Vallon VMH3CS Mine Detector

[crane]

Sampling that portion of an under damped waveform should result in a significant response from a lump of soft ferrite or non-viscous magnetite. Probably doesn't matter though if the detector is only used on non-mineralised beach sand.

[Ferric Toes]

This is getting way off topic, but OK, it is relevant in many ways and would no doubt improve the detection of mines with only a tiny amounts of metal.

I couldn't resist getting out my ultra-fast PI and taking a picture of the response. The coil is 5.5in diameter, 330uH and wound with Teflon insulated 0.25mm solid wire. Resistance is 5 ohms and the coil is shielded with copper fabric tape spaced from the coil with polythene spiral wrap.

TX pulse width is 12uS, delay 4uS and 1st sample 4uS. pulse rate 25,000. Coil switch on TC 10uS.

Included in the the damping circuit is a cermet preset resistor for fine trimming.

This only works at such a short delay if there there is no cable to add extra capacitance. With 1m of good cable 5uS should still be attainable. Alternatively, the inductance could be dropped to, say, 200uH.

Eric.

[green]

This is getting way off topic, but OK, it is relevant in many ways and would no doubt improve the detection of mines with only a tiny amounts of metal.

I couldn't resist getting out my ultra-fast PI and taking a picture of the response. The coil is 5.5in diameter, 330uH and wound with Teflon insulated 0.25mm solid wire. Resistance is 5 ohms and the coil is shielded with copper fabric tape spaced from the coil with polythene spiral wrap.

TX pulse width is 12uS, delay 4uS and 1st sample 4uS. pulse rate 25,000. Coil switch on TC 10uS. [ATTACH]37137[/ATTACH]

Included in the the damping circuit is a cermet preset resistor for fine trimming.

This only works at such a short delay if there there is no cable to add extra capacitance. With 1m of good cable 5uS should still be attainable. Alternatively, the inductance could be dropped to, say, 200uH.

Eric.

Nice looking trace. With a mono coil mine looks closer to Teleno's. Would you mind listing peak coil current, amplifier gain, amplifier frequency response, damping resistance, coil self resonance and anything else that might help us get closer to a perfect decay?

[Teleno]

Nice looking trace. With a mono coil mine looks closer to Teleno's. Would you mind listing peak coil current, amplifier gain, amplifier frequency response, damping resistance, coil self resonance and anything else that might help us get closer to a perfect decay?

I can tell you what my case is:

Circuit: MPP rev-D,
DD balanced coil.
Tx pulse: 100us,
Tx Current: 1.1A,
Tx coil: 580uH - 1 Ohm, damping: 700 1Ohm,
Rx coil: 500uH - 5.4 Ohm, damping: 1000 1Ohm,

The MOSFET (IRF740) is avalanching at 400V for a few usecs. Not good. I'll replace it with a 900V MOSFET and see how it improves the decay.

[KingJL]

Curiouser and curiouser... what a puzzlement!!

Previous excerpts:

I wonder if they are using the high voltage spike from one pulse to kick start the next opposite polarity pulse in order to get the fast rise time.
....
All I have gleaned from a quick look at the boards is that the bipolar TX is probably an H bridge arrangement as there are four Mosfets close together.
...
By the way, they use a 7 pin connector for a mono coil. I think I figured out why, very elegant.

I think I have too! Plus I think I have found the use for 2 of the 4 MOSFETS... like maybe disconnecting virtual ground from the coil and kickstart capacitor during flyback, breaking the damping R path and providing voltage doubling charge to the kickstart capacitor while it is de-referenced from virtual ground... Plus, I think I am missing about 3 ohms (remember 6 Volts at 1 amp) in the TX coil to virtual ground (probably in the flyback ground de-referencing MOSFETS)... I need to do some modeling!!! Any chance you can read the P/N of the MOSFETS?

[Ferric Toes]

Nice looking trace. With a mono coil mine looks closer to Teleno's. Would you mind listing peak coil current, amplifier gain, amplifier frequency response, damping resistance, coil self resonance and anything else that might help us get closer to a perfect decay?

Peak coil current is 0.3A

Coil self resonance 1Mhz without cable.

