Series/Parallel V1 question

[lonote]

I am wanting to try something not completely Fender for a small shop amp/guinea pig & I am thinking I will try cobbling up a Matchless Spitfire.

I have never done an amp with a paralleled V1 & while reading up online, I came across any number of adaptations of parallel, single/parallel & cascading V1 circuits. The most appealing of the variations was a parallel to series version using a switch, as it would mostly keep the Spitfire circuit intact. Parallel V1 sections switched to the sections in series.

The switching version had separate cathodes & plate resistors (unlike typical parallel sections), but there was switching of the plate feeds that I was unsure about, as it is high voltage. So I then started wondering if switching just the grids & outputs post CCap would net more or less the same functionality?

OR is part of the desirability of parallel V1 stages dependent of shared plate & cathode components?

It is all an academic question at this point as I will just follow the Matchless schematic to start, easy enough to mod it later.

I did a quick cut & paste of my proposed signal-only switching. I would not hot-switch so not really concerned with popping, etc.

Revised drawing

Series-Parallel V1, ver2.JPG

[printer2]

You do not have a resistor to ground on the second triode's grid. When switching you most likely will get a pop unless you put a 4.7M (as an example) after the capacitor on the first triode's plate. Going straight in from the first stage into the second might be a bit too much signal level. You can stick a 470k in the wire above the DPDT label you have on the schematic giving you a voltage divider using the 470k and 1M grid leak (I took the liberty of selecting the value for you ). With no capacitor bypass on the second triode will result in 6 dB less signal from this triode as compared to the other. This can give you a mix of distortion and clean if that is what you are after.

[Helmholtz]

Yes, the second triode can't work properly without a grid leak resistor.
Another issue is that the 2 triodes have different gains.
Only the first one has a cathode bypass cap increasing gain by around 5dB.
It's no good idea to parallel 2 stages having different gain because the lower gain stage will load down the higher gain stage.
Also an input triode without a cathode bypass cap may introduce heater buzz.
Generally there's not much benefit from paralleling triodes. Maybe 2dB less tube noise and around 0.3dB more gain.

[lonote]

Thank you both for the replies, I did indeed forget to add the grid leak, which as pointed out, can be added into the jumper between the switching points. I will revise the sketch tomorrow as not to have an unusable schematic out there.

As I mentioned, I have never built anything with a parallel front-end, so I can't really speak to whether or not is is worth the trouble. As there are some popular designs out there using them, I thought it would be interesting to try.

Below is one of the Spitfire schematics I have been looking at.

SpitfireSchematic.bmp

[nuke]

If you want to try paralleling the V1 stage, just tie the plates, grids, cathodes together. Use a 47k 1W plate resistor, and a 680-ohm cathode resistor.

You can see an example in the Deluxe Reverb II schematic (Rivera design), though, most copies of the schematic have an error showing the plate resistor as ".047 1W" which is nonsensical. It is 47k.

You may or not like the 1000pf from plate to cathode.

The effect is to halve the plate resistance, which reduces the output impedance looking into the tone stack. Some say it reduces noise.

[lonote]

If you want to try paralleling the V1 stage, just tie the plates, grids, cathodes together. Use a 47k 1W plate resistor, and a 680-ohm cathode resistor.

Right, that is actually reflected in the first attached schematic. They didn't draw out the second section of the 12AX7, but instead double-numbered each pin; 1-6, 2-7, 3-8. A bit lazy really, took me a second to notice it.

This second one attached below has V1 fully represented.

matchless_spitfire.pdf_2.png

[nuke]

Ahh, I missed the commas as well.

That's an interesting choice of plate and cathode resistors with a paralleled stage, which also led me to think it was a single stage.

Have you plotted the load line and operating points for the supply voltage at these values vs. the 47K/680 ohm values Rivera used?