Building Power Supplies for 2 X 12ax7 pedals

[didit]

Hello Matt --

That module you've already got on order looks promising. The HT 15mA current spec will be adequate for a single 12AX7, not sure I'd be comfortable going above. Our designs don't tend to set triode's at the 8mA MAX. Been too long since I last closely modelled it to offer precisely. My crude rule of thumb is 10mA per pair. Perhaps Martin can share true and accurate numbers from his Spice models.

I don't see sizeable inductors but still worry it might be noisy. Nixie tubes won't be affected by high frequency junk in either the filament or HT power. Might find it necessary to add a few small off board donut chokes to suppress the feeds and trust that the module PCB is designed properly to avoid it emitting junk.

A charge pump is a SMPS. SMPS broadly have various modes of design for doing various jobs. Fun topics but not worth getting further into those weeds today.

Liking the Dr. Balls approach, though PCB assembly feels simpler. Perhaps there's a finished CAD design floating around one could download and send off to a fab for cheap.

..

[bcmatt]

Not quite the same thing but it's food for thought. I've been building pedals with subminiature tubes and using a "nixie" type power supply for the HT (~100V). Two 6.3V filament tubes will run just peachy off a 12VDC supply. I did that with my tube muff

OK, these are really cool. You've got my full ADHD attention. Now I want to experiment with these. Aside from the lower plate voltages, what other different characteristics do they have that you would compensate for in the circuit? I'm going to look up their tube charts.

I'm already thinking about building a version of the Hot Cat Preamp using a couple of 6112s...

[bcmatt]

That module you've already got on order looks promising. The HT 15mA current spec will be adequate for a single 12AX7, not sure I'd be comfortable going above. Our designs don't tend to set triode's at the 8mA MAX. Been too long since I last closely modelled it to offer precisely. My crude rule of thumb is 10mA per pair. Perhaps Martin can share true and accurate numbers from his Spice models.

I don't see sizeable inductors but still worry it might be noisy. Nixie tubes won't be affected by high frequency junk in either the filament or HT power. Might find it necessary to add a few small off board donut chokes to suppress the feeds and trust that the module PCB is designed properly to avoid it emitting junk.

The module claims 15mA, and shouldn't each triode be under about 1.5mA each?
Although, I really don't understand how these SMPS things work, and if the output is dialed to 235V, does that reduce the amount of mA it can handle?

As far as I understand it, I should maybe be putting a small film cap in parallel with each of the filter caps to help resolve any high frequency noise... again, I have 0 experience with this... I just feel like I read it somewhere...

[Dr Tony Balls]

Not quite the same thing but it's food for thought. I've been building pedals with subminiature tubes and using a "nixie" type power supply for the HT (~100V). Two 6.3V filament tubes will run just peachy off a 12VDC supply. I did that with my tube muff

OK, these are really cool. You've got my full ADHD attention. Now I want to experiment with these. Aside from the lower plate voltages, what other different characteristics do they have that you would compensate for in the circuit? I'm going to look up their tube charts.

I'm already thinking about building a version of the Hot Cat Preamp using a couple of 6112s...

6112s are what I used, as they seem to be about the highest gain dual triode available. Its a little less gain than a 12AX7 so not exactly one to one but it's worked out just fine for me. I pretty much just set em up per the datasheet with 100K plate resistor, 1.5K cathode, and a 1M grid leak.

[bcmatt]

6112s are what I used, as they seem to be about the highest gain dual triode available. Its a little less gain than a 12AX7 so not exactly one to one but it's worked out just fine for me. I pretty much just set em up per the datasheet with 100K plate resistor, 1.5K cathode, and a 1M grid leak.

Interesting! You just answered what I was researching. Google AI suggests doubling to 220K plate resisters and around 3K cathode resisters... But I can't always trust that guy...

