Power Circuit Advice

[JJH0906]

Hey group,

Has anyone got any suggestions of how to construct a power circuit, specifically the "voltage divider" to supply the various voltages needed for plates, screens, etc.? I've been using the Amp Books from Richard Kuehnel and while he's got some good stuff, I'm having a hard time making sense of how/why he's doing what he's doing.

Here's an LTSpice example of what he's showing in his second book. When running it in Spice, I do not get the results he lists in the book!

RK_PowerCircuit.JPG

[Phil_S]

I don't use Spice, so I can't comment the result. I might speculate that the inputs provided are somehow off from what they need to be or one or more inputs are missing from the model.

Try modeling this in Duncan's power supply modeling software. You can d/l it here: https://www.duncanamps.com/psud2/ last item on the web page. What's nice about psud2 is that it is geared towards audio amplifiers. You'll be able to choose a model and modify it, so you don't start from scratch.

[sluckey]

Has anyone got any suggestions of how to construct a power circuit, specifically the "voltage divider" to supply the various voltages needed for plates, screens, etc.?

You cannot calculate that voltage divider network without knowing the current that will be flowing in the V1, V2, V3, V4, and V5 branches. Once you have that info you can figure out everything else.

[JJH0906]

I don't use Spice, so I can't comment the result. I might speculate that the inputs provided are somehow off from what they need to be or one or more inputs are missing from the model.

Try modeling this in Duncan's power supply modeling software. You can d/l it here: https://www.duncanamps.com/psud2/ last item on the web page. What's nice about psud2 is that it is geared towards audio amplifiers. You'll be able to choose a model and modify it, so you don't start from scratch.

Is the Duncan PSUD available online? I have had trouble downloading it, possibly because it's a 32-bit program (I'm using Win10).

[sluckey]

I don't use Spice, so I can't comment the result. I might speculate that the inputs provided are somehow off from what they need to be or one or more inputs are missing from the model.

Try modeling this in Duncan's power supply modeling software. You can d/l it here: https://www.duncanamps.com/psud2/ last item on the web page. What's nice about psud2 is that it is geared towards audio amplifiers. You'll be able to choose a model and modify it, so you don't start from scratch.

Is the Duncan PSUD available online? I have had trouble downloading it, possibly because it's a 32-bit program (I'm using Win10).

It's available through the link provided by Phil_S. Win10 is NOT a factor.

[JJH0906]

I installed it but then it barked at me about not having any rectifier files. I ignored that and then it had some kind of a violation error. Could just be my PC but it seemed like it was a operating system kind of thing.

I'll try and take another look at it and see if I can get it going. I don't mind doing it out by hand but it would be nice to have some of the heavy lifting done by software.

[sluckey]

You cannot calculate that voltage divider network without knowing the current that will be flowing in the V1, V2, V3, V4, and V5 branches. Once you have that info you can figure out everything else.

[JJH0906]

You cannot calculate that voltage divider network without knowing the current that will be flowing in the V1, V2, V3, V4, and V5 branches. Once you have that info you can figure out everything else.

Yes, that's the other part I need to work on, to get all the different currents from the various branches of the rest of the circuit.

[R.G.]

Some general comments that might help you...

- 12AX7 plates are generally run at about 0.5 to 1 milliampere.

- Self-biased triode circuits as common in guitar amps are fairly insensitive to the plate supply voltage.

- Setting a specific plate voltage is not what those dropping resistors are for. They're there with the decoupling caps to prevent current changes in one tube from affecting another tube.

- The order of the tubes connected those decoupled points is important. You want the most-filtered voltage connecting to the most sensitive point on the string, so the input-i-est voltage on the right of the string goes to the input tubes. The output tubes connect to the least filtered voltages; next the phase inverters, etc. from left to right in your picture.

It's common to see a single resistor value used between the rightmost points in the string. This is often 4.7K to 10K. A 10K resistor drops 10Vdc for each milliampere of current through it, so the rightmost, input-most stages, being run at low currents and needing the least B+ overhead, can generally use the biggest resistors.

[JJH0906]

Some general comments that might help you...

- 12AX7 plates are generally run at about 0.5 to 1 milliampere.

- Self-biased triode circuits as common in guitar amps are fairly insensitive to the plate supply voltage.

- Setting a specific plate voltage is not what those dropping resistors are for. They're there with the decoupling caps to prevent current changes in one tube from affecting another tube.

- The order of the tubes connected those decoupled points is important. You want the most-filtered voltage connecting to the most sensitive point on the string, so the input-i-est voltage on the right of the string goes to the input tubes. The output tubes connect to the least filtered voltages; next the phase inverters, etc. from left to right in your picture.

It's common to see a single resistor value used between the rightmost points in the string. This is often 4.7K to 10K. A 10K resistor drops 10Vdc for each milliampere of current through it, so the rightmost, input-most stages, being run at low currents and needing the least B+ overhead, can generally use the biggest resistors.

These are all great points and I appreciate you sending those. Eventually I'm going to get this all figured out, haha!

Attached is a reference from one of the books that I'm using for my build. You can see that it's a full wave rectified circuit with some diodes that feed a bunch of node resistors. I realized that those are connected to tubes and other components that are drawing current and that's where the voltage drop comes from, I'm just having a little bit of a challenge getting that all sorted out still.

I was just trying to work with the 12 AX7 tube so I could throw a signal generator on there to watch things work, but as shown in the attachment, there is a crap ton of a resistor network going on that makes it a little more tricky to just provide some reasonable amount of voltage to the tubes.