Perhaps I should have applied for a patent?

[green]

Green, Where did 2 nf come from? Given the circuit, discounting the effect of the fabric, I would expect the resonant freq to be in the range of 17-20 MHz assuming the coss of the IRFP220 @ ~150p and a gap capacitance of less than 5p (of course this capacitance would increase when he (if he does) adds cable). This is probably why the fabric is so effective at damping/absorbing/attenuating the flyback.

Adding the diode has allowed sampling sooner when coil inductance is around 300uH. Isolates the Mosfet capacitance. A 300uH coil with leads has a SRF is around 1Mhz and with the diode SRF changes very little. I could sample near 5usec, without the diode the resonance is closer to 400kHz requiring a longer delay time. 2nf is just a value higher than what I would expect so I would expect resonance to be higher than 5MHz. Rd=pi*L*SRF, 3.14*.5*5=8ohms. The reason I asked about adding a 5 or 10 ohm damping resistor to see if it had any effect? I'm guessing Rd of 10 or 20 ohms wouldn't effect the circuit. Just trying to learn something or see if I know anything. Have thought SRF might effect initial delay time but if Eric can sample at 1usec with Rx SRF=650kHz probably not.

Can't find IRFP220 data sheet but IRFP250 has Coss of 315pf@25v and over 2500pf@0v. What voltage level should I be looking at when reading capacitance to calculate SRF?

[green]

[What voltage level should I be looking at when reading capacitance to calculate SRF?]

That might have been a dumb question. Supply voltage?

[KingJL]

Can't find IRFP220 data sheet but IRFP250 has Coss of 315pf@25v and over 2500pf@0v.

COSS for IRFP220N is 150pf typical, 300pf MAX per data sheet (I found data sheets for IRFP220 from both Fairchild and Samsung). Looking at the charts, it looks like about 200pf @ 12V, 210pf @ 10V, 250pf @ 7V.

[Ferric Toes]

COSS for IRFP220N is 150pf typical, 300pf MAX per data sheet (I found data sheets for IRFP220 from both Fairchild and Samsung). Looking at the charts, it looks like about 200pf @ 12V, 210pf @ 10V, 250pf @ 7V.

I knew something was wrong when I looked at the Samsung data sheet where it stated Rds(on) as being 0.8 ohms. It is not the Mosfet that I used for this experiment, which should be IRFP250N which has an Rds(on) of 0.075 ohms, but still 200V avalanche. Previously, I had used IRFP22N50A with Rds(on) 0.23 ohms and 500V, but reckoned it was better to use the one with the low on resistance as there are no high voltages here. Capacitances are also considerably lower for the IRFP250N.
Sorry for the confusion which was my mistake.

Eric.

[KingJL]

Eric,
Thank you for the clarification... I like the specs on this MOSFET and think I can use it in some other projects. As pertaining to the subject of this thread, have you run any tests on RX performance/range? My sense is that this would be more suited to an industrial application such as materials processing or food processing for detecting minute metals contamination where the high TX currents and possible reduced range would be an acceptable trade-off. I suspect that the type of metal will effect both the polarity and form of the return waveform.
Thank you for starting this thread, as it has it has made me consider some concepts that otherwise would never have been contemplated, namely a TX with the RX inductor contained within the conductor to eliminate RX/TX inductor coupling , and using the TX conductor properties to achieve self damping of the inductor.

[Idefix]

HI,
I also see this old data-sheet from Samsung, and may be it is an obsolete part.
So I use an IRFP264 in the sim. 250V 0.075 ohms.
And no doubt that the markings of the mos were also a little erased and not very legible. In my stock, I have sometimes troubles to recognize some components, mostly from TI int CI.

I have trimmed R7 to 0.44 to fit 10 Amps.

