DIY Digital microphones

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mhelin

Well-known member
Joined
Mar 12, 2005
Messages
748
Location
Tampere, Finland
Apart from usual analog capacitance to voltage conversion techniques (high voltage biased capsule to impedance converter or various "RF bias" methods) you could also convert the capacitance of capsule to digital directly by measuring the charge time and compare it to an internal reference (fixed capacitor). That is how for an example the ams PICOCAP capacitance-to-digital converter (CDC) works.

https://ams.com/picocap-capacitive-sensing

This chip (less than $10 in single piece quantities) can sample at 50 kHz rate with 20 bits output, and has SPI and I2C outputs:

https://www.mouser.fi/datasheet/2/588/PCap04_DS000574_2-00-1379780.pdf

I don't know any mic using  that kind of digital conversion. AFAIK, Neumann's Solution-D uses a pair of conventional A/D converters 24 dB apart from each other to increase the DR.

The evaluation kit for above Picocap is too expensive, but maybe one could design own PCB for the chips available from Mouser. There is one problem though, the chip includes DSP which needs some programming. It might though be possible to use the exact procedure of capacitance measurement using some microcontroller which has fast enough GPIO and comparators available (old tiny Atmel AVRs for and example could be used to measure resistance that way overcome missing A/D converters).

There exist some Arduino applications for capacitance measurement, though those are too slow (but would be nice to have):
http://www.circuitbasics.com/how-to-make-an-arduino-capacitance-meter/

Now, the forthcoming Raspberry Pi will contain an user programmable FPGA chip with GPIO so that could be one application to try on it, though you could obviouslyuse any other FPGA. There are also many academic papers on the fast CDC's, but not being an IEEE or ACM member can't read them (for nothing), for an example:

https://ieeexplore.ieee.org/document/7833782
https://ieeexplore.ieee.org/document/5415404
https://dl.acm.org/citation.cfm?id=3057039.3057104
https://www.electronicdesign.com/analog/measure-capacitive-sensors-sigma-delta-modulator


 
Neat idea, i suppose, but... What exactly is the upside to all that complexity? :eek:

Well, direct to S/PDIF, TOSLINK or AES/EBU might be one, i suppose. And/or a stereo mic down a single standard mic cable...
 
but then you have to deal with syncing the digital signal, or at least src'ing... it opens a huge can of worms.

Plus, so far, not a single (profesional) manufacturer could "sell" the advantages of digital microphones to me, especiall as the analogue technologies are so far advanced that the sonical improvement is negliglible if existant, but on the other hand using them makes it both a smaller choice of microphones and increased difficulty/complexity in operations
 
This has been suggested by others in another forum. The idea was suggested to me three years ago by an instrumentation designer when I was first trying to investigate DIY RF condenser mics. I looked into the method at the time, but it was way beyond my comfort zone (I am a mechanical engineer; I like high pressure steam and combustion - pass the matches ;)).

50 kHz sampling rate will be able to measure to a maximum frequency of 25 kHz; good enough for CD quality (44.1 kHz sampling rate) or 48 kHz sampling rate. Certainly good enough for me, as an amateur, but not fast enough for 96 kHz or 192 kHz sampling rates.

But is it worth the development effort, to be limited to 44.1 kHz and 48 kHz recordngs? If you think so, then I will follow this topic with interest.

Incidentally, there are a lot of papers on sigma-delta measurement techniques that are publicly accessible. Most are to do with MEMS microphones or touch screens, but the theory is applicable and some of the papers do have a more direct relevance to what you are suggesting. I read a lot of them, but ditched them as the chips available at the time were too slow.
 
Back
Top