SRPP

[shooter]

I've been following the discussion on the 4watt critique.  I have a PCB pre that uses the srpp, channel 1 is working fine, ch2 was/is an XLR out but I want to do a PP el84 instead.  I don't really want to cut the board all up so was wondering if I can use the srpp, (snippet attached) as a pseudo PI by using the normal out, pins 3/6 and adding pin 2 as the inverted out?  got enough gain to drive the el's just need 2 outta phase signals with minimum hacking.

thanks

[PRR]

> wondering if I can use the srpp

That is not a SRPP.

It is a White Cathode Follower.

(And not optimally detailed.)

[shooter]

Thanks PRR, I googled it and found a great side by side of srpp n white by tube cad journal on the differences.  I still need an inverted signal to drive the EL's  the article gave Rplate as 1/gm.  my guess is i'll have to mess with the plate R to get similar amplitude signals??? 

[HotBluePlates]

... I have a PCB pre that uses the srpp ... but I want to do a PP el84 instead.  I don't really want to cut the board all up so was wondering if I can use the srpp ...

Ignoring the SRPP vs White Cathode Follower issue, what you need is 2 opposite-polarity drive signals. To get that, you need a phase inverter.

SRPP and the White Cathode Follower each have only a single output. There's no getting around having to install a phase inverter somewhere (or a zero-gain inverting amplifier, a la the 2nd stage of a paraphrase inverter).

Looks like you'll have to cut up the board, or get (or build) a different board.

[shooter]

Guess that's the real question, why can't I tap the inverted signal I need off the plate (pin 2), maybe split the plate R into 2.  I already have the non-inverted (pin 3&6).  i expect i'll need something besides the 2.5k that's there to develop a usable signal.  I did something similar on a previous build where i used the plate to drive a SE tube and tapped off the cathode for a "pre-out" line,(driving another PA), granted it was just a single triode n not driving PP but if my notes are correct they were outta phase and i "balanced" the plate / cathode r's for about equal signal amplitude.  Guess that's what I'm wondering here, swap the plate n cathode R's, balance the signal without losing to much signal, i have about 12vpp out, thinking i need about 5-6v P to drive the EL's as a clean amp.

Thanks for the time

[HotBluePlates]

Guess that's the real question, why can't I tap the inverted signal I need off the plate (pin 2) ...

The purpose of a White Cathode Follower is to provide an active load for the follower, to reduce output impedance. You don't need low output impedance to drive push-pull output tubes (if all are on the same chassis). So you're much better off using 2 triodes as a split-load inverter and its pre-gain stage instead, just like umpteen other push-pull amplifiers.

What is the bias voltage of the EL84's? That's the real peak output voltage you need from each output. Depending on what it is, you may not be able to derive the needed output signal with any permutation of a White Cathode Follower or SRPP stage, because you need a B+ voltage equal to 2x output signals plus enough voltage across 2x triodes for them to operate properly (i.e., amplify cleanly), plus voltage to drop across self-bias components, etc.

This is also a prime consideration of the split-load inverter, which has to develop 2x output signals, but only drop voltage across 1 triode and 1 bias resistor.

[sluckey]

You have the circuit. Just inject an AC sine wave signal into pin 1. Measure the AC signal on pin 3. Now measure the AC signal on pin 2. You may just have what you want already. If so, just add a coupling cap to pin 2. Let us know what measurements you get and we can probably get them matched up if necessary.

[shooter]

Thanks guys,
so let me get terminology 1st, split load PI is essentially single triode, inv sig from plate, non-inv from cathode?

I was thinking of starting somewhere around 30 - 40k top n bottom AFTER I measure cathode current in the current config.  I should have numbers by wed.   The B+ is 290ish at the 1st tap, bias will be cathode. The quad SE I did liked around -5.5v so i'll probably use that as a start, but was gonna run it hot, maybe 95% max plate then dial down is needed.  I think i'll have around 12vac PP for drive, (after PI loss) which, I'm guessing this;    " which has to develop 2x output signals"   is the 2X? assuming it biases below 6v?

again, thanks

[PRR]

The WCF is a double-feedback amplifier.

When you really get into it, it is a mystery.

