Wire and Cable for an Amateur Geomagnetic Observatory

Journal notes, Wire and Cable

GELLER Labs "Backyard Science"

Thoughts on a proton precession magnetometer design - a Proton Magnetometer Project

The goal of this project is a low cost high performance proton magnetometer (a digital magnetometer) kit for amateur scientists to be able to accurately measure and monitor changes in the Earth's total magnetic F field and to observe geomagnetic storms. There is a regular daily (diurnal) variation in the Earth's magnetic field. During events related to solar activity, there can be sudden changes in the field (such as a sudden impulse) as well as large excursions in the field which can be more than ten times the regular diurnal variation caused by magnetic storms.

(be sure to hit refresh to pick up our latest changes and entries)

Tuesday, May 11, 2010:

Overnight run: PDF, several days PDF, TXT, 24 hour USGS PDF. The plots from the desktop are much nicer, I need to understand why. Note that the txt file has all of the measurements, including those rejected in the auto-retry process as having a FOM (figure of merit) higher than 2e-6. Only the accepted readings (FOM < 2e-6) are plotted at a rate of about one every 2 minutes.

Yesterday a reader asked why I use #24 wire with a 1.2A polarize current. It works because of the relatively low duty cycle, however the use of such relatively small diameter wire does leave one susceptable to an over-heat condition if the polarization power supply ends up continuously on, such as when I terminated the program in the middle of a polarization cycle. I have been arguing that smaller diameter wire is lower cost and easier to wind, and that it can be protected with a timed interlock.

On further thought, however, perhaps it does make sense to go to with a more robust #18 wire. The thicker diameter of the #18 wire would allow for a lower polarization power supply voltage as well. hmm, maybe it is time to order some wire. (Note that even #18 wire is derated in current capacity for a tightly wound magnet coil.).

The second coil of the counter-wound pair is only present for noise rejection purposes, and never carries the polarization current. I suppose only the one powered coil of the counter-wound coils needs to be wound with #18 wire and that the second coil could remain wound with #24 wire. Probably (might need more investigation) if both coils have the same inductance, ambient noise cancellation will still work as before even with the two coils having different sized wires.

Wednesday, May 12, 2010:

Overnight run: PDF, TXT, (with some relatively large offsets from vehicles) USGS PDF, last few days: PDF. The event at VA (USGS plot) appears to be a local disturbance and not a geomagnetic event. Interesting, I am still focused on the electronics design and working towards the experiment kit, however I think I have noticed a decrease in the local average field from around 53825 nT to now around 53810 nT over several months.

Mouser Electronics kindly offered to sell 100' of the Alpha 86702CY SupraShield (Premium Foil/Braid) XTRA GUARD at the old $99 price (see April 26 entry, now ~$200 at several distributors ), so I guess I cannot resist the temptation to try a length of this most impressive cable along with the re-wound powered coil of the counter-wound sensor coil pair (when the #18 wire arrives). Belden's 5541P1WATERBLOCK direct burial cable might be a more cost effective alternative. Later, I will buy 1000' for testing. ANIXTER appears to have the best price for the Belden cable.

The gyromagnetic constant for geomagnetic observatories has been published by the IAGA. We added a brief article on how to use the Larmor equation and the gyromagnetic constant on our articles page.

Here are a couple of snapshots of the current prototype setup: FDM magnetometer, FDM mag with computer. No reason to have the isolating output transformer in its own box (lower left), it just worked out that way for now. Note that analog and digital grounds combine and make earth ground at the power supply (right most post). From some of the backyard testing last fall to the initial lab test stand, the FDM magnetometer is a lot closer now to a production experiment kit.

Unfortunately, the FDM Magnetometer project is probably going to cost more than the $150 I had hoped for, since in addition to our kit, the user is going to need to buy the copper wire for the counter-wound coils, the cable, some PVC and wood parts for a stand, a USB 6008 (or equivalent for experimenters who branch off with their own designs), and a power supply of some sort (an hp lab supply is certainly not needed). Any suitable existing PC can be used. A dedicated computer should not be needed. I still think that this is an experiment that a lot of amateur scientists, hams, and engineering and physics students will want to build; time will tell.

