Research proper Electret mic driving requirements

[SmittyHalibut]

5vDC Bias through 2.2k is WAY TOO HOT for every radio I've used it on. Below is an example of what I measured by another radio.

With the FHM-3 hand-mic plugged in, the Flex 6400 has 1.8vDC at the Mic connector port.

Internally, the Mic+ goes through R1, a 2k series resistor, then branches two directions:

1. Mic: The electret mic element, and a small chip capacitor (unknown value, but probably very small), are in parallel to the Mic- pin.
2. Power filter: A .33uF electrolytic capo, and R2, another 2k resistor, in parallel, to Ground. (separate from Mic-).

The DC voltage across R1 is 1.10v (0.55mA), and the DC voltage across R2 is 0.70v (0.35mA). This means there's .20mA across the sum of leakage current on the caps, and through the electret mic element.

I don't know what the internal-to-the-Flex resistance and supply voltage are, but given the 0.55mA total current measured above, these are the common combinations that would work:

• 1k, 2.3v
• 2k, 2.9v
• 3.33k, 3.3v
• 5v would be 5.818k

Wait a sec, the "way too hot" here is the headset mic going straight into the Flex. This isn't my Electret driving circuit. I'm still filing this ticket to document the above stuff about the FHM-3.

[SmittyHalibut]

Oh, look at this:

https://edge.flexradio.com/www/uploads/20200818185014/FLEX-6400M-and-FLEX-6600M-User-Guide-v3.x.pdf#%5B%7B%22num%22%3A92%2C%22gen%22%3A0%7D%2C%7B%22name%22%3A%22XYZ%22%7D%2C69%2C496%2C0%5D

aka: https://www.flexradio.com/documentation/flex-6400m-and-flex-6600m-user-guide-pdf/ Page 32

Looks like it's a 3.0v and 2.1k. Which matches my measurements above, pretty closely at least. 3-1.8v = 1.2v/.00055A = 2.181k. Close enough for me.

[SmittyHalibut]

http://mynixworld.info/2017/09/01/simple-electret-microphone-circuit/

Looks like the goal is: 2v across the Electret, .5mA through it. R = (V_cc-2v)/0.55mA

[SmittyHalibut]

2v across the Electret, .5mA through it, gives:

• 5vDC: 6k
• 3.3vDC: 2.6k

TODO

• Update the OHIS document accordingly.

[SmittyHalibut]

Putting these notes here for lack of a better place to put them:

http://www.uneeda-audio.com/pads/

Shows how to pad an electret down to dynamic level. Specifically:

• Output impedance of the pad needs to be low, 100R or so. Expect a load impedance of 600 to 1.5k ohm.
• That means that the shunt resistor needs to be that impedance. Then the series resistor is selected for the appropriate dB loss.
• Because I like operating in powers of 10 when I can, and I don't need it to be EXACTLY 20dB... I'm going to 100R and 1k for my attenuator resistors.

[SmittyHalibut]

That didn't work. The 100R shunt across the output just loaded something too much and the output was low and distorted. I think we need to trim first, then buffer (again).

[SmittyHalibut]

OR, and here me out here, a 100R shunt resistor on an electret circuit just pulls too much DC. If I put a DC blocking cap in series with it, it'll not pull the DC too low, but will still attenuate the AC.

Problem is, 100R needs 100uF to pass 16Hz. This is effectively a high pass filter, because whatever we don't shunt will get less attenuation and appear louder at the mic input to the radio. So if this shunt cap is too small, we'll be sending too much rumble to the input. So we have to be careful here.

[SmittyHalibut]

Testing shows a 47uF electrolytic cap in series with a 1k trim pot configured as an adjustable resistor, across the OHIS Mic+ and MicGnd, will attenuate a signal, of both the electret mic element I tried (CTIA headset), and the 5 component dynamic mic amp (with Heil Pro 7 mic element). The cap blocks DC and only attenuates audio. The adjustable resistance creates a voltage divider with the source impedance of the electret mic/mic amp output.

So, the simple case of an adjustable User device is the bypassable mic amp, with the 47uF/1k pot across the output. I think I like this design.

[SmittyHalibut]

Ok. Summary of what we have here:

• DC Bias Voltage with series resistance applied to drive about .5mA through a 2V drop. eg:
  • 5vDC: 6k
  • 3.3vDC: 2.6k
• Expected output:
  • Thankfully and coincidentally, mic sensitivity is measured at about comfortable speaking voice level. So the sensitivity rating (units usually something like "dBV at 1Pa at 1kHz") is about what you can expect to see from the microphone when used like we are using it.
  • RMS, About -42dBV (8mV) to -48dBV (4mV).
  • Peak voltages, between 11.3mV and 5.6mV. (Peak voltage becomes important when designing a VU meter.)
• Some electrets are hotter than this. An attenuator can be added across the Mic+ and Mic-/MicGnd lines. A 47uF (or greater) capacitor, in series with a 1k adjustable resistor (or pot with wiper tied to the clockwise end). The capacitor prevents DC from passing through the attenuator. The cut-off frequency is 1/2PiRC. R is the 1k pot, possibly turned down to ~100R or lower. C is the 47uF. 100R47uF = 34Hz high pass. 10R47uF = 338Hz.

Works with dynamic mic preamp, as documented here: https://github.com/Halibut-Electronics/Open-Headset-Interconnect-Standard/issues/3

From a standards point of view, I think the critical numbers are the -42dBV to -48dBV, and the resultant 11.3mV and 5.6mV peak values. In addition to the Mic Preamp in https://github.com/Halibut-Electronics/Open-Headset-Interconnect-Standard/issues/3, the standard should include a VU meter design:

Specifically, the voltage divider ladder values.

• Use a 1.2v regulator or voltage reference to define the top of your voltage divider ladder RELATIVE TO THE AUDIO REFERENCE POINT! Note that the "Gnd" pin on U2 is tied to Bias, so it's 1.2vDC above the audio center, NOT above 0v or Vdd.
• Three resistors: 5.1k, 4.7k, 91k
• The junctions between those values produce: 0.0607v and 0.117v, just about right for the -42dBV to -48dBV range (see below).
• Note these are 10x expected AC peak values! Notice that the precision rectifier on the left side also has 10x gain (100k input, 1M feedback). This just makes the voltage divider values larger, easier to get precision from a 1.2v reference.
• Peak hold time on the VU meter is defined by the RC of C7 and R16. These values (1M and 1uF) are pretty good (qualitative, I know...)

Ok. I think that's the summary of everything I've learned about Electrets. This will need to get put in the OHIS standard doc. See https://github.com/Halibut-Electronics/Open-Headset-Interconnect-Standard/issues/10

Resolving.