Getting ready to build the MPP

[KRinAZ]

Hello KRinAZ,Sorry for the mistake in your name. In SAMPLE INTEGRATOR (I think it's DIFFERENTIAL SAMPLEINTEGRATOR) there accurate resistors, for example 1%? In SURFMASTER, BARACUDA, HAMMERHEAD resistors determiningSAMPLE gain in both channels are accurate.

Hi etwert, I'm just using standard 5% resistors - my gain seems to be good, I modded my MPP to add a Gain/Sensitivity control so I can reduce gain as needed - and it's rare that I have my gain set higher than 50% - since the ground iron mineralization is so strong where I'm detecting - the only time I have gain turned all the way to most sensitive is when I'm measuring detection depths of test targets - & I usually do those tests on more mild ground. Thanks - KRinAZ

[KRinAZ]

--------------------------Hi Kyle,I know this has been discussed a lot before but the shape of your lead target and its orientation to the coil makes a big difference in detection distance. Is your lead target 'pure lead' or some unknown bullet/fishing sinker alloy? Is it a sphere , cylinder, or flat, thin shape?I just took a piece of pure lead that was nipped off a block with a wire cutter weighing exactly 5.0 grains and measuring 4mm X 4mm X about 4mm thick with beveled sides from the cutting and trimming process. It is repeatably detected in air with my CHANCE PI at 1/2". Then pounded it into the following dimensions with the corresponding repeatable detection distances in air using my newest 4" X 12" 3DSS 328uh Racetrack coil with all targets presented flat to the coil:5 X 5 X 1.5mm -- 19mm6 X 6 X 1.0mm -- 35mm7.5 X 7.5 X 0.7mm -- 42mmThe flat target was then nipped into a 6.5 X 6.2 X .7 mm shape weighing 3.98 grains or .257 gram with a detection distance of 19mm.Note: The Racetrack coil is designed for fast/wide coverage of the ground and sacrifices some detection depth compared to a round coil. Thanks,Dan

Hi Dan, nice job on creating the test targets & performing the measurements & on the creation of your new coil. I'm curious what field/detection pattern it has (e.g.cone shaped) and what delay (in uS) you're using, the ohms of your coil, and also the specs of the wire you used for your coil? I'm still thinking my resistance at 3.1 is too high, slowing my coil down (too long of a spike recovery time) - making it miss very short TC targets...we'll see when I build & compare to the next coil...next coil will be identical except with more optimum wire... I agree, the shape of the target & it's orientation to the coil (both air testing and in the ground) make a dramatic difference to detection depth - for all types of detector technologies I've used. So what I'm using are lead bullets & fragments that I recovered from the ground - with oxidization - that are deformed to similar shapes of the natural gold nuggets I have found. I only use soft lead bullets (ones I can easily scratch with my fingernail) - I don't use bullets that are jacketed with any other metal. This way I end up with test lead targets with as close as possible TC's to actual in the ground gold nuggets. As for orientation I always present the largest surface area of the target to the coil - the best case scenario for that target - for measurement purposes. Thanks - Kyle

[etwert]

Hello KRinAZ,
Sorry for the mistake in your name. In SAMPLE INTEGRATOR (I think it's DIFFERENTIAL SAMPLE
INTEGRATOR) there accurate resistors, for example 1%? In SURF
MASTER, BARACUDA, HAMMERHEAD resistors determining
SAMPLE gain in both channels are accurate.

[etwert]

Hello KPinAZ,
In SAMPLE INTEGRATOR (I think it's DIFFERENTIAL SAMPLE
INTEGRATOR) there accurate resistors, for example 1%. In SURF
MASTER, BARACUDA, HAMMERHEAD resistors determining
SAMPLE gain in both channels are accurate.

[baum7154]

Hi Dan, yes, I kept the 1k resistor in place, & thx for pointing that out - I guess it makes sense when thinking about why I don't detect a piece of lead smaller than about .3 gram - the coil just isn't fast enough. I think one reason it's not faster is the wire resistance - I think I need to use a significantly larger wire (smaller AWG number) to get the resistance down (& impedance also) to well below 1 ohm. I think the SRF would go up a lot and the spike recovery would be at least 1 or 2 uS quicker. Any thoughts on an optimum wire size? - I'm running 26AWG stranded teflon coated now and it's too small/resistance is too high. I would like to get the right gauge wire and wind another coil exactly like this one (it's nice & tight). I'm able to adjust my MPP min main delay to match - it is capable of sub uS delay timing if I ever get a coil that fast...

