Large PI Coils ????

[Tinkerer]

parasitic capacitance of coils

Hi bbsailor,

thanks for the info. It is very useful. I do not have a schematics of the CS6PI, but I suspect that probably the Tx pulse might get a little bit short at 13.8 Kpps. this would mean that there would still be eddy currents present in the coil at switch off.
Also the integration time might be a bit long.

I totally agree with you that cutting down on the parasitic capacitance is primordial and the best way to measure it, is the resonant frequency of the coil, including the coax and the preamp.
The way I see it (and please correct me, if I am wrong) The L/R time constant of the coil means that increasing the R, reduces the time constant. however, there is also the RC of the parasitic capacitance to consider, and there, the increase of R increases the time constant. If the RC time constant is longer than the L/R time constant then we get the ringing.
Now, for the L/R time constant we have, lets say 300uH / 30 Ohm = a time constant of 10uS. However, for the RC time constant of the RX, we have the damping resistor of lets say 1k Ohm which means 1000 times xC for the time constant of RC and a peak voltage of lets say 320 volt that have to be reduced to near 0 volt for the preamp. The formula is: t= 2.3 RC log V/v. This seems to give a fairy accurate idea of the time needed for the first delay.
I like your idea for the testing of different Mosfets. As soon as I have a scope again, I also want to look closer at the possibility of using the Avalanche diode to speed up the Flyback. It is built to do that job. The way I understand it, it would mean that using a Mosfet of 100 Drain /Source voltage, the peak voltage of the example above would be reduced to 100 V and should therefore speed up the damping.

I never did a lot of beach hunting. My experience was mostly diving in 20 to 70 feet of sea water. there the delay time limit seemed to be around 25 to 30uS. However, on hind sight, with the slow machines used at that time, we did not pay much attention to coil capacitance. The large amount of sea water surrounding the coil probably also produces more parasitic capacitance, something to look at. I remember well, that in mid water, there were eddy currents present that diminished when the coil got closer to the bottom. Less water, more sand.
Tinkerer

[bbsailor]

The CS6PI uses my homemade 5 ohm coil and a 39 ohm series resistance plus what ever the on-state resistance of the MOSFET is. With about a 45 ohm total resistance on a 300 uH coil, the TC should be pretty short, about 6.6uS.

While you are waiting for your scope to arrive here are a few thoughts to ponder and research that will help you understand the dynamics of coil performance. I only learned some of this stuff myself in the last few years as I began tinkering with PI coils. I researched a lot of theory, wound a lot of coils, made lots of notes and tried to understand the theory and the practice. With the help of many on this forum and the Findmall PI forum, I started connecting the dots. I still have a way to go but I have learned a few things that I am willing to share.

The formula for calculating the critical damping resistance (theory) is 2 times the square root of the quanties of L over C under the Square Root sign. To visualize it easier, draw it out with L divided by C under a Square root sign with the result multiplied by 2.

Your know your coil inductance. You can know the combined coil capacitance, shield capacitance and coax capacitance by knowing the coil's self resonance frequency and calculate what value of capacitance would result in your known resonant frequency. So you calculate the ideal critical damping resistance value based on the above formula.

You fit your coil to you PI machine and use a variable pot to observe the critical damping on you scope. Lets assume you get a observed value of 750 ohms for a critically damped coil. But what if you calculated an ideal value of 2250 ohms using the above formula? What accounts for the difference?

The answer to this question will be very educational.

Hint: Do an inventory of all the things that contribute to the resistive and/or capacitive loading on the TX coil circuit.

bbsailor

[Tinkerer]

Separate RX coil

Hi bbsailor,



This is what I find so great on this forum. People willing too share their knowledge and experience.



If all electronics design could be done by calculations, we would not need a scope anymore. For myself, being very weak in Maths, I prefer to do a lot of trial and error in the detector design, seems that I learn better to understand the dynamics that way.

One major capacitive load on the TX is the output capacitance of the Mosfet. We can avoid this on the RX by making a separate RX coil.

What have been your experiences with separate RX coils? Should the RX coil be on the inside or outside or inside and outside like Woody.au’s?

On the CS6PI, A TC of 6.6 uS looks good to me for the TX if the TX time is at least 20uS.

How high does the flyback voltage go?

Is it clipped by the avalanche diode of the Mosfet?

Am I right when I think that this clipping can reduce the delay?

