Getting ready to build the MPP

[KRinAZ]

[All timings are read from where the Tx pulse at TP1 just transitions to it's downward slew at the end of the pulse]Hi KR, I've been using the mosfet gate turnoff as time zero. I think start of flyback is normally used for zero time. Again I'm liking you experimenting. Trying to compare your results with my on the bench testing and the PI simulator. Timing is important. I'm not understanding why increasing pulse rate makes it better unless it's a smother signal, less noise. Just what Davor said in reply 258 plus mosfet avalanche current if enough coil energy.Thanks green

Hi green, thx! - I'm having fun with the experimenting. As far as the depth increase with higher pulse rate - I've actually made a lot of mods to my MPP & it may be that those mods are needed to get a benefit from the increase to the TX 5K PPS. One important change was shortening the TX pulse width to just where the coil charge voltage flattens out (30uS in my case), also shortening the sample widths & the EFE delay. I have also added a Sensitivity & a SAT Speed control, & changed how the (Audio)Threshold pot/resistor network are balanced so that the full turn of the (Audio)Threshold control is useful. In it's original form only about 1/5 of the pot's sweep was useful & it was hard to fine adjust the Threshold. These IMO are needed on the MPP just for general use - and as a bonus - with these controls & mods I have no problem using my MPP on highly iron mineralized ground - with no GEB. GE is there, but doesn't interfere with target responses & if I need to - by decreasing Sensitivity, decreasing (Audio)Threshold, & increasing SAT speed (actually increasing SAT speed A Lot) I can tune out the GE, with a little loss in depth/sensitivity but not bad - it's as George & Carl say - there's no free lunch - haha.

[KRinAZ]

So a MOSFET conducts, and a diode conducts together with it. Then the MOSFET stops conducting, and a coil tries hard pushing the current through it, resulting in a flyback, while a diode continues conducting shortly until the Mosfet capacitance is brim full with charge at high voltage (that's why we use fast high voltage diodes there), and only after a coil voltage drops below its peak voltage the diode goes to reverse polarisation, keeping the Mosfet charge from discharging through coil.

Ok, many thx Davor for that explanation, sounds like the diode does not "turn off" fast enough to block the flyback from getting briefly to the mosfet. This is one of those things I think I will grasp better when I see it happening on my o'scope. Next time I get a chance I will take a look at the signals at the mosfet. Again thx, I learned something new here...

[green]

[All timings are read from where the Tx pulse at TP1 just transitions to it's downward slew at the end of the pulse]

Hi KR, I've been using the mosfet gate turnoff as time zero. I think start of flyback is normally used for zero time. Again I'm liking you experimenting. Trying to compare your results with my on the bench testing and the PI simulator. Timing is important. I'm not understanding why increasing pulse rate makes it better unless it's a smother signal, less noise. Just what Davor said in reply 258 plus mosfet avalanche current if enough coil energy.

Thanks green

[Davor]

Hmmm, ok, so you're saying the flyback voltage passes back thru the diode in the direction the diode normally blocks current flow & voltage & makes it to the mosfet? I'm not understanding that, seems like the diode then isn't doing what diodes do, or am I just completely not getting it? - thx

So a MOSFET conducts, and a diode conducts together with it. Then the MOSFET stops conducting, and a coil tries hard pushing the current through it, resulting in a flyback, while a diode continues conducting shortly until the Mosfet capacitance is brim full with charge at high voltage (that's why we use fast high voltage diodes there), and only after a coil voltage drops below its peak voltage the diode goes to reverse polarisation, keeping the Mosfet charge from discharging through coil.

[Achillion]

Good work KR. Sincerely hope you will pick up a nugget. Would be the sugar on the cake. Good luck mate!

[6666]

7 "basketweave" turns means 7 groups of 5 windings

Got it, thanks.

[KRinAZ]

Very interesting KRwhen you say 7 basketweave turnsdo you mean 7 turns then move forward one slot rather than wind the normal 5 turns then move forward one slot ?

Thx 6666, & no I mean the normal 5 turns then move forward one slot - like in the pic you supplied so that 7 "basketweave" turns means 7 groups of 5 windings, for a total of 35 laps around the form. By the way, in a thread in the Coils forum it was decided to call this coil construction a spiderweave & refer to it as a 3D Self Shielding (I write 3DSS for short), so going forward that's how I'll refer to it - just a fyi

[6666]

Very interesting KR
when you say 7 basketweave turns
do you mean 7 turns then move forward one slot rather than wind the normal 5 turns then move forward one slot ?

[KRinAZ]

OK, this new coil wind is proving sensitive, it appears to be able to rival the Minelab Commander 8" mono coil (their small gold coil) & exceed the current design of the MPP. I made some timing changes & did air tests today on very low TC targets.

But first, hooked up my o'scope & read all the timings again to record where it is right now, those in a minute. All timings are read from where the Tx pulse at TP1 just transitions to it's downward slew at the end of the pulse.

