NBLNA Gain Increased
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 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:
Saturday, April 2, 2011
Overnight: PDF, TXT, Polarization Time Controller PDF. There were a series of relatively slow disturbances overnight including this morning. Afternoon: quiet, normal diurnal cycle PDF, Polarization Time Controller PDF. It is only about 45 F outside, however, the sensor package is nice and warm in the direct sunlight in a transparent container under a translucent plastic cover at 75 F! Evening: shut down for a time, but took some screen shots before power down. There has been significant geomagnetic activity this afternoon, albeit so far well below magnetic storm levels. PDF, TXT, Polarization Time Controller PDF (unfortunately complicated by offsets from two or more nearby vehicles). The spike in the red polarization curve is where I went momentarily to a 20 second fast cycle with the PEV servo to see how far off the Tau 2 calculation was (roughly after 550 on the x axis). Quite an active late evnening: PDF.
Here is the latest excel template for post processing plotting of the polarization controller data XLSX.
Sunday, April 3, 2011
Overnight: PDF, TXT, Polarization Time Controller PDF. Continued sub-storm level geomagnetic activity overnight and this morning. Two relatively fast pulses <10 nT were followed by a 10 nT negative going pulse about 11 pm local. Three negative going half sine wave like disturbances followed at around 1 am, 4 am, and 7 am. There was one vehicle offset, fortunately well defined, just after 12 am. And, yet another small pulse, just after 10 am (interesting that they are all about 3 hours apart) PDF.
Monday, April 4, 2011
Overnight: PDF, TXT, Polarization Time Controller PDF. After an impulse between about 8 pm and 9 pm last evening (local NY, US time), the field was somewhat quiet overnight. The polarization time controller was changed over from the Tau 2 feed forward mode to the PEV servo mode about point 600 to pick up more data for curve fitting at fluid temperatures above 60 F. Note that on occasion we switch to fast cycling (1 minute cycling) or a modified fast cycle (as low as 10 to 20 seconds in the current configuration). Since the polarization time controller graph is by point, time is not always the same between points on the x axis. The actual times can be found in the TXT file. The Tau 2 disturbance which happens once every couple of days or so (e.g. around point 900 on the polarization time graph) remains a mystery. It is most likely due to EMI/RFI interference which distorts to some degree the precession waveform envelope causing an error in the tau 2 calculation. With the polarization time over 2.5 seconds with the fluid temperature under 70 F, it is likely that the NBLNA gain will be increased with warmer outdoor temperatures to maintain a peak envelope voltage of about 2 volts peak. There is plenty of gain margin to go; the extra gain needed, plus more, is available. The experimentation with the polarization time controller of the past few weeks has finally answered the question of how to set the NBLNA gain for nearly optimal system performance.
Couple of day view magnetogram: PDF
Tuesday, April 5, 2011
Overnight: PDF, TXT, Polarization Time Controller PDF.
I am still working on a procedure to determine constants for the relationships between NMR fluid Tau 2 and fluid temperature (natural logarithm) and NMR fluid Tau 2 and polarization time (exponential). The constants (for a given NMR fluid type) for polarization time are also dependent on NBLNA gain, and for the present, we calculate them for a 2 V peak filtered waveform envelope. For the present gain setting and Prestone DeIcer windshield washer fluid as our working NMR fluid, I prepared another two point exponential fit for a plot of polarization time versus Tau 2 PDF. Not surprisingly, approaching summer time extremes (for Upstate NY, US) of 100 F, for the present NBLNA gain setting, it would take polarization times of over 8 seconds to maintain our 2 V peak envelope amplitude. So, it is clear that the NBLNA gain should be increased moving towards warmer seasonal temperatures. Probably the Tau 2 equations should eventually be rewritten to include NBLNA gain as a variable.
All this work of the past few weeks of Tau 2 measurement with ambient temperature calls into question our original fluid studies from November, 2009. Since Tau 2 is a strong function of temperature, fluid comparisons need to be made with fluid temperature in mind. Also, since we now better understand the relationship between Tau 2 and polarization time, it is clear that there can be "apparent" amplitude affects for fixed polarization time comparisons. For example, for a fixed polarization time, a fluid with a short Tau 1 (related to Tau 2) will certainly produce a higher amplitude free induction decay (FID) signal compared with a fluid having a longer Tau 1. So my initial thoughts on comparing amplitude might have been in part comparing amplitudes as a function of Tau 1 (the NMR time constant affecting polarization). Also those early results were probably skewed by fluid temperature.
Wednesday, April 6, 2011 - G1 Minor Magnetic Storm!
Overnight: PDF, mid day PDF, (200 nT vertical scale) TXT, NOAA Space Weather Now PDF, USGS PDF, NRCan PDF.