Tinkerer

[bbsailor]

Tinkerer,

If all electronics design could be done by calculations, we would not need a scope anymore. For myself, being very weak in Maths, I prefer to do a lot of trial and error in the detector design, seems that I learn better to understand the dynamics that way.

I believe you need both math and trial end error (experiment). The math just helps you understand the relationship of the variables so when you choose to experiment, you work in the area where the greatest gains can be made and ingnore the things that stay relatively constant.

I will simplify the math for you. All you need is a simple calaculator that can do basic functions plus do a square root.

The basic math is simple but the leading decimal places with zeros can sometimes get confusing. The formula is L in Henries and C in Farads. You need to take into account getting all the decimal places right, so do the following. Note: 1pf = 0.000001uf.

Lets assume (you can change the values to fit your own coil) that your coil is 275 uH and the coil's self resonance with a shield and coax results in a 167 pf effective capacitance. Enter 275 into the calculator then divide by .000167 = 1646706.5. Now push the Square Root sign to get the square root of 1646706.5 which is 1283.2, multiply this times 2 and you get 2566.48 ohms for critical damping. Go to Gary's web site for a good LC calculator to figure out the effective capacitance of a coil with a know self-resonance.

But if you adjusted the damping resistor pot (across the coil) to 750 ohms on your coil and observed on the scope a smooth response with no ringing, you have just achieved critical damping. Take this measurement after the first preamp to account for the preamp to recover after being saturated which will add 1 or 2 additional uS. Using the above formula you can figure out that it takes 1950 pf capacitance with a 275 uH coil to critically damp at 750 ohms. Now you see that the circuit sees 1783 pf equivalent capacitance (1950 - 167). Most of that comes from the MOSFET so obtaining a lower capacitance MOSFET can make a big difference. The capacitance of the input clamping diodes add very little extra capacitance (only a few pfs), so take this as a given. The input resistor to the first preamplifier stage, helps set the gain but also appears as a partial load for the coil, so raising it to twice the value will reduce (half) the gain but also change the effective loading on the coil.


One major capacitive load on the TX is the output capacitance of the Mosfet. We can avoid this on the RX by making a separate RX coil.

Yes, the MOSFET is the single largest source of additional capacitance in the TX circuit. It is still there in the TX coil of a DD circuit, but the RX of the DD coils does not see it. In fact, if the DD coils is balanced or closely balanced with a null, the preamp stage saturates less and recovers slightly faster. Since the RX coil also needs a damping resistor and it does not see the MOSFET capacitance, it's critical damping value will be higher than the TX coil damping resistor value. However, the input preamp stage will still add some effective loading so the critical damping will not even come close to the calculated formula value. The RX coil critical damping might be about 1.2 to 1.5 times the value of the TX coil critical damping.

What have been your experiences with separate RX coils? Should the RX coil be on the inside or outside or inside and outside like Woody.au’s?

I have limited DD coil experience, as I have only made a few and the above results seem to be consistent with my calculations and measurements.

On the CS6PI, A TC of 6.6 uS looks good to me for the TX if the TX time is at least 20uS.

How high does the flyback voltage go? Only to about 80 volts.

Is it clipped by the avalanche diode of the Mosfet? Not by very much, if at all.

Am I right when I think that this clipping can reduce the delay?

Clipping the natural peak pulse tends to extend it and also contributes to the heat generated in the MOSFET.

I hope this helps.

bbsailor

[Tinkerer]

Calculated delay time

Hi bbsailor,



I really appreciate your patience in explaining the dynamics of the PI detectors to me and hope there are many others that also benefit from your knowledge.

I did a Google on critical damping and only found formulae applying to waves. I look at the short TX pulse, with a long interval, more as a repeated single transient event. Could this be the reason for the discrepancy of the calculated damping resistor and the best R found by T+R?

Thanks also for the detailed calculation procedure.

I spent hours calculating all kinds of formulae, this time not trying to calculate R, but what the damping time for a given R might be. Here is an interesting result: When I use 1950pf with a damping 750 + coil 50 + 1K preamp input resistor, and 80V peak flyback, I get a time of 8.5us to the point of 4.5V on the preamp (Gain 1000), just where the signal of the coil itself, comes out of saturation. (no time added for saturation, noise or eddy currents) How does this compare to your observations?

Could you give me some more numbers of different coils to calculate, to see how close they correspond with your results observed on the scope?

