PI DAMPING EQUATION CHALLENGE

[Gunghouk]

I think this is close

[moodz]

I think this is close

You have a way to go yet ... your solution does not damp for about 4 microseconds ... the Teleno solution would be about 650 nanoseconds.

[Teleno]

You have a way to go yet ... your solution does not damp for about 4 microseconds ... the Teleno solution would be about 650 nanoseconds.

Have you figured out how to physically realize the formula? I'm thinking of a MOSFET that turns on when the current in the coil approaches zero, shorting the coil to its reference ground.

[Qiaozhi]

Have you figured out how to physically realize the formula? I'm thinking of a MOSFET that turns on when the current in the coil approaches zero, shorting the coil to its reference ground.

I'm not sure I understand the point of this challenge. If you artificially force-damp the coil flyback to zero, then surely you also kill off the target signal?
Is there really a free lunch here, or is this just a wild goose chase?

[Teleno]

I'm not sure I understand the point of this challenge. If you artificially force-damp the coil flyback to zero, then surely you also kill off the target signal?
Is there really a free lunch here, or is this just a wild goose chase?

The target signal is alive in the Rx coil. The scheme is for a balanced PI.

[Qiaozhi]

The target signal is alive in the Rx coil. The scheme is for a balanced PI.

OK - with a balanced coil I can see you're suggesting the eddy currents from the target will still be present in the RX coil.
I'm still not certain this will actually work in practice, as there will effectively be a short across the TX coil which will upset the system.
This is one of those occasions where it needs to get out of the simulator and into the real world. Let's see ...

[Teleno]

OK - with a balanced coil I can see you're suggesting the eddy currents from the target will still be present in the RX coil.
I'm still not certain this will actually work in practice, as there will effectively be a short across the TX coil which will upset the system.
This is one of those occasions where it needs to get out of the simulator and into the real world. Let's see ...

The Tx transient is like the transient of a bipolar CC system, except that the TX coil is clamped to GND instead of to a negative voltage, so if the CC works then this will work as well.

[KingJL]

... This is one of those occasions where it needs to get out of the simulator and into the real world. Let's see ...

Actually Vallon did something somewhat similar to this in the VMH3CS detector... and that was with a mono coil. The short across the coil only lasted ~400 - 500 ns just at the zero coil current point... the coil damping R was also witched in at the zero current point. The patent said it was to discharge the residual charge of coil capacitance.

[moodz]

Well done everyone ! HERE IS THE SOLUTION
The one I got anyway. I will give Teleno first prize with honorable mention to KingJL for insight and encouragement awards for everyone else.

The formula I arrived at gets the coil current down to 10s of microamps ( practically zero ). I am pretty sure the object of the Vallon patent was to detect the extremely small amounts of metal in land mines and they were successful in this AFAIK. MY understanding would be that the energy in the coil ( after TX off ) is transferred to the capacitance and at I = 0 and peak flyback voltage then all the energy is stored in the cap. By dumping the cap into a resistance R the energy is dissipated. You cant actually just short the coil or no ( insufficient ) energy will be absorbed ... there has to be an R to absorb the energy ( C and L cant absorb energy by themselves ) ... the R is proximated to root(2) in my formula.
If you sim the solution ( there could be other solutions ) you will see that the current in the cap almost exactly matches the current in R. The time constant CR will be very small as the capacitance is 300p x 0.7071 ohms.

The implementation of a real solution is now fairly easy ... its all a matter of timing

[KingJL]

Actually Vallon did something somewhat similar to this in the VMH3CS detector... and that was with a mono coil. The short across the coil only lasted ~400 - 500 ns just at the zero coil current point... the coil damping R was also witched in at the zero current point. The patent said it was to discharge the residual charge of coil capacitance.

And they (Vallon) did this ~28 years ago!!!

[moodz]

And they (Vallon) did this ~28 years ago!!!

WOW ... I am not sure how they implemented it but circuit I am looking at uses a diode switch with say nanosecond timing to dump the cap energy. ( a 5$ arm chip has nanosecond time resolution switching )

The diode only stays on for the duration of the dump so the coil does not remain shorted. I will probably have to slip in an additional R to clean up residual ringing.
In the pic below 0.546 microseconds practically nails it... a feedback loop will optimise it.

[Carl-NC]

I suspect there will be some Devilish Details here. The flyback peak occurs at t = τ = sqrt(L/C_L) but the problem is that L depends on permeability, which varies with ground. So you will need a way to dynamically detect the voltage peak (or the zero current) to adjust the clamp timing. Even then, swinging over a decent hot rock could saturate the preamp.

[KingJL]

WOW ... I am not sure how they implemented it but ...

Here is an LTSpice of the actual Vallon TX (traced and verified) from an actual board.

[moodz]

I suspect there will be some Devilish Details here. The flyback peak occurs at t = τ = sqrt(L/C_L) but the problem is that L depends on permeability, which varies with ground. So you will need a way to dynamically detect the voltage peak (or the zero current) to adjust the clamp timing. Even then, swinging over a decent hot rock could saturate the preamp.

No doubt ... and the Vallon method posted above has a fair complexity. I am working on a high power version of the MAGPI and the method used there may solve this problem also. The current MAGPI depends on a sampling feedback loop to damp the coil. An improved version uses autodamping so no CPU required for the damping function. Using the new method very high "diode speed" response times can be achieved and damping is optimised much faster than target swing speeds. Relatively low gain high voltage preamps with larger dynamic range and a more robust ananlogue front end AFE deal with overload conditions. ( hopefully )

[KingJL]

... and the Vallon method posted above has a fair complexity...

Totally agree... it was not posted to be "here is how to do it"... only to say "here how it was done 28 years ago". For 28 years ago, I would say it was pretty much state of the art and an elegant solution. Today, it is "complex" and should be able to be simplified (greatly)... but the principles remain the same!