Large PI Coils ????

[chemelec]

Here is a Test I just did:

One of my poorer Radial Coils, from my junk box.
Its 8 inches in diameter and measures 310uH
Its wound with about 20 turns (Guessing) and # 20AWG enamel wire.

Measuring it as in my previous post, this one has a Free Air Resonance of 1,250 Khz. (1.25 Mhz.)
According to my Inductance chart that means the coil has an Internal (interwire) capacitance of about 55pF.

Putting a 100pF cap in Parallel with the coil, the Resonance changes to 740 Khz.
In reality this would represent the 55 pF + 100 pF = 155 pF, and looking again at my chart, it comes Very Close to being correct at 740 Khz.

The Chart I am using is one I made up many years ago. Basically it is simular to "Inductance/Capacitance Reactance chart" found in the ARRL Handbooks.
Except my chart is a Bit Bigger. I Expanded it to 20 Inches Square and filled it in with more lines for better accuracy.
Save me a lot of calculations for close approximations.
Gary

[Unregistered]

single turn 1x1,2x2... pi coil

Enybody know how to do a single turn 1x1,2x2... pi coil ?

[bbsailor]

Gary,

That was a good tip about putting a 1 pf cap in series with the scope input.

I have made many different kinds of PI coils from simple PVC stranded hookup wire, Litz wire, thick magnet wire in a basket weave and have come to a few design conclusions. It must securly fit into a convenient coil housing, be waterproof for beach hunting or getting the coil wet, and be shielded to operate at lower pulse delays. Being light is a big plus.

Unshielded coils, at low delays, can detect your hand near the coil and become unstable as the coil approaches ground and sees a change in capacitance. Plus, unshielded coils tend to be more subject to noise.

The quest for a high self resonance coil is primarily for making faster coils to detect lower conductivity targets like gold nuggets or thin jewelry. People seeking coins and larger targets can use longer delays, higher power and beyond about 30uS delay can use unshielded coils.

For anyone using coil housings from Hays Electronics or from Charles (Upstate NY) or who improvise coil housings from Frisbees or plastic dinner plates, fitting and securing a shielded coil inside the housing is critical for a stable operation. That is what sent me on the quest to wind fast coils that could be easily coved with a low dielectric spacer then add a shield that does not sustain eddy currents at the time of sampling. As sampling gets sooner and sooner, eddy currents in the shield material starts being detected.

Teflon has a dielectric of 2 and is one of the best insulators compared to PVC which can range from a dielectric of 4 to 8. Poluethylene (PE) has a dielectric of about 2.4, only slightly higher than Teflon, but is much cheaper. Use PE spiral wrap. A coil that is wound on a coil form for may have 250 uH of measured inductance, but when the spiral wrap is applied, it squeezes the wire together and the inductance goes up to about 290uH, but so to does the capacitance. Gary's spiral coils have such a high-self resonance because they have minimal capacitance between turns but require 20 percent more turns to obtain the same inductance as a wire bundle. When I add the shield and take the inductance measurement again with the shield attached to one coil lead I noticed that the inductance go up slightly. My 290uH coil with a shield added becomes 303 uH. Another interesting thing is that there is a different coil to shield capacitance between each lead of the coil that can be up to 25 pf different.

Here are a few coil making rules of thumb that I have discovered.

You need to:

1. accurately measure or calculate the inductance of your coil.

2. choose materials that will give you the best performance.

3. validate that your coil has a high self resonance with shield added and coax added.

4. secure the coil in the coil housing.

5. set the delay of you PI detctor as low as practical for low conductivity targets.

6. make sure you coil operates at the lowest delay. If it does not operate, or suddenly shuts down, you need to reduce the total coil capacitance for the delay you are operating.

Obtaing accurate measurements about your coils is the first place to start seeing if any improvements can be easily made. The numbers are just guideposts.

Gary, I realize that your coils are unshielded and you designed your PI machine for longer delays. However, your latest revision has modifications for getting a delay under 10uS. To operate a shielded coil at this low delay takes come careful coil design so it does not lock up.

I hope to have your latest PI circuit built pretty soon. I'll let you know how it operates with some of my coils at the low delay settings.

bbsailor

[chemelec]

bbsailor, Did you recieve those boards yet?