So far, this event includes relatively slow field excursions. Note that we changed our magnetogram X axis from 100 nT to 200 nT to better view larger field changes, should they occur. There were somewhat large changes in the "H" component of the field as seen in the NRCan magnetograms (pdf listed above). It will be interesting to see what developes during the day. So far, this minor magnetic storm has none of the characteristic fast changes and fast patterns seen in recent previous storm over the last year or so.
12:13 pm (100 nT scale) PDF, 12:50 pm PDF, 4:22 pm (200 nT scale) PDF. , 6:04 pm PDF. This storm caused only modest relately slow excursions of the F scalar here in upstate NY. Compared with some of the fast moving swings of recent past storms, it was relatively gentle.
NOAA Space Weather has switched back over to the new Wing Kp prediction index from the Costello Index, Wing Kp 7 day PDF. If I am reading the graph correctly, the "phase shifted" predictions indicate that it might still need some more fine tuning.
Two day view showing today's minor geomagnetic storm as observed in upstate NY, US PDF.
Thursday, April 7, 2011
I answered a question regarding pre-release sales of FDM magnetometer kits at our Yahoo Groups Proton Magnetometer Group . A pre-release kit ($300 plus shipping) would include all of the parts needed to build the NBLNA and the SWCTRL modules (including the Hammond aluminum boxes (not drilled)). The software is also included (Windows 7 and Vista machines only, at this time). There is more to buy, please see the post and our preliminary project articles. Some reasonable technical support (generally by email) as well as software upgrades are also included. The free Win 7 compatable software is only available with a kit purchase. If one or more readers want to go ahead with a pre-release kit, shipping will be delayed about two weeks for more PCBs (latest version, see the Docs page) to be ordered and made. While still in development and not fully characterized, as seen from our daily postings, the FDM proton precession low cost geomagnetic observatory has certainly reached a very respectable level of performance.
Afternote: By not adding to our present company debt just now, such as by printing more printed circuit boards, I am not suggesting that the FDM proton precession magnetometer geomagnetic observatory project is on hold or in jeopordy. However, without further financial input, such as perhaps some folks or a group who might want to get started with the project early by buying a $300 pre-release kit, it will be some months before kits are offered as a regular product for sale.
Just a reminder, this is a somewhat advanced kit, not suitable for new kit builders or those without electronics experience. For those without sufficient electronic construction experience who are interested in building and/or operating an amateur geomagnetic observatory, it might be practical to team up with a local engineering student, professor, retired engineer or experienced electronics technician. The FDM module is a "black box" software module which converts an acquired precession waveform to an estimated frequency. There is no need for advanced mathematics knowledge.
Friday, April 8, 2011
Overnight: PDF, TXT, I added a brief paragraph to our introductory remarks (top of the page) on the Wadsworth project.
Early evening: Interesting fast positive rise in the local F scalar PDF, the Wing Kp Index appears, so far, less useful as an early predictor as compared with the older, now retired, Costello index PDF, NOAA Space Weather Now did timely predict a K4 event PDF. Ahh, I never seem to leave the fast mode (measurements at 2x or once per minute) on long enough! There was a notable fast risetime impulse just after 7 pm local (2300 UTC) PDF, NRCan Ottowa (OTT) PDF, USGS PDF. Late evening: looks that is about it for this event, 11 pm PDF. I somehow managed to do fast cycle on both sides of all that was fast moving and interesting. It would be really nice to have an automatic fast-cycle mode some day (auto fast-cycling has been discussed at length months past). As always, the spikes are not noise, but actual measurements of vehicles large and small that stopped momentarily too close to the counter-wound sensor coil pair.
Saturday, April 9, 2011
Overnight: PDF, TXT , sample Spectra PDF, Log spectra PDF, some minor disturbances about 2 am local (UTC -4), all well below storm levels.
As we have discussed, the PEV servo is a sort of automatic gain control. It automatically adjusts the polarization time to maintain a desired precession waveform peak envelope voltage. It turns out the PEV servo has just about answered the question of how to set the fixed amplifier gain of the narrow band low noise amplifier (NBLNA) module. Since longer NMR fluid Tau 2 times (related to the fall time of the precession waveform) come with warmer fluid temperatures, more gain is needed to achieve a desired peak voltage for a given polarization time. As we have been noticing, with our present gain, it appears that the polarization times needed at the warmest summer time temperatures would be out past 5 seconds.
Today I shutdown for a bit to set a higher NBLNA gain. Initially, the NBLNA gain was set to give about about a 1 second polarization time for roughly 70 degree F fluid temperature (as calculated by the instrument from the Tau 2 measurement).
