Need some ideas on P.I front end

[woody]

Just about every P.I has the input signal rammed up the input of a super low noise opamp. This is usually fed via a 1 or 2K resistor with a couple of 1n4148 dides acting as a clamp staight into the input of the opamp.
Guy's there has to be a better way as this is 1960's technology.

I have made a small improvement by clamping the back emf from the coil via a 270 ohm resistor shunted into an 18 volt zener, the output of this then goes via another 270 ohm R across the clamp diodes into the opamp.

I used an 18 volt zener as I just happened to have a few lying around.

My thoughts on the matter are red hot limiting resistors generating thermal noise. All diode clamps add capacitance to the circuit. Has anybody experimented with a syncronised fet shorting the input while the TX stage is firing? What about transformer coupling? There must be some bright sparks that have previously looked at this problem. Any thoughts on this most appreciated.

[Carl-NC]

I briefly looked at using an active FET clamp... after burning up a few opamps and still not getting the timing right, I dropped it. But that was in the early stages of the HH design, and I wanted a relatively foolproof circuit... I'd like to revisit the active clamp at some point.

- Carl

[woody]

p.i front ends

It is an interesting part of the circuit and thus can give the most gain and depth increases. I like the idea of using a clamping fet but one would need to take into account the output capacitance and see if we could still sample at 10us. As an experiment I used 2 coils, one on TX and another wound in the opposite direction as to not couple the TX pulse into the receive coil. I used a 10 ohm resistor into the opamp an used a single FDH300 low leakage type diode in series with 10 ohms as a clamp. sensitivity went up around 30% but I really want to see if I can do the same using a Mono coil.

[Unregistered]

PI Front Ends

I have wondered the same thing for years. I would have thought some enterprising engineers would have taken the balanced coil approach to PI's similar to that used in VLF's. Sure you would still need to eliminate the back emf as quickly as possible using the normal methods. But by separating the TX and Rx signals and using a balanced coil you no longer need to worry about popping the input opamp or having clamping diodes on the input. You could also increase the number turns in the Rx winding making the detector more sensitive. VLF's have thier Tx and Rx sections both turned on full time yet with a properly tuned coil there is little or no signal transfer between the Tx and Rx sections. You should be able to do this for a PI using either DD or Concentric coil designs. Maybe someone will try it someday! JMHO

[Geo]

What about transformer coupling? There must be some bright sparks that have previously looked at this problem. Any thoughts on this most appreciated.

Hi. I experimented with transformer coupling without better results.I agree with you about higher gain but at tests the difference is very small. For example now i play with two projects, Delta Pulse and one similar to Gary's. Delta pulse has total gain 30000 about and the other has 10.000.000. The air test gives about the same distance detecting. So why so much labour for high gain?
Regards

[Ferric Toes]

I'm fairly certain I wasn't dreaming, but I recall seeing a patent in which each turn of the separate receiver coil in a proposed PI mine detector had a cmos switch. So, for 30 turns, you would have 30 switches, which for all intents and purposes, made the Rx coil non-existant. A short delay after the Tx switch off and back emf spike, all the switches were simultaneously closed, and the coil connected to the preamp.

Eric.

[Paul]

I'm fairly certain I wasn't dreaming, but I recall seeing a patent in which each turn of the separate receiver coil in a proposed PI mine detector had a cmos switch. So, for 30 turns, you would have 30 switches, which for all intents and purposes, made the Rx coil non-existant. A short delay after the Tx switch off and back emf spike, all the switches were simultaneously closed, and the coil connected to the preamp.

Eric.

I played around with that idea in my head last year. In some instances it could be beneficial. Obviously you'll get less interaction from the Tx coil, but 30 switches adds noise to the preamp.

I would be more than happy to share some of my design details from previous and present design, but I doubt anyone's interested. I will do so if certain people agree to be civilized.

Paul

[Ferric Toes]

but I doubt anyone's interested.

Correct!

[pml]

Eric,

One day I will post the exact details. One thing you have to lose is a lot of sales since I am willing to freely give away such information.

Paul

[Qiaozhi]

Eric,

One day I will post the exact details. One thing you have to lose is a lot of sales since I am willing to freely give away such information.

Paul

Please do so, and report back with your results. You'll be away a long long time - hopefully.

[pml]

Please do so, and report back with your results. You'll be away a long long time - hopefully.

That's OK. You people have been a great personal steppingstone as it's given me a lot more compassion for you people.

God bless you Qiaozhi!
Paul

[Charles (Upstate NY)]

That's OK. You people have been a great personal steppingstone as it's given me a lot more compassion for you people.

God bless you Qiaozhi!
Paul

YAWNNNNNNNN...gawd you are so boring, whatever!

[Unregistered]

JC1

Gee whiz,,, the cap of 4148 is what 4 pf?

If I remember my math two would be 8 pf?

Glup,,, you can make this lower? ok

Whatever happen to things like disconnecting the input amp

from the coil? Corbyn did this when? in the 70s.

diodes are fine, if the thermal noise of the damping resistor

is bothering you, then you have done well. Congrats.

So disconnect it, the damp resistor, that is, when you

connect the rec coil.

the noise of one analog switch is not too bad, and besides,

good ol random white noise can be dealt with fairly well,

its that external stuff that start "beating" with your

sampling that will drive you mad.

[Unregistered]

JC1

Hi Toes,

I'm not the sharpest knife in the drawer, and I don't get

the thirty switches thingy. disconnect every turn? what for?

huh? really? this does something? gets government funding?

Hey did you get your demod going? it works better

down stream, and not to replace the diff integ in regular location.

I also had it removing dc off drift in previous gain stages,

allows NoMotion operation of the detector.

And boy does that make testing/using easier.

[Ferric Toes]

Hi JC,

Having some holiday "light reading" looking at the recent posts?

The problem is not with the parallel damping resistor getting hot, but also the input resistor to the opamp. Using a 1k, and the coil developing 400V on switchoff, the input resistor and the relevant diode have a 0.4A peak current. If the prf is about 3kHz then even a 2W input resistor gets too hot to touch. This is not too bad as the noise increases as sq.rt. of temp. Going from 25C to 100C only doubles the resistor noise.

As you say, the diode capacitance does not amount to much, and you can alway halve it by using 1N4448's.

Eric.