Tube preamp PSU...

[alk509]

Hey guys,

I'm designing a power supply that will power a microphone preamplifier. I need an unregulated high voltage DC output, an LM317-based regulated 12.6V DC output for heaters, and a 48V phantom power output.

I'm going to use a Hammond transformer with a high voltage secondary and a filament secondary. I had thought about using two back-to-back 44V Amvecos, and use the first transformers secondary for both the heater supply and phantom, but I'm going to be pulling up to 1.2A with these heaters, and the Amvecos can't handle that much current...

So my question is this: Is it cool to get phantom power from the unregulated ~300V B+ supply? I was thinking of using another 3 or 4 stages of RC filtering, and then a zener regulator, maybe two 24V zeners in series... What do you think?

Peace,
Al.

 

[bcarso]

That's a fair amount of power to be burning up for phantom, and zeners are awfully noisy, but otherwise I don't see what's wrong with it in principle. Whether zeners or some other shunt regulator you will want to have some idling current at max draw from the phantom load, so the dissipation starting from 300V really mounts up in a hurry---maybe 5W or more.

 

[PRR]

> I need an unregulated high voltage DC output

How much current???

> 12.6V DC output for heaters ... up to 1.2A

15 Watts for preamp heaters? I have good power-amps that use less heater.

> and a 48V phantom power output.

15W of heater suggests ~10W of B+ power. 25Watts, but doubel that for AC VA rating. 25W of total power seems like a lot to take from back-back transformers: two 50VA 15% regulation trannies will yield total regulation in the 30% range, which is pretty poor. You might have to go to two 100VA cores to get good regulation.

Write your loads so you can see which one(s) are the worst problem.

300V DC at ???mA
12V at 1.2A = 15W
48V reg, ~70V unreg at 10mA = 7W

Anyway you slick it, you have three loads at significant power and very different voltages, so far apart that droping one from another means huge waste.

One old trick: raise the B+ about 48V, use a push-pull output stage, and tap some 48VDC from the cathode resistor/Zener.

 

[CJ]

I have done a phantom off the B+, but like Brad says, it's going to generate some heat. And the phantom voltage will vary depending on what model mic you use. This is due to the fact that the volage dropping resistors will cause different voltage drops, depending on the current draw of the mic. You can not afford to have a solid 100 ma flowing thru a phantom supply of this type to keep things steady, as your pwr diss. will be 0. x 300 = 30 watts = 60 watts of power resistors needed. Also, the transformer will proboaly overheat, depending on it's VA rating. Plus, it's just plain wasteful, and you do not want to smell the dust cooking off power resistors every time you come back from vacation. It will generat a false alarm.

But, if you know the current draw of every mic in your cupboard, and it falls within a fairly tight bracket, you can get away with the B+>phantom stepdown.

So if 10 ma is enough for every mic, 0.01 x 300 is 3 watts, not an unreasonable amount of heat, especcialy if you build a four section RC filter network and split up the load amongst all the resistors.

If you are using a transformer in a dual diode centertapped cofiguration,
you can simply put a 1N4004 diode off one of the red wires. This will knock your voltage into the first cap in half. Going from 150 to 48 is easier, no?

If you really want to get sneaky, you can order a transformer that has a tap for a fixed bias supply for vac tubes, which will really make life easier

One thing to consider: I do not know if you are using an Apex med. diaphram, or a 1952 Neumann U-67 that has been rebuilt by Stephen Paul, but if something happens to go wrong with your 300>phantom circuit, or if you drop a screwdriver into the supply while it's on, you have the potential of getting 300 volts dc on your favorite mic..

cj

 

[alk509]

Thanks for the comments, guys! This zener thing is too much of a waste, I didn't like it from the start - hence my asking the question...

[quote author="PRR"]> 12.6V DC output for heaters ... up to 1.2A

15 Watts for preamp heaters? I have good power-amps that use less heater. [/quote]

That should've read 6.3V @ 1.2A, so more like ~7.5W. Sorry ! That should be the most I'll be pulling (a 6N1P and a 12BH7 running at 6.3V). The 300V B+ was also a ballpark figure...

OK, so I've thought about it a little bit more, and here's where I'm going with this now: I'm trying to come up with a board that I can use for different pres by just using different components... get 20 of those made and I'll be all set for power supplies for my tube pre experiments. So using different transformers back-to-back and different component values, I'm thinking:

- B+ will be anywhere from ~120V to ~300V at 30mA TOPS. Different pres will use different resistors in the 3 or 4 stage RC filter bank.

- 6.3V and 12.6V heaters will be possible by changing the resistor values around a LM317 regulator. The regulator will be preceeded by an RC filter to drop some voltage for 6.3V operation.

- Phantom power will be regulated by a TL783, preceeded by a tripler, which in turn is fed from the first transformer's secondary.

I found these 50VA Amvecos at DigiKey that might work... I'm thinking something like this...

... but I have a funny feeling by the time you're all done with it, it won't be nothing like this! :grin:

Peace,
Al.

 

[bcarso]

"If you are using a transformer in a dual diode centertapped cofiguration,
you can simply put a 1N4004 diode off one of the red wires. This will knock your voltage into the first cap in half. Going from 150 to 48 is easier, no? "

You just get half-wave rectified that way, CJ---not half voltage. Would that it worked that way ;-).