Also, I should now mention I'm using a 10 pak of AA NiMH's to have a power supply similar to a SLA battery - the pack is full charged at 14 volts and full discharged at 10 volts, 2400mah batts. This compared to a AA Alkaline 10 pak where it's fresh (full) charge is 15 volts & full discharge at 10 volts. An AGM SLA full charge is 13.6 volts and full discharge 10.6 volts. This matters, as I found that as I vary the power supply from NiMH full charge to full discharge I lose about an inch in air test detection distance. I also found (as I expected) that the coil voltage spike drops as the power voltage drops - it acts like (actually is I believe) a step up transformer. The batt pack at the beginning of air tests was at 13.44 volts.

The timings I read today before doing anything else:

Tx pulse frequency 2500pps

Tx pulse width 30uS (the minimum time to fully saturate this particular coil's magnetic field)

Coil flyback stabilization where it just flattens after the spike recovery knee 2.5uS (wow!)

Primary sample delay 6uS (if I go to 5.75uS the signal flatlines so this as we know is MP14538 chip's limitation - better than spec anyway, I would love to start sampling at the 2.5uS the coil is capable of, I'll be working on that) & range via the delay control if I remember correctly (forgot to write it down) 6uS to 35uS

Sample width's 22uS (longer than I thought, I could probably replace the resistor with a pot and adjust this down to around 12uS to 15uS, I need to understand better the implication of changing the sample widths, maybe we'll discuss this...)

EFE (secondary) sample delay 148uS

Then I did air tests on detection depth on very small to moderate sized pieces of lead (.05 gram to 10.52 gram, a copper Canadian penny, and a pulltab with tail intact and straight, more below...

Then I changed timings, anything not mentioned below stayed unchanged as above:

Tx pulse frequency 5000pps

EFE delay 140 (basically adjusted it so that the end of the EFE sample is complete about 20uS before the next Tx pulse, this much time may not actually be needed...)

I got a very consistent extra inch / 2.54cm of detection depth increase across the board on targets I could detect.

At 5000pps I got a good target signal on a .33 gram lead piece at 3 inches / 7.62cm, and all targets smaller got absolutely no signal, so I think the response cliff the MPP falls off is related to TC, not receive gain, especially since the smallest target depth is 3 inches (& not less) and if I could sample earlier I believe I could detect smaller targets.

The lead piece target responses:
.05 gram no response
.12 gram no response
.19 gram no response
.33 gram 3" / 7.62 cm
.45 gram 3.5" / 8.89cm
1.99 gram 7" / 17.78cm
2.50 gram 7" / 17.78cm
4.44 gram 8.5" / 21.59cm
10.52 gram 9" / 22.86cm
Canadian 99% copper penny 8.5" / 21.59cm
The ubiquitous pull tab with straight tail 12.5" / 31.75cm

I'm pleased that the MPP can detect a .33 gram nugget at 3", and that a larger nugget's (4.44 gram & up) detection depth surpasses the coil width.

Then, just to test it's cache hunting ability (for such a small coil) (think Forest Fenn's Treasure!) - my Jeep as a test target - 5 feet / 152.4cm - wow, didn't expect that kind of distance for any target...

I think the coil is performing well & I will now concentrate on the electronics. I am a big fan of keeping the target signal path analog, seems to me to have more of a "direct connect with and feel for" Mother Earth, but I think better control of the timings - especially via a fast PIC (it's been a while now but I can probably still program in C or Assembly pretty well) - and the capability of additional samples at various points in the complete cycle - would greatly benefit the MPP and make it a challenger of some commercial products. my2c fwiw

But for now I am going out with this wonderful MPP to search for gold for a bit!

[KRinAZ]

Eric Foster's CS6-PI design also changes the TX pulse width along with changing the PPS frequency and the delay. This design keeps the power in the flyback energy approximately the same over a wide range of frequency adjustments. Eric's high frequency PI machines (3000 PPS) only adjust the frequency range about 10 percent to stay away from other PI machines operating on the same frequency or be near a harmonic of the power line frequency. This small range of TX PPS adjustment keeps the energy in the damping resistor near the same level for optimum delay sampling.

The practical aspect of this observation is to keep the energy level in the damping resistor the same over the full range of pulse width adjustments or have a way to fine tune the damping resistor for an optimum value for the earliest possible sampling. The coil discharge time constant really matters when stimulating a target. This is calulated by dividing the coil inductance by the damping resistor value. Optimum transfer of energy to a target is to turn off the TX pulse 5 times faster than the TC of the target sought. This way you can switch into the RX mode as fast as possible and still have some time to detect the quickly decaying eddy currents in that small target. Balanced DD coils typically sample a little faster than mono coils as there is less voltage on the RX side that needs to wait to die down. On DD coil circuits the first amplifier stage can come out of saturation a little faster than a mono circuit and the input resistor is no longer in parallel with the damping resistor down to 0.6 volts. Because there is less energy in the RX coil in DD coils the damping resistor on the RX side is typically a higher value than the TX damping value. If the TX and RX damping values are near the same value they probably put more turns on the RX coil to make it more sensitive and thus use a little lower RX damping resistor value to damp the extra RX turns.

This all gets down to making trade-offs between target size (time constant), search coil swing speed, battery power consumed (hunting time), TX coil current, coil type, coil size and ground conditions. Most good beach hunting coils are not good for seeking small gold nuggets. Select the right tools for the primary targets sought and ground conditions.

Joseph J. Rogowski

Hmmm, excellent info, many thx bbsailor!