[green]

Additionally, the tape is 1in wide and it is impossible to wind it on helically without small creases and folds. Each turn overlaps slightly resulting in double thickness at the overlap, so electrically I agree that it would be almost impossible to model. A small component of the collapsing field will interact with any crease or fold and absorb some of the energy, hence the apparent self damping.
I might get a reel of 1/2in tape as this will wind on smoother or make a larger RX winding; say 10 or 12in diameter for the 1in tape. One other advantage of the tape TX is that the large surface area helps to dissipate heat at high pulse currents.

Eric.

Just a question, not a suggestion. If the tape was wide enough or use two pieces of the 1inch to wrap around Rx without helical winding would you expect similar results?

[Ferric Toes]

Just a question, not a suggestion. If the tape was wide enough or use two pieces of the 1inch to wrap around Rx without helical winding would you expect similar results?

Yes. I would expect similar results but one can't be absolutely sure till it is tried. I have both 25mm and 30mm tape and I have used your suggested method on both 10in and 12 inch mono coils where the shield is doing its normal job and not acting as a TX. It is more difficult to apply the tape and also you get more rucks and creases than winding it helically.

I have just ordered a roll of each 6mm and 12mm of the same tape to try further experiments in using the material as a TX coil. With the 6mm, I will wind some on a tube to make a solenoid with perhaps 10 turns and see how fast it is. This may be useful in my magnetic viscosity meter for measuring soil samples. The shortest delay I can do at the moment is 10uS with normal wire, at which point the 30uS current pulse must be flat topped by the start of the sample. If I can get down to 1uS (or even 2uS) that would give greater high frequency detail. Luckily this tape is relatively inexpensive when compared to 3M or other makes of woven copper on fabric tapes, so if it doesn't work it's no great loss.

Eric.

[green]

Many PI metal detectors give good detection ranges where the final peak TX current is 1A or less. With this pulsed shield idea I have a flat top current of 10A and which could be increased to 15 or even 20A, except for the limitation of 2A steady current from my bench power supply. I would have to reduce the pulse rate in proportion to achieve this. These higher levels of pulse current compensate for the lack of turns.
This is all at a very preliminary stage and done within the past week. The idea is to investigate an alternative way of making a coil for a PI that detects much smaller and less conductive objects. It is a bit too early to discuss range or performance.

Eric.

Been using a 5usec time constant target with spice. Not very low. Is there a specific target or low time constant you are thinking about, maybe 1usec(higher or lower)?

[bbsailor]

Been using a 5usec time constant target with spice. Not very low. Is there a specific target or low time constant you are thinking about, maybe 1usec(higher or lower)?

Green and Eric,

This is a great question because you can always reverse engineer the best TX, RX times and coil parameters for a specific target size measured in micro-seconds. The problem is that if you move in too close to a particular target shape and size you minimize the response for other targets away from that particular optimization. Seeking small gold nuggets in a stream bed requires a different approach from finding jewelry on the beach.

One issue that needs more attention is that the target must be assumed to be fully stimulated so that other variables can be examined and isolated for optimization. This full stimulation requires that the TX turn off TC be five times faster than the target TC. This now gets us into other issues such as:
1. Damping resistor value as this controls the coil discharge TC to fully stimulate the target
2. Coil circuit capacitance and related connected components that add capacitance energy to the TX discharge spike and affect damping resistor value
3. Any residual eddy currents remaining in the coil wire itself that must now be eliminated to be able to detect targets with a lower TC than the wire TC itself
4. Coil size that is optimized for the target size and desired detecting distance from coil/probe
5. The search environment that may dictate changes in other design areas and issues

Here are some of my thoughts about Eric's new design that may help develop it further or stimulate further discussions.

Figure out how far the TX pulse effectively stimulates the desired target. If the distance is small in the 5 to 10 mm range, you may want to make an oblong coil (5mm to 10mm wide by 100mm to 200mm long) to cover more search ares in a single sweep. When a target is detected just rotate the coil 90 degrees to pin point it or switch to a probe. Possibly using a ferrite core probe will help project the detecting distance. Just make sure that the ferrite is a soft ferrite so as not to act as a target.

