coil to plus input

[green]

Ordered some parts to disconnect amplifier + input from coil during coil decay for faster decay. If they don't work, any ideas on how to do it? Advantages to + input over - input if any? I was thinking board layout would be easier if all the decay current went thru R damping.

[Qiaozhi]

Ordered some parts to disconnect amplifier + input from coil during coil decay for faster decay. If they don't work, any ideas on how to do it? Advantages to + input over - input if any? I was thinking board layout would be easier if all the decay current went thru R damping.

The coil current decay curve can be split into three sections:
1) MOSFET avalanche.
2) Preamp input resistor in parallel with damping resistor (down to forward voltage drop of protection diodes).
3) Damping resistor alone.

#1 can easily be excluded by not allowing the MOSFET to avalanche by either, using a higher voltage MOSFET, reducing the TX on-time, or adding a suitable series resistance.

#2 means that the damping resistance is decreased for a large part of the curve. For example - if your damping resistor is 680R and the input resistor is 1k, then the resistance is actually 405R. Significantly lower than expected. Since you're using a low gain in the preamp, then it would be possible to increase the input resistor to a higher value, let's say 100k. In this case, the overall damping resistance drops to 675R.

With the decay of section #2 now closer to #3, the calculation of the target decay constant should be more accurate. This could be a simpler alternative to the more complicated method of disconnecting the amplifier input.

[green]

The coil current decay curve can be split into three sections:
1) MOSFET avalanche.
2) Preamp input resistor in parallel with damping resistor (down to forward voltage drop of protection diodes).
3) Damping resistor alone.

#1 can easily be excluded by not allowing the MOSFET to avalanche by either, using a higher voltage MOSFET, reducing the TX on-time, or adding a suitable series resistance.

#2 means that the damping resistance is decreased for a large part of the curve. For example - if your damping resistor is 680R and the input resistor is 1k, then the resistance is actually 405R. Significantly lower than expected. Since you're using a low gain in the preamp, then it would be possible to increase the input resistor to a higher value, let's say 100k. In this case, the overall damping resistance drops to 675R.

With the decay of section #2 now closer to #3, the calculation of the target decay constant should be more accurate. This could be a simpler alternative to the more complicated method of disconnecting the amplifier input.

I've been using a couple formulas. R damping=coil inductance(uH) times(pi) times circuit(resonance Mhz),300*3.142*1.1=1036 ohms and approximate critically damped decay time=3/circuit resonance. Disconnect the damping and input resistor from the circuit and place scope probe near coil or mosfet to measure resonance. With spice the decay looks best with just R damping connected. I was thinking lower input resistor lower noise, maybe not important. Formulas seem to work with spice. Maybe not practical to disconnect input but something to try. Worth a try for short delay times. Spice circuit to play with. Connect input to coil and change R1, R2 or coil specs and see what happens. If I could disconnect the input during the decay, R damping could go with the coil making it easier to get best response when changing coils.

[Qiaozhi]

I've been using a couple formulas. R damping=coil inductance(uH) times(pi) times circuit(resonance Mhz),300*3.142*1.1=1036 ohms and approximate critically damped decay time=3/circuit resonance. Disconnect the damping and input resistor from the circuit and place scope probe near coil or mosfet to measure resonance. With spice the decay looks best with just R damping connected. I was thinking lower input resistor lower noise, maybe not important. Formulas seem to work with spice. Maybe not practical to disconnect input but something to try. Worth a try for short delay times. Spice circuit to play with. Connect input to coil and change R1, R2 or coil specs and see what happens. If I could disconnect the input during the decay, R damping could go with the coil making it easier to get best response when changing coils.

Are you sure your 300uH coil has a DC resistance of 3 ohms? Seems a bit high to me.

[Davor]

I'm not quite following your + over - thing, but if it comes to choice between P-channel or N-channel MOSFETs, the P-channel ones have significantly lower leakage current in off state. Spice models generally do not model leakage at all.

[green]

Are you sure your 300uH coil has a DC resistance of 3 ohms? Seems a bit high to me.

I have a large spool of awg28 enamel coated wire that I use to make the coils until I know what I'm doing. The resistances vary. 200mmID spider web 23 turns, 300uH is about 3.3 ohms. The coil resistance, power supply and Tx time were adjusted to keep the voltage from avalanching for the spice model. I have been connecting the coil to the minus input using input resistor for R damping(depends on the op amp holding minus input at ground). Just wanted to try the plus input to see if could be better.

[green]

I'm not quite following your + over - thing, but if it comes to choice between P-channel or N-channel MOSFETs, the P-channel ones have significantly lower leakage current in off state. Spice models generally do not model leakage at all.

N channel MOSFET. The +and- thing is connecting the coil to the + or - input of the preamp. When connecting to the + input I want to open the connection to the + input during coil decay so it damps with R damping only, not with R damping and input resistor and diodes.

[green]

I have two different N channel depletion mode mosfets and a p channel Jfet on order to try. Looks to me like it might work, maybe not(won't be the first time) not sure about noise. Was wondering if it didn't was there something else to try. I'll post what I tried whether it works or not.

[crane]

An RX switch causes ringing when it reconnects the coil to the op-amp and this requires a longer delay than conventional methods.

Minelab use an RX switch to eliminate the input resistor and minimise resistor noise.

[greylourie]

Which is the lesser of two evils ? If the RX switch ringing is only for a short period ? Commercial examples using both examples show that sensitivity is finer with RX switches. Seems like there is a limit when using big R on the input. (Using the old R on the input is proven for finding fresh drops, run of the mill objects, even fruit dishes deeper at burial sites. But for that extra edge, particularly on sensitivity to smaller or deeper objects, maybe, having the input R is the barrier ?)

I look forward to seeing any future measurements from green. I very much appreciate all he has done so far. It makes quite a difference for a dummy like me, when I can see tables with figures and comparisons like he has produced. It helps a lot, to understand or reinforce what everyone else points out.

[Qiaozhi]

I have a large spool of awg28 enamel coated wire that I use to make the coils until I know what I'm doing. The resistances vary. 200mmID spider web 23 turns, 300uH is about 3.3 ohms.

Enamelled wire - that's why the resistance is higher.
I have a 450uH coil here with a DC resistance of 1R5, but that's wound with PVC coated wire.

[Qiaozhi]

Since you're using a low gain in the preamp, then it would be possible to increase the input resistor to a higher value, let's say 100k. In this case, the overall damping resistance drops to 675R.

I decided to actually try this in practice, but the results were disappointing, as the preamp output became very unstable. Adjusting the damping resistor did not help at all, so I would say that this idea is a failure. We now have another way not to make a light bulb (so to speak).

[mickstv]

I decided to actually try this in practice, but the results were disappointing, as the preamp output became very unstable. Adjusting the damping resistor did not help at all, so I would say that this idea is a failure. We now have another way not to make a light bulb (so to speak).

A 100k resistor is going to have approx 40nV/√Hz noise at the input to the first amp, then after you amplify it the noise would swamp target detection.

[Qiaozhi]

A 100k resistor is going to have approx 40nV/√Hz noise at the input to the first amp, then after you amplify it the noise would swamp target detection.

I'm not sure if that's the reason for the instability. I also tried simulating this in LTSPICE, and you can see evidence of the same thing happening there, even though the resistor models are ideal.

[green]

I'm not sure if that's the reason for the instability. I also tried simulating this in LTSPICE, and you can see evidence of the same thing happening there, even though the resistor models are ideal.

A 10k or 15k resistor is faster than 100k with spice for me. Got the parts, been busy and haven't had time to try yet. Hope to soon. I'll try larger input resistors also.