Tinkerer

[bbsailor]

Tinkerer,


See this web link for a good discription on critical damping. You will find the formula that I explained. http://www.dartmouth.edu/~physics/labs/p16/lab6.pdf

My CS6Pi has been modified by doubling the input resistance to 2K which cuts the gain in half but more than makes up for it in being able to sample a little earlier. Last Summer I took a cruise to Alaska, and while there picked up two nugget samples to use on my CS6PI. I can detect a 1.2 gram (irregularly shaped something like the letter Y) nugget at 5" from my 11" (actually 10.5" ID) mono coil. I suspect I am sampling at between 7.5 and 8uS. Just look at the slope of the exponential target decay and see how much more signal I can detect by just sampling 1 uS sooner.

If you don't have a scope, one way you can cheat is to put a fixed 1.5K ohm (1 watt) resistor on the circuit board add a 5K in parallel with 1.5K . Put a 1K resistor in series with the 5K pot to act as a minimum limit resistor. This gives you 6K in parallel with 1.5K at one end of the pot rotation and 1K in parallel with 1.5K at the other end of the rotation. This will give you an effective range of damping of 600 to 1200 ohms without the danger of burning up your pot as most of the power is across the fixed 1.5K resistor (beyond mid rotation). Mount this pot next to your coil connector and use a screwdriver (preferred) or knob to adjust while waving a small nugget in front of your coil and adjust for the most sensitive response as the nugget is moved slightly father away from the coil. Absent a scope this will help you optimize the critical damping, assuming the circuit is working. It is necessary that your minimum delay can go as low as you want to sample. This may mean changing (reducing) some resistor values that limit the minimum RX sampling range. Read the Hammerhead article by Carl to see how he suggests modifying the delay range.

The 600 to 1200 effective ohm range pretty much covers most coils that I have seen on other PI machines and have made myself, excluding very low frequency PPS, high current models designed to look for large, deep targets which have a critical damping range of 200 to 600 ohms and use unshielded large coils. Look at various schematics in the Metal Detector Projects section of this web site to get a feel for the size of damping resistor and coil inductance.

The whole purpose of this exercise is for you to get a feel for how the critical damping value changes as the coil peak self-resonance changes due to changes in effective capacitance. Coils that have lower self resonances have more stored energy in capacitance that needs to be disapated in a lower value resistor, which also tends to reduce the sensitivity of the coil. Build what you consider to be your fastest coil design and compare it to the calculated ideal value. Use what you consider to be your slowest coil. Calculate that one also. Considering that PI coils are normally slightly overdamped, you can begin to appreciate how much effective capacitance is fixed in you circuit; constantly reflected by the MOSFET, and other circuit components. This effective capacitance value is fixed and as your coils begin to approach this value, improvements in coils will have some speed improvement but beyond this point, coil improvements have less effect. This is like the typical effective parallel resistance S curve graph. If the internal effective capacitance is a fixed 500 pf and your total (coil, shield and coax) coil is also 500pf, any improvements in coil performance will be dominated by the fixed 500pf in the PI circuit. That is why it pays to look for new MOSFETS and other coil designs (like DD) with lower effective output capacitance. If you replace the MOSFET and now have a 300 pf internal fixed capacitance, you can more easily see the effects of reducing the total coil capacitance. You can see this effect even for a 30 to 60 pf reduction in total coil capacitance. To see the effect of a 100 pf change, do this simple experiment. Set your coil to the most sensitive critical damping to a small gold target. Temporally add a 100 pf capacitor across the coil. Readjust the variable critical damping pot. It will probably be about 100 ohms lower to damp the extra 100 pf of energy storage in the TX circuit. This will give you an intuitive feel for achieving a balance between competing variables.

What you are ultimately trying to do is make the best coils you can build given the material you have on hand or can easily get. You may also need to modify the PI circuitry to be optimized for the targets you seek. Gold nuggets require higher frequencies, lower flyback voltages and a lower peak coil current. One thing you can do is to increase the series resistance value between the coil and MOSFET to just beyond the avalanche point of your chosen MOSFET.

Pay attention to the dielectric constant of wire insulation and coil to shield spacer. PVC is worse than Polyethylene (PE). Teflon is one of the best wire insulations and makes coils with consistently higher-self resonances.

Every little bit of coil improvement helps, but you need to see the overall performance of your total PI circuit realtive to the targets you are seeking. Once you get beyond analysis, you then need to apply that analysis to obtaining some measeurements from your coils and PI circuit. Progress is based on a continuous loop of analysing, measuring, calculating, adjusting, remeasuring, understanding, and strating the loop over again.