Flyback volts 130 with no avalanching and Mosfet is IRFD210 (200V)

Preamp AD8055, not particularly low noise but OK for the application I did this for, which was broken off needle tips in clothing manufacture, but using a 3in coil and an even shorter delay.
High speed
300 MHz, −3 dB bandwidth (G = +1)
1400 V/μs slew rate
Low noise: 6 nV/√Hz

Low pulse current is compensated by much higher pulse rate of 25K to give good S/N.

I'm not sure about preamp gain but I think it was 100 or less. Gain made up later on in dc amplifier stages. I will have blow off 17 years dust to examine the board with a lens to see the feedback and input resistors.

Eric.

[Ferric Toes]

Previous excerpts:

I think I have too! Plus I think I have found the use for 2 of the 4 MOSFETS... like maybe disconnecting virtual ground from the coil and kickstart capacitor during flyback, breaking the damping R path and providing voltage doubling charge to the kickstart capacitor while it is de-referenced from virtual ground... Plus, I think I am missing about 3 ohms (remember 6 Volts at 1 amp) in the TX coil to virtual ground (probably in the flyback ground de-referencing MOSFETS)... I need to do some modeling!!! Any chance you can read the P/N of the MOSFETS?

I will have a look the next time I open up a unit. At least they dont erase part numbers but they are difficult to see through the varnish coating which I assume is to give additional environmental protection.

Eric.

[green]

Peak coil current is 0.3A

Coil self resonance 1Mhz without cable.

Flyback volts 130 with no avalanching and Mosfet is IRFD210 (200V)

Preamp AD8055, not particularly low noise but OK for the application I did this for, which was broken off needle tips in clothing manufacture, but using a 3in coil and an even shorter delay.
High speed
300 MHz, −3 dB bandwidth (G = +1)
1400 V/μs slew rate
Low noise: 6 nV/√Hz

Low pulse current is compensated by much higher pulse rate of 25K to give good S/N.

I'm not sure about preamp gain but I think it was 100 or less. Gain made up later on in dc amplifier stages. I will have blow off 17 years dust to examine the board with a lens to see the feedback and input resistors.

Eric.

Thanks for the info. I've had it where I could sample at 4usec with a 300uH mono coil, amplifier gain 300 by connecting the coil to the -input and using the input resistor for Rd but the decay trace is no where near as clean as yours. Might have to play some more.

[KingJL]

... Plus I think I have found the use for 2 of the 4 MOSFETS... like maybe disconnecting virtual ground from the coil and kickstart capacitor during flyback, breaking the damping R path and providing voltage doubling charge to the kickstart capacitor while it is de-referenced from virtual ground... Plus, I think I am missing about 3 ohms (remember 6 Volts at 1 amp) in the TX coil to virtual ground (probably in the flyback ground de-referencing MOSFETS)...

Plus an elegant means for accurate current limiting!

[Ferric Toes]

Thanks for the info. I've had it where I could sample at 4usec with a 300uH mono coil, amplifier gain 300 by connecting the coil to the -input and using the input resistor for Rd but the decay trace is no where near as clean as yours. Might have to play some more.

I mostly use the -ve input resistor with a pair of back to back diodes to protect the IC. As you say, this resistor acts as Rd in parallel with any other damping resistor across the coil. Do you only use the -ve input R as Rd? and, presuming you use diodes, which do you use? I mainly use 1N4148 and am surprised that they stand up to quite high peak currents from 400 - 500V flyback. The NE5534 has input protection diodes which quickly failed when running without external ones. You can run with just one diode i.e the one clipping the flyback. The other just clips the TX volts which is OK. On some standard PI arrangements, I use a 2W metal oxide resistor at the input because it can get quite hot.

I like the low peak current and do it more often approach as it avoids a number of problems. One being that a cermet trimmer can be included in the damping circuit for fine adjustment, which I find quite critical when you get down to <10uS delay. High flyback voltages cause arcing between wiper and track and you end up with a serious N/S level.

I suppose the best method to connect the coil to the preamp is to have series and parallel gates in front of the preamp so that it is isolated until the flyback period is over. Then the preamp can be connected to the coil without any noisy resistors in the way.

Eric.

[green]

I mostly use the -ve input resistor with a pair of back to back diodes to protect the IC. As you say, this resistor acts as Rd in parallel with any other damping resistor across the coil. Do you only use the -ve input R as Rd? and, presuming you use diodes, which do you use? I mainly use 1N4148 and am surprised that they stand up to quite high peak currents from 400 - 500V flyback. The NE5534 has input protection diodes which quickly failed when running without external ones. You can run with just one diode i.e the one clipping the flyback. The other just clips the TX volts which is OK. On some standard PI arrangements, I use a 2W metal oxide resistor at the input because it can get quite hot.

