LM317 power supply design (~33V DC) from 50V AC tap (tube preamp content)

[owlexifry]

i am trying to design a power supply. but i am not very smart.

i have a few limits with the design (and there’s a few reasons why..)

1) i’d like to use a LM317 voltage regulator since it’s available everywhere and low cost.

2) i have to use a 50V AC tap (reasons below) - unless i try to use a separate 150ma (24VAC sec) transformer (would rather not)

3) aiming for ~+33V DC supply

4) the circuit to be supplied apparently draws 84ma (per multimeter ammeter)

5) i did some measurements - the ‘50V AC tap’ rectifies to about 64V DC (no load) - (with a bridge rectifier + 1uF cap)

my understanding is that LM317 voltage max rating is a differential of Vin to Vout (40v max) rather than an absolute 40V max at Vin.
so, I’m assuming the LM317 won’t melt if it sees 64V DC, and also assuming that a 39V zener at Vin can limit this input voltage to 39V anyway…?

here’s my first draft schematic:

would really appreciate advice on this design.

a lot of spitballing on the values here… i’ve been trying to follow datasheets / examples i’ve been seeing online.

another problem/annoyance
- the rectified 64V DC is going to require at least 100V rated (or more preferably) filter caps before it gets to the regulator.
that's getting expensive - almost becomes cheaper and simpler to just use a separate 150ma 24VAC transformer.
…unless a zener diode can clamp voltage before the filter caps?
(not sure if that would even work)

my understanding of zener performance is limited and i’m not sure if they will truly and reliably protect the regulator in the way i understand they do/might…

this is all for a project i started about 18 months ago and have been neglecting (for pedal builds of course) because this power supply problem has been a real challenge for me to address / understand.

the project:
MK2C+ inspired preamp build

- lead channel only. (no rhythm channel / switching)


- with 5 band EQ based on original (but with transistor gyrators instead of inductors)
(this is what the power supply is for)


- tube preamp runs off 300VAC + 6.3VAC windings (hammond 369BX transformer)
- and yep, there’s a spare ‘50VAC bias tap’, so i thought hey why don’t i just use this tap instead of trying to squeeze in a (24V) transformer - for the EQ circuit.

the preamp sounds amazing.
when i first finished it, i finally i got to hear that mesa mark lead overdrive sound. it's astounding. a unique complex flavour, different to the marshall/SLO/5150 topology that i'd been used to.
(into FX return, 4xEL34 100W Peavey))

but for high gain/metal riffs/styles, it really needs the 5 band EQ, to bring back the low end (and some top end), after cutting it at the front end (necessary to keep the overdrive tight)
and the EQ kinda needs to be high(ish) voltage.

i first tried an opamp based eq circuit.
- TL074 @ 9V - not too good, @12V improved, but still lacking. could hear the lost dynamic, definition (vs no EQ). and the highs were a bit 'harsh'
(into FX return)

then i got hold of some OPA4134 ICs.
and damn, they’re pretty good. noticeable improvement over TL074.
(into FX return)

then i found a schematic for a transistor gyrator based circuit, intended to run at ~+33V, that looked pretty much bang on vs what i’ve seen in mesa boogie schematics for the 5 band EQ.
so i went ahead, did a vero layout, and built it.

- using a separate 9V supply with chargepump circuit to boost to +33V - and heck yes, this sounds so much better than the opamp circuit (no demo yet, sorry)

and since then i’ve been kinda stumped. just gotta figure out a power supply for this EQ.

i built the ‘chassis' for the EQ a few months ago, turned out ok:


hoping to integrate the EQ onto the chassis sorta like this:


and getting to the end of this thread, ive just realized/queried something... can the PT sustain all this?
the 369BX PT secondary windings are rated 300V @ 86ma. (= 25.4 watts ?)
- my understanding is that 12AX7 triode current draw rating is [1.2ma at 250V], so 6 triodes = 7.2ma (x 250V) = ?~1.8W total?
- EQ circuit apparently draws 84ma, @ 33V , (= 2.8W ?)
to me, looks like the PT should be just fine.....
(all of my terrible maths/understanding on display here)

yes, this is a stupidly oversized build for a tube preamp.
it's no fancy 1RU build with torroidal transformers, printed circuit boards, and all the nice tidy things you'd expect to see with properly printed faceplates etc.
a lot of that stuff is beyond my understanding. i just do i what i can

 

[jwin615]

If you still have the opa4134 EQ intact, try running it at +- 18v(if the other components can take it). Your mention of dynamics is likely due to decreased headroom in the circuit.
I have some thoughts on your current situation, but I'll let others more experienced in this chime in because I'm not 100% on them or, since it's Friday night, not 100% in general.
That preamp rips as it is though. Good work.

 

[owlexifry]

If you still have the opa4134 EQ intact, try running it at +- 18v(if the other components can take it). Your mention of dynamics is likely due to decreased headroom in the circuit.
I have some thoughts on your current situation, but I'll let others more experienced in this chime in because I'm not 100% on them or, since it's Friday night, not 100% in general.
That preamp rips as it is though. Good work.

thanks!
yep, still got the opamp EQ (other than that chargepump circuit, this was my first ever vero build, and i had no idea what i was doing)


i could try the bipolar 18v supply idea, but regardless of whatever EQ circuit i choose, i still have the problem of designing a power supply from that 50V AC tap (or another transformer)

 

[xefned]

Max input voltage on LM317 is 40v sez datasheet. I don't know if that will be a problem.
Might want to use a TL783 instead. (?)

 

[owlexifry]

Max input voltage on LM317 is 40v sez datasheet. I don't know if that will be a problem.
Might want to use a TL783 instead. (?)

Vin 40v max yes. (although other datasheets state 40v max between Vin and Vout)
but could a 39V zener do anything about that? idk.

TL783 sounds like a good idea. getting hold of one though…looks like mouser is it.

 

[Paradox916]

I really don’t see a lot of people tackling high gain stuff like this, I don’t know about building amps, one day I would like to try and do a similar desk top pre amp as I’m just a headphone player into a daw, I haven’t seen a lot of high gain tube preamp kits out there so this is cool.

 

[owlexifry]

I really don’t see a lot of people tackling high gain stuff like this, I don’t know about building amps, one day I would like to try and do a similar desk top pre amp as I’m just a headphone player into a daw, I haven’t seen a lot of high gain tube preamp kits out there so this is cool.

from what ive seen, some lads here have been getting onto the rosamp kit SLOs and doing some super clean builds with great success.

point being, if you can build a complex modulation pedal - building an amp with a kit isn't gonna be harder or more complex. it's just more expensive, potentially more dangerous (high voltage), and on a larger scale.

pretty sure @vigilante398 is doing some amazing work with high gain tube preamps right now, with the brown eye (on my list), and supposedly a uberschall is in the works...