Hi,

New to the forum, first post! I'm a mechanical engineer who dabbles in electronics (sorry, I'm a newb!). I'm building a phantom power supply box for an electret microphone. Simply applying 9-18vdc to an electret (2 pin) mic is all that is needed for it to work, unless you need more gain. A 9v battery works great on it's own. But when a 9v power supply is used, it brings AC noise. Shure has a suggested circuit here.

As I understand, R3-R5 and C3-C5 are for AC noise elimination. But my reading of bias capacitors for AC noise elimination suggests those caps should be different. And I've never seen 3 in a row like that. What is a better way to do this?

 

I mocked up this circuit using stuff I had laying around: 47 mfd's and 470 ohm resistors. It seemed to work, but there did appear to be a little LF noise, maybe a ground loop. I know, this is ugly.

 

It is a really crappy 3 section RC low pass filter. Each of the RC sections load the previous sections and the corner frequency is at 4.6 Hz. This is absolutely pointless for the intended purpose. It would help to know the purpose of the microphone (voice or music) to propose an alternative.



A good place to start would be to identify the range of frequencies that you wish to pass through the filter and how much attenuation you want in the stopband. The difference between the end of the passband and the beginning of the stopband will tell us about the order of the filter you need. For audio work it is unlikely that a passive filter can be constructed to do what you want it to do.

 

Yes, audio. Guitar. 80hz - 10k. But it's an internal mic so I'll only be using 300hz and up.

Without this 3 section RC filter, I got a lot of hiss. After adding the 3 section RC filter (above picture), it definitely helped a lot. I just seemed from my reading that I could do better.

 

Use a Transformer-based Power-Supply instead of a Switching-Power-Supply.
.
.
.

 

Transformer-based Power-Supply instead of a Switching-Power-Supply

This is interesting. Most if not ALL power supplies available to guitarists and their pedal boards are switching based. I would think linear transformer based supplies would be a fair bit more expensive. Not sure.

 

Here is an article written from the guitar world on the difference. Evidently, modern switching power supplies are better than linear ones.

Effects Pedal Power Supplies - White Paper - Strymon

 

So, it should be possible to construct a 3rd order bandpass filter for a passband of 300 Hz -10 kHz with a single opamp.

 

Your 2-wires electret mic is completely different to Shure's 3-wires electret mic so you need a different circuit.
An ordinary cheap 2-wires electret mic is powered from filtered 9VDC in series with a 10k ohms resistor and it needs a preamp circuit. Without the 10k series resistor then the mic is shorted by the battery or by the power supply.

Your wiring is a group of antennas that pickup interference. All microphone wires must use shielded cables with very short bare wires. That is why Shure said, "in a metal box".

 

"" Here is an article written from the guitar world on the difference.
Evidently, modern switching power supplies are better than linear ones. ""

The things that are "better" about a SMPS is that they are .......
Smaller, more light-weight, cheaper to manufacture, and usually have higher efficiency.
These things are not deal-breakers in most projects.

The problem is, they may require much more sophisticated Output-Filtering,
because they are naturally extremely "Noisy" Electrically.
Sophisticated Output-Filtering costs Money and Circuit-Board-Space,
so, your mileage may vary, depending upon which Model and Manufacturer You choose.

Transformer-based Power-Supplies have
only 50, or 60hz "Hum" which needs to be filtered,
and are otherwise, generally, Electrically-quiet.

They're ~twice as big, and weigh ~3-times as much,
and usually don't present You with any difficult surprises.

There are some very high-performance Circuit-Designs that will do what You want,
but the Spec-Sheet for the Microphone will be required,
or possibly, a more appropriate Microphone may need to be chosen.

A bare Electret-Condenser-Microphone is not going to do well pushing a signal through a long Cable.

A 9-Volt-Battery is going to be the most trouble-free choice.
.
.
.

 

Your 2-wires electret mic should be connected like this:

 

Going to larger capacitors in your filter network will probably reduce the noise without affecting DC performance.

