is why I left them out.



What do you think?

[kevinkr]

You have a little work ahead of you to get good balance in your SRPP phase splitter. I used these once upon a long time ago in line stages I built and I can tell you a ccs in the cathode circuit is your friend. The current configuration will be highly asymmetrical ac - wise. (Not to mention the input as drawn doesn't quite make sense unless differentially driven and is missing a needed ground connection. If not the bias network isn't right either. ) I would model this circuit in spice if you can, and for good performance a ccs should have a compliance voltage of about 12V or so across it - either use a -12V supply or ac couple the inputs and raise them above ground by the required amount. (Say 12V) Look at how some typical LTP splitters using direct coupling to the preceding stage are configured as a possible inspiration.

Should you decide for some reason to go with a very high negative supply instead of a ccs you must clamp the cathodes of the srpp that are connected to the tail resistor to a safe voltage until they warm up.. Hint: Diodes and neon lamps in series. This may or may not be practical depending on the current required, the voltage required would certainly give me pause.

Done carefully it is possible to get good balance with this topology.

The power supply design is another matter altogether.. You've got at least two, two diode drop dead shorts across the windings. You really will need to rethink this or use several separate transformers.

On reflection Spice really is your friend as it would show you the design flaws prior to building the thing, of course peer review as found here helps in that regard as well..

[sorenj07]

Thanks a lot for the advice. I'm having trouble visualizing how a short would be created, but I accept that it's entirely possible given my scanty knowledge of EE. What course would you recommend for a power supply then? The 6S19P's want either 220V for fixed-bias or 305V for cathode bias. Is 300V or so enough to power the 6N2P's to the voltage swing required, or are we back to a 5687 gainstage tacked in between? Could I try a 6CG7 as well?

[Yvesm]

Hi Sorenj,

Do not confuse gain and swing.

The bias on the 6S19P being, according to your schemo, some 85Volts means you need to apply 170Vpp of signal on each grid.
This is near impossible with just 220V power supply unless you use an interstage transfo, but this is a completly different approach.

Considering the 6AS7 PP schematic you refer to, note that the phase spiltter is not a LTP but a paraphase design.
Just the common cathode resistor is not bypassed but it does not participate so much to the phase split process.

Note also that the SRPPs are powered at 400V while the final stage runs at 200V.
This allows for the large grid voltage swing needed.

If you target an 1V rms sensitivity (that is 2.8V pp) the gain you need is (170/2.8) = 60.
This is obtainable with the 6N2P in an SRPP configuration but you have no headroom left for gNFB (superfluous here IMHO).

Yves.

[sorenj07]

This makes sense, and sorry for the errors. I knew something wasn't quite right with that "LTP" setup.. I can't wait to try and build something, but the OPT's I'm planning on using probably aren't the best call. I was wondering if the "real deal" would come and discuss his own amp. How is it holding up, by the way? I can't wait to try and put something together but I have very little cash and a cheaper, smaller triode amp to build with my cheaper, smaller OPT's... Hopefully I can fight my way up that price vs performance curve

p.s. yves si c'est jamais plus facile de t'expliquer en francais, mon pere est francais donc je me debrouille assez bien avec la langue.. quelques mots clefs: polarisation, resistance, lampe, condensateur, transformateur mais comment dit-on "impedance" et "swing"?

[kevinkr]

Originally posted by sorenj07
This makes sense, and sorry for the errors. I knew something wasn't quite right with that "LTP" setup.. I can't wait to try and build something, but the OPT's I'm planning on using probably aren't the best call. I was wondering if the "real deal" would come and discuss his own amp. How is it holding up, by the way? I can't wait to try and put something together but I have very little cash and a cheaper, smaller triode amp to build with my cheaper, smaller OPT's... Hopefully I can fight my way up that price vs performance curve

p.s. yves si c'est jamais plus facile de t'expliquer en francais, mon pere est francais donc je me debrouille assez bien avec la langue.. quelques mots clefs: polarisation, resistance, lampe, condensateur, transformateur mais comment dit-on "impedance" et "swing"?

Something like: excursion de la tension? I think impedance is the same in French, but unlike you, je ne parle pas francais courament.. Il ya plus de 30 ans depuis le dernier fois que j'avais parle francais frequemment. You can tell eh?

[Miles Prower]

Originally posted by sorenj07
What do you think?

The Vpp looks to be awfully thin. You're not gonna get the drive voltage the finals need from that version of an LTP. You have a dead short across one of the input tubes that'll put it into cutoff and make gobs of distortion. The input side of that also needs a DC blocking capacitor so that whatever gets connected to it doesn't either short out the bias completely and/or poof from being over-volted.

Furthermore, I'm not a big fan of the SRPP. It's a wierd-looking sucker that's gotten trendy because of that, but the thing will only be balanced for one specific load for the series resistor. That's OK for what it was originated to do: drive a T-line at low SWR and its characteristic impedance. Let the load depart from the optimum and THD rises rapidly as it goes out of balance. As a gain stage working into a Hi-Z load, it also ain't much. I'd cascode the LTP if you really need the gain, or just replace the "fancy" with a plain vanilla LTP. I'd also use a solid state CCS as an active tail load. The high impedance of a CCS helps greatly with AC balance and harmonic balance as well. It's better than trying to force balance by unbalancing the plate resistors.

I would also include grid drivers. That's especially important when the finals are triodes with their higher Cmiller. You'll need to slam some current into that Ci if you are to avoid slewing.