Phantom pwr from a DC to DC converter?

Is this possible?...to just tap off of your Preamp's incoming DC (+18v or whatever)?
I have not seen this option. (Maybe there is a reason :idea: )

http://www.maxim-ic.com/appnotes.cfm/appnote_number/911

=FB=

good question, want to know that too

I think Sony made a +24 to +48 DC-DC converter. I know that a few people used them when racking old modules.

Sony part
1-464-758-11
Converter ,DC-DC (CD-54)

there was a picture somewhere on the old GDIY site but I can't find it now

I think it can lower than 24 volts input

Ah Sony.
Problem is that neither Farnell, RS components stocks them in Europe, or am I wrong ?

I don't think anyone bar Sony will stock it ?

sony parts don't grow on trees :sad:
some repairmen have good contacts with Sony and Sony Parts suppliers and generally they are not in catalogues.
But
often if you ask for it specifically you can get them

yes there are many small DCDC converters for 48 since that is a telephone standard, however most output less than 200ma..

i'll find some when i get a chance.

My BA - N*VE supply uses a 24-48v module from Endicott Research Group , 1000uf on the input, and 1000 and a .1 on the output.

It's markings are:

ERG INC
E728-.4850U
0440A C

how about a small boost converter? I have a design at work that will be fine for this.. let me dig it up and iron it out.

could be an IC, a few caps an inductor and a schottky then also be adjustable.

is low cost or ease of build more important? I could do it with a PWM IC that costs .50$ or a ready made IC for a couple of $. using the PWM ic takes a few more supporting parts but makes the layout bigger and more complex.

we have to define what voltage and current we expect to use for the input of this thing.

which would be less of a problem.. operation at 300khz, or 700khz?

m-Audio's MobilePre USB gets Phantom from the USB +5V rail with a good old 555 and about 8 stages of voltage multiplier.

I have not found my magnifier to see how it is wired, nor put a meter on it to see what the voltage really is, but I see the rows of diodes and caps, and it lit-up the Marshall 990 mike which is apparently external-bias.

The Maxim plan leads to 48V at 1 Amp! We only need 10mA per mike, and we don't have 100 mikes. Don't reach for the big hammer when simple tricks work.

yes the maxim circuit is also entirely too complex.

check out the national lm5000 series. just a few caps, an inductor and diode. Fet is internal.

there are some others with even fewer parts.

you can do it with a hexinverter/schmitt trigger array too.

but I'd prefer something with less parts myself.

the 555 is a quick and dirty way of doing it but in my experience the 555 jitters entirely too much although this is not really a problem here but there is another outstanding problem here, unless you have some other kind of comparator or schmitt trigger to condition the output, you get a funkly triangle wave which will drive the FET switch nuts.. or is this driving the V multiplier directly??

http://www.paia.com/tubehsch.pdf

something I have been meaning to do tests on. Some of the premade 24V power supplies use a voltage X2 diode cap circuit for a 2nd higher voltage supply.

So inside the power supply before the regulation circuit there will be two raw voltages the higher current for the 24 output and a lower current added to it, or the base drive of the NPN series pass section. I guess this is done because when the circuit was designed PNP powers might have cost more and NPNs and the cap and diode cost less. You would not need the 2nd higher voltage if a PNP series pass was used.

Now I wonder if the cap was increased in value for the base drive if one could add a reg for 48V for the phantom.

> in my experience the 555 jitters entirely too much

You have too much experience, you think too much.

> get a funkly triangle wave which will drive the FET switch nuts.. or is this driving the V multiplier directly??

FET switch???? I don't see no steekin FET in there. A genuine 555 (I understand they are obsolete) will throw 200mA each way, what more do you want? 20mA at 50V multiplies up from 200mA at 5V, which is inadequate (we can't do 100% duty cycle), but true 10mA mikes are fairly rare and maybe it does not multiply up to a full 48V.

There is a resistor and a backward diode on the output so it is probably a dumb shunt Zener regulator. Again, simple trumps fancy.