 

Your 2-wires electret mic should be connected like this:

Heyyyy, that works! And it has an almost imperceptible noise floor. I breadboard it up, didn't have a .33mfd, so I put a .1 and a .22 in parallel. Same on the resistors, had two 470's in series.

So help me understand what's going on here. The 100 and 47 mfd are similar to the triad of Shure 47 mfd's, right? They bleed off ac ripple voltage to ground (my terms might be wrong). R1 is to get the right current to the mic. And C1 is maybe a HPF for the signal?

So what if I wanted to change some parameters? Like 12-18vdc? Or full range guitar, 80 Hz and up? What equations did you use?

Thanks man, this is awesome!

 

The datasheets for many of the 800 electret mics sold at Digikey show that the mic draws a little less than the maximum current of 500uA and needs a voltage across the mic of at least 1.5V so I guessed a current of 450uA and a dead 9V battery at 6V.
Then for 1.5V across the mic then the resistor gets the remaining 4.5V at 450uA which is 4.5V/450uA=10k ohms. Then when the battery is 9V for the mic to have a current of 450uA, the voltage across the resistor is 450uA x 10k ohm= 4.5V and the mic also has 4.5V across it.

 

So the 33nf, that's a HPF right? My mic has an output impedance of 1700 Ohms. So using a .33mfd gives a cutoff frequency of 284 Hz. To get full range guitar of 80 Hz, I'd want 1.2 mfd. Am I missing anything?

 

The output capacitor FEEDS the amplifier's input impedance.
I used 330nF (not 33nF) so that the low frequency response is flat down to a hifi 20Hz when feeding an amplifier input that is 100k ohms. 330nF is 0.33uF.
For the response to be flat (not -3dB) at 80Hz then you calculate a cutoff frequency of (80/5=)16 Hz.
What is the input resistance of the mic input on your amplifier?

 

Right, typo on the 330 nfd, yes equals .33 mfd.

The guitar mic (Shure WL50, output impedance 1.7k) goes into a Line6 Helix Stomp (multi-purpose pedal) that has auto-z input. But it can be manually set to anywhere from 10k to 1M. The other pickup in that guitar prefers 1M... and they have to be set the same so that is what I'd suggest using: 1M. So if that's the case, a cutoff of 16 is a 10 nF. Or if I wanted a higher HPF like flat at 300, that's 300/5 for cutoff and a cap of 2.7 nF. I think.

What happens without a cap there?

 

The output level from the pickup on a guitar is much higher than the tiny output level from a nicrophone. Usually a microphone connects to the mic input on an amplifier or to the mic input on a mic preamp.

Without a coupling capacitor from the microphone then the 450uA of current goes into the input resistor of the mic preamp instead of to the electret mic. Also the few volts of DC from the 10k resistor that powers the Jfet in the mic goes into the mic input on the amplifier, possibly muting that input.

 

I cut a small piece of prototyping pcb and soldered up this circuit, it's about 1" x 3/4". Try as I might, I couldn't get it in a 1590L size box (2"x2"). But I printed a 2x2.5" and that worked. I changed the (yellow) cap to a 10 nf, which gives a cutoff frequency of 16 hz and a high pass filter of 80 hz... assuming 1M input impedance on my Helix (which I set it as).

Pedalboard below that. The Cioks power supply is held in place by a 3D printed bracket; white filament, with top spray painted black.

 

I'd like to add an opamp to this, here's a circuit I'm considering I found on University of Washington site:


Before the opamp, it is very similar to what Audioguru Again suggested. I'm new to opamps. As I understand it, inverting opamps are only recommended for AC applications as it flips the phase, but not needed for DC?

Also, the gain would be the 100k pot / 2.2k to ground, right? When the pot is at zero, it's unity gain. When the pot is at 100k, the gain would be 45x. So if I wanted 10x, I'd pick say 100k pot and 10k. Right?