Safe to put V1a coupling capacitor on a switch for variety?

[Dave H]

Another consideration is the cap value. The .68uf value will NOT pass the same frequencies across the the 820R resistor as it would across a 1k or a 1.5k. And certainly not the same as a 2.7k which is the "classic" sound. So if that is your goal you need to adjust the cap value to knee at the same frequency for the 820R as the .68u value does for a 2.7k

The knee frequency doesn't change as much as you might think with change in resistor value because the resistor is effectively in parallel with the "resistance looking into the cathode" which is a lower value (1/Gm?) but the low frequency (un-bypassed) gain changes a lot.

[Chuck H]

Wait. Serious question. How is my drawing in post 26 different from Dave's drawing in post 5?

In post #5 the caps are referenced to ground with the 10M AFTER being isolated from DC. In the cathode circuit you drew the DC is across the 100k resistor which is simply parallel to the 820R resistor, which is just the same as if the cathode resistor were 813k with no other resistor in the circuit. Ideally Dave's circuit (and mine) would have the 10M at the switch end of the caps. But being as there are multiple caps and poles this is cumbersome. For the cathode bypass cap you need the resistor across the switch to keep the cap in the circuit and equalize the charge for when the switch is flipped. For Dave's circuit in post #5 the caps are referenced to 0V at either end in a strictly AC environment. They are different circuits.

[Mr. Bill]

Oh okay. I overlooked that they were isolated by the cap before the switching network. I'm learning more this week than I have in months. You guys will never get rid of me now, ha.

[Mr. Bill]

Just triple checking that I have my head wrapped around this. So if I go with Dave's switching idea since I have all those parts in shop, then follow that with a single mini toggle to take the 470k/500p in and out of circuit before the gain knob, I'm fine with the second switch because the .022 decouples the signal before any other switching? Here's my crude drawing. I really need to get a better grip on LTspice so my simulations actually work.


Further still, my switch to take the the .0047 in and out of the NFB loop needs nothing more than a switch since the cap on the presence knob is already isolating it from DC taking the scratchies out of the pot?

[Chuck H]

Good to go on the three way switch. Connect the two way so that it shorts either end of the 470k/500p circuit together. Effectively "shorting" it rather than trying to bypass it.

[Mr. Bill]

Like this?

Man, I need to quit music altogether and go into graphic design. A bit redundant using a DPDT, but I have some of those. I guess I should order more switches.

[Chuck H]

Too complicated. Think simpler

EDIT: Imagine the amp is built on an eyelet board (as I assume it will be?). You just have your parallel components in place on the board and run two leads from either end to the switch as shown.

And more to keep in mind. When adding leads in the signal chain that fly from the board to the panel you'll want to design your layout so that these leads can be kept short and away from like other phase circuits if possible. If you get too jiggy with switchable doodads and don't pay careful attention to lead dress you can end up with an unstable amp very quickly.

[Chuck H]

And... You mentioned having the three way switch on hand.?. If it is a three way toggle you'll want to be certain it's an on/on/on since those are sort of hard to find. Most three way toggles are on/both/on or on/off/on and would require more circuit consideration for wiring.

[Mr. Bill]

I just found an SPST in the drawer, so that simplifies things and is less waste. I'm on mouser right now ordering some more along with some rotary switches. This is going to be one crowded chasis for sure. I think at this point I'm going to build it in a 4 hole box and eliminate the low sensitivity jack to make room. That'll give me 1 Jack and 3 rotaries in the 4 hole and I'll drill for a few mini toggles.

[Chuck H]

Like I mentioned above. It's a dangerous business running control leads back and forth across the circuit to the front panel. Controls should occur on the panel as they occur in the circuit as much as possible and layout should be designed to keep leads as short as practical. This is less critical in non cascade type vintage amps, but the 800 type circuit IS a cascade preamp AND you have a hotter cathode resistor in V2 which increases gain. Too much spaghetti and your amp will whistle like it forgot the words.

[Mr. Bill]

I absolutely hear you on that one. If I eliminate the low sensitivity jack that I've never seen used anyway, then there is zero extra wire length to run the first two switchable mods. Those wires are already running to the face plate anyway. The depth switch, likewise can be right next to the Z selector needing no extra wire length or routing.

[Mr. Bill]

Quick silly question, maybe? Doesn't the 2203 cold clipping stage clip the negative lobe of the waveform? I'm Working in LT Spice as suggested by Chuck, and it seems to be clipping the positive lobe in my simulation. I guess dumb question #2 is....I'm supposed to probe the plates to see the output waveform of each stage, correct?

[eschertron]

Quick silly question, maybe? Doesn't the 2203 cold clipping stage clip the negative lobe of the waveform? I'm Working in LT Spice as suggested by Chuck, and it seems to be clipping the positive lobe in my simulation. I guess dumb question #2 is....I'm supposed to probe the plates to see the output waveform of each stage, correct?

The plate of the triode will be the "output" of the transfer function, so yes, probe the plate to see the effect the stage has on the signal. Keep in mind that each triode stage is inverting, so the plate signal appears with the effect of cutoff on the positive half of the waveform. I think you're doing it right.

[Dave H]

If it's cold biased the bias current is lower so the plate voltage is closer to B+ voltage and it will clip on the positive lobe when you probe the plate.

[Mr. Bill]

Thanks. And if I may add, is there a way to tell the program to keep the heaters warm so you don't have to wait a long time for it to run each simulation?