For really short delays on my detector, you will probably want to try some different Mosfets. Lower Capacitance types.
You Might also Slightly Reduce the values of C15 and C19.
On the Pulse-1 they are 220 pF.
On the Pulse-2 they are 47 pF.
You Might try 39 pF or even 33 pF.

Gary

[chemelec]

bbsailor, Correction on the Previous post. C19 and C22.

[chemelec]

Don't know Why, But the last few days I am having EXTREME Problems accessing this site. No Problems for Any other site.
Anyway, bbsailor, The Best way to Reduce Capacitance Effects is to place a Recieve, Non-Inverting Amplifier, Right at the coil and Using a Gain of about 10.

However this also requires additional send and recieve cables back to the main detector. Also a Reduction in the gain of IC9, if your using my detector.

I have thought about posting a detector of this type in my projects. But it can get a bit complicated and I already get too many Inexperienced persons, causing me too many headaches trying to help them in Emails.

Some 30 plus emails a day is already too many.

Take care........Gary

[bbsailor]

Gary,


I thought about a new breed of PI DD (meaning a separate TX and TX coil) where the RX coil is attached to a 10X preamp right in the coil housing. This would be sort of like Minelab uses in their Sovereign DD coil. You are right in that a special dual coax cable would be needed with one or two smaller lines to carry preamp voltage. The necessity for additional power line(s) could be eliminated if one could build the preamp like a mast mounted RF amplifier with phantom power coming from a remotely located power supply in the metal detector circuit board housing. This might be a good project to add to your PI project and could even be a useful addition to supplement others PI machines with minimal modifications.

I suspect that one would need to be careful about what materials the components and leads are made with so they don't get detected or overload the front end. I'll bet that a postage stamp size device could be made using a few surface mount components that would provide a 10X gain and also allow the main PI preamp gain to be reduced to extend the bandwith and assist in a faster recovery. Maybe one could keep the main preamp the same and enjoy a little better sensitivity with 10X more gain.

This whole amplifier, once designed could also be encapsulated in a small epoxy module to keep it stable. Only three connections (excluding power if it is remotely supplied on the coax) would be needed: (1) Input, (2) Output, and (3) Ground. If phantom power cannot be designed, a fourth or fifth connection for DC preamp power would be needed.

Another alternative that I have contemplated is putting this preamp module about one foot up on the shaft or just far enough away to not be detected. It should pass the TX signal right through but be in series with the RX signal. It should be in a metal or shielded box of some type. To make it easier to build, a 9 volt battery could power the preamp and be located in the box with an on/off switch combined with a bypass switch. At a minimum, this would eliminate about 50pf of coax cable capacitance and start pushing the limits of coil design. The preamp output should be 50 or 70 ohms (or some convenient value to allow good battery operation time) to match the impedance of the cable or have a clean, fast response.

I have not figured out, nor am I sure how or if this could work with mono coils.

Does this sound like it has possibilities?

bbsailor

[NorthWestern]

Gary,



I have not figured out, nor am I sure how or if this could work with mono coils.

Does this sound like it has possibilities?

bbsailor

This type of switching and pre-amps are used in a lot in Amatuer Radio projects and a believe pin diode switching is the prefered method as any losses can be compansated for in the design.

[chemelec]

Here are some Tests on coils I have made up.
These tests are Quite Accurate and varified with a 100 pF, 1% Silver Mica Cap.

I used my "LC Calculator" to determine the Actual Capacitance and Reactance values. Its located in my Calculator section on my site.

http://www3.telus.net/chemelec/Proje.../Reactance.htm

I will also update this with more info in the near future.

I'm in a rush right now, but will be back shorly to look at the posts since my last visit.

One Final Note: Has Anyone considered using 300 Ohm Twin Lead instead of Coax?
I Measured a piece and it only has about 4.5 pF Per Foot.
If it Works at Television Frequencies, Why not Here?

Take care........Gary

[chemelec]

bbsailor


I thought about a new breed of PI DD (meaning a separate TX and TX coil) where the RX coil is attached to a 10X preamp right in the coil housing.

**This has already been done on Numerous detectors. Even one Tube design many years ago.

This might be a good project to add to your PI project and could even be a useful addition to supplement others PI machines with minimal modifications.

