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)
I created a new Proton Magnetometer Group at Yahoo groups for those interested in discussing proton magnetometers with an emphasis on Earth's field measurements. I will try to keep it as open as possible without getting attacked by spam. I will also try first with no review of posts, let's see what happens. Please keep it friendly and professional. Use of real names is preferred, possibly required in the future.
September, 2009
Well, it seems every couple to few years I start thinking about the famous C. L. Stong Scientific American Magazine, "The Amateur Scientist" column on Nicholas Wadsworth's proton magnetometer (more to follow on the Wadsworth article). Every time I go back and re-read the project, I marvel over how competent the description is and what a challenging design it must have been back then. Here is a low resolution snap-shot view of page 1 solely for purposes of identification; we are actively working to obtain copyright re-print rights. (Scientific American told us that they do not hold the copyright re-reprint rights for the column. Please contact us at the email address below if you are or know of the person or entity who is the present owner of these rights.) The full article can be still found in many public libraries or private collections. Also, the February, 1968 Scientific American issue occassionally shows up on eBay.
NOAA's Space Weather Prediction Center posts a webpage on Solar Cycle Progression. During this lull in solar activity before the solar cycle maximum, GELLER (Geller Labs) is working towards a low-cost proton magnetometer kit. The goal of our project is a low cost / relatively easy to build / high performance proton precession magnetometer suitable for home and school magnetic observatories. Probably (too early to know for certain) we will offer a box of parts including one or more printed circuit boards, most or all active and passive components, and/or the code and plans to build a stand and pick up coils (the sensor assembly). At present, we are leaning towards the National Instruments USB 6008 DAQ module for digital control and data acquisition (no decision yet). If we use a commercial DAQ module, users will buy the module directly from the manufacturer. National Instruments offers an academic discount. Users will probably also provide a PC or laptop, requirements to be determined.
Project Documentation, Links and References (very early stages)
Journal Notes:
Monday, February 7, 2011
Overnight: PDF, TXT The geomagnetic field was quiet overnight. Here is my latest data for the powered 2.5" coil circuit resistance including the cable to the outdoor counter-wound sensor coil pair PDF.
Tuesday, February 8, 2011
Overnight: PDF, TXT Interesting sinewave shaped disturbance after 10 pm local, 8 nT pk to pk. This USGS plot shows how the same disturbance (after 10 pm EST / 0300 UTC (+5 hours)) looked across the US PDF.
Wednesday, February 9, 2011
Thursday, February 10, 2011
NBLNA Johnson Noise Testing and the Noise Worksheet: I have been thinking more about our Johnson noise tesing method for the narrow band low noise amplifier (NBLNA). The Johnson noise testing method uses only a short and one or more resistors. We can publish a list of nominal effective noise bandwidth values for a selected center frequency for our design. Then, if the circuit is built correctly and is operating properly, users can enter that nominal effective noise bandwidth value as a first pass approximation into the noise worksheet and use the worksheet to determine and/or set any particular NBLNA gain. This approach might be helpful for those who do not have available gear (e.g. a lock-in amplifier) that can reliably measure test signals down to 10 microvolts or less. To fill in the noise worksheet, all that is needed beyond a short and one or more test resistors, is an AC voltmeter that can measure down to about 5 to 10 millivolts. The NBLNA output during Johnson noise testing is a somewhat noisy (the amplitude of the noise signals over time periods of about 0.1 to 1 seconds), so an AC voltmeter with a time averaging feature is even better (not bandwidth limiting of the input noise signal, but time averaging of the output amplitude value), however, one can also "eyeball" the average. True RMS is better, otherwise multiply the output of an average reading AC voltmeter by about 1.13 to get the equivalent RMS noise value. Some older hp 400 series meters (e.g. the older hp 400E and 400EL analog models) have a rear panel DC output representing the AC amplitude which could then be filtered by a simple relatively long time constant RC filter, then measured on a DMM as a DC voltage representative of the AC amplitude. (The link to the current noise worksheet can be found with the Part V article on our Articles page.) This Johnson noise worksheet technique is believed to suitable for quite general use for characterizing amplifiers. I hope to write a design note soon discussing the Johnson Noise Worksheet approach for one of the professional magazines.