Glad I can help!


bbsailor

[Reg]

Hi Tinkerer,

The more you get into this PI experimenting thing, the more you will realize that sampling at less than 10 usec is extremely difficult. One advantage the CS 6PI has is it is a low power PI, so sampling sooner than 10 usec is easier to do. It is very difficult to sample less than 10 on this machine so it will be even more difficult on other PI's such as the Hammerhead.

One trick you can do on higher powered units is to add a series resistor in series with the FET and the coil to reduce the coil current. If you select the right FET, then sampling at shorter delays becomes easier also. This works for the Hammerhead quite well. If you do not reduce the current, you are going to have a very difficult time getting the delay even down to 10 usec. It can be done, but it is tricky.

So, if you are trying to build a PI that detects very small nuggets then you may have to deviate from the norm.

Finally, the best way to figure things out is to simply build the circuit or coil and try it. If it doesn't work as planned, don't be afraid of experimenting with what you already got and comparing the results. Make sure to keep good notes. For example, if a coil doesn't work quite right, don't be afraid to try different things along the lines of shielding. You might try spacing the shielding farther out from the windings, or try a different shielding material all together. Again, keep track of just what happens. This will give you some ideas of what might have to be done to get from where you are at to where you want to be.

As it has been pointed out, a lot of the math doesn't quite fit, unless you have all the factors in place. As such, experimenting is the ultimate way to build what you want.

Now, BB Sailor mentioned he increased the input resistor to a 2K. The CS 6 PI has a 2.2K resistor to begin with, so I suspect there is a slight error here, but the technique is correct. That technique is you can reduce the gain of the preamp and increase the bandwith among other things. This will also allow for a faster sampling. Also, the CS 6 uses the old 709 chip. This chip works fine, but changing it to a NE 5534A is by far one of the best things you can do. Make sure to eliminate any IC compensation since this will slow things down.

So, you can reduce the feedback resistor also if necessary and make up the gain in later stages. Just don't lower it too much.

When I first started I was extremely happy with the 10 usec sampling and thought I was up against the ultimate "wall". That is what I thought until I was sent some "invisible" nuggets that couldn't be detected by PI's. Even with a 10 usec sampling, I could only get a whisper when the nuggets were rubbed across the bottom of the coil.

So, the project was on to get the delay down even more to see if they could be detected. This is where otherthings such things as the shielding, different types of wire, as well as gain changes, came into play. Very few stones were left unturned.

When done, I was able to detect the invisible nuggets very well and the sampling was pushed back to the 6 to 8 usec range.

Now, here is one more little secret that makes things a little easier. You do not have to have a perfect looking decay curve. What has to happen is the decay curve signal has to be stable at the time of sampling. In fact, if you search back through all the info that has been posted over the last few years, you will find a pic I took of the decay curve of a ML coil used on a different machine. What it shows is the decay curve is quite ugly, but does stabilize well before sampling. I call this cheating, but it works just fine.

Now, BB Sailor has mentioned that each coil can have a different ideal damping resistance. Ok, how do you build one machine that will allow for this? Easy, just use the highest resistance value that works consistently on the pc board, and simply add any additional damping resistance as needed to each coil. This can be done inside the coil housing or even in the connector. This way, the machine remains constant and the coils vary accordingly. Anyway, this works for me.

Hope this helps.

Reg

[bbsailor]

Tinterer, Reg

Reg is right. I did the modifications quite a while ago and was working from memory, which can be a problem at times. The input resistor is 2.2K. I changed the first amplifier to an NE5534, removed compensation capacitors, and reduced the first amplifier gain by cutting the feedback resistor in half. The net result is the same...speeding up the response.

The MOSFET on the CS6PI I believe is an IRFD113 (HEXDIP package), 60V peak voltage and .8A current. So my 80 volt statement in the previous post should be changes to be 60 V. Again, a foggy memory. The point is still true; this is a low voltage, high frequency, low current pulse type machine. This type is easier to me made to operate at low delays than higher power types.

Here are some numbers that will help demonsrate my point about fixed internal effective capacitance and loading.

300 uH coil, 167 pf resonating at 711KHz See the calaculator below http://www3.telus.net/chemelec/Calculators/LC-Calculator.htm

Just that coil alone would need about a 2680 ohm critical damping (CD) resistor.

