Build a Geomagnetic Observatory ! 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 magnet
ometer) 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)
Journal Notes:
Saturday, December 10, 2011
On software: It is proper to end a LabView program in a controlled manner, where files are properly closed and memory is properly cleared. Most of us have been using the terminate execution button because my previous implementations of a "stop" bottom often meant a two minute cycle wait for the program to end. With the help of National Instruments tech support, the FDM PPM program now has a proper stop function that acts within a second of button operation. Previously, the program flow for the various delay cycles was nested within the auto-retry case structure PDF. Also, the auto-retry function was a loop within a loop. Now that section of the program has been simplified and streamlined as "flat sequence" structure with frames, just a cascaded serial sequence of code steps.
In the first frame, if the measurement was found to be acceptable (passes the auto-retry criteria), the number of plotted measurements and the success rate is updated. Also, the plotted data is sent to the plotted data file on hard disk. PDF
In the next frame, the Boolean 0 or 1 of both the auto-retry bit (green line, upper left) and the fast cycle bit (the front panel switch) are combined to form a two bit binary number which is converted to an integer 0 to 3. The integer is used to select one of four cycle delay times (the orange "double" from the front panel control) using a case structure. PDF
Finally in the last frame, a one second resolution loop runs a time delay until the current cycle delay time passed from the earlier frame is reached. Now, with one common cycle delay timer, a "stop" button, located within the delay loop, can easily interrupt the delay and cause a proper, near instantaneous, program termination. Presently, this frame is also set to properly terminate the program on a LabView error (the error status line). PDF
We are working on a possible TCP/IP functionality (perhaps an option, or a function that can be switched on or off) so that there can be an informal geomagnetic network for those who wish to publish their local field data 24/7. This new cycle auto-retry and delay path should be more compatible with TCP connection and cycle to cycle reporting functions.
Sunday, December 11, 2011
Overnight: PDF.
Monday, December 12, 2011
A well built coil stand by Mark Haun in Washington state: JPG1, JPG2, JPG3 . Excellent! Mark is using paralleled pairs in a shielded outdoor CAT-5 ethernet cable for the cable between his sensor stand and the indoor switch control module.
Tuesday, December 13, 2011
Overnight: PDF, TXT. Instrument health: Polarization controller chart PDF, FDM figure of merit (FOM) and Amplitude PDF. Note that with cooler weather, polarization times of 500 milliseconds and less are providing the desired 0.85 V precession signal at the output of the narrow band low noise amplifier (NBLNA).
Wednesday, December 14, 2011
More software edits: As discussed above, Dec. 10, I have been working lately in the Auto-Retry section of the supervisory LabView program. I edited the first frame further, so that for a "good" measurement (i.e. one that survives the Auto-Retry filters), in addition to concatenating the latest mesaurement data into the overall data plotted text file, there is now another small file written that has only one measurement, the latestest ASCII data string. PDF Many thanks to Tim at National Instruments for more tech. support today! Also, unlike the overall plot file, which remains "open" for continuous writing, the new single line file is opened, written to, then closed each cycle. Note that within the loop that writes the small file, the "error wire" is used in a somewhat unusual way to account for the possibility that another program is reading the small file when the main FDM PPM program tries to write the new data. Since some other program using the small data file will be running asynchronously with the main program, this could happen, although probably not too often. In the even of such a "crash", the loop waits a second and tries to write the data again, up to five tries, then it quits and moves on with the next measurement cycle.
My current thinking is that any added communications capability, probably TCP/IP, linking our geomagnetic observatories to a common data display site should run as a separate program and read and periodically transmit only the small file. I am leaning in this direction for two reasons, first I am concerned that embedded TCP functions might cause the main program to be less stable, and second there are probably some experimenters who don't want to transmit any data to a central station.
In the following frame, which determines the cycle delay for a current measurement cycle, I eliminated an unnecessary frame structure within the frame. PDF. There are two booleans (T or F) going into this frame, whether there was auto-retry success or failure for the current measurement cycle and if the system is in the normal or "fast" measurment mode (a front panel "soft" switch). So, there are four possible combinations. I was looking for way to improve my program documentation through better labeling (e.g. a label on the frame). However, it turns out you can attach a label to a function as well (here, the array function which combines the two boolean values to make a binary number). PDF.
Thursday, December 15, 2011
Friday, December 16, 2011
Overnight: PDF, TXT. New PC boards are due in today. We will be building and fully testing one set (NBLNA and SWCTRL) which will be available for sale. The changes are very minor, existing units do not need to be updated.
Saturday, December 17, 2011
Built and tested the first Ver. 0.9.4 NBLNA board today. The noise floor appears to be very slightly higher at about 3 nV/rt Hz. The minor change was in the frequency adjustment in the multiple feedback bandpass filter. The corresponding changed positions in parts in the filter was supposed to also improve the isolation from the input stage. There could be an odd component issue, or possibly the increased noise can be attributed to the new layout. Several parts changes made no significant difference in the noise floor, on the order of 14 mV rms with a gain of 500,000. Since we normally operate now at a gain of about 250,000 and our NBLNA output signal peaks at about 0.85 V, there is little reason for concern and this is just another question of academic interest. Also, to put the question of a noise floor of 12 mV versus 14 mV (Gain 500,000) in perspective (6 mV to 7 mV at our present operating gain of about 250,000), the best actual observed noise floor values from the counter-wound coils has been on the order of 75 mV to 150 mV (polarization power supply "off"). I switched over to operating on the new NBLNA board. Instrument performance appears to be unchanged. I will post the polarization controller chart and FDM fom and amplitude plots tomorrow.