--------------------------

Hi Kyle,

I know this has been discussed a lot before but the shape of your lead target and its orientation to the coil makes a big difference in detection distance. Is your lead target 'pure lead' or some unknown bullet/fishing sinker alloy? Is it a sphere , cylinder, or flat, thin shape?

I just took a piece of pure lead that was nipped off a block with a wire cutter weighing exactly 5.0 grains and measuring 4mm X 4mm X about 4mm thick with beveled sides from the cutting and trimming process. It is repeatably detected in air with my CHANCE PI at 1/2". Then pounded it into the following dimensions with the corresponding repeatable detection distances in air using my newest 4" X 12" 3DSS 328uh Racetrack coil with all targets presented flat to the coil:

5 X 5 X 1.5mm -- 19mm

6 X 6 X 1.0mm -- 35mm

7.5 X 7.5 X 0.7mm -- 42mm

The flat target was then nipped into a 6.5 X 6.2 X .7 mm shape weighing 3.98 grains or .257 gram with a detection distance of 19mm.


Note: The Racetrack coil is designed for fast/wide coverage of the ground and sacrifices some detection depth compared to a round coil.

Thanks,
Dan

[baum7154]

Hi Dan, yes, I kept the 1k resistor in place, & thx for pointing that out - I guess it makes sense when thinking about why I don't detect a piece of lead smaller than about .3 gram - the coil just isn't fast enough. I think one reason it's not faster is the wire resistance - I think I need to use a significantly larger wire (smaller AWG number) to get the resistance down (& impedance also) to well below 1 ohm. I think the SRF would go up a lot and the spike recovery would be at least 1 or 2 uS quicker. Any thoughts on an optimum wire size? - I'm running 26AWG stranded teflon coated now and it's too small/resistance is too high. I would like to get the right gauge wire and wind another coil exactly like this one (it's nice & tight). I'm able to adjust my MPP min main delay to match - it is capable of sub uS delay timing if I ever get a coil that fast...

-------------------------------------------------------------------------

I've been using 24 AWG. Tried 26 AWG but did not like the result. Since each jump in wire gauge doubles the cross sectional area and lowers resistance perhaps you should give the 24 AWG a try. There is a trade off with increasing wire sizes in terms of greater eddy currents, inter-winding capacitance, weight, and physical wire bulk in the 3DSS coil form slots.

Regards,

Dan

[greylourie]

Some more inspiration.

http://www.geotech1.com/forums/showthread.php?19940-Pulse-Induction-History-and-Theory&p=165128#post165128

[Qiaozhi]

http://www.findmall.com/read.php?34,...978#msg-171978
Here he mentions the details of tx width/repetition rate of lower powered P.I (Goldquest 10 000 PPS max), and also sample width. He points out the type/size of targets that could be detected. And mentions that copper and silver coins wont be found at great range due to the tx width, and sampling regime.

It appears from the above link, to a post made by Eric Foster, that he uses the magic number of 100,000. In other words, the product of the TX pulse rate and the sample pulse width needs to equal 100,000.
For example:
Goldquest SS: 10,000 pps x 10us sample = 100,000
Textile Detector: 100,000 pps x 1us sample = 100,000
EF's 5us unit: 20,000 pps x 5us sample = 100,000

You could also try reducing the coil inductance as a means of speeding up the decay time.

[greylourie]

On the tx side, it all revolves around speed to switch off. You could go ahead and use long tx pulse, and use some means to reduce the turn-off / flyback time period. But if you are targeting a particular size range of metal bits, then just how long does your tx need to be ? How much power, and what duration ? Maybe increasing PPS improves the odds too, in other ways. I have heard posters here mention S/N ratio is better.

Its funny they mention 3kHz in that article. The TDI range of detectors run at a similar rate. Companies are constrained, they have to produce product that will fulfill a wide range of expectations. A hobbyist with knowledge of the size and TC of what he is looking for, and the depth - can try to tailor tx power and duration.

Eric Foster provides some good examples.

http://www.geotech1.com/forums/showthread.php?16763-I-bought-a-second-hand-deepstar-and

Here he mentions how later version evolved, reducing pulse width and upping PPS but still getting the goods. These were high power P.I. All sizes of jewelry and more, at great depth.

http://www.findmall.com/read.php?34,171965,171978#msg-171978

Here he mentions the details of tx width/repetition rate of lower powered P.I (Goldquest 10 000 PPS max), and also sample width. He points out the type/size of targets that could be detected. And mentions that copper and silver coins wont be found at great range due to the tx width, and sampling regime.

