Simulating Target Response

[Dean Sarelius]

Starting this thread as I would like to encourage thoughts around techniques for simulating a target on the search coil when running simualtions.
Normally when checking the waveform response from a target I personally either add or subtract 5uH from the 300uH coil in the simulation.
My feeling is that this is generally a good value for a small target response say a 1 euro coin about 20cm from the coil.
Was wanting to know if anyone else has some experience and may suggest that this value is either too large or too small for a practical setup.

[Teleno]

Starting this thread as I would like to encourage thoughts around techniques for simulating a target on the search coil when running simualtions.
Normally when checking the waveform response from a target I personally either add or subtract 5uH from the 300uH coil in the simulation.
My feeling is that this is generally a good value for a small target response say a 1 euro coin about 20cm from the coil.
Was wanting to know if anyone else has some experience and may suggest that this value is either too large or too small for a practical setup.

The usual technique is to add an RL circuit with the desired tine constant and couple the L to the transmit coil with a K coefficient. Usual values are 1uH and 1 Ohm for a nugget or 10-20 ohms for a coin. K can be from 0.01 up. One end of the RL circuit must be attached to GND.

[Carl-NC]

Here is what Teleno is talking about:



I usually set R_TGT to 1Ω and use L_TGT to set the tau. This will model any non-ferrous target between 90° and 180°. It will not model magnetic responses (0° - 90°) like ground or iron.

[Dean Sarelius]

Thanks guys this is very helpfull.
So if you make the gold nugget a value of 1uH and 1 ohm with K 0.01 and you wish to compare the no target response do you simply change the inductance to zero for example..?

[Dean Sarelius]

In addition to my previous post what parramters would you use typically for the Rx coil..?

[Carl-NC]

My drawing is a generalization of a system with a separate TX & RX coil. In a mono PI simulation you only have the TX coil and the target.

Yes, you can set L_TGT=0 or k = 0 for a no-target reference.

[Teleno]

Thanks guys this is very helpfull.
So if you make the gold nugget a value of 1uH and 1 ohm with K 0.01 and you wish to compare the no target response do you simply change the inductance to zero for example..?

Use the SPICE command .step to vary the resistance, use a very high value for no target, like 100 MOhm. It will run a simulation for every value of R in the list.

[Dean Sarelius]

Thanks thats perfect. If I were to go dual coil what parramters do you suggest for an ideal RX coil..?

[Carl-NC]

In sims, just about anything will work. I tend to use 1mH for no particular good reason.

[Dean Sarelius]

Hi Carl,
For a flat spiral wound receive coil with 80mm diameter does this look about right to you or would you chose different dimensions perhaps..?
https://www.circuits.dk/calculator_f...l_inductor.htm

[Carl-NC]

If you are building a coil then it depends on the overall coil design. I can't answer to that.

[Dean Sarelius]

Thanks Carl
On a side note
In order to optimise the target respnse with your expereince what would you consider the best timing for the SAT versus target S&H.
For example lets say I have a very fast mono setup with sub 10us flyback, bi-polar pulses spaced at 100uS with repeat interval of 300us.
Sampling on the second pulse and ignoring the flyback the approximate frequency would be 2500 samples per second.
Maybe this sample rate is too high and will need to be adjusted but for argumants sake how would you divide the samples per second for the SAT would you go with averaging 10 or 50 samples for the SAT or more perhaps..?

[waltr]

Thanks Carl
On a side note
In order to optimise the target respnse with your expereince what would you consider the best timing for the SAT versus target S&H.
For example lets say I have a very fast mono setup with sub 10us flyback, bi-polar pulses spaced at 100uS with repeat interval of 300us.
Sampling on the second pulse and ignoring the flyback the approximate frequency would be 2500 samples per second.
Maybe this sample rate is too high and will need to be adjusted but for argumants sake how would you divide the samples per second for the SAT would you go with averaging 10 or 50 samples for the SAT or more perhaps..?

Is there an integrator circuit before the SAT circuit??
What is it's timing?

From other forum discussions start with coil sweep rate and time of target under coil. This gives basic time required parameters in which to design the integration and Sat timing.

[Carl-NC]

I generally try to set the sampling integrator tau for 100-500 integrated samples. For 2500pps this means a tau between 40ms and 200ms. Faster gives a zippier response, slower gives more depth. Maybe start with 100ms.
You generally want the SAT to be slower than the demod. If it's faster, you are wasting samples. Nominally I would make it 2X slower but I also like a variable SAT, so I might make it adjustable from 1X to 10X.

[Dean Sarelius]

Thanks Carl this is very helpfull...cheers