Al, I think you will want a quadrupler, as the loaded trafo, diode drops, and low line may have things sagging below what you can regulate to 48 with. And these multiplers have really terrible regulation and ripple too. For example a quadrupler with 470uF caps fed from 12V 60Hz with a ~30mA load manages about 56.5V with 3V p-p ripple.

 

[adamasd]

I have been thinking about this very thing for a couple days now. I was wondering about gas regulator tubes for this job? Are there any around 50 volts, it seems like it would be a nice simple way to get B+ down to phantom. Or at least useing one of the numorous ones that are around 100 volt before filtering down to 48. but I know nothing about these things, will they just pull what current is needed by the circuitry after them, or do they eat up some on their own? are they at all suitable for phantom power? If so are there any running around 50 volts.

adam

 

[CJ]

 

[bcarso]

The 991 operates at 59V but only good to about 2mA. The OA3 is rated up to 40mA at 75V.

I don't know why you would want to do this, but at least they look cool.

 

[alk509]

[quote author="bcarso"]Al, I think you will want a quadrupler, as the loaded trafo, diode drops, and low line may have things sagging below what you can regulate to 48 with.[/quote]

Absolutely, Brad. At first I was playing with a 15V secondary, so a tripler was good enough...

[quote author="bcarso"]And these multiplers have really terrible regulation and ripple too. For example a quadrupler with 470uF caps fed from 12V 60Hz with a ~30mA load manages about 56.5V with 3V p-p ripple.[/quote]

True... And even after the regulator, 3mV of ripple is still kind of a lot in a phantom power supply... but short of using a separate transformer for one of the supplies, I really can't think of any other way to do this! I could use 15V-secondary transformers, which coupled with a tripler would give me enough voltage and reduce ripple a bit... However, the cheap Amvecos only go up to 50VA, or 1.6A (AC) per secondary - a little low if you're pulling 1.2A DC for heaters...

And having to use three transformers in one box would suck so much ass!!!

What to do, what to do? :?

Peace,
Al.

 

[bcarso]

Well I can do a regulator with a few more (ok, a fair number more) parts that will handle the input ripple. So you could still use the multiplier.

 

[adamasd]

I don't know why you would want to do this, but at least they look cool.

I do not know why either, thats why I asked. They are cheap easy to get tubes, and you are already useing tubes for the pre so it is most likely not a problem to throw in another. As I said i know just about nothing about these tubes yet. Something like the 991 with more current seem like a really simple way to get that phantom power when all you have is 300 volts.

adam

 

[pstamler]

I have done the drop from 325V B+ to a 47V 5W zener, and have gotten good results. However, a friend has pointed out that if the zener fails open, my microphones will get fried by the B+. So I'm considering installing an MOV across the zener, the MOV being in series with an LED. If the zener fails, the MOV conducts and passes current to the LED, which lights up angrily and I yank the plug muy pronto.

Or I just go with another transformer and to heck with it.

Peace,
Paul

 

[alk509]

Thanks for all the replies, guys... I think I'm going to have to bite the bullet and use a separate 15V transformer for heaters. That way I can use two 44V Amvecos back-to-back to get B+ and phantom...

Peace,
Al.

 

[bcarso]

FWIW, there is an interesting article about shunt and other regulator noise here: http://www.tnt-audio.com/clinica/regulators_noise3_e.html

The designs could be readily extended to 48V.

I am still looking for a link for another article about active power filters, which are intended to be inserted in series with the basic regulated output voltage. Thanks to Norm Thagard for showing me these a while ago.

 

[bcarso]

Here is the link for the power filter stuff:

http://www.qsl.net/yo5ofh/projects/voltage_regulator_noise/finesse%20voltage%20regulator%20noise.htm

 

[alk509]

Thanks for the links, Brad! :thumb:

However, I'm starting to get used to the idea of using a separate transformer for the heaters... I'll take advantage of the situation and use the 12V supply to drive some pad and polarity relays as well :grin:.

But back to the Vreg noise article: So he's putting an AC current sink, which senses the AC (noise) voltage across a low value resistor which is in series with the regulator's output, and sinking exactly that amount of AC current so that it won't show up at the load... Am I getting this right? :?

Man, I'm stupid when it comes to these solid state things! :roll:

Thanks again!

Peace,
Al.

 

[bcarso]

You've got it Al. It's an active filter that is shunting the noise to ground. You lose a little bit of d.c. regulation but that's usually a small issue compared to the noise. One of the cool things is the energy required to reduce the noise is drastically lower with this approach compared to a second series regulator that has to pass all of the load current and incur a significant additional voltage drop.

There is also a topology that amounts to the same thing but with a two-pole response, but it is much trickier to implement.

I think your move to another trafo is the correct one, I just wanted to take the opportunity to let people know some of their options.

 

[alk509]

Great info. Thanks again!

Peace,
Al.

 

[ToobieSnack]

Al did i understand you correctly?
are you building 20 of these?
wow you must have some DIY schedule/aspirations :shock: :grin:
are you building a tube console?
if so... why not build ONE massive power supply: :wink:
any way ...nice thread... :grin:
later
ts