Journal notes, New Coil Forms, 2.5" Coils
GELLER Labs "Backyard Science"
Thoughts on a proton precession magnetometer design - a Proton Magnetometer Project. Build an Earth's field magnetometer.
The FDM MAGNETOMETER1 project is a low cost high performance proton magnetometer (a digital magnetometer) kit under development for universities and amateur scientists to be able to accurately measure and monitor changes in the Earth's total magnetic F field and to observe geomagnetic storms. Magnetic storms can cause large excursions in the field and are of concern to interests ranging from electrical power grids, radio communications, and satellite operations, to aurora watchers and amateur radio operators.
1 Filter Diagonalization Method "FDM" (harmonic inversion), see Jan 21 and Jan 23 entries, based on: Vladimir A. Mandelshtam, Howard S. Taylor, Harmonic inversion of time signals and its applications, Journal of Chemical Physics (1997), Volume 107, Issue 17, 1997, Pages 6756-6769
(be sure to hit refresh to pick up our latest changes and entries)
Project Documentation, Links and References (very early stages)
Journal Notes:
Monday, January 24, 2011
Overnight: PDF, TXT. The duration of the precession signal is reduced at temperatures of 0 F PDF (the spikes are the monday morning garbage trucks). The FDM figure of merit (FOM) held up fine overnight (-20 F, -29 C) PDF, even though the amplitude of the FID was noticably reduced at these cold temperatrues PDF. The working fluid is still Prestone DeIcer windshield washer fluid, said to be usable to -30 F. Temperatures reached -35 F (-37 C) to the north of here at the Canadian border overnight, the record low at our location is about -28 F.
(Afternote: This version of the coil form was later rejected, see Jan. 25) I decided to make this new test coil form with the two part (cleaner followed by solvent) PVC cement. By an accident of calculation, I ended up glued with a 2.5" coil form length JPG. This form uses two open 2" couplings as described Jan. 23, with a 4" long piece of 2" PVC pipe (actual OD about 2.37"). This might be a bit of serendipity, since 2.5" better covers a 125 mL Nalgene bottle. I am leaning towards winding over 600 turns this time, perhaps ending with an even layer (to be determined). A partial winding layer probably contributes some to a non-uniformity of the polarizing field. I drilled a hole through the coupling on one side for the inner winding connection JPG1 JPG2.
The test coil ended up about 2.5 inches long, 2.37" coil ID (2" PVC actual OD), 686 turns, 3.29 ohms at about 22c, and about 21.2 mH with a Q of about 38 at 1 kHz. Initial rough measurements put the field at about 145 Gauss (0.0145T) near the center of the coil at 1.6 Amps JPG. Recall that while inductance varies as the number of turns squared, magnetic field varies linearly with current. Here is a graph of the magnetic field at the center of the coil versus current (over range of current 1.0 to 1.8 Amp most suitable for the FDM magnetometer) PDF. (While this coil has more turns than our working design (and operational unit), it also longer (2.5" versus 2"), so the center fields are probably very similar. Now that I have a working axial hall probe, time allowing, I will wind another 2" coil, and/or bring in the working counter-wound coil for similar measurements of a 2" 600 turn coil.
The coil is somewhat similar to our shorter 2" long, 600 turn coil in operation in terms of center field. I believe my earlier statements calling for about 200 Gauss for the polarization field were in error. I will go back and edit those comments and our Part II article accordingly. It appears that about 100 Gauss or more is perfectly sufficient for Earth's field NMR (EFNMR). Note that many of the online calculators that give the inductance and the magnetic field of a solenoid appear -not- to work well with the relatively short multilayer solenoid that we use in this experiment. Also, note that while there might be more optimal coil designs, the design in service has performed well for many months now.
Tuesday, January 25, 2011
In reviewing the recent coil form made from parts on hand, I recall why we chose the 2" PVC F cleanout adapter that over-laps the 2" PVC pipe versus the recent type of cleanout plug adapter that fits into the end of a 2" coupling. These NIBCO series 4800 fittings are PVC (Polyvinyl Chloride) C-DWV (drain, waste and vent) fittings used for residential and commercial sanitary systems. In our exising design, with an over-lapping PVC 2" F adapter, the cleanout plug screws in upside down with the square head (for a wrench) inside the coil structure. A 3/4" wood disc resting on top of the square heard provides proper support for the Nalgene 125 mL sample bottle to position the sample bottle just about solenoid coil center. JPG With the coupler version of the cleanout, there is a much longer space from the top of the inverted plug to the bottom of the Nalgene bottle, and therefore much wasted length (the coil form is much longer than it needs to be). On the other hand, it might be better to use a coupling on the top of the coil form, where the bottles slide in and out (e.g. for testing the NMR characteristics of a great number different types of fluids). At 400' to 500' per coil, I guess I need to recycle my test coil copper wire for a try at the next coil form design, a threaded over-lapping coupling on the bottom and a standard 2" coupling on the top.