The new NBLNA board has a wider adjustment range with less dependence between frequency and other parameters. The effective noise bandwidth is now slightly narrower at about 100 Hz, down from 160 Hz. The -3 dB points are somewhat less critical with the PEV servo now being the normal mode of operation. If an extremely large excursion causes a very large frequency swing, the automatic increase in polarization time should be sufficient to maintain a useable precession signal amplitude.
The weather is turning colder now. This evening (~11 pm local) the air temperature is down to about 12 F (-11 C) and the NMR tau 2 of our working NMR fluid (Prestone De-Icer windshield washer) has fallen to near 0.5 seconds PDF.
Sunday, December 18, 2011
Overnight: PDF, TXT. Instrument health (excellent): Polarization controller PDF, FDM FOM and amplitude PDF.
Monday, December 19, 2011
Tuesday, December 20, 2011
Wednesday, December 21, 2011
Thursday, December 22, 2011
Overnight: PDF, TXT. Instrument health: polarization controller chart PDF, FDM FOM and amplitude PDF.
Friday, December 23, 2011
On the NBLNA design, we used a 100 ohm to 600 ohm output transformer, Mouser Xicon part no. 42TL031. The main intent of the output xfmr is to provide a galvanic ground-break between the NBLNA output and the psuedo differential USB 6008 inputs (each input ground referenced to the computer ground via the USB connection). The 100 ohm "secondary" winding (used as the primary connection in our application) provides a good match for the LT 1357 output driver stage. However, there is little need for the 600 ohm step-up into the USB 6008 25 k input Rs. It turns out that the 100-600 ohm xfmr is often out of stock at Mouser. Fortunately, they seem to keep a stock of some 10s of thousands of the 100-100 ohm version, part no. 42TL030. I ordered some of those and will soon build and test a NBLNA board with the new xfmr. The voltage gain attributable to the output transformer is modest (sqrt(600)/sqrt(100)), so we should be able to easily compensate with our available gain trim for which the high range is already many times the gain ever needed, even since we reduced the gain trimpot from 20k to 10k ohms. There could be a very slight advanage to the 100-100 xfmr, perhaps with less capacitance between the primary and secondary with less windings (not sure, something to look at). Afternote: Using an Agilent U1733C, two 600 ohm to 100 ohm 42TL031 boards measured 247 pf and 253 pf output winding to board analog common. A new 42TL030 board (100 ohm 100 ohm) measures 217 pf output winding to board common, so there is a small improvement in output isolation.
Saturday, December 24, 2011
Sunday, December 25, 2011
Wednesday, December 28, 2011
Overnight: PDF, TXT. We have been running for some days now (since the evening of the 24th) on the newest version of the NBLNA board. We have noted (12/17) that the NBLNA noise floor is very slightly higher than earlier versions, at about 3 nV/rt Hz (was ~ 2 nV/rt Hz). Also, the effective noise bandwidth is slightly narrower at ~100 Hz (was ~150 Hz). There is some possibility of an error in the calculation: Perhaps the new measured noise floor is more indicative of the new eff noise bandwidth than an actual change in the NBLNA (ACV) noise floor, more review is needed.
Also, we have noticed in past brief studies that the most dominent parameter related to the percent success rate is the auto-retry FOM parameter. For many months, we have seen about 82% with an FOM threshold of 2e-6 (~0.1 nT). Now, with the new NBLNA modules (two tested so far), our percent success rate for the same 2e-6 FOM threshold has improved by about 5% to 87%. - yet another interesting aspect of the FDM proton precession magnetometer to investigate later.
Thursday, December 29, 2011
Overnight: PDF, TXT. The spikes and probably some of the activity is from a large snow plow (I'm at an angled corner and he stops to maneuver, so these values get past the auto-retry FOM filter as a stationary ferrous mass. Also, if you view our daytime magnetogram, you will see a number of small positive and negative offsets as parked cars come and go. Ah, the perils of a small yard suburban geomagnetic observatory.
Testing and balancing FDM Proton Precession Magnetometer Counter-Wound Coils for Dave Eubanks of Iowa and Alaska JPG, JPG. Balancing was done prior to this picture out in the middle of a room on pair of plastic saw horses and 3/4" sheet of plywood using an Agilent 1733C LCR meter.
Friday, December 30, 2011
Saturday, December 31, 2011
Want to build your own FDM Proton Precession Magnetometer?
Project Documentation, Links and References
QUESTIONS/COMMENTS/notice of typos, etc. send email to joegeller @ gellerlabs dot com
COPYRIGHT © 2009, 2010, 2011 JOSEPH M. GELLER, All rights reserved.