How can I improve this switching circuit for PI?

[moodz]

I would be using bipolar pulsing ... but if you want monopolar then go big and consider a more radical design like this. Make one of the inductors your storage core ..literally a ferrite core inductor L1 ... and the other inductor L7 is your transmit coil.
Uses a lot less power also. Adjust C1 and C2 for rise time, peak flyback voltage, and tilt etc. No need for neg supplies.

[CrizzyD]

I like the design, the issue I'm going to have is that the coil characteristics are going to change considerably from unit to unit. In addition to this, the coils are connected to the main circuit with coaxial cable which can be any length (say 3m up to 20m) so things like cable capacitance/inductance have to be taken into account. My attempt above is to effectively just brute force a large magnetic field irrespective of the coil or any parasitic characteristics. I still wonder if there's maybe an "on the PCB" coil that could do something similar to what you suggest however...

[moodz]

I like the design, the issue I'm going to have is that the coil characteristics are going to change considerably from unit to unit. In addition to this, the coils are connected to the main circuit with coaxial cable which can be any length (say 3m up to 20m) so things like cable capacitance/inductance have to be taken into account. My attempt above is to effectively just brute force a large magnetic field irrespective of the coil or any parasitic characteristics. I still wonder if there's maybe an "on the PCB" coil that could do something similar to what you suggest however...

Hmm .. if your TX coil varies so much in tolerances ..you are going to have a problem with detecting small objects at great distances as you indicated previously was your aim because the RX will have trouble with the variation in TX parameters to be optimal.

However the circuit I posted is fairly immune to tolerance changes. I doubled all the coil values ( inductance x 2 , cap x 2 , resistance x 2 ie 100 % ) and you get the same result ( had to lift the supply voltage from 5 to 6 volts though ... no biggie.

[CrizzyD]

Agreed that the number 1 issue is the coil tolerance differences. It's like having a bad TV aerial and still expecting your reception to be perfect - hence the reason why I'm simply going for maximum energy output and accounting for the losses. I'll have a stab at simulating your circuit and get back with any comments. Thanks!