[didit]

That module you've already got on order looks promising. The HT 15mA current spec will be adequate for a single 12AX7, not sure I'd be comfortable going above. Our designs don't tend to set triode's at the 8mA MAX. Been too long since I last closely modelled it to offer precisely. My crude rule of thumb is 10mA per pair. Perhaps Martin can share true and accurate numbers from his Spice models.

I don't see sizeable inductors but still worry it might be noisy. Nixie tubes won't be affected by high frequency junk in either the filament or HT power. Might find it necessary to add a few small off board donut chokes to suppress the feeds and trust that the module PCB is designed properly to avoid it emitting junk.

The module claims 15mA, and shouldn't each triode be under about 1.5mA each?
Although, I really don't understand how these SMPS things work, and if the output is dialed to 235V, does that reduce the amount of mA it can handle?

You're correct -- any single 12AX7 triode is almost certainly running at a typical 1.5-2.5mA. So your module's 15mA doubtless will handle 4 or more triodes. That said, our 12AX7 spec sheet calls out 8mA Max current. The spec for 12AT7 18mA Max, with a typical operating current 8-10mA. The rest of the family generally settles in between. The wary engineering persona controlling my typing fingers tends toward caution. Hence 10mA per pair as rule of thumb.

I take your presumptive intuition that lowering the module output voltage should trade some fraction of available current -- assume 200V setting you might hope for ~17mA, so improve a tad and not reduces. This swag is based on what the circuit design appears to be from the photos. Would be cool if they'd properly spec'ed it but for that price etc one cannot expect 1975 era HP quality documentation.

As far as I understand it, I should maybe be putting a small film cap in parallel with each of the filter caps to help resolve any high frequency noise... again, I have 0 experience with this... I just feel like I read it somewhere...

Inductors in series is the conventional way to catch HF noise. A film cap in parallel with bulk e-caps is an OK idea, but should be unnecessary.

Hope that helps a bit.

Meanwhile, seems you might be off following a missile tubes approach. That's fun too.

Best ..

[bcmatt]

That module you've already got on order looks promising. The HT 15mA current spec will be adequate for a single 12AX7, not sure I'd be comfortable going above. Our designs don't tend to set triode's at the 8mA MAX. Been too long since I last closely modelled it to offer precisely. My crude rule of thumb is 10mA per pair. Perhaps Martin can share true and accurate numbers from his Spice models.

Ok, I didn't realize the potential there. Thanks for the word of warning. I'll contemplate this a bit. I"m thinking I will feed the heaters directly off the 12VDC input with the hopes of not taxing the nixie supply needlessly....

I don't see sizeable inductors but still worry it might be noisy. Nixie tubes won't be affected by high frequency junk in either the filament or HT power. Might find it necessary to add a few small off board donut chokes to suppress the feeds and trust that the module PCB is designed properly to avoid it emitting junk.

I am not very experienced with inductors in this setting, so if you know of appropriate ones I should look at using, please let me know.

Meanwhile, seems you might be off following a missile tubes approach. That's fun too.

I am definitely leaning in that direction... thanks to the inspiration of Dr Balls here. I just ordered a couple 6N21Bs and some 6N17Bs from Ukraine with the idea that I want to experiment building some circuits with them. I know I will have to tweak things a bit to make them work in a circuit originally designed for 12ax7s but for this Hot Cat preamp inspired build, I was already planning to play with it a bit anyways... So I'm happy to forge a new path there and get a little experimental.
However, they seem to require more current for the heaters and plates... so that might become a stress point...

[Dr Tony Balls]

However, they seem to require more current for the heaters and plates... so that might become a stress point...

FWIW, on the Emporium pedal (single 6112) I measured ~350mA for the whole unit.

[didit]

Missed a messed construct with that "don't" -- I do/did see sizeable lumps in the photos that are certainly inductors.

I keep a box of smallish ferrite donuts for when HF noise is suspected. I'll wrap a few loops of wire around and that's usually enough. I anticipate your Amazon module will be well behaved. If not, try a donut.