Eventually I need to start thinking about laying out one or two printed circuit boards. I suppose I'm stalling a little to give the overall design some more thought. Immediate jobs are to re-wind at least the powered coil if not both of the counterwound pair in #18 wire and to re-run the sensor cable testing with the Extreme Alpha shielded cable. Since all has been running fine on an old surplus two pair shielded run that cost $9 (eBay surplus cable) the high end cable is not needed. I suppose I am just curious to see if it further attenuates some of the AC harmonic spurs on the FDM spectra.

Thursday, May 13, 2010:

The laptop either shut down LabView or rebooted at about 3AM. It appears that the incident was related to a failed Microsoft Security Essentials update (the wireless router was off). Not so much of a problem during these short prototype runs. I moved the system back over to the Win7 desktop, the display and PDFs are so much clearer, still not sure why.

I took a quick look at tau 1 times for Prestone De-Icer this morning by graphing precession amplitude for a range of polarization times: PDF. I took an average of five data points at each polarization time. I might revisit the number of turns and polarization currents since I need to re-wind the coils. With lower resistance, it would be relatively easy to increase the number of windings and/or the polarization current. The graph shows that the present operating point (1.2 A, 447 turns, about 2" ID and 1.8" long coil, 1.8 seconds) is viable, yet more sample polarization could be had with longer polarization times. On the other hand, once reliable operation is attained, is there any advantage to more sample polarization?

Also, I do not believe there is any advantage to fast cycling or a higher data rate for an amateur geomagnetic observatory. Since we are not relying on moving averages, but rather taking one high quality point (through auto-retry using FOM filtering), there does not seem to be any reason for oversampling in a basic magnetic measurment interval of about two minutes (a different question than oversampling of the precession waveform). There might be a field of study of micro changes of the geomagnetic field over very short periods of time <<2 minutes, however, I think this very specialized area is probably well beyond the scope of magnetic storm monitoring that most amateur observatories will be looking for. Note that the relatively slow measurements of an amateur magnetic observatory are also very different from some amateur archeological applications, where an increased measurement rate might allow for better ground coverage in a shorter time.

Some good news, zero current switching with the FET-small signal relay hybrid appears to be working well. Here is the latest autopsy photograph of a relay in service from about April 18 to May 12 (no burning or arc marks): JPG.

Friday, May 14, 2010:

Overnight: PDF, TXT, USGS PDF, pretty uneventful; the car going out is about it (the +18 nT offsets).

Waiting for the #18 wire and new Xtreme shielded cable for the sensor run (about 100 feet).

Saturday, May 15, 2010:

Last couple of days the downturn in the diurnal cycle started earlier (was around 10 am EST): PDF. Cable arrived today JPG. Hope to get it installed in the next few days.

Sunday, May 16, 2010:

The new cable is in. It does appear at first glance that the filtered envelope waveform is cleaner and that the number of auto-retries is down. We'll see what the overnight run looks like. The shield is open at the sensor end and connected to the center-tap common of the counterwound coils at the FDM magnetometer connection.

With warmer weather, went back to distilled water and a 2.2 second polarization time: PDF. Evening PDF.

I am thinking when the #18 wire gets here next week, it might be worth using a higher polarization field around 100 Gauss (was 200 Gauss in error) or more with the new coils. I might try about 600 turns, staying with a relatively low polarization current around 1A to 1.5A. I suppose more turns should give slightly larger precession voltage as well.

Interesting to note that since the inductance goes up as N squared, with 600 turns, the inductance will be pushing 20 mH from the present 10 mH per coil. That means a review of the discharge time. Since series coils add in inductance, as opposed to N squared, I wonder if inductance is what was really bothering Wadsworth when he wound his coils with breaks, rather than the concern about self capacitance that he spoke of in his article? I still don't see the effects of self resonance of his coils down in the 1.5 kHz to 2.5 kHz frequncy ranges needed for Earth's field magnetometry.