Wednesday, January 26, 2011
Overnight: PDF, TXT The geomagnetic field was very quiet overnight. Costello Index 7 day PDF.
While I have been discussing how to provide a higher polarization field, I have not been asking the related question:
How low can the polarization field go?
I still need to bring the current 2" coils back into the lab to plot solenoid center field strength versus polarization current using our "new" axial Hall probe. However, in the mean time, this morning I was wondering what would happen if gradually lowered the polarization current over many cycles (in the fast cycle mode). Well, the results are somewhat surprising. With our present working fluid of Prestone De-Icer, and an outdoor air temperature of about 30 F (there is a heat wave today!), I ran about 5 measurements at a time and then lowered the polarization current by 0.1 A. I performed this experiment from our present working polarization current of 1.6 A down to 0.3 A. The system stopped working at 0.3 A because of a low amplitude threshold in our working version of the FDM frequency estimator, not because there was a problem calculating the fundamental free induction decay FID signal frequency. Here is the graph of FID peak envelope amplitude versus time PDF . Also, here is a PDF of somewhat legible notes showing at what time each reduction in the polarization current was made. The complete list of data is in this txt file TXT, and the data that was plotted (measurements that survived the auto-retry process) here TXT . Finally the plotted data PDF (1nT per div) and a corresponding plot from NRCan OTT PDF (0.5 nT per div). This is quite interesting. Perhaps at 1.5 A or 1.6 A, a 2" 600 turn coil is already quite conservative? The FDM narrow band S/N versus time with a polynomial trendline is interesting and suggests that the higher polarization currents give a better chance of higher FDM S/N values. PDF So, perhaps once there is adequate polarization time (e.g. our present polarization time of 2 seconds), S/N is the reason to run at 1.5 A or 1.6 A? Also, it might be good to have some margin for extremely low temperatures, when the amplitude is reduced. Otherwise, the field continues to be accurately calculated at polarization currents of well below 1 A.
I made new 2.5" bobbins this afternoon. This time I marked the plastic 3M electrical tape cover with a pencil using a 2" PVC pipe as a guide, then cut the hole with a small router bit (for side cutting) in a handheld Dremel tool. I used a thicker and wider Frost King insulation (3/4" x 7/16") to support the covers on the 2" PVC threaded end fittings. JPG The three parts (coupler at the top, threaded fitting to accept a plug that supports the Nalgene bottle, and a 4" length of 2" PVC pipe) JPG were prepared and all gently dry fit, before gluing. JPG I will let them dry overnight and hopefully wind them tomorrow. The 2.5" version at this point, is more of an experiment than anything else. As evidenced by this morning's experiments on polarization current, the 2" 600 turn version is more than adequate. To make a similar 2" long coil, use a 3.5" length of pipe.
Thursday, January 27, 2011
Overnight: PDF, TXT The geomagnetic field was very quiet again overnight. Evening, except for a couple of vehicle induced offsets, there was a text book diurnal cycle today PDF.
I wound the new 2.5" long counter-wound coil pair today. JPG We used the forms JPG described January 25, 26. I should note that there is nothing wrong with the original 2" design. Although some improvements, such as the larger foam insulation size (3/4" x 7/16") and using an open coupling on one side and a threaded coupling fitting on the other can be incorporated into the 2" design. The open coupling should allow for faster sample bottle change out during fluid studies. {Portable applications might want to use threaded couplings on both sides to hold the sample bottle in during movement in the field, however such users will need to address the safety concern of a pressure relief valve or mechanism, especially when using volatile fluids. Portable adaptations are still beyond the scope of our present geomagnetic observatory application, although a future gradiometer version is of some possible interest for later.}
While I have been counseling to wind the coils tight and neat, I recall today as I wound these, that some alignment errors will happen, particularly in the later layers. Also, you might not end up with a complete final layer. If possible place an incomplete layer towards the middle, and, well just do the best you can. However, do not scatter wind these coils. Also, just wind back and forth, like a fishing reel, except with flat layers. (Do not go directly back to a starting side than wind forward again.) I use a small winding machine (modified to reverse direction) with a hefty old hp supply. However, by the later windings and near the ends of each layer, I do a lot of hand turning and hand winding. Others have reported a similar approach to achieve the most uniform layers.
