ELLIPTICAL COIL QUESTION

[baum7154]

Regarding the targets made of 1/8" diameter slugs of aluminum or any aluminum it is more accurate to say that every .7 grain of any kind or aluminum is equal in volume to 5 grains or 1/3 gram of gold. All you need is an accurate scale to measure in grains. This is derived from the specific gravity of the two metals, gold being 19.32 grams/cu. cm and aluminum being 2.712 grams/ cu.cm a 7.12:1 ratio. 5 grains gold / 7.12 = 0.701 grains of aluminum. My 1/4" X 1/4" piece of .004" aluminum can weighs just 0.16 grain or about the volumetric equivalent of 1.141 grains of gold. Note that this is for pure gold not nugget gold alloy.

By the way CHANCE PI detector with 8" fast coil sees a 0.7 grain piece of .090" X 1/8" aluminum rod at 4" for first detection and at 2.5" for a pretty solid tone always indicated solidly as gold/aluminum on it's VDI.

Regards,

Dan

[green]

I recorded some data. Made a single layer coil 10.3 inch diameter. 21 turns awg 28 enamel coated wire. Made it so I could slide the coil off the round to put it on the other coil forms. Same coil for all tests. Plotted change in amplitude out of integrator vs target distance from coil. Coil not shielded. Circuit not stable as it needs to be to plot the lower amplitudes. Will do again and add the rectangle if I solve the drift problem. Probably not stable enough to define .25 x .25 in aluminum can. Wondering if I should have done the test different.

[baum7154]

I recorded some data. Made a single layer coil 10.3 inch diameter. 21 turns awg 28 enamel coated wire. Made it so I could slide the coil off the round to put it on the other coil forms. Same coil for all tests. Plotted change in amplitude out of integrator vs target distance from coil. Coil not shielded. Circuit not stable as it needs to be to plot the lower amplitudes. Will do again and add the rectangle if I solve the drift problem. Probably not stable enough to define .25 x .25 in aluminum can. Wondering if I should have done the test different.

--------
Good work!!
1X1 to 1/4X1/4 is a pretty big reduction. You might want to add a 1/2 X 1/2 can target. It is much easier to detect than the 1/4 X 1/4.

[baum7154]

Doing a little analysis on the tests it looks like the 4.5 X 13.5 coil really does not give up much in distance to the other two coils and it had higher output on the 1/4" target @ 5" than the other two coils did. Encouraging results with the round end 3:1 coil trading a broad detection blade for 7% less depth and higher signal strength on the small target. It would be good to see what the 1/2 X 1/2 target would look like. Also anxious to see if the rectangle form improves performance over the round end coil.

Thanks for the tests,

Dan

[baum7154]

Built a 3.5" X 11", 306uh, racetrack or OVAL coil with 1/2 circle ends and 7.5" straight parallel sides, and in first tests it doesn't seem to give up any appreciable detection depth on small fast targets compared to the 8" circular coil. The 'blade' search field of this new coil is readily apparent with a slightly deeper detection at the midpoint of the 'blade' but overall pretty consistent detection along the 'blade'. It does detect my vial of small gold flakes (totaling about the volume of a book match head) from about 1" away, the same as the circular coil. A 4.5 grain (>1/3 gram) piece of 14 kt jewelry gold is detected in air at 2 5/8". I need to get this coil critically damped and run some more tests. Will post them here.


Dan

[baum7154]

Correction:

4.5 grains is < 1/3 gram.

[Fabian]

Doing a little analysis on the tests it looks like the 4.5 X 13.5 coil really does not give up much in distance to the other two coils and it had higher output on the 1/4" target @ 5" than the other two coils did. Encouraging results with the round end 3:1 coil trading a broad detection blade for 7% less depth and higher signal strength on the small target. It would be good to see what the 1/2 X 1/2 target would look like. Also anxious to see if the rectangle form improves performance over the round end coil.

Thanks for the tests,

Dan

Tesoro's Sq Sq coil

[bernte_one]

hello green

very interesting coil comparison.PNG

can you post the excel chart as zip file i see you have collected some more datas?

regards

[green]

hello green

very interesting coil comparison.PNG

can you post the excel chart as zip file i see you have collected some more datas?


regards

I still haven't learned how to send a zip file. Still trying to get less noise, so I haven't recorded any more data for the coils. Like I said in reply #32 the lower amplitude numbers aren't as good as I would like. I'm just trying to learn metal detectors. Keep getting side tracked on something different. Been trying to plot time constant for a hypodermic needle. Each time I try something different I learn something, but don't solve much. Moved the data on the chart to get it all on a PNG

[baum7154]

[QUOTE=green;193625] Still trying to get less noise, so I haven't recorded any more data for the coils. Like I said in reply #32 the lower amplitude numbers aren't as good as I would like.
___________________________________

Have you tried to put some RC filters in your amps to reduce noise? Can you post a schematic of the amps you are using?

Regards,

Dan

[green]

[Have you tried to put some RC filters in your amps to reduce noise? Can you post a schematic of the amps you are using?]

A copy of the schematic I'm using for a test circuit today. Other than changing the value of C1 I haven't done anything with amplifier filtering. With a 1k damping resistor the total gain is 560. The digital voltmeter does some filtering. With no target what voltage change should I expect with a good PI detector? Any suggestions appreciated.

[baum7154]

[Have you tried to put some RC filters in your amps to reduce noise? Can you post a schematic of the amps you are using?]

A copy of the schematic I'm using for a test circuit today. Other than changing the value of C1 I haven't done anything with amplifier filtering. With a 1k damping resistor the total gain is 560. The digital voltmeter does some filtering. With no target what voltage change should I expect with a good PI detector? Any suggestions appreciated.

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

C1 and R9 already form a 1st order filter. What is the value of C1? I'm sorry but I don't understand your question on voltage change in a good detector with no target, versus what? Voltage change will be dependent on target material and size and proximity etc.

Thanks,

Dan

[golfnut]

I was thinking of how to get a geometry that gives a good linear blade or fan of energy into the ground without having a large 'footprint for susceptability to ground effect.

Something like this where you essentially have numerous paralell isolated wires ( like litz ) without the burden of 'full loops' picking up all in sundry. S

[green]

[C1 and R9 already form a 1st order filter. What is the value of C1? I'm sorry but I don't understand your question on voltage change in a good detector with no target, versus what? Voltage change will be dependent on target material and size and proximity etc.]

Most of the time C1 is 10pf or open. The circuit is non motion. The voltage change I'm talking about Is how much it changes without a change in target while watching it for 5 seconds. To plot the data I record the change in volts with and without the target at the different distances. At longer distances the change approaches zero. I have more confidence in defining a 3 count change if the target or no target reading isn't changing, or at least changing less than the voltage change I'm trying to record.

[baum7154]

If R9 is 14K and C1 is 10pf then the Fc of the filter is in the neighborhood of 1137 kHz. Since there is only one 1st order filter it attenuates 6db/octave which I believe that at twice the Fc, 2274khz it gives 12 db. If you lower the Fc by adding more capacitance you can cut the noise a bit more. Maybe you can try C1 @ 30 pf and that should get the Fc to 379 kHz or so and give about 18 db of attenuation at 1137 kHz. This is true for sinusoidal noise. The downside of lowering Fc too far is if it begins to cut off the desired signal. If the desired signal has a rep rate well below Fc then I think just the 'pass loss' through the filter is the penalty. Anyway I'd give it a try just to see what happens.

I'll have to see what the stability of my detector is with no target.

Good luck,

Dan