Capacitance-to-Digital converters

Hi all...

Has anyone out there had any experience using Capacitance-to-Digital converter chips (CDCs) as the basis of a Theremin, or similar?

If you are familiar with any, perhaps you could recommend one for me to use: i want to try constructing a single-antenna (at this stage...) "theremin" capable of outputting digital proximity data in 16-or-more-bit resolution, with a data rate of at least 1000Hz. I've thought about MIDI, but as far as I'm aware the format only allows for 7 or 8-bit resolution (correct me if I'm wrong...?).

It's the high data rate that I think is the critical factor here, since the digital signal will need further processing before being fed to some kind of oscillator or software synth, thus creating sluggishness in response. From experience, I think that any delay longer than about 1 or 2 100ths of a second between the player's movement and the audible output is too long...

Some chips I've looked at are the Analog Devices AD7747/AD7142/AD7150etc range, and things like Quantum's QT300. The AD chips look like they might not be fast enought (?), but has anyone used the QT300 and know what the data rate from that is?? I've checked the spec sheet (http://www.qprox.com/downloads/datasheets/qt300_102.pdf) but i'm afraid it's a little bit over my head at this stage... it looks promising though.

any thoughts?

Hi Jon,

The range of capacitance is relatively narrow on a Theremin. Consider that the frequency change that results from moving towards the antenna is only about 5000hz. On an oscillator that is running at 500khz, the percentage change in frequency is only 1%. Thus, your capacitance to digital converter would have to be sensitive to a tiny change in capacitance.

The other issue is latency -- if the converter has much latency (more than about 5ms) the theremin will have noticeably sluggish response.

Of course, a capacitance-to-digital design is compelling because one wouldn't need to use oscillators or heterodyne design -- just measure the capacitance and that's it!

The most attractive feature of such a design is the prospect of implementing linearity in software.

All the best to you. Keep us posted.

exactly. actually, from my small amount of technical knowledge, it seems that both the Quantum and AD sensors are extremely sensitive to very small changes in capacitance.

problem is that, at least judging from Quantum's replies to my emails, these chips are really targeted at applications on small hand-held consumer devices, such as the slider wheels or touch buttons on iPods, etc. they don't really have the capability (that i know of) to measure capacitance in air to a distance of, say, up to 2 feet as would be required for a digital theremin.

so, some things i need to figure out are:

- can readily available CDC chips somehow have their sensitivity adjusted to work at larger distances? maybe something to do with antenna design or tank capacitance???

- if so, will they still be accurate?

- do they have a high enough data rate and low enough latency to be of use in the first place?

as you can tell, i'm a bit over my head here. but i think the possible end result is worth trying for, eh? :)

in case anyone's interested, i've had a little bit of success using a QProx QT300 capacitance-to-digital converter to make a kind of digital theremin. you can read a bit about the project on this forum thread:

http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1169728880

i'm using a programmable arduino board (based on the Atmel ATMega8 chip) to communicate with the sensor, then relay data to a PC for further processing. it's exciting, because the theremin can now be used to control... well, virtually anything, rather than just pitch and volume. but i'm still only just scratching the surface - i need to do a lot of work to improve the accuracy of the sensor if it's going to be at all useful as a theremin. i also need to figure out a way to increase the sensitive distance of my antenna.
if anyone has any comments or suggestions, i'd love to hear them!

jon.

Sounds like what's probably going into all those cheap cap meters out there on the market (or are they still using the time constant method?)

There's a company called Almost All Digital Electronics, www.aade.com that makes a neat little LC meter kit. In fact I recommend building it (just watch that tantalum cap, positive has to face toward the LCD) it's obviously doing C to digital, and it's a through-hole hobbyist level kit.