Monday, May 17, 2010

Overnight: PDF, TXT, USGS PDF, sample spectra this am: s1, s2. Spikes are from vehicles (newspaper delivery, garbage trucks, shool bus, etc.)

Running on the new Alpha cable. Still convinced that a far more inexpensive cable will do, however, no question that some additional noise has been suppressed. The surplus cable I was running on was two twisted pair with one pair having an additional foil shield. That was less than ideal, since the ADI balanced amplifier input achieves best noise rejection with a matched input impedance for each coil feed. The capacitance was imbalanced with the old shield arrangement. Also, there were noisier spectra yesterday afternoon with the new cable.

Presently running on distilled water with a relatively long polarization cycle (2.5s). Time between measurements is 2 minutes, with an auto-retry cycling at about 5 seconds for a figure of merit (FOM) > 2e-6 (~0.1 nT). If the FOM is too high, the auto-retry feature continues to take data until a "good" measurement is achieved, then a new 2 minute interval begins. Among other advantages, the FOM filter (part of the FDM executable software module) rejects passing vehicles since the field is changing during a digitization interval (1.1 seconds) causing a relatively high FOM. Very slow moving vehicles or momentarily stopped vehicles are not rejected (e.g. garbage trucks).

Waiting for the #18 wire to re-wire the coils (presently #24 wire). Planning to increase the number of turns on the counter-wound coils from about 450 turns to about 600 turns, while maintaining a polarization current in the range of 1 to 1.5 Amperes.

Wednesday, May 19, 2010 - COILS!

Wound the new counter-wound coils with 600 turns of #18 AWG magnet wire yesterday and and balanced and installed them today.

I used 5" long 2" PVC sections (about 2.3" diameter) with threaded cleanouts on either end. The threaded cleanout provides both an end stop for my 2" long solenoid coils, as well as a mount to affix the coil form to the coil winder. To make a crude coil bobbin, I cut a 2.5" diameter hole in one cover of a 3M electrical tape package with a hole saw in the drill press. The hole saw generally did not finish the cut, so I cleaned them up a bit with a razor blade. Next, I would vinyl foam insulation around one end of the threaded PVC cleanout to support the 3M electrical tape cover. Then I placed the tape cover over the insulation on the theaded cleanout. Here are a couple of more pics: pic1, pic2. Then I tapped (no PVC cement) the assembly together to make the 2" coil form bobbin. I used some rubber cement to better seat the covers (not shown). Then I screwed each bobbin onto the coil winder (winding could be done by hand without a coil winding machine). Note that I run the coil winder from the large hp supply. Also, I modified the winder motor by bringing out wires from the series connected motor to a reversing switch that I added to the winder (reversing polarity to a series connected DC motor does not reverse the direction because the polarity of both the stator and the rotor windings change). To reverse a series connected DC motor, you literally need to break the series connection and reverse either the stator or rotor winding with respect to the other. I only use the counter on the winder and stop it by reaching across and turning down the current limit knob on the hp supply. Note that counter-wound coils are not the same as like-wound coils with one coil flipped! I did not get this for some time.

I wound the counter-wound coils, marking the direction of each wind. It was remarkably easy this time to balance the coils (match the inductance) using my like new, but very old, hp 4562B Impedance Bridge (a super nice piece of equipment!). I did the balancing on a wood floor in the middle of a room. It only took one turn to match the inductance.

The counter-clockwise (CCW) red coil is 17.61 mH, Q 36, R 2.84 ohms, weight 1346 grams. The clockwise (CW) blue coil is 17.61 mH, Q 36, R 2.85 ohms, weight 1359 grams. I applied power and marked the leads for N-S field with a "+" label. The "C" marker designates what will be the center-tap back at the polarization PCB (both coils are brought back on their own twisted pair and later joined during switching at the polarization circuit (see April 17, 2010). The coils still fit in the storage bin on the coil test stand, pic1, pic2, pic3. (It is possible that better cancellation could be achieved with some more separation between the counter-wound coils, perhaps 1/2 to one radius, I hope to get back to this question.)