Once the coils are wound, one intentionally with three or more turns than the first, the inductance can be matched, such as with an LCR meter. Those without an LCR meter can use a resonating capacitor, signal generator, and an AC voltmeter or scope and work from first principles of LC resonance to balance the coils. If using an old hp (we have our favorite ancient hp 4265B Universal Bridge), be sure to re-check your reference coil every time you remove wire from the trim coil to allow for warm-up drift. Our coils came in at 21.1 mH each with about 686 turns.
Next, tape up the coils, preferably with two different colors of electrical tape. Measure the resistance, preferably with a 4 wire ohms function. Our coils both came in at 3.26 ohms (at about 74 F (23 C)).
Then, with an indelible marker, draw the direction of the outer winding with an arrow showing how it turn to end the last winding of the coil and mark each coil counter-clockwise and clockwise.
Finally, with the coil horizontal, apply a small current with the positive terminal of the power supply connected to the outer winding connection. Mark the outer connection with a "+", then with a compass nearby and the coil at a right angle to the local North - South line, observe and mark the ends of the coil N-S appropriately. If you properly counter-wound your coils and applied power supply plus to the outer winding, the N-S will be opposite between the coils. Here is a picture of our new counter wound pair JPG. In operation, the threaded fitting which accepts the plug goes on the bottom and the bottle (in only the one "powered" coil) sits on a 3/4" wood spacer on top of the square part of the plug. JPG
10 pounds of 200C #18 enameled wire is running around $100 (two coils take about 5 pounds) and the pipe, 3M electrical tape, the Frost King insulation, and the fitting runs about another $35 (parts for four coils was $42, including 8' of 2" pipe). So, budget around $100 for two coils (or ~$150 for four (two pair)) if buying new parts and materials at a home center, such as a Home Depot here in the US. (Note that we use only the "large" half of the 3M tape plastic case for the bobbin, so you need to buy four rolls of 3M electrical tape for each coil pair (two coil bobbins) JPG JPG ). Time needed to construct the coil depends on skill level and available equipment. Estimate two to three hours to assemble the forms, and a half to a full day to match the counter-wound coil pair and to completely characterize both coils. Assembly in the outdoor sensor box (wiring the coils) can be included with the time estimate to build the sensor stand and run the sensor cable to the building.
I need to bring in the 2" working coils to study the magnetic field strength with our "new" axial Hall probe. So, rather than wind another duplicate set, I chose the 2.5" long coils to think about how better coverage of the volume of the 125 mL Nalgene bottles would affect the signal strength of the free induction (FID) precession signal. If all goes well, I hope to change out the working coils with this new set tomorrow.
In the mean time, I measured the center field strength versus current for the new 2.5" windings. JPG CW PDF, CCW PDF Excel. It does not matter which coil is powered and which coil is present only for noise cancellation. Of course, the NMR fluid sample bottle must be in the powered coil.
Use an axial Hall probe to measure the solenoid center field strength JPG. You can make the measurements at 1/10th current, say 10 mA to 200 mA and scale the results if your Hall probe only has a 20 Gauss range. Probably a fluxgate magnetometer probe would interact with the solenoid and therefore such magnetometers are not recommended for this application. Check the data sheet for your Hall probe to see where the active chip is. On our Hall sensor, the active area is far off to one side of the small plastic Hall sensor package. Finding center field, while taking many iterations, is not that difficult. Within the solenoid pipe, side to side and top to bottom (with the solenoid sitting horizontally on the table) raises the field as the probe moves towards an inner pipe wall. Dead center is minimum field. Then, move the probe in or out of the pipe to find a maximum along the long (longitudinal) axis of the solenoid. Use a clamp system to fix your probe in that spot for the duration of the field strength testing. If you use a ferrous clamp, be sure to use a probe holder long enough to keep the clamp or small vise from affecting your measurements of the solenoid field (also, clear the immediate area of large ferrous objects and any permanet magnets).
The NBLNA gain was reduced several times, following each of the many system improvments over the last few weeks. Before we change out the coils tomorrow, I finally got back to measure the present gain of the NBLNA using a Stanford SR510 lock-in amplifier. The SR510 easily measures 1 uV for a calibrated input test voltage, then the output voltage was measured. The working gain was found to be 273,000 (.273 V rms output for 1 uV rms input) at 2288 Hz. In the early days of the experiment, we needed a gain of 500,000 or more. As long as we had the NBLNA box open, we reset the center frequency to our current nominal center field Larmor frequency of 2286.6 Hz and the gain to 273,000.