There are many ways to measure the NBLNA gain. I use an hp 3325B synthesizer set to our center frequency, currently about 2286 Hz. The hp 3325B output goes into a Kay 837 attenuator and then to one side of a pair of 50 ohm resistors across the NBLNA input. The sync out from the 3325B goes to the sync in of a Stanford SR510 lock-in amplifier. When such sync reference is available, a lock-in amplifier is a great way to measure tiny input signals of 1 microVolt or less. First I set the 3325B output to about 10 mV and the Kay attenuator to give an NBLNA output voltage just below 500 mV, the max input value for the SR510. Then I recorded that value (about 435 mV) and measured the input voltage across the 50 ohm resistor in use (about .79 uV). So, our new NBLNA gain, previously set to give about a 1 second polarization time for roughly 70 degree F fluid temperature, is about 551,000. This is actually kind of ironic, since some months back we were using a NBLNA gain of about 500,000, long before we began to understood the relationship between fluid Tau 2 and polarization time. Our polarization current is 1.5 Amps. Not surprisingly, with the higher NBLNA gain, the PEV servo started showing signs of oscillatory overshoot at start up. Accordingly, the PEV servo gain set was lowered from 5 to about 4 and the moving average to 4 points.
Note that once a FDM magnetometer is up and running, while it is nice to be able to measure NBLNA gain, the gain adjustment can be made using only the FDM magnetometer instrument. In fact, that is what we did. We first set the NBLNA gain to give about a 1 second polarization time (PEV servo on) for a fluid temperature of about 70 F. For Prestone DeIcer, that put Tau 2 around .95 seconds. Only later did we shutdown and remove the NBLNA board temporarily for gain testing on the bench.
Evening: All is well with the new NBLNA gain (~550,000) and the new PEV servo values PDF. I am starting to rough in the Tau 2 FF constants by plotting polarization time (in the PEV servo mode) versus Tau 2 PDF. The first polarization time controller plots look fine as well PDF. 10:56 pm Tau 2 FF cal graph PDF Pt controller PDF . This is really pretty remarkable. Early on in the project, I wondered if 1.5 A for two seconds would be sufficient. Now as we better understand the relationship between Tau 2, temperature, and polarization time, at cooler winter temperatures, a 1.5 A polarization current for well under 500 milliseconds is fine (for a 2 V peak precession waveform envelope). It should be possible keep the polarization time below 3 or 4 seconds on the hottest days (100 F). However, keep in mind that the relationship is exponential, so the problem of what minumum gain to set (for wide temperature range operation) becomes somewhat more difficult, particluarly when higher ambient temperature operation >100 F (>38 C) is considered (e.g. for the US southwest, or any location for that matter closer to the equator). There might be cases where far higher NBLNA gains are appropriate (e.g. 600,000 to 1 million) and/or longer polarization times (e.g. 3-6 seconds are used). Such extended conditions should be completely okay for the fixed base applications of an FDM magnetometer. All of this discussion pertains to our present working NMR fluid, the all weather (-30 C capable) Prestone DeIcer. It is probably the case there are types of fluids more optimal for high ambient temperature operation. It will be most interesting to get back to fluid type testing someday in light of our improved understanding and ability now to better characterized Tau 2.
Sunday, April 10, 2011
Overnight: PDF, TXT The geomagnetic field was very quiet overnight.
I realized that some of the errors in the Tau 2 calculation were being caused by too long a time window. The noise floor (predominent out at longer times), particularly when elevated, was causing an error in the exponential fit to the precession waveform filtered envelope. I re-coded the Tau 2 measurement routine to auto-size the time window as a subset of the acquired Envelope window (presently 2.5 seconds) based on the moving average of the Tau 2 values. A factor of (1.4*Tau 2) is being used to determine the desired window subset length for the exponential curve fit. At cooler temperatures, the window will be relatively short, as short as 600 milliseconds. As temperatures climb, the window will automatically extend, perhaps as wide as 1.8 seconds of the available 2.5 second digitized signal. This new auto-sized window for the expential fit should eliminate many of the errors in the Tau 2 measurement that were being caused by noise further out on the decay waveform.
Afternoon: So far, so good! The precession waveform filtered envelope graph (upper right) shows the new auto-sized reduced time window used to fit the envelope and compute Tau 2 PDF. The polarization controller curves (post processed in Excel) shows good performance and no significant glitches PDF. We are back to a figure of merit threshold of 2e-6 (perhaps a more effective filter of slowly moving vehicles) which explains the somewhat lower percent success rate of about 77% (which is completely normal for this FOM threshold). The small steps are distant departed and then returned vehicle for which, unfortunately, there is no easy remedy. Evening: PDF Even with the new Tau 2 auto-window, we picked up a disturbance in the Tau 2 record PDF.
Readers should note that wore related to the Tau 2 measurement is very fine tuning "around the edges". There is no impact on the quality of the magnetic measurement of the F scalar, the main goal of the project. The Tau 2 measurement, beyond providing the measurement of the temperature of the working NMR fluid, is probably most useful for related NMR fluid studies. Although not entirely clear yet, it is looking as though the PEV servo might offer superior performance for peak envelope amplitude control over the Tau 2 feed forward technique.
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.