Effective Noise Bandwidth for our present NBLNA design is a function of NBLNA center frequency. Here are the estimated effective noise bandwidth values for a number of NBLNA center frequencies: 1900 Hz effective Noise Bandwidth : measured 130 Hz (linear fit 128 Hz); 2000 Hz 133 Hz (140 Hz); 2100 Hz 159 Hz (153 Hz); 2200 Hz 165 Hz (165 Hz); 2300 Hz 175 Hz (177 Hz); 2400 Hz 191 Hz (190 Hz) With component tolerances, these effective noise bandwidth values are within about +/- 7% of what your actual value might be. These values very roughly follow the linear fit effNBW = (0.1243 (center frequency) - 108.4) Hz The actual curve is non-linear. PDF
Friday, February 11, 2011
Overnight: PDF, TXT With our present NBLNA gain of about 305,000, the input exponentially decaying free induction (FID) or precession signal is running about 6 microvolts peak with our present 686 turn powered 2.5" coil (Prestone De-Icer fluid sample in one coil only) in a resonant circuit.
Saturday, February 12, 2011
Overnight: PDF, TXT, one nearby parked vehicle offset overnight. I notice the percent success rate has settled in around 75% to 77%. This rate is entirely acceptable, however puzzling none-the-less. It should never be near 100% since the system is designed to reject unacceptable measurements via the fast cycling auto-retry process (e.g. disturbances caused by passing vehicles). What is not yet known is what is an optimal success rate, or what are all of the factors that can improve the rate. On the other hand, this rate is much improved from the 60% values that were common months ago. There could be some relation to shorter NMR time constants and/or reduced FID amplitude associated with the cooler winter temperatures (typically from 20 F to 0 F, occassionally to - 25 F).
Sunday, February 13, 2011
Overnight: PDF, TXT. Except for one or two very minor slow distubances after 1am (<5 nT) and some increased scatter before 8 am (<3 nT), the geomagnetic field was very quiet overnight. Sample Spectra PDF1, PDF2.
Monday, February 14, 2011 - Kp 4
Overnight: PDF, TXT. Ran overnight with delay 2, relay close, advanced by 5 ms (from 20 ms to 15 ms) and delay 3, advanced by 5 ms (45 ms from 50 ms). There was no apparent change in the success rate, which converged once again to about 77% after about 608 plotted measurements. okay, so here is the view showing the Monday morning garbage trucks PDF. The spikes are somewhat shorter than usual (~7 nT), probably the truck was parked further away this time and the fellow walked to my driveway to pick up the bags.
11 am (1600 UTC) A very fast impulse (here in upstate NY, relatively small amplitude, ~5 nT ) was observed throughout North America, PDF USGS PDF. USGS PDF (1606 UTC). As seen by NRCan OTT (Ottowa) PDF. While the high latitude values recovered, at Puerto Rico, there remains almost a 20 nT offset, interesting. USGS PDF. Expanded view of event as viewed here PDF, USGS 1626 UTC PDF.
4:20 pm (2120 UTC): PDF, NOAA Costello index (prediction) PDF, there is some significant activity this afternoon, albeit still at sub-storm levels. 5:39 pm Kp 4, PDF, NOAA Costello PDF. The aurora oval looks a little expanded, but is not quite to the United States mainland NOAA SWN PDF. 8:38 pm, this event winds down PDF. The total F field remained elevated at Puerto Rico for four hours USGS PDF.
Tuesday, February 14, 2011
Overnight: PDF, TXT Costello 7day view PDF.
