Yet another option would be cleaver use of Capsense in a high modulus, extremely flexible silicone rubber.

Your idea of a slitted tube is similar to the idea I was intending of using a split ring or a structure similar to a pair of tweezers with the strain gauge mounted on the joint which would be thinned down thus generating a high strain concentration.

As for coupling. I agree that AC coupling is actually quite appropriate here. muscles can only respond at about 20-25Hz (interesting how that number comes up in all things bioengineering, max discernible frame rate is 25fps-ish and the lower limit on hearing is 20Hz-ish) and something round 0.2-0.5Hz is probably the lower limit needed here. Anyhoo. just have to build it and get someone to test it. (amazing how much work can be glossed over with \”just\”)

P

Considering how handy these sensors are, especially for human-oriented applications, I’m surprised they’re not more widely known and used. They’re not precision/long-term stable to tight tolerances, but they are very sensitive, and a single op-amp can linearize the signal pretty well (The shematic’s available somewhere in the interlink docs; email me if you want it and cannot find it online.) These work on a slightly different principle than classic piezo-resistance; these are closer to percolation (basically a non-tunneling version of what the QTMs do — pressure causes an increase in many physically small resistors in parallel.)

I’ve read about the QTM-based sensors, but haven’t gotten ahold of any. I’m pretty sure I have another interlink sensor somewhere; would you be interested in a swap by mail?

When I was considering using strain-gauges for a similar application, I was thinking about using a thinwall, slit cylinder as the structure — with the gauges mounted opposite the slit. Given the force levels (and thus small strains), one would need either Silicon strain gauges ($$) and/or lots of gain. One thing that might help is that, for this application (unlike others), DC accuracy/stability isn’t needed, so one could AC couple the amplifier, making it easier to get high gain without saturation. IMHO, most of the relevant signal would be in the decade 0.5 Hz — 5 Hz.

HTH,

Larry
PS I’m reachable at ursine at gma1l d0t c0m — with the obvious corrections.

Interesting, not seen this type type of sensor in this form on the market before. There are other piezoresistive technologies I am wondering about using since a strain gauge would be need to be attached to a structure to create stress from the stress generated by the muscular force acting on said structure. I was therefore looking at Quantum Tunnelling Materials (QTMs) which are mainly made by http://www.peratech.co.uk/ or a MEMS pressure transducer in a cast block of high modulus silicone. Also PVDF, PZT (and other ceramic) and quartz Pressure and displacement transducers are an option. But I like your idea.

P

I’ve had good luck using “force sensing resistors” from
http://www.interlinkelectronics.com/ or
http://www.tekscan.com/flexiforce.html

These are nonlinear (resistance roughly inverse with force) but the sensitivity is *huge* compared to strain gages, and are quite sensitive — very useful for this application!
The small ones are just over a cm in diameter.

Note that interlink’s website is not IMHO very hobbiest-friendly (minimum purchases, etc. so you may have better luck buying small quantities from a reseller.)

Second note: the whole structure (including the leads) is printed, and can easily be damaged by folding/creasing.
However, bending around a cylinder seems OK, just avoid sharp creases.

Hope this helps,

– Larry