The counter-wound coils are wired into the short cable that goes to the water-proof connector, pic1, pic2. The new pair resonated with about 131.5 nF. The 125 ml sample plastic bottle sits on a wood plug on top of a backwards mounted PVC threaded cleanout plug, which places the center of the sample volume about in the middle of the powered (red) coil. Note that as described earlier, only one coil is powered, while the other (blue) is only switched in for ambient noise cancellation during a digitization interval.

It will take some study to optimize adjustments and settings, however, all seems fine with Prestone windshield washer fluid, 1.2 A polarization current for 2 seconds. Here is the afternoon record so far (the measurement scatter is because the geomagnetic field is a slightly active this afternoon): PDF. Too early to tell much about performance. The 2220 60 Hz harmonic has been relatively hight at times (even after cancellation): sample spectra1, spectra2, spectra3. I might have the resonating capacitor tuned a little low (there appears to be a little bias below the center frequency for our nominal field, presently around 2290 Hz). There seem to be a lot of FOMs in the e-7 and e-8 range, however, there are auto-retries as before too. Also, I lowered the amplitude threshold for a good measurement when I got a number of zero returns. The amplitude is higher than before as expected with 600 turns versus 447 for the original prototype coils. At least I no longer need to worry about overheating if the polarize system fails and leave the polarization current on steady state (such as when I terminated the program during the current on part of a polarization cycle some months back). Still looking good this pm: PDF.

I would like to try raising the polarization current to about 1.5A to present a polarization field of around 100 Gauss (was 200 Gauss in error) or more, but that will have to wait for another power supply (this one maxes out at 1.2A). However, it is not clear to me a higher current or higher degree (percentage) of polarization of the sample fluid would give better results (e.g. less auto-retries). The reason is that if the gradient or rate of change of the field at the sensors is too high at any given moment, it is not clear that more polarization would help. It is possible, in those cases, that more polarzation would simply yield a higher magnitude signal with an unacceptable FOM.

As an aside, I noted some days back that while I am not doing any serious data taking yet, I have noticed the total field fall off by some tens of nT during the months that I have been developing the experiment. While sorting through some old pages today, I remembered the NOAA geomagnetic field calculator. Turns out our local average total field is falling off at about -114 nT / year!

Thursday, May 20, 2010

Overnight: PDF, TXT, USGS PDF. late morning: PDF.

Evening: PDF, TXT, USGS PDF. Getting a few too many failed readings at times (below the FDM amplitude threshold), otherwise, honestly, I do not see the results getting much better than this. Resolution is below 0.1 nT and figure of merit (FOM) readings in the e-7 and even the e-8 range are not uncommon. It is probably time to step back and look where the project is. I do not think it is quite ready for laying out PC boards yet, but I do not think there is a lot of room for further improvement in the basic operation ... Maybe I will start a first draft of the construction article.

Friday, May 21, 2010

Overnight: PDF, TXT, USGS PDF.

Project Articles!

Project Documentation (very early stages)

Past Project Journal Notes

QUESTIONS/COMMENTS/notice of typos, etc. send email to joegeller @ gellerlabs dot com