We made another set of detailed measurements for the working NBLNA board. While precise testing of gain is not needed, we begain with a combination of the SR510 lock-in amplifier and an older hp 3403C True RMS AC Voltmeter. The SR510 can directly measure voltages at 10 uV, however has a known gain error and it can only measure signals to 500 mV. We have been wanting to try out the old hp 3403C and used it to make the same measurements where the hp 3403C and the SR510 overlap. After that all voltage measurements for the gain data were made with the hp 3403C in the slow mode. Here is the tabular data showing each voltage measurement and the corresponding gain of each NBLNA stage. PDF The blue uV input voltage value in the upper left corner under the hp 3403C column is actually a SR510 corrected value. Next we annotated a schematic diagram of the NBLNA board with the measured voltage and gain data (marked up schematic PDF). And, finally we repeated the Johnson noise worksheet using the measured gain data. To complete the Johnson noise worksheet PDF, we use an Agilent 34410A in the ACV mode with the statistics average function enabled for the second display and wait several tens of seconds to minutes for an average value for each resistor. The average is reset after each resistor is installed at the input terminals. Note that the Johnson noise worksheet can be started with either an estimated gain or an estimated effective noise bandwidth value (See Feb. 10, 2011). An alternative approach is to take and plot many non-averaged AC voltage readings. The scatter of the measured noise data should similarly center over the the theoretical curve with a variance (such as can be represented by error bars) related to your AC measurement time window. In such cases, the graph can be changed to remove the lines connecting the data points for the measured data.
Wednesday, February 16, 2011
Thursday, February 17, 2011 - Kp5!
As you have probably heard, it is common today for banks to de-fund small businesses. For many years, we have enjoyed a generous credit line from Bank of America, despite reports of credit being tight. However yesterday (even though we never missed a payment ... typical), Bank of America de-funded us by lowering our credit lines to present value. Even though I've never used more than 50% of the limit of any account, all of our Bank of America accounts are now by definition at 98% or 99% which presumably will destroy our credit rating until I can pay them down. If the engineering company survives (I am hopeful), FDM Proton Magnetometer kits might be delayed now until about mid November to December of this year (2011), when more funds become available.
There are some interesting fast transitions in the field this afternoon, will check later against NRCan OTT, I don't think all of these were vehicle related. PDF Another event around 8:30 pm PDF (there is also now a K5 warning). There is still significant activity PDF, yet somehow the NOAA Costello index (which otherwise has been an extremely reliable predictor of geomagnetic activity) so far has completely missed this event (??) PDF. (Afternote: Carl Olsen's plot from 10 pm in Colorado PDF)
~11 pm (0400 UTC) NOAA Costello is catching up PDF. Finally, I remembered to turn on fast cycling (at a relatively slow 1 minute rate) to catch an event with fast edges PDF in more detail PDF. A two minute measurement rate, while perfect for quiet conditions, is too slow for the fast edges of an even like this. Eventually, we need some heuristics to automatically switch between slow (2 minute) and fast (perhaps 1 minute is good) cycling. This event is a bit strange, very stormlike in shape, however of very limited amplitude at only about 20 nT pk to pk here in upstate NY. A micro storm?
Friday, February 18, 2011 - G1 geomagnetic storm
Overnight: PDF, TXT. This event had an interesting and complex pattern, however was of relatively low amplitude as geomagnetic storms go. For example, the lead-in very fast positive going impulse was under 10 nT in peak amplitude. We ran at "fast cycle" with a 1 minute measurement rate overnight. Having come into the event after about 8:30 pm local est at about 53,714 nT, despite the complex and varied structure of the event, the field stayed within about +/- 5 nT until a larger fast series of events just before 1 am (0600 UTC) PDF ; 6:59 am expanded vertical scale PDF ; 11:49 am PDF. Two groups of very deep spikes PDF might have been event related, or a large truck or bus might have stopped near the sensor (Unfortunately a school bus seems to time his pickups lately by stopping for a bit not far from our counter-wound sensor coil pair). NRCan OTT, recorded a very deep spike, however just before 0600 UTC (1 am local est) PDF. NOAA Costello 7 day plot PDF.
Saturday, February 19, 2011
Overnight: PDF, TXT. It has not been a good late week period for our favorite NOAA Costello index. The model was late to the G1 storm (albeit minor as storms go) and has been predicting K4 activity which hasn't been happening 7 day PDF. 9:55 pm est, the NOAA Costello model is in serious need of a tune up PDF.
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.