Friday, January 28, 2011
I changed out the coils today. The new 2.5" 686 turn coils JPG were exchanged for the previously working 2" 600 turn counter-wound pair (to reclaim the 2" coils for magnetic field measurement). The new counter-wound coil pair JPG resonated with about 112 nF on our SWCTRL board (the 100' cable also adds some parallel capacitance). The FDM magnetometer continues to prove to be a robust system, and came right back on with no difficulties.
Here is the magnetic field at the center of the solenoid plot for the the previously working CCW 2" coil PDF. The two inch coils reach a 100 Gauss solenoid center field at about 1.2 Amps. At the working polarization current of 1.6 A, center field was measured to be about 133.5 Gauss. To position an axial probe on a mounted long holder, move the probe up and down and side to side for minimum field (as in the center of a circuilar slice of the solenoid), then in and out along the longitudinal direction for maximum field.
The 2" threaded PVC fittings from Home Depot (See Jan. 25 for links to the parts) are significantly shorter than those I originally got from Lowes. These NIBCO series 4800 fittings are PVC (Polyvinyl Chloride) C-DWV (drain, waste and vent) fittings used for residential and commercial sanitary systems. The new coil bobbin ( JPG described January 26) is not only shorter, but with a simple straight coupling on the top, it will be significantly easier to swap out 125 mL Nalgene bottles for NMR fluid property testing. In fact, the bobbin for the new 2.5" long coil (Orange) is considerably shorter than the older 2" coil (Red) JPG. Flat styrafoam sheets were added to raise the bottles slightly from the bottom of the stand container to clear a plastic feed-through that routes the cable into the container. Here is a top view of the newly mounted coils JPG and an end view JPG.
The noise floor is about the same, sample measurement (polarizaton power supply off, Exp FV 1 (window off)) PDF, Log Spectra PDF, Spectra PDF.
I performed another quick test of the dynamic range of the polarization current (outdoor air temperature about 28 F (-2 C)). This time I ran 0.4 Amp steps from 2.0 A to 0.8 A. Data was taken at 2 A (180 Gauss), 1.6 A (144 G), 1.2A (108 G), and 0.8 A (72 G). The magnetometer ran okay at each of these currents: amplitude PDF, figure of merit (FOM) PDF, sample front panel PDF at 0.8 A showing the gathered field data points and TXT file. Between 1.2 A and 0.8 A, the narrow band S/N dropped off considerably PDF. This is a relatively large range, 108 G to 72 G, so more detailed testing is needed to investigate the drop off. The polarization field should probably be higher than some minimum in this range for this FDM magnetometer as configured. There was some automobile traffic as well as possibly relatively high EMI/RFI levels. The system did not work well today below about 0.6 A (54 G) to 0.7A (63 G). This test was performed with an older, now abused (currents intentionally too high during some testing) small signal relay (The small signal relay was changed out 1/30 with no observable change in instrument performance). This type of testing needs further study.
Saturday, January 29, 2011
Overnight: PDF, TXT The geomagnetic field was very quiet (flat) overnight. Based on yesterday's field measurements, we reducedthe polarzation current to 1.4 Amps. NBLNA gain was adjusted a few times during the coil changeover, and is believed presently to be between 275,000 and 300,000. Here are PDFs of the FDM amplitude PDF and FDM figure of merit (FOM) PDF from overnight. We will run overnight with a 1.5 A polarization current.
Sunday, January 30, 2011
Overnight: PDF, TXT, Amplitude PDF (temperature dependent), FOM PDF. The geomagnetic field was very quiet again overnight. Going into this morning's down-turn of the normal quiet diurnal variation of the geomagnetic field, our measurment success rate is running about 80%. PDF Outside air temperature is about 26 F (-3 C). The NOAA Costello Index 7 day shows how quiet the geomagnetic field has been over the last week. PDF
We are running now with the new 2.5" 686 turn counter-wound coil pair JPG JPG. The powered coil is operating at 1.5 Amps for 2 seconds during each polarization cycle. The performance of the original 2" 600 turn counter wound coils was fine, this is just more experimentation. Whether you use 2", 2.5" or some other coil length, we do recommend our newer coil bobbin design. The new coil bobbins JPG were described on January 25 and 26.
Project Documentation, Links and References (very early stages)
QUESTIONS/COMMENTS/notice of typos, etc. send email to joegeller @ gellerlabs dot com
COPYRIGHT © 2009, 2010, 2011 JOSEPH M. GELLER, All rights reserved.