I like the low peak current and do it more often approach as it avoids a number of problems. One being that a cermet trimmer can be included in the damping circuit for fine adjustment, which I find quite critical when you get down to <10uS delay. High flyback voltages cause arcing between wiper and track and you end up with a serious N/S level.

I suppose the best method to connect the coil to the preamp is to have series and parallel gates in front of the preamp so that it is isolated until the flyback period is over. Then the preamp can be connected to the coil without any noisy resistors in the way.

Eric.

I use back to back 1N4148's -input to ground with a 1k input resistor for Rd. The fet avalanches at about 450v limiting the current to less than 1/2 amp. The 1k damping resistor is close to critical damping if the coil SRF is a little above 1MHz. I usually limit the peak coil current to 1 amp. with 1kHz sample rate. Two 1/4 watt 500 ohm or three 1/4 watt 332 ohm metal film resistors in series for Rd. I have trimmed Rd but still can't a decay as clean as yours.

[Ferric Toes]

I use back to back 1N4148's -input to ground with a 1k input resistor for Rd. The fet avalanches at about 450v limiting the current to less than 1/2 amp. The 1k damping resistor is close to critical damping if the coil SRF is a little above 1MHz. I usually limit the peak coil current to 1 amp. with 1kHz sample rate. Two 1/4 watt 500 ohm or three 1/4 watt 332 ohm metal film resistors in series for Rd. I have trimmed Rd but still can't a decay as clean as yours.

I have tried a variety of different IC's for the preamp, many of which look good from the specs, but do not perform good in this type of application. Some overshoot, some ring, some have slope on the dc base level. so there are only a few that perform well in all respects. The AD8055 is good, as is the LME49990 but the AD is noisy and the LME is current hungry which can be an issue in AA battery powered detectors.

Is your SRF including some cable capacitance? It does give large benefits to get the SRF as high as possible by using a very short cable by placing the TX/RX as close as possible to the coil and the sampling/processing/display further back up the shaft. I remember reading somewhere that for critical damping the SRF is damped in 1 period. Is that correct I wonder?

Eric.

[green]

I have tried a variety of different IC's for the preamp, many of which look good from the specs, but do not perform good in this type of application. Some overshoot, some ring, some have slope on the dc base level. so there are only a few that perform well in all respects. The AD8055 is good, as is the LME49990 but the AD is noisy and the LME is current hungry which can be an issue in AA battery powered detectors.

Is your SRF including some cable capacitance? It does give large benefits to get the SRF as high as possible by using a very short cable by placing the TX/RX as close as possible to the coil and the sampling/processing/display further back up the shaft. I remember reading somewhere that for critical damping the SRF is damped in 1 period. Is that correct I wonder?

Eric.

I have used OPA1612 or LT6237. Using +- 2.5v for the analog circuit and 5v for the digital. Winding the coils spider web(flat basket) to keep the SRF high. Including a test I did awhile back. The SRF of the coils without lead wires is higher than shown. My measurement circuit was adding some capacitance. If I excite the test coil with an operating PI and measure the frequency with the scope probe near the test coil they test 3.5 to 4 MHz. http://www.geotech1.com/forums/showt...255#post214255

LT spice simulation of a critically damped coil. SRF(circuit resonance) about 1MHz. Takes about 3usec to decay to 100uv. Avalanche time would add to that. Traces test point a

Probably not thinking correctly but I'll try anyway. As the coil decays the voltage across the coil determines the decay rate. (I/T=E/L) The reactive signal is proportional to I/T. How low does the decay voltage need to drop for the reactive signal to not cause a problem with a PI detector?

[green]

Hi Eric

Made a chart to compare the differences with 1k and 25k pulse rate. Wasn't expecting average current to be lower at 25k. Signal is proportional to peak current. Do you have a formula for reduction in noise for pulse rate increase? 10k/1k, 25k/1k

[6666]

The AD8055 is good, as is the LME49990 but the AD is noisy and the LME is current hungry which can be an issue in AA battery powered detectors.

I've ordered a LME , will have a play.