Phase splitter/inverter maths/calculations!

[stokes]

Nice rant Alexander.But quite unimpressive.The idea of this forum is to exchange ideas,to learn and to help the new guys along,no?We all have different approaches and dont always agree,which is actually beneficial most times.Now I'm not saying that trying to be innovative is a total waste of time,but tube amps have been around along time and coming up with something that hasnt been done before is tough for even the seasoned pro.I'll give you an example.Back in the '70's I built a 5E3,before anybody called them "clones",it was just "something to do".I showed it to an older guy,who has long since past away.He was a radio designer in the '40's and '50's for RCA,knew nothing about guitar amps per se,but he hated the sound of it,way too distorted.I showed him the schem I used and he drew me up a "mod" to the PI that he said would clean it up,and it did.Now fast forward to the '90's and the internet age and some guy named Paul C comes up with a new "innovative" mod.Now I dont know Paul C and I am in no way knocking him and he may very well have come up with this idea on his own,but it was exactly what I was shown 20 something years ago.I have a dozen amps I have built for myself and my sons that are based on various established designs,but none of them would be considered "clones" as they all have been modded extensively,but they are still not any earthshattering innovative "originals",they are tube amps.I have built and sold many of them to local guys and even tho they may think they are "special",to me they are just basic tube amps.So now if you've cooled down,maybe you will share some of your "non-clone" ideas with the rest of us and try to be somewhat constructive.I've looked at your web page and I see you refer to "Point-to-Part"TM,I assume you claim some kind of trade mark for this as being "new".I've taken apart many '50's PA amps to convert,that were built just that way,every capacitor was wired from tube socket to tube socket,very few wires,Chicago Webster is one that comes to mind,not a tag board or circuit board in sight.No you dont see any builds with my name on them,but not to single you out either,I havent seen the name Retrodyne anywhere but your Myspace page.Again I am not knocking your amps,I've never heard one,I am sure they are fine amps,but if a suspended circuit board is the wave of the future,why are there still so many fine old amps still working.Okay I'll stop now.I really do wish you the best with your amps,but dont be so quick to rant about the old ways,if it werent for the old ways,we wouldnt be here to discuss this today.

[Steve Conner]

You guys crack me up with your animated sarcasm meters.

Anyway, here's how to design a LTPI. You start with the values of Rp and Rk that make the tube work optimally as a regular RC coupled amplifier stage. These may already be in the tube datasheet. If they're not, maybe the tube was never designed for RC coupled audio service? Remember the idle cathode current too.

Then, you use the same Rp as above, but cut the Rk in half, because there's only one Rk shared between the two tube sections.

Now, you need to choose the tail resistor. This is a tradeoff. A bigger tail resistor gives better balance, but hurts your output voltage swing. The more volts you burn in the tail resistor, the less there is left between plate and cathode for the tubes to work with.

You probably want to aim to get about 100V across the tail resistor, I guess, with the combined cathode currents (which you remembered from the first step above) passing through it. With this knowledge of V and I, Ohm's law will give you R.

Once you put all this together, you should be in the ballpark, and you can tweak further to taste. Use an oscilloscope to check that it's putting out enough swing to ram the power tube grids right up to 0V. The power tube grids should clip before the PI does. Also check the balance, and change one of the plate resistors (to... shock horror... 82k?) to improve it if needed. If you don't have an oscilloscope, buy a 5E3 kit from Bruce! ;-)

My contribution to the cloning vs. innovation debate was the Ninja Toaster

http://scopeboy.com/toastpix.html

which might even still be in the Ampage homebrew gallery I built it in 2000, still play it almost daily, and have gigged with it several times.

It is slightly relevant, in that I designed it to take either a 12AT7, 12AX7 or 12AU7 in the PI. So I ended up with a PI circuit that had a balance pot:

http://scopeboy.com/poweramp.gif

I'm currently working on another original design, but I've got a lot less tinkering time since I graduated. This one is all tubes, no MOSFETs or op-amps this time...

[Old Tele man]

It is slightly relevant, in that I designed it to take either a 12AT7, 12AX7 or 12AU7 in the PI. So I ended up with a PI circuit that had a balance pot:

http://scopeboy.com/poweramp.gif

...two suggestions:

(1) to "balance" the idle DC-characteristics of the two triodes, replace the two separate 1.0K-ohm cathode resistors with a single, common 2.0K-ohm pot with the wiper connected to ground.

(2) to "balance" the dynamic AC-characteristics of the two triodes, replace the two separate 47K-ohm plate resistors with a single common 100K-ohm pot with the wiper connected to the same, original, plate voltage source.

...YMMV.

[Steve Conner]

I built this thing years ago, and it works fine. The DC conditions are naturally balanced anyway, and the balance pot deals with the AC conditions.

It would be hard to find a pot rated to take the DC voltage and dissipation of a plate resistor, and IME, pots used as cathode resistors go noisy sooner or later.

[MattT]

Coming up with 'new' circuits is fun and you do learn...but I've often asked myself; "Am I building this new circuit to just to get the same tone that some older design gets?" Because let's face it...the older designs defined what we consider good tone. I'm sure my 'new' circuits aren't really new anyway...several times I've later found something extremely similar to what I had tried earlier.

And with all these different circuits I try...I still end up preferring the old standards. Because they are our definition of good tone. Which is probably why so many can make a profit selling 'clones' of old designs.

[Earl]

I agree Matt. It's fun to mess with stuff, but I doubt there's anything tube-wise that's gonna be better than the original 'foundations' we love to hear.

Did you know that the ultra-linear (tapped primary) output circuit was patented in 1936? .. :-)

[Old Tele man]

...which was the same year that the 6L6 beam tetrode was developed (April 1936) by RCA.

[slideman82]

The Vox AC50/2 uses a larger tail resistor (47K), which balances the phase inverter outputs much better than the standard Marshall 10K value, so equal value plate resistors work fine. The Vox AC50/4 uses local negative feedback to balance the second input (via a voltage divider from the outputs back to the second input), so it isn't a standard "long-tail pair" inverter.

Randall Aiken

Thanks, mr. Aiken! That will help even more! I have to take a look at the schematics I've got. I think I've only got the /4. Anyway, I'll mod it a bit. Regards.

[slideman82]

well, it seems I can't find this schematic, or am I wrong? I've got some of them: Piazza's drawing, the one from 1976 and the one from 1965 (I think they just differ in the first stage), and there is an AC50/4 on firebottle, I think, but coud be wrong. I'm looking to the one with that paraphase type splitter. It must be the 1965 circuit, isn't it?

Thanks to all who share their knowledge.