** I could if you can find me the Extra Time.

I suspect that one would need to be careful about what materials the components and leads are made with so they don't get detected or overload the front end.

**Typically a standard IC is Not a Problem. Noo need for surface mount.

Maybe one could keep the main preamp the same and enjoy a little better sensitivity with 10X more gain.

**Probably Too Much. It would add a lot of Noise.

This whole amplifier, once designed could also be encapsulated in a small epoxy module to keep it stable. Only three connections (excluding power if it is remotely supplied on the coax) would be needed: (1) Input, (2) Output, and (3) Ground. If phantom power cannot be designed, a fourth or fifth connection for DC preamp power would be needed.

** You Need + and - Supply Lines, Plus Two Seperately Shielded lines. One for recieve and one for transmit.
**But it Is Possible to Super-impose the DC Supply onto the recieve line, reducing it to just two shielded cables.

Another alternative that I have contemplated is putting this preamp module about one foot up on the shaft or just far enough away to not be detected.

**Not near as much Advantage as putting it on the coil.

Take care.......Gary

[chemelec]

My New Link to testing coils:

http://www3.telus.net/chemelec/Proje.../Reactance.htm

Comments Welcome.
Add-on? Changes? Errors?

[Esteban]

Hi

Some years ago I post on PI forum the posibility for to use the pair 300 ohm TV cable for his very low capacitance respect others cables. Capacitance in cable affects PI performance in great manner.

Try this low capacitance cable. You'll obtain more sensitivity. Now, connect accross this cable in paralell only 3 pF and you'll sense the difference.

Another "improvement" (but I don't try) is to make holes with a punch along this cable for to reduce the capacitance.

[bbsailor]

Esteban,

You said "Some years ago I post on PI forum the posibility for to use the pair 300 ohm TV cable for his very low capacitance respect others cables. Capacitance in cable affects PI performance in great manner."

I agree that the coax cable adds to the total capacitance that the TX circuit sees and thus affects potential coil speed. If the 300 ohm twin-lead cable is unshielded, how did you prevent the cable from being part of the coil circuit and causing false signals when it moves against the metal upper shaft? Was the damping resistor, when using the twin-lead, in the coil housing or near the printed circuit board?

Thanks

bbsailor

[chemelec]

Esteban,

You said "Some years ago I post on PI forum the posibility for to use the pair 300 ohm TV cable for his very low capacitance respect others cables. Capacitance in cable affects PI performance in great manner."

I agree that the coax cable adds to the total capacitance that the TX circuit sees and thus affects potential coil speed. If the 300 ohm twin-lead cable is unshielded, how did you prevent the cable from being part of the coil circuit and causing false signals when it moves against the metal upper shaft? Was the damping resistor, when using the twin-lead, in the coil housing or near the printed circuit board?

Thanks

bbsailor

I was Curious so I changed my detector and put the 300 ohm twin lead on it.
But It Didn't seem to affect the Sensiivity at all.

My Detector has No Metal Shafts. However Whan I bring a Large Piece of metal Near the Twin Lead, It does Detect it Just Very Slighty. Obviously the twin lead has Inductance and therefore acts as a Unshielded Coil.

Also in my Experimenting, I Replaced the coil with a 10 Ohm Resistor. Than I connected a Series Resonant LC Circuit, using a 300 uH Coil and a 0.47 uF Cap, in parallel wih the resistor. It Works almost as good as Just the coil.
I have No Explaination for this Capacitive Coupled Coil.

Additionally using a 650 uH coil and a .068 cap, Also in Parallel with that resistor and Also Works well.

I Don't have a Capacitor substitution box, but I did try some various values. The ones above worked best, But Possibly this could be fine tuned for optimum performance, If someone else is Interested in playing with it.

Part of the prooblem is also the Other Adjustments. Pulse width, Dead Time and recieve time. There are Litterly Millions of settings for any given combination of coils and caps.

[Esteban]

Most of the time the shielded cable on PI no prevent the interference since also the coil colect the interference. Another idea for to study is how affects the lenght or type of cable or the both parameters if you build separate circuits: the pulse generator + the power transistor + amplifier in a plastic box near the coil, 20-25 cm. Electrolitic and another metalic components must be glued, because the vibration of this components is detected, so false signals. Anyone experiment in this way?