Oscillator inestability with some coils

[samu]

Hi everybody,

I´ve built a BFO metal detector to be placed under a conveyor belt to find big metals mixed with the product. The coil i used was 9 turns 0,9 mm thickness and elliptic shape (260x600mm). I don´t know anything about the inductance or resistance of the coil ¿any suggestion?.

Also, the metal detector is based in a common base Colpitts oscillator with a PNP BJT. One of the capacitors (the one connected between collector and transmiter) has 10nF, and the other, has 100nF.
The oscillation frecuency is 180Khz.

All is working OK, but before that, i was making tests with diferent coils and although all of them worked OK there was problems. The problem was inestability: sometimes the circuit make fake detections.
I think it´s important to choose the capacitances to make the circuit "with more Q" but i´ve no idea about doing it. Im almost sure that the problem is the coil-capacitance set that its not propertly selected because i use the same caps for all the coils. Maybe there are too many armonics.

I´ve read about Q, noise in oscillators, frecuencies of metal detectors but i didn´t find the answer im looking for.


Thanks in advance for any answer and for the forum. Sorry for my english, im trying to improve it.

[Qiaozhi]

Hi everybody,

I´ve built a BFO metal detector to be placed under a conveyor belt to find big metals mixed with the product. The coil i used was 9 turns 0,9 mm thickness and elliptic shape (260x600mm). I don´t know anything about the inductance or resistance of the coil ¿any suggestion?.

Also, the metal detector is based in a common base Colpitts oscillator with a PNP BJT. One of the capacitors (the one connected between collector and transmiter) has 10nF, and the other, has 100nF.
The oscillation frecuency is 180Khz.

All is working OK, but before that, i was making tests with diferent coils and although all of them worked OK there was problems. The problem was inestability: sometimes the circuit make fake detections.
I think it´s important to choose the capacitances to make the circuit "with more Q" but i´ve no idea about doing it. Im almost sure that the problem is the coil-capacitance set that its not propertly selected because i use the same caps for all the coils. Maybe there are too many armonics.

I´ve read about Q, noise in oscillators, frecuencies of metal detectors but i didn´t find the answer im looking for.


Thanks in advance for any answer and for the forum. Sorry for my english, im trying to improve it.

Could you post the circuit diagram?
Based on the frequency of 180KHz, and the two capacitors (10nF and 100nF) the coil inductance is 86uH.

[samu]

Thank you for the information about the inductance.

Here you have the scheme, after that there is a OPAMP so there is almost no current comsuption.

The power source is a typical 7824.

I used this oscillator because i found it in a scheme but i don´t know if its better to use other kind of oscillator, or FET better than BJT, or NPN better than PNP.

Im looking for stability in short time. it doesn´t mind in long time, the circuit stabilices itself. Also i don´t have a prefered frecuency because i use a rectifier and then a level meter.

I hope you have enought information. Thank you.

[Max]

Thank you for the information about the inductance.

Here you have the scheme, after that there is a OPAMP so there is almost no current comsuption.

The power source is a typical 7824.

I used this oscillator because i found it in a scheme but i don´t know if its better to use other kind of oscillator, or FET better than BJT, or NPN better than PNP.

Im looking for stability in short time. it doesn´t mind in long time, the circuit stabilices itself. Also i don´t have a prefered frecuency because i use a rectifier and then a level meter.

I hope you have enought information. Thank you.

Hi,
the problem is probably at capacitors. What kind of capacitor are you using in that circuit ?

If capacitance vary due e.g. to thermal effects you'll get wide frequency drift from that circuit.

The very best of the best are mica-based caps... difficault to find today, they are extremely stable and could give you real improved stability.

They are often made with silver over copper foil or just silver... and totally filled with epoxy: are enormously stable respect to e.g. common poly caps.

Problem is the fact you'll not find them easy... and costs. But are really worth in BFO applications.

Kind regards,
Max

[Qiaozhi]

Thank you for the information about the inductance.

Here you have the scheme, after that there is a OPAMP so there is almost no current comsuption.

The power source is a typical 7824.

I used this oscillator because i found it in a scheme but i don´t know if its better to use other kind of oscillator, or FET better than BJT, or NPN better than PNP.

Im looking for stability in short time. it doesn´t mind in long time, the circuit stabilices itself. Also i don´t have a prefered frecuency because i use a rectifier and then a level meter.

I hope you have enought information. Thank you.

Hi Samu,

There is a much better circuit, which almost as simple here ->
http://geotech.thunting.com/cgi-bin/...akes/index.dat
Have a look at "Metal Detecting Circuits II" and in particular the circuit in figure 5 - it's a crystal-filter metal detector.

[samu]

The Cap im using are made in polyester. Remember that my MD works perfect with one of the coils. I tested at least 10 more coils without changing the caps.

The problem is not the sensibility or the long time stability. Sometimes there are oscillations in the rectified signal that cause fake detections. Once, i saw the oscillation signal through a spectrum analizer and there was a lot of spurious frecuencies. I guess those frecuencies cause the fails.

I took a look to the scheme by Rakes and of course this scheme seems to be better for sensitivity but it´s also better for stability? However, i still ignoring why to use C2 and C3 with those capacitances and not others. Notice that the coil i´m using is really different from the coil in the project. I think C1 and C2 need to be recalculated if the coil change so much.

Thank you for your help , every suggestion will be studied.