Would this work?

[lucifer]

Hi,
I've been wondering if the following PI mono coil variant would work? Basically it's a center-tapped coil which is driven at the tap or a step-up autotransformer. This allows the RX part to have more turns than the TX part and hence possibly increase the sensitivity. I'm aware it won't allow for earlier sampling. Just curious if it's worth doing any further experiments with that configuration. Thanks.

[Carl-NC]

Yes, it will work. The additional RX turns will actually slow things down somewhat, which is the trade-off here.

[lucifer]

Yes, it will work. The additional RX turns will actually slow things down somewhat, which is the trade-off here.

Thanks, Carl! That's what I thought. It's likely comparable to separate TX and RX coils which are not inductively balanced.

[kt315]

what your idea? splitted system gets small delta pulse signal on RX. it gives go-at-left main sample you can choice.
in your circuit you have full delta pulse on left end of RX coil. i do not see a sense in the connection.

[lucifer]

what your idea? splitted system gets small delta pulse signal on RX. it gives go-at-left main sample you can choice.
in your circuit you have full delta pulse on left end of RX coil. i do not see a sense in the connection.

The idea is that TX pulse drives L2 only (300 uH) and the the RX signal is received by L1+L2 (900 uH). In other words the coil has more turns when in RX mode. The additional turns will add some parasitic capacitance and will make the overall coil performance somewhat slower. However, the received signal might be slightly stronger.

[kt315]

good, what we have.
1. ordinary coil system. monocoil. on hot end of coil full delta pulse. on cold end a signal limited by diodes.
2. your coil system. two coils. on hot end of Rx coil full delta pulse. on cold end a signal limited by diodes.

correct me.

[lucifer]

good, what we have.
1. ordinary coil system. monocoil. on hot end of coil full delta pulse. on cold end a signal limited by diodes.
2. your coil system. two coils. on hot end of Rx coil full delta pulse. on cold end a signal limited by diodes.

correct me.

Hi kt315,
I see your point. What you say is correct but the full delta pulse, as you call it, is different in amplitude between a monocoil and the posted variant. If we compare a 300uH (say 25 turns) monocoil to a center-tapped 300uH + 600uH coil (say 25 turns + 25 turns) same wire gauge, the TX voltage and current will be identical but the amplitude on the hot RX side will be higher for the center-tapped coil, having all other parameters the same (supply voltage, mosfet switch, etc.). That can be best understood if you consider the center-tapped coil as a step-up autotransformer.



The output voltage of the transformer is increased by the ratio of coil turns N2/N1 = 2 in the above example. This should be valid for any target signals too. So you get a slightly better sensitivity at the expense of slightly delayed sampling.
I'm aware that autotransformers are usually wound on a magnetic core and an air-core coil will have significant losses. I have some ideas for improvement as well as reducing the parasitic capacitance somewhat but will see how that ends.

If you want to compare a 900uH monocoil to the same center-tapped coil then you'll get different current/voltage waveforms for the TX. A 900uH monocoil will have a higher resistance and parasitic capacitance compared to the 300uH part center-tapped coil. They will be equally fast and equally sensitive in RX mode.

To summarize I think the proposed variant has a tiny advantage over the ordinary monocoil: additional sensitivity at the expense for the slightly delayed sampling.

[kt315]

with the big signal in RX coil you can not move at left, on X axis, imagine this picture.
big amplitude has not relation to sensitivity.
for example, Eric did 1mH coil detector to decrease TX pulse to get more sens - like i wrote above.
his idea in-generally to get 5us MAIN DELAY, to take a sample at left of X, while you want to get just bigger RX amplitude.
so main rule - than you take a sample more to the left the more you get the sensitivity.
if your idea is just a depth.

[lucifer]

with the big signal in RX coil you can not move at left, on X axis, imagine this picture.

Agree, I'm aware of that and mentioned it in my first post and in my next posts again. RX signal is increased but you should sample later (more to the right on X or time axis) because additional coil turns add inductance and parasitic capacitance - so coil tau is slower. You have a stronger target response but you need to sample later. And fast tau targets response is weaker or none. This is the trade-off, nothing is perfect.

big amplitude has not relation to sensitivity.

I don't quite agree with that. After all in classic PI we sample for changes in amplitude. That's why we have a preamp too. BTW, the preamp also shifts the sampling point to the right. It amplifies the signal but adds some additional delay before you can sample depending on opamp's tau.

[Carl-NC]

A PI mono coil is a trade-off between TX ampere-turns and RX induction-turns, plus parasitic issues. On the TX side fewer turns is better, and on the RX side more turns is better. The autotransformer coil lets you have it both ways.