On the CS6PI I have adjusted the damping resistor to 1000 ohms. My assumption: there must be some effective internal capacitance and loading that requires a real damping resistance value that is much smaller than the one calculated.

It turns out that the internal circuit capacitance of the MOSFET is 80 pf and each input diode is 4 pf , 8 total, providing a known capacitative load of (167 + 80 +4 + 4) 255 pf. But if I look for a CD value for this combined capacitance, I only get 2169 ohms. It is moving in the right direction but still far away.

If I add the input resistor loading and other capacitive effects, the circuit represents a capacitive addition to the TX circuit that would need about 1600 ohms of damping to achieve critical damping (1600 ohms in parallel with 2680 = 1001 ohms). Even though reducing coil capacitance improves the total circuit capacitance, there is still a substantial fixed amount represented by the internal circuit and circuit components. Let's assume the we can see the impact of placing the first amplifier series resistance of 2200 ohms in parallel with the calculated CD value of 2169 ohms and get 1092 ohms. That is close enough (within component tolerance values) to get an appreciation to see how the various components may be interacting.

This analysis, possibly imperfect, was just my way to attempt to visualize the contribution of coil, shield and coax to the effective internal capacitance and other loading so I could see how close I was getting to the physical wall when I attempted to shave a few more 10s of pfs off my coil designs or attempt to optimize circuit components.

Getting newer, low capacitance MOSFETS with about 50 pf output capacitance (COSS) compared to 200pfs of other MOSFETS is a very fast way to improve the situation. Increasing the value of the input resistor, reducing gain to help the first amplifier settle faster, adding series resistance to reduce coil current, matching pulse TC to desired target TC, operating at faster pulses per second and finding an appropriate integration time, all contribute to improving the sensitivity to gold.

Everything interacts! It is my belief that doing this exercise for your particular PI circuit and coil will shed some light on where your design energy will be well spent. I would need to increase the speed of my CS6PI from 13.5Kpps (7.4 uS) to 15Kpps to start sampling at 6.66 uS.

I hope this analysis shelds some light on this very elusive subject. I could also be totally wrong and, in that case, would welcome another way to view this interesting subject.

bbsailor

[chemelec]

bbsailor vbmenu_register("postmenu_40313", true);
Member
Join Date: Feb 2005
Location: Barnegat, NJ
Posts: 42



Tinkerer and All,

Answers to Your Questions.

The Kynar AWG 30 is single stranded with a .019 inch outside diameter.
The Teflon AWG is single stranded with a.024 inch outside diameter.

Bbsailor, Question on Tuflon wire that you used:

Is it like a Magnet Wire coating?
Or is it like a Plastic Coated Hook-up wire?

Also, Have you tried Both types?

Gary

[bbsailor]

Gary,

The AWG 30 is Teflon coated, like hook up wire. The dielectric constant of Teflon is very low compared to PVC so it provides a better turn-to-turn insulation and thus a higher self-resonance. The self-resonant frequency of a Teflon-wound coil is almost twice what similar Kynar-wound coil is.

What I do is screw 16 2" tall "C" hooks around my desired circumference. Use a screw in the center to secure the start and finish coil leads. Then wind 18 turns for a 300 uH inductance (10.6" diameter). Next, I add the 1/16" ID spiral wrap while the coil is still being held on the hooks. I can actually get the spiral wrap around the coil at the crossover points where the coil rests on the C hooks. The 1/16" ID spiral wrap will expand to about 1/8" so the spiral wrap needs to be about 1.75 times the circumference of the coil to fit in one piece. I recommend not using PVC spiral wrap as it had a higher dielectric than Polyethelyne (PE) which would provide less capacitance between the coil and shield. You should be getting a coil-to-shield capacitance of about 125 to 140 pfs. Measure it to each lead of the coil as it will be different. The self resonance of the 300 uH coil alone, without a shield, shoulds be between 1.1MHz and 1.2 MHz and with a shield should be beween 900 KHz and 950 Khz. Be careful to compensate for the capacitance of the scope probe. Use the 10:1 probe as it has less capacitance.

When I use my normal 10:1 scope probe I get a reading that is about 30Khz lower (in the 1Mhz area) than if I use my Tektroniz FET probe that has a 10M loading impedance. To compensate for the scope probe loading, just add add a capacitance equal to the scope probe loading (this value should be published for your probe on the instruction sheet) and take another measurement, which will be lower. Then just add the difference of this lower amount to your previous reading to get an unloaded resonance reading.