The Paia is similar. They just feed a tube which scorns regulation, so no Zener. They start with more voltage so fewer stages. They need less current and have more voltage so the 4000 series CMOS was good enough for them, and maybe if we get Phantom from 18V, but not from 5V.

The inductor for a flyback booster will be an odd value, either not commonly stocked or regular stock except somebody else bought a million yesterday and all the distributors are showing empty. The caps for a multiplier are so dead common and unfussy that you should never have trouble finding something that works, and you get the kilo-quantity discount.

You have too much experience, you think too much.

Click to expand...

:shock:

I didn't think there was such a thing.. oh wait there I go thinking again..

FET switch, I'm used to seeing these in most switching supplies and most switchmode psus i have dealt with are much higher current than we need.. I guess i'm over engineering again.

The inductor isn't that hard to find in such a small cored part. Most are a function of the inductance, the reverse recovery of the diode, and the cap itself. you can fudge a little on the inductance if you have a large enough diode.. but this is bad form for a production device.. but great for DIY!

let me do some math when I get back to work with all my paperwork and I'll see if it's even cost effective.. I'm thinking less than 6$ for all the parts.

And you can get a *near* square output from a 556 timer, using the other half as a comparator. works OK as long as you don't vary the duty cycle when it's working. you get the strange artifacts I was hinting at before.

If PRR could build this supply out of a tube and a transformer I bet he would.. :green:

The POE (power over ethernet) is done pretty much the same way. Look at the Linear Tech chips for this. Lots of ways to make a 48V output (from just about anything). I'd suggest stocking with an inductor-based type, the switched capacitor and multipliers types are very noisy thanks to the current spikes.

But you will definately need to add some passive filtering stages.

jh

> If PRR could build this supply out of a tube and a transformer I bet he would..

It's hard at 5V, or even 18V. At 130V, it was once common: every other US home had a hollow-state 20KV flyback boost converter in the TV.

> you can get a *near* square output from a 556

And...? We don't need a square, we need any kinda wiggle that we can jack-up. Square is often more efficient, but we only talkin' a half-watt per mike, nothing's going to burn-up even at 50% efficiency. And square's efficiency also means it throws spikes all over: a falttened-top sine-ish wave might throw less hash.

IIRC, you build those large motor controllers. They have to be just-right or they burn up or sell at a loss. This is a different world. No doubt there are sophisticated little step-ups in some audio boxes, built by sophisticated guys like you. But in DIY Audio we just stick things together and the simpler the better.

> an inductor-based type, the switched capacitor and multipliers types are very noisy

Hmmmm. I ran some noise plots on that MobilePre USB, and noted a small spike near 3KHz. Only open-input, totally masked by room and mike noise. I blamed the room, full of PCs. But now I'm thinking: 555, lowest-price diodes, that could have been the Phantom supply's 555 wiggle frequency. Yeah it is in the audio band but, without any sealed box, it was near-enough neglgible for a $150 box.

At a gross level, the switching spikes are the same. The input current toggles between the supply rails and is equal to output current times multiplication times about 2 for 50:50 duty cycle. The coil likes higher frequencies and probably won't be 3KHz, so may be easier to filter. Cheap diodes don't like to break 20KHz so a many-diode/cap multiplier may end up in Audio. I'm sure the diodes don't all switch at the same instant so it may leave a ripple like a crippled centipede. Both ways are ugly; are they uglier than running to the store for five 9V batteries every 8 hours? Or in wall-power work, doing something at 50/60Hz?

The real answer, since we are stuck with +48V, is to design the preamp for +48V.

Hi,

The LM2577adj in a TO5 package ($6 Digikey) looks like a very good candidate for the voltage step-up converter to +48V (or for a converter from 24 to +/-15V also).

Regards,
Milan

Here is a further proposal.

This step-up circuit
http://web.telia.com/~u31641623/48%20V%20phantom%20from%209%20V%20battery.jpg
have I use for long time ago, when it was hard to find IC step-up circuits.
What I remeber, this design function and work ok.
(I think I have find it in a audio magazine, R-e/p or dB around 1970/80)

--Bo