Paraphase Phase Inverter...and other things 5C3

[EETStudent]

So if I hadn't gone old school enough with my 5E3, I decided to compare its tones to the 5C3. Mainly the difference in the phase inverters. I LOVE the change in tone from the paraphase inverter. Somehow added a deeper, tighter, more articulate low end. Now I want to perfect this PI circuit. From I understand this PI is fairly unbalanced, or at least it can be, and may require some tweaks to balance both sides. I've also noticed a similar circuit in the 5E9-a schematic, but the values there are a little different.

So my question is, what are the optimum values(those in the 5C3 vs 5e9-a) for this type of PI circuit?

[ThermionicScott]

Kudos on exploring further back than the 5E3... I built a 5B3/5B5 Pro clone about a year ago that I love. (It has only two 6SC7s like the 5B3, but no grid-stoppers on the inputs, 2x6L6 output tubes, but only 20uF in the first filter stage, and it drives a 15" speaker using a Deluxe Reverb OT. Kind of a "mutt" based on what I see as the essentials of both circuits.)

Anyways, there will be a little imbalance due to resistor tolerances, imprecision in the math used to obtain the divider values, and imbalances in the tube itself.

Basically, the procedure used when designing one of these things is to figure out the gain of the first stage, come up with a divider that is the reciprocal of that but is still a good value for the power tube's grid-leak, then feed the second stage with that. If balance is a concern for you, you might try the floating paraphase from the "D" series Fenders, which has some ability to correct its imbalances.

That said, my whole point in building the 5B3/5B5 clone was to explore the grungy primitive lo-fi circuits on which Fender built their reputation. And it's my favorite amp now. The cleans are rich and beautiful and tonally balanced from top to bottom, the onset of breakup is smooth and subtle owing to the cathode bias, lack of negative feedback, and the buildup of harmonics in the paraphase, and the full-out distortion is grungy and wild. I shudder at the thought of "perfecting" any part of that!

- Scott

[MWJB]

Why do you think it is unbalanced? Measure the AC at each 6V6 grid at various outputs to see if it is....I wouldn't worry about a few %. The signal to the 2nd/bottom/out of phase tube goes through an additional triode, so this side may be a little more distorted.

Tweak the 25K in the 5C3 for balance.

[EETStudent]

lets compare the 5C3 to the 5D3:

-5C3 has 250k plate resistors, 5D3 has 100k. I assume the effect here is to set the bias point, in conjunction with the cathode resistor. My setup uses a 12AX7 here, so 100k might be better suited

-Both circuits have .05uf coupling caps. Why not .1uF?

- After the coupling caps, the two circuits get a little cloudy to me. The 5D3 seems to have two grid resistors, one 220k one 270k, followed by a 56k resistor to ground. The 5C3 seems to have a 1M grid resistor, a 25k to ground, and one 220k resistor to ground after the coupling cap. There is a 1M feedback resistor. At this point in the both circuits, what are the differences?

[MWJB]

.05 was common accross the board for 5B/C/D Fenders.

The 56K in the 5D3 helps the PI to self balance. No NFB loop on the 5D3. Getting the 1M/25K voltage divider/grid load in the 5C3 is more critical for balance. 220K in the 5C3 is th grid load for the out of phase power tube.

[EETStudent]

I see. Which of these circuits is more balanced by nature?

[MWJB]

5D3

[ThermionicScott]

lets compare the 5C3 to the 5D3:

-5C3 has 250k plate resistors, 5D3 has 100k. I assume the effect here is to set the bias point, in conjunction with the cathode resistor. My setup uses a 12AX7 here, so 100k might be better suited

No, the plate resistor determines the gain of the stage. The cathode resistor is what sets the bias point. With the higher gain of the 12AX7, a lower-value plate resistor can be used.

-Both circuits have .05uf coupling caps. Why not .1uF?

The coupling cap, along with the next stage's grid-leak resistor, forms a high-pass filter. When you do the math, a 0.05uF coupling cap with a 220K grid-leak resistor has a -3dB corner frequency of 15Hz. Compared with the guitar's low-E of 82Hz, why should you need a bigger capacitor?

Oh, here's some good reading for how floating paraphases work. You can click forward and backward through that online document for other PI's: http://www.vias.org/crowhurstba/crow..._vol2_063.html

Seems like the "first page/next page" links are screwed up. Here's the link for "fixed" paraphases: http://www.vias.org/crowhurstba/crow..._vol2_062.html

- Scott

[jmaf]

No, the plate resistor determines the gain of the stage. The cathode resistor is what sets the bias point. With the higher gain of the 12AX7, a lower-value plate resistor can be used.

In practice both the plate and cathode resistors set the BIAS; it's the same current path. The higher plate resistor will limit the current through Rk and will lower the BIAS voltage.

[ThermionicScott]

In practice both the plate and cathode resistors set the BIAS; it's the same current path. The higher plate resistor will limit the current through Rk and will lower the BIAS voltage.

Changing the plate resistor value will affect the bias point, this is true. That's not a legitimate reason to change it, though. Fender used 220k plate resistors on the 6SC7's in order to wring as much gain as they reasonably could from them. With the switch to 12AX7's, it made sense to use 100k plate resistors instead. "Gain-monster" amps weren't yet in demand.

- Scott

[jmaf]

Changing the plate resistor value will affect the bias point, this is true. That's not a legitimate reason to change it, though.

I don't know what you mean by legitimate reason. The Rp / Rk relation is crucial for several reasons, the operating point set by the BIAS being one of them.

[ThermionicScott]

I don't know what you mean by legitimate reason. The Rp / Rk relation is crucial for several reasons, the operating point set by the BIAS being one of them.

What I'm talking about is circuit design. Properly, you should first decide how much gain you want from the stage, and choose the plate resistor that will achieve that amount of gain. Then, you choose the cathode resistor that sets the bias point where you want it (center bias for maximum swing, or off-center for a particular breakup characteristic).

If you wanted to, you could fiddle with the plate resistor to change the bias point, but this would also change the gain of the stage, perhaps drastically, so it's not really the right way to do it.

Why do you keep typing "bias" in all-caps, anyway?

- Scott

[EETStudent]

Great replies fellas.

One more question regarding the various input jack wiring of the deluxe models. Is the 5C3 grid-leak bias? How does that vary tonally from the typical input stage?

[MWJB]

It can't take big signals, some harp players prefer it for crystal mics, but otherwise I would skip it...with 2 channels yuo could always make one cathode biased & the other grid leak?

Jmaf, the cathode resistor predominantly sets the bias (hence the term "cathode biased"), yes the plate resistor has an effect...just like B+ on an output tube, but when one generally talks about rebiasing an output stage, one would change the cathode resistor value & live with the voltage they get (within reason), rather than specifically aim to alter the B+. Fender dropped resistor values to achieve better fidelity, as time went on they ran tubes at higher & higher voltages, so values in the B+ rail and at plate loads dropped....actual tube type is pretty moot.

[jmaf]

Why do you keep typing "bias" in all-caps, anyway?

LOL! Good question. I don't know - I've always typed BIAS in all caps, I think due to those old cassettes which had "High BIAS" printed on them....

The cathode resistor has 60 to 100 times more effect in setting the bias for the stage, that is true. Though you can't simply separate things like that, tube transconductance and those 2 resistors are all intimately tied. That is, if you're doing nitty picky preamp design. And I just nittypicked on you for saying one resistor sets bias and the other gain, it's not as clear cut as that.