I think your idea to build a tx experiment board is a good one.

[KRinAZ]

Thinking of building & measuring fast coils - I think I'm going to get another MPP board & just build the power sources [needed] and just the TX circuit - to purely observe the coil with no RX circuit influence - I'll do that before the next coil build. Also thinking of fast coil sampling - anyone read the KB article about the MPF technology on Minelab's new SDC 2300? It's linked to on the product page and quite an interesting read!

[KRinAZ]

Hi Dan, yes, I kept the 1k resistor in place, & thx for pointing that out - I guess it makes sense when thinking about why I don't detect a piece of lead smaller than about .3 gram - the coil just isn't fast enough. I think one reason it's not faster is the wire resistance - I think I need to use a significantly larger wire (smaller AWG number) to get the resistance down (& impedance also) to well below 1 ohm. I think the SRF would go up a lot and the spike recovery would be at least 1 or 2 uS quicker. Any thoughts on an optimum wire size? - I'm running 26AWG stranded teflon coated now and it's too small/resistance is too high. I would like to get the right gauge wire and wind another coil exactly like this one (it's nice & tight). I'm able to adjust my MPP min main delay to match - it is capable of sub uS delay timing if I ever get a coil that fast...

[baum7154]

OK, I've build yet another 8" round SSBW (self-shielding basketweave) coil and I'm going to stick with it as I'm finally happy with the coil - I finally got it as tight as I wanted - to improve the magnetic field's strength, quality, and consistency. It uses exactly the same length of wire as the last one so has the same resistance, just wound tighter, and gave me a wonderful 20% increase in inductance (& presumably magnetic field strength) - without sacrificing even 1uS in speed, very nice. Previous coil was 300uH.
Here are it's spec's:
7 basketweave turns
360 uH
3.0 ohms (due to the small gauge 26AWG wire)
2455 ohms critical damping
Flyback spike recovered at 4uS after end of TX pulse - by monitoring the coil, and also at TP3.
Flyback spike now at 500 volts due to the fast coil & I'm getting a bit worried I'll fry the mosfet, anyone have suggestions on a good higher voltage replacement? 800 volts would be nice...

Also since this coil has has relatively high resistance & about the same ohms as current limiting R3 (3.3 ohms) I removed & jumpered R3.

This coil should be very good for trying out the timings I'm going to experiment with. I am almost done putting the MPP on a pole setup, so more soon...

[ATTACH]33301[/ATTACH]

----------------------------------------------
Hi Kyle, did you keep the original 1K input resistor on the input to the preamp? If so the actual 'in circuit' critical damping resistance with the 2455 ohm coil resistor in parallel with that 1k input resistor is 710.5 ohms. Working the formula back for the srf/critical damping puts the SRF of this 360uh coil at about 628.2 kHz in circuit with the feed line.

Regards,

Dan

[KRinAZ]

That would be interesting. I'm sure the metal detector manufacturers must have this information, but I cannot find anything in the public domain.

This is interesting - I like what green has measured and established.
I have a growing interest in the establishing of a TC for various metals, been thinking along the lines of unit of time per unit of weight for specific metals, using the radioactive decay model as a guide - I'm thinking this is what green is doing - I think it would be nice to make a standardized chart of signal decay for various metals and alloys per unit of weight e.g. for say 1 gram of the given material, how many microseconds to decay to .37 target signal strength.

[green]

Thanks. For my understanding it is time between switch off and lower point of (back EMF) voltage drop. So I should correct myself for target No. 7 (gold nugget 2) - this way TC is about 6 not 10. OK, probably I will never touch some field in theory.

One time constant is the time for the signal to drop to 37% of it's value. 100 to 37, 8 to 3, 1 t0 .37. The signal should have an exponential decay, straight line on a lin-log chart.

[WM6]

One of two ways. Time between amplitude of 3 and 8 for 1 TC or time between 5 and 100/3(3 TC 1/20 or 5%), hope it's right.

Thanks. For my understanding it is time between switch off and lower point of (back EMF) voltage drop. So I should correct myself for target No. 7 (gold nugget 2) - this way TC is about 6 not 10. OK, probably I will never touch some field in theory.