freddylocks wrote:Hi Pijoy
The following should be correct. Can you cast your eye over it and see what you think or any thoughts:
http://www.falstad.com/circuit/#%24+1+5 ... istance%0A
I am hoping that the above should show you a circuit diagram that I drew up this evening.
I have not entered the correct figures for the Highpass, lowpass or amplifier parts of the circuit.
Any thoughts?
Freddylocks,
I don't know that circuit simulator, and it's showing an error for me.
I hope you'll take my comments below in the spirit intended, namely to help you avoid pitfalls at the design stage, so you don't have to re-solder everything to remedy mistakes.
If the leftmost symbol is a battery (standing in for the excitation voltage), then you have the sensor linearizing stage correct. If it's a ground symbol (hard for me to tell, with the low-rez drawing), then the FSR will see ground on both ends, so it won't conduct any current, so constant output.
I'd put a buffer stage after your second filter section, and I suspect you'll need lots more gain.
Your R/C values for your filters should be re scaled for the OpAmps, and their knee frequencies need to be revised to get them anywhere close to what you want. For an R/C filter, the knee is at 1/(2 * pi *R * C) Hz. For your high-pass stage, your values give F = 1.57*10^4 Hz(?!?) -- way, way too high! Your low-pass is at 159 Hz. I'd put the low-pass below 60 Hz, and you need to have F_high_pass < F_low_pass in order to make a band-pass filter. As is, virtually nothing will come out of this circuit, because your high pass will only pass frequencies above 1.57*10^4 Hz, and your low pass will block everything it gets that's above 159 Hz, leaving nothing. You want the high-pass to block DC, but not much else, so bring that knee frequency way, way down.
To re-scale your R/C values, think about how much current an op-amp would have to source through those 100 ohm resistors if the OpAmp output were at either supply rail -- more than most OpAmps are happy with. I try to keep resistors loading OpAmp outputs (or the equivalent impedance of whatever network is loading the output) above 1K or (preferably) higher, so the OpAmps are sourcing or sinking only 12 mA max. and won't cook. This is slightly less critical if you're running the OpAmps from a lower voltage, like the Arduino's 5 volts. For a robust design, pay attention to both current limits for each amp as well as the worst-case power dissipation in each package -- including those 100 ohm resistors!
Finally, remember that textbook circuits designed for dual-supplies for the OpAmps need appropriate mods to work right from a single supply.
Please don't take my comments too hard. You're making progress.
Best regards,
PiJoy