For sourcing ferrite cores or even some prefab inductors check out :: https://www.a1parts.ca/Chokes&Coils.htm


..

[bcmatt]

Missed a messed construct with that "don't" -- I do/did see sizeable lumps in the photos that are certainly inductors.

I keep a box of smallish ferrite donuts for when HF noise is suspected. I'll wrap a few loops of wire around and that's usually enough. I anticipate your Amazon module will be well behaved. If not, try a donut.

For sourcing ferrite cores or even some prefab inductors check out :: https://www.a1parts.ca/Chokes&Coils.htm
..

Yes, I missed them too, haha! Until I realized that's what L1 and L2 are.
I'm currently redrawing the schematic to how I might approach it using my Nixie supply board and a 6N17B tube for V1 and a 6N21B tube for V2. Google (AI) suggests I tackle any High frequency noise with an RF pi Network (2x 100nF caps to ground with a 100ohm resister between them) right after the nixie board... but then I can also reduce my filter caps significantly in order to strain the Nixie board less. I'll also reduce the 2x 22K dropping resisters to about 2.2k so that I can dial down the voltage from the Nixie board to about 200V

[bcmatt]

This is the current iteration of the plan for this Hot Cat preamp pedal. Any thoughts?

Dark Mini-Sub Schematic.png

[bcmatt]

Well, the plot thickens...
A friend stopped by and dropped this off... He apparently got them off a random guy he met that he was buying some vinyl records from.

1b17c2f1-c68b-4c85-9a03-d701e9e18442.jpg

Must of these are weird radio and TV tubes it seems... Not that I've been through them all yet.
But there are lots of weird combo tubes... Like a pentode with a couple diodes, or a pentode with a triode... But I feel like they would potentially make some interesting pedal designs.
Already have a couple design ideas for some of the metal ones....

4ef5127f-b089-4d3e-abcf-a25925e803cb.jpg

[didit]

This is the current iteration of the plan for this Hot Cat preamp pedal. Any thoughts?

I've no experience with this specific circuit, but it looks straightforward. Curious why experimenting with two unknown and slightly different types of sub-mini vs learning characteristics of one, and then applying that knowledge to the cascading stages. The Hot Cat schematic I found shows a pair of 12AX7s.

I'd skip a Pi filter 'til you really discover whether it's needed; however, given this hopefully avoidable eventually, would recommend a single small film cap into a modest inductor (50-100mH and 10 Ohms is a sloppy swag), then directly into the initial bulk e-cap. That would be a clean and more effective Pi, and consumes less of the power you'll want for amplification.

..

[bcmatt]

This is the current iteration of the plan for this Hot Cat preamp pedal. Any thoughts?

I've no experience with this specific circuit, but it looks straightforward. Curious why experimenting with two unknown and slightly different types of sub-mini vs learning characteristics of one, and then applying that knowledge to the cascading stages. The Hot Cat schematic I found shows a pair of 12AX7s.

I'd skip a Pi filter 'til you really discover whether it's needed; however, given this hopefully avoidable eventually, would recommend a single small film cap into a modest inductor (50-100mH and 10 Ohms is a sloppy swag), then directly into the initial bulk e-cap. That would be a clean and more effective Pi, and consumes less of the power you'll want for amplification.

Thanks, I'll give that a try.

As far as the two different sub-minis approach, I ordered 2x 6N21Bs, which the supposed upgrade of the 6N17B (which I got 4 of). But the 6N17B has the potential to be higher gain, which might be ideal for the 1st tube position. But, you may be right that it would be better to stick with one for the whole circuit, then try they other in a second version perhaps.

[bcmatt]

I'm still contemplating things. I haven't shown the Heaters yet, but this is more of a proper Schematic and closer to what I will try:
WARNING: Tube PIN Numbers are wrong!!!

Dark Mini-Sub Schematic.png