What's happening here?

[tubenit]

I've not seen a cathode cap/resistor set up like this?   Any idea why it would be done that way?

With respect, Tubenit

[sluckey]

You've seen it before. Look at the PI in the Princeton Reverb. Imagine the circuit without the two bypass caps. It's called bootstrap biasing and provides a very high impedance for the grid. I don't know why they chose a bootstrap circuit. Maybe a quest for tone? Just want to be different? 

There are some other amps that also use a bootstrap circuit but the PR came to mind first.
http://www.el34world.com/charts/Schematics/files/fender/PRINCETON_REVERB_AA1164.pdf

[jjasilli]

Reminds me somewhat of the Paul C mod for the Fender Princeton.  See:  http://www.blueguitar.org/new/articles/blue_gtr/amps/john_paul_rick_PRmods.pdf

"The Paul C. mod involves changing the bias of the cathodyne PI from cathode to fixed bias.  Paul found that it reduces the assymetrical nature of distortion from an amp that uses a cathodyne phase inverter.  The net result of this mod is a more powerful sound from the amp and an apparent increase in the bass response."

But the Paul C mod is pure fixed bias.  This seems to be a hybrid.  There also seems to be positive feedback:  the incoming signal appears on the grid & on the cathode (in phase), reduced by the R12/R15 voltage divider, and partially??? bled to ground through C12.

[HotBluePlates]

I don't know why they chose a bootstrap circuit. Maybe a quest for tone? Just want to be different?  

Probably to maximize the gain of the 1st stage.

We could analyze the circuit with/without the bootstrapping to discover what difference there would be.

In a nutshell, the input impedance of the 2nd stage is seen by the output of the first stage as an additional load in parallel with the tone circuit and volume control. That's 3 loads, all in parallel, acting as a load, right along with the plate load resistor. The net effect is to reduce the gain delivered by the 1st stage.

There's a catch: The bootstrapping works by virtue of feedback. You might be used to seeing a cap in the C11 position (which bypasses the biasing resistor, R8). You're likely not used to seeing a cap in the C12 position (bypassing the bootstrap resistor, R15). Bootstrapping works to raise the input impedance, but does so due to feedback across the bootstrap resistor.

Therein lies either the catch, or the coolness of the circuit. With a bypass cap across the bootstrap resistor, the effective feedback is eliminated above the turnover frequency implied by the resistor and cap. So the circuit has a higher input impedance at d.c., but at some frequency, feedback reduces to zero, input impedance drops to 1M (due to R12), and the boosted gain goes away.

Someone check my math, but 3.9k against 2.2uF looks like a -3dB point of ~18Hz to me. Which means this addition does nothing, except maybe add noise. The cap should be much smaller, or the resistor bigger. A bigger resistor would also increase the apparent input impedance, by making the bootstrapping more effective; I'll go out on a limb and, without calculating, say it would probably take a lot more than 3.9k to get reasonably effective bootstrapping.

[PRR]

> to get reasonably effective bootstrapping.

But there's no need for bootstrapping. The 1Meg grid resistor is not a heavy load on the 500K pot. There's no shortage of gain; in fact there's a 220K+500K loss at the volume control. And certainly 18Hz is not bass-cut for guitar.

It may be that the OVERload recovery time is affected. Put a voltmeter on the 3K9 and beat on it.... is it steady or does it droop? Do different caps give less-pleasant overload tails?

Or it may be a this-and-that design, never "optimized for minimum parts". They may do nothing but they sure don't cost much in DIY quantity.

[DummyLoad]

heathkit did something similar within their aa32 amp - have a look at 3rd stage amp in schematic in the link below.


http://www.nijnkonijn.nl/machmat/schema/power/ppp/heatha32.gif

--DL

[PRR]

> something similar

Only one cap. 0.01u. Gain is proportion of 2,180 for low freq, 1,500 above the audio band; it gives a mild boost to extend the overall -3db point well above 20KHz.

[tubeswell]

Hi tubenit, that looks like bootstrapping for higher freqs. (It if was regular bootstrapping you wouldn't have the bypass cap || with the tail resistor)