Find a way to use a microprocessor and the coil to analyze the search area soil to optimize TX and RX parameters within the range of the desired target parameters. Use a graphic display to show ground signals and the TX and RX signals on a particular target with the ability to store results for later comparisons and sharing with others using a similar detector product.

In my many years of publishing automation, I have found as any one technology advances farther, it must be willing to adopt adjacent technologies. Here is where a community of probing minds, like here on Geotech1 Forums, comes into play.

Thanks,

Joseph J. Rogowski

[KingJL]

"... and I suspect that at frequencies of upwards of 5MHz the nickel/copper fabric has damping properties of its own. After all, it is advertised as an EMI preventative."

Eric.

Some interesting specifications that I have dug up:
polyester core material plated with a layer of nickel, a heavy layer of copper, and a layer of nickel
200 threads per inch
impedance 0 - 100 kHz < 0.5
attenuation >60 db 10 MHz - 3 GHz (80 db@10 Mhz, 93 db@100MHz, 65 db@3 GHz)
resistance 0.04 (specified as ohms/(square symbol), assumption is per sq mm)
aggressively conductive acrylic adhesive

[Digger429]

Figure out how far the TX pulse effectively stimulates the desired target.

That would be nice considering all the weeks of simulation in spice and yet nobody (not even Eric) has actually shown the design to work in the real world, at all, and isnt just the equivalent of a dummy load.

So my only question is the same as back on page 1: Does it actually work?

[green]

Not wanting to side track the thread. My interest is comparing spice simulation with Eric bench circuit. Eric replied he was doing something else for a few days before he could do some testing. So a small side track. I tried a coil in spice to be able to sample at 1usec awhile back. Made one, could sample at 1.5usec, 1usec looks like it could be difficult for me. Reply #54 asked Eric what target TC he was trying to detect. Did a test with three different diameter coils awhile back. Charted target distance vs signal strength. Straight line on linear log chart similar to time(after Tx off) vs signal strength. Time constant(TC) is the time for signal to decay to 37% of the start value. Distance constant (DC) is the distance for the signal to decay to 37% of the start value. Slope appears to be different at distances above and below coil diameter. A (DC) .2 times coil diameter would be approximate value. I think a delay time increase equal to the target time constant would cause a detection loss equal to the coil (DC). Example1: 200mm coil, 1usec TC target, detection distance 150mm@1usec delay, 150-36=114mm@2usec delay,114-36=78mm@3usec delay. Example2: 200mm coil, 1usec TC target, detection distance 72mm@1usec delay, 36mm@2usec delay, not detectable@3usec delay.

[green]

Not wanting to side track the thread. My interest is comparing spice simulation with Eric bench circuit. Eric replied he was doing something else for a few days before he could do some testing. So a small side track. I tried a coil in spice to be able to sample at 1usec awhile back. Made one, could sample at 1.5usec, 1usec looks like it could be difficult for me. Reply #54 asked Eric what target TC he was trying to detect. Did a test with three different diameter coils awhile back. Charted target distance vs signal strength. Straight line on linear log chart similar to time(after Tx off) vs signal strength. Time constant(TC) is the time for signal to decay to 37% of the start value. Distance constant (DC) is the distance for the signal to decay to 37% of the start value. Slope appears to be different at distances above and below coil diameter. A (DC) .2 times coil diameter would be approximate value. I think a delay time increase equal to the target time constant would cause a detection loss equal to the coil (DC). Example1: 200mm coil, 1usec TC target, detection distance 150mm@1usec delay, 150-36=114mm@2usec delay,114-36=78mm@3usec delay. Example2: 200mm coil, 1usec TC target, detection distance 72mm@1usec delay, 36mm@2usec delay, not detectable@3usec delay.

Tried to add another example but the hour ran out. Example3: 200mm coil, 1usec TC target, detection distance 72mm@6usec delay, 36mm@7usec delay, not detectable@8usec delay.

[green]

Should have added to replies #58 and #59, valid for targets that have a straight line decay on a linear log chart. Small or thin targets with no skin effect.