Use a shield material where the eddy currents will not be flowing long enough to be detected. This will reduce the sensitivity of your coil to short delays. I use 3M lead tape and also 3M copper fabric tape. House hold alumium foil and aluminum tape are too thick to be useful at short delays.

Right now some of my fastest 10.6" ID coils (that fit inside an 11" coil housing) I estimate to be operating at between 7.5 and 8 uS (13.5 Kpps) and can detect a 1.2 gram gold nugget at about 5"; a U.S. Nickel at 16" and a U.S. quarter at about 8". To detect quarters and other coins (not Nickels) at a greater distance, I would need a longer TX pulse, lower PPS frequency, and operate at a longer delay with higher TX power.

Look on ebay for a good selection of AWG 30 Teflon wire. Right now you can find some single strand yellow AWG 30 Teflon wire that will make a 5 ohm 300 uH coil with 18 turns on a 10.6" ID coil form. With about a 1 ohm MOSFET-on resistance, you will be generating about 2 amps peak current in your coil. This should be good for a PI machine working in about the 1 to 3Kpps frequency range.

For a lower frequency PI machine, look on ebay for some stranded (19/36) AWG 24 Teflon wire .053 OD. Right now you can find some red Teflon wire that will make a 10.25" ID (to fit an 11" coil housing" 20 turn coil that is about 300uH.

Here is a little tip that will help coil winders approximate the cross section diameter of a wire coil bundle in the 18 to 20 turn range to fit inside your coil housing. Just accurately mesure the OD of a single strand of wire and multiple it by 5. That will be the approximate cross section diameter of your coil bundle for planning purposes. Just add the tickness of you coil spacer and shield to snugly fit your coil housing.

bbsailor

[chemelec]

bbsailor, What is wrong with this setup?

[chemelec]

Darn Picture won't Upload.

Go Here:

http://www3.telus.net/chemelec/Junk/Resonant.png

Using this setup my 8 inch, Aprox 500uH Radial coils measure a Resonance of between 1.3 and 1.4 Mhz.

[bbsailor]

Gary,

Nothing is wrong with that set up for obtaining relative resonance peak measurements. However, absolute resonant frequency measurements require the least amount of loading on the coil, especially coils with a high Q. Most scope probes add a little more than 1 pf of capacitance loading. Any coax cable used in the measurement will have a minimum of 12 pf to 17 pf of additional loading.

I am curious about the self-resonant frequency of your spiral coils with a shield added? Can you measure and post?

Who knows, maybe a spiral wound coil with Teflon wire will make the ultimate coil, if a shield can be easily added and not affect the performance too much.

bbsailor

[Esteban]

Hi

Are this arrangaments useful for you?

http://www.mytempdir.com/421661

[chemelec]

Using my Inductance meter, I measure 29pF of capacitance going into the Scope with the probe set at X10.
With the Probe set at X1, I measure 198pF of capacitance going into the Scope.
This Capaciance is obviously both in the probe and scope itself.

By putting that 1pf Cap in series with the probe, The Total capacitance across the coil Can't Exceed 1pF.
So in My Mind, I Should be measuring only the Coil without any real amount of added capacitance. (Series or Parallel.)
**I Believe this should be Actual Self-Resonance.

I do understand that a Shielded cable will change this, Adding Capacitance.
But thats also a function of What cable and What length of cable you use.
There are some good, Fairly low capacitance cables available.

Sorry bbsailor, I Can't do any Shielded Coil Tests.
I have Never shielded any of my coils and have no materials to do it.

What I can possibly do: Is load the coil with a 100pf cap, than Recheck the resonance and see if it works out correctly to what it should be, using a mathmatical calculation.

Take care........Gary
************************************************** ********

Gary,

Nothing is wrong with that set up for obtaining relative resonance peak measurements. However, absolute resonant frequency measurements require the least amount of loading on the coil, especially coils with a high Q. Most scope probes add a little more than 1 pf of capacitance loading. Any coax cable used in the measurement will have a minimum of 12 pf to 17 pf of additional loading.

I am curious about the self-resonant frequency of your spiral coils with a shield added? Can you measure and post?

Who knows, maybe a spiral wound coil with Teflon wire will make the ultimate coil, if a shield can be easily added and not affect the performance too much.

bbsailor