resolution

Thread Starter

shreyas_bhat

Joined Jul 26, 2004
47
Hi Folks
Thanks for all the responses so far.

I've created an oscillator and the two (out of the three) capacitors that form the Twin T oscillator are actually sensors that I have created for a project.

The final aim is that as the capacitance changes due to certain parameters, the oscillator frequency changes, which has to be picked up. Now, the reality is that the change in capacitance is pretty small (0.01 pF), and I would still want an appreciable change in frequency because of that. (I am operating at 1 MHz).

Any tip or ideas, on how to achieve that ?
 

pebe

Joined Oct 11, 2004
626
Originally posted by shreyas_bhat@Feb 14 2005, 03:22 PM
Hi Folks
Thanks for all the responses so far.

I've created an oscillator and the two (out of the three) capacitors that form the Twin T oscillator are actually sensors that I have created for a project.

The final aim is that as the capacitance changes due to certain parameters, the oscillator frequency changes, which has to be picked up. Now, the reality is that the change in capacitance is pretty small (0.01 pF), and I would still want an appreciable change in frequency because of that. (I am operating at 1 MHz).

Any tip or ideas, on how to achieve that ?
[post=5324]Quoted post[/post]​
One way would be to use an arrangement similar to the simple BFO metal detector. You use a fixed, stable oscillator - xtal preferably.

You then have a VFO which uses you variable capacitor to change its frequency. Mix the two together and get out a variable beat frequency, which you can measure. It might be simpler to use an LC oscillator, rather than your twin-tee though.
 

Thread Starter

shreyas_bhat

Joined Jul 26, 2004
47
hi Pebe
i wud really appreciate if u cud send me a block diagram of the same.

Thanks
Shreyas

Originally posted by pebe@Feb 14 2005, 04:40 PM
One way would be to use an arrangement similar to the simple BFO metal detector. You use a fixed, stable oscillator - xtal preferably.

You then have a VFO which uses you variable capacitor to change its frequency. Mix the two together and get out a variable beat frequency, which you can measure. It might be simpler to use an LC oscillator, rather than your twin-tee though.
[post=5327]Quoted post[/post]​
 

pebe

Joined Oct 11, 2004
626
Originally posted by shreyas_bhat@Feb 15 2005, 12:59 PM
hi Pebe
i wud really appreciate if u cud send me a block diagram of the same.

Thanks
Shreyas
[post=5337]Quoted post[/post]​
Hi shreyas_bhat,
This link will give you plenty of info on BFO metal detectors and you should be able to sort out a block diagram from them.

http://www.thunting.com/cgi-bin/geotech/pa...le=projects.dat

Metal detectors use an LC combination for the sensing oscillator, with a fixed capacitor and a variable inductor. In your case you would need to use a fixed inductor and your variable capacitor.

Suitable frequencies for the oscillators would depend on your application. You mentioned cap changes of .01pf, but not the nominal value of the caps.

Also most of the simple BFO detectors output to a loudspeaker. I assume you would probably want some sort of display? The best way to do that depends on your application. Can you give any more details?
 

Thread Starter

shreyas_bhat

Joined Jul 26, 2004
47
Thanks for the quick response.

My application requires me to use my capacitive sensors as the two capacitors in a Twin T network. Assuming that I am not permitted to deviate from this, what I need now is that as my sensor capacitance varies, the Twin T oscillator frequency varies. This shift in frequency is what I am aiming to measure. The BFO that u suggested is a good idea. Yes, I would need some kind of a display instead of an audible output. The sensor capacitance is around 100pF, and the variation in the capacitance is around .01pF.

I hope this info helps.


Originally posted by pebe@Feb 15 2005, 05:21 PM
Hi shreyas_bhat,
This link will give you plenty of info on BFO metal detectors and you should be able to sort out a block diagram from them.

http://www.thunting.com/cgi-bin/geotech/pa...le=projects.dat

Metal detectors use an LC combination for the sensing oscillator, with a fixed capacitor and a variable inductor. In your case you would need to use a fixed inductor and your variable capacitor.

Suitable frequencies for the oscillators would depend on your application. You mentioned cap changes of .01pf, but not the nominal value of the caps.

Also most of the simple BFO detectors output to a loudspeaker. I assume you would probably want some sort of display? The best way to do that depends on your application. Can you give any more details?
[post=5339]Quoted post[/post]​
 

pebe

Joined Oct 11, 2004
626
Originally posted by shreyas_bhat@Feb 16 2005, 02:51 PM
Thanks for the quick response.

My application requires me to use my capacitive sensors as the two capacitors in a Twin T network. Assuming that I am not permitted to deviate from this, what I need now is that as my sensor capacitance varies, the Twin T oscillator frequency varies. This shift in frequency is what I am aiming to measure. The BFO that u suggested is a good idea. Yes, I would need some kind of a display instead of an audible output. The sensor capacitance is around 100pF, and the variation in the capacitance is around .01pF.

I hope this info helps.
[post=5357]Quoted post[/post]​
OK. What frequency does your Twin-T oscillator run at?
 

Thread Starter

shreyas_bhat

Joined Jul 26, 2004
47
1 MHz.

Also, if you could suggest the mixer design, life wud be a lot better.

quote=pebe,Feb 16 2005, 10:59 AM]
OK. What frequency does your Twin-T oscillator run at?
[post=5364]Quoted post[/post]​
[/quote]
 

pebe

Joined Oct 11, 2004
626
Hi shreyas_bhat,
The capacitor variation of .01pF should give you a shift of only about 100Hz, so your reference oscillator will need to be crystal controlled for stability. You will also need to trim your Twin T osc to give zero beat from the mixer.

I’ve Googled for mixers but most are for mixing audio channels. There appears to be a dearth of info on RF mixers. I’ve come back to this article from EDN - you could modify the circuit.

http://www.thunting.com/geotech/pages/metd...cts/edn/edn.pdf

The mixer (IC1) is an LM389 amp similar to the LM386. It accepts inputs from two oscillators running at 370KHz. It should be OK for IMHz oscillators. The output will vary between DC and somewhere in the audio frequencies – depending the maximum change in your sensors.

You could measure using a frequency meter, or feed it to a 555 to generate a fixed width pulse each cycle. The mean DC level at the 555 output would then give an analogue voltage proportional to frequency.

Hope that helps.
 
Top