7

Ever play with high end condenser microphones? Like the ones used in test facilities for vehicles, and other stuff?

Well I do and it's not a cheap hobby, I have taken care of all the other issues (amplification, power supply, High-end ADC, etc.) but haven't figured out how to generate a +200V polarization voltage that is very stable. Any noise will significantly affect the output of the microphone. I should need at most a few milliamps, but in all reality, I'm dealing with microamps - the milliamps just makes it easier to deal with losses.

Sorry for the weird question, I'm a ME, not an EE. Also yes, I know they make power supplies specifically for microphones, but I can't justify spending $4k + wait for these long lead times.

Summary: I have various power input options, +/- 24V, +/- 12V, +12V, +5V, or 120V AC. My desired output is +200V (+ or - 0.5V, with a ripple of less than 50mV if not better) capable of supplying about 100 microamps (including losses at circuit) so I can measure it with my Agilent 34401A (10 MΩ internal resistance / about 20 microamp draw) while connecting to my microphone circuit.

DiskBiskit
  • 115
  • 5
  • 5
    Slightly left field but workable answer: giant stack of batteries. – pjc50 May 06 '22 at 21:38
  • 4
    As above, snapping 20 or 25 9V batteries together would be a low noise source. A bit of a shock hazard. – Spehro Pefhany May 06 '22 at 22:03
  • @SpehroPefhany aren't those known for thermal sensitivity? I have NO idea how much of a difference in mV would it do, but 20-25x multiplies any fluctuations by x20, making it a little more than one order of magnitude worse. I also think that stack-of-9V could do the trick, but I would try searching.. one source I remember is https://youtu.be/XpbDMo8an5w?t=1150 where the person mentiones "small gust of wind" and oscilloscope reading plummets, though I think it was in micro-Volt range, so seems far from acceptable 50mV even with /25 scale.. – quetzalcoatl May 07 '22 at 12:18
  • @quetzalcoatl https://tf.nist.gov/general/pdf/1133.pdf – Spehro Pefhany May 07 '22 at 15:51
  • Another battery stack option is button cells in plastic pipe. Using standard plumbing fittings makes life easier. I saw this in use about ten years ago when there was interest in modulated light communication to supply ~30v biasing currents with negligible noise. Using plumbing pipe would make heat insulation easier, although you'd need a fair few sections. Follow the batteries (of any kind) with a linear regulator and gusts of wind shouldn't matter too much, providing your voltage source is stable---which is easy with an 'IC zener', or indeed a floating lm317 setup. – 2e0byo May 07 '22 at 15:52
  • Are you sure you need 200 VDC +/- 0.5 volt? As long as the level is stable, would a wider nominal range be acceptable? And thermal issues should not be problem. You are not asking for DC response (or at least, since this is for microphones, I assume it's not,) Low frequency drift simply will not produce an audible output. – WhatRoughBeast May 07 '22 at 16:20

2 Answers2

6

The laziest approach is to simply buy a DC-DC converter. More like $50 one-off than $4K. An isolated one will give you the most flexibility. I see 10mV p-p ripple as a standard spec. Eg. Bellnix BYH series, but do your own searching at distributors. Of course you can add filtering to further reduce the noise.

Making one would probably best be done with a small custom transformer for that kind of specs, though other approaches are possible.

Spehro Pefhany
  • 376,485
  • 21
  • 320
  • 842
  • This is incredible, I've looked for DC-DC converters but got lost in the wash of just how many options that didn't meet my needs. I'll see if I can throw a quick circuit together and get it running. – DiskBiskit May 08 '22 at 01:39
0

Given the small current needed, you could use a voltage multiplier to avoid needing an unusual / custom transformer (this is common practice for 48V microphone supplies). This needs an AC voltage input, which could be a transformer winding, or a square wave oscillator running from a DC supply. I have seen a logic gate oscillator used to drive a multiplier for 48V phantom power - I think it was a hex inverter, one oscillated and then the other 5 in parallel provided the current to feed the multiplier circuit, running on something like 9V DC. Using a higher frequency than 50Hz will reduce the size of capacitors you need, both in the multiplier and the output filtering.

An unregulated power supply won't get the 0.25% voltage tolerance you specify (which seems a lot tighter than what B&K used to specify for their measurement microphones), so you will have to use a voltage regulator. Given the very small current, a shunt regulator would do (but a suitable 200V Zener diode could be hard to find), or you could use one of the adjustable linear voltage regulators like a LM317 (but bear in mind that an output short will destroy the chip, because it relies upon operating without connection to ground at these voltages).

RichardAshAudacity
  • 21
  • 3
  • A 200 V "zener diode" is called a neon lamp! (or two in series, perhaps.) Glow discharge tubes (*not* arc discharge tubes like neon signs; the little glow lamps like the NE-2) were the usual form of voltage reference before zener diodes were available, and they're still quite usable as such. They're just not used that often because they only work at high voltages, above around 70 or 80 volts. But here, that's an advantage! – Hearth May 07 '22 at 22:06