Here's the insight you do not want to hear:

> using the normal out, pins 3/6 and adding pin 2 as the inverted out?

Pin 2 (upper plate) is also pin 7 (lower grid). The gain from pin 2 to the "normal" output pins 3+6 really needs to be high, nearly the Mu of the tube.

In a happy WCF, the signal at pin 2 is much smaller than the input or output signal. Like 0.182V versus 9.8V. (Very nearly the Mu of the 12AT7 which I used.)

When I raise the "2.5K" to try to equalize the outputs, say to 100K, first the DC bias goes all to heck because of the large drop in that resistor. When I tediously find a new bias point for grid pin 1, I *still* get 0.185V versus 9.8V. When I take the absurd value of 500K, there's no good bias point (it becomes hyper-sensitive in the sim, and won't be happy in real life); AND I still get outputs like 0.188V against 9.8V.

The two outputs you propose WILL have different gains, by roughly Mu, until you beat it out of action with absurd values.

[HotBluePlates]

so let me get terminology 1st, split load PI is essentially single triode, inv sig from plate, non-inv from cathode?

Yes, but they're almost always used with another triode in front, as the inverter itself has no gain due to the large cathode load. Any negative feedback is generally returned to the cathode of that gain stage ahead of the split-load inverter, which I tend to call a "pre-gain stage."

Look at a Fender Princeton Reverb for an example of a split-load inverter with a pre-gain stage with output stage negative feedback injected at the cathode.

The B+ is 290ish at the 1st tap, bias will be cathode. The quad SE I did liked around -5.5v so i'll probably use that as a start, but was gonna run it hot, maybe 95% max plate then dial down is needed.

Are you certain of those numbers? Mullard's EL84 data sheet shows a push-pull output stage with 300v plate and screen, 40mA of total cathode current and a 270Ω cathode resistor per tube (top of page 2). That gives 0.04A * 270Ω = 10.8v of bias, and Mullard shows total output stage drive (grid-to-grid) of 20v peak-to-peak.

There's some implied statements in Mullard's data, that you have 300v from plate & screen to cathode after the drop across the bias resistor, so your bias will be a little lower if you start with a 290v B+ and reduce the amount of plate & screen voltage by the amount dropped across your bias resistor. That will reduce the bias voltage a little, but not a whole lot. The 250v push-pull condition shows a bias of ~9.3v, once you do the same math to figure it.

Anyway, you have to have the B+, a few stages of power supply filtering, and the output stage design known before you will know the requirements for phase inverter design. Which is why I'm asking about output stage particulars  when you asked about a phase inverter.

[HotBluePlates]

For Dr Gonzo

[shooter]

Wow, a lot to digest, I see what you're saying PRR, I'm still fuzzy on HBP's retort, at the end of the day I think I'm gonna have to dermal it but i'll play around 1st when I get the amp n scope together in the same room.  I attached the schematic, all the values are straight from the kit with the exception of gain - tone control, pot on the NFB and 22uf on v1B ch1 along with the xlr out on ch2, which has yet to be confirmed other than swapping channels into the SS amp to make sure they both work n scoped about the same.  that's about the time my son n friends took the amp, they all want to know what's inside to makes that sound!!
I just want to keep it as simple as possible till everything works, then go tweak as needed.  I'm even open to a parallel SE EL84, I don't have the OT yet anyway.  the last PSE EL84 was a good sounding amp but just a prototype.

thanks for your time

dave

[shooter]

HBP thanks for pointing out the Princeton ref.  If I cut up the WCF I can do an extra gain stage and the splitphase then I should have enough of everything.  except maybe space!  I'm gonna attempt to mount the ELs on an L shaped tube bracket so they set horizontal inside the  chassis then still gotta get the OT a home.

[shooter]

I made it to the amp, in Indy, but wound up with a couple amp repairs, while I was unpacking  parts I found I did indeed have a OT, but it's a Hammond 125 SE, sooooo... since one of the repairs was a GA-8 6V6 PSE, I decided to change the PA to a PSE EL84 for my amp, that way nothing from the pre-amp needs changing n don't need to buy anything.  Looking at 250, 250, -7.5 and 2500ohm as my starting point.  hopefully i'll have it ballparked later today.  Thanks again for all the help learning what a WCF is.