Geller Labs	Geller Labs
Products	Journal notes, Wire and Cable GELLER Labs "Backyard Science" Thoughts on a proton precession magnetometer design - a Proton Magnetometer Project The goal of this project is a low cost high performance proton magnetometer (a digital magnetometer) kit for amateur scientists to be able to accurately measure and monitor changes in the Earth's total magnetic F field and to observe geomagnetic storms. There is a regular daily (diurnal) variation in the Earth's magnetic field. During events related to solar activity, there can be sudden changes in the field (such as a sudden impulse) as well as large excursions in the field which can be more than ten times the regular diurnal variation caused by magnetic storms. (be sure to hit refresh to pick up our latest changes and entries) Tuesday, May 11, 2010: Overnight run: PDF, several days PDF, TXT, 24 hour USGS PDF. The plots from the desktop are much nicer, I need to understand why. Note that the txt file has all of the measurements, including those rejected in the auto-retry process as having a FOM (figure of merit) higher than 2e-6. Only the accepted readings (FOM < 2e-6) are plotted at a rate of about one every 2 minutes. Yesterday a reader asked why I use #24 wire with a 1.2A polarize current. It works because of the relatively low duty cycle, however the use of such relatively small diameter wire does leave one susceptable to an over-heat condition if the polarization power supply ends up continuously on, such as when I terminated the program in the middle of a polarization cycle. I have been arguing that smaller diameter wire is lower cost and easier to wind, and that it can be protected with a timed interlock. On further thought, however, perhaps it does make sense to go to with a more robust #18 wire. The thicker diameter of the #18 wire would allow for a lower polarization power supply voltage as well. hmm, maybe it is time to order some wire. (Note that even #18 wire is derated in current capacity for a tightly wound magnet coil.). The second coil of the counter-wound pair is only present for noise rejection purposes, and never carries the polarization current. I suppose only the one powered coil of the counter-wound coils needs to be wound with #18 wire and that the second coil could remain wound with #24 wire. Probably (might need more investigation) if both coils have the same inductance, ambient noise cancellation will still work as before even with the two coils having different sized wires. Wednesday, May 12, 2010: Overnight run: PDF, TXT, (with some relatively large offsets from vehicles) USGS PDF, last few days: PDF. The event at VA (USGS plot) appears to be a local disturbance and not a geomagnetic event. Interesting, I am still focused on the electronics design and working towards the experiment kit, however I think I have noticed a decrease in the local average field from around 53825 nT to now around 53810 nT over several months. Mouser Electronics kindly offered to sell 100' of the Alpha 86702CY SupraShield (Premium Foil/Braid) XTRA GUARD at the old $99 price (see April 26 entry, now ~$200 at several distributors ), so I guess I cannot resist the temptation to try a length of this most impressive cable along with the re-wound powered coil of the counter-wound sensor coil pair (when the #18 wire arrives). Belden's 5541P1WATERBLOCK direct burial cable might be a more cost effective alternative. Later, I will buy 1000' for testing. ANIXTER appears to have the best price for the Belden cable. The gyromagnetic constant for geomagnetic observatories has been published by the IAGA. We added a brief article on how to use the Larmor equation and the gyromagnetic constant on our articles page. Here are a couple of snapshots of the current prototype setup: FDM magnetometer, FDM mag with computer. No reason to have the isolating output transformer in its own box (lower left), it just worked out that way for now. Note that analog and digital grounds combine and make earth ground at the power supply (right most post). From some of the backyard testing last fall to the initial lab test stand, the FDM magnetometer is a lot closer now to a production experiment kit. Unfortunately, the FDM Magnetometer project is probably going to cost more than the $150 I had hoped for, since in addition to our kit, the user is going to need to buy the copper wire for the counter-wound coils, the cable, some PVC and wood parts for a stand, a USB 6008 (or equivalent for experimenters who branch off with their own designs), and a power supply of some sort (an hp lab supply is certainly not needed). Any suitable existing PC can be used. A dedicated computer should not be needed. I still think that this is an experiment that a lot of amateur scientists, hams, and engineering and physics students will want to build; time will tell. Eventually I need to start thinking about laying out one or two printed circuit boards. I suppose I'm stalling a little to give the overall design some more thought. Immediate jobs are to re-wind at least the powered coil if not both of the counterwound pair in #18 wire and to re-run the sensor cable testing with the Extreme Alpha shielded cable. Since all has been running fine on an old surplus two pair shielded run that cost $9 (eBay surplus cable) the high end cable is not needed. I suppose I am just curious to see if it further attenuates some of the AC harmonic spurs on the FDM spectra. Thursday, May 13, 2010: The laptop either shut down LabView or rebooted at about 3AM. It appears that the incident was related to a failed Microsoft Security Essentials update (the wireless router was off). Not so much of a problem during these short prototype runs. I moved the system back over to the Win7 desktop, the display and PDFs are so much clearer, still not sure why. I took a quick look at tau 1 times for Prestone De-Icer this morning by graphing precession amplitude for a range of polarization times: PDF. I took an average of five data points at each polarization time. I might revisit the number of turns and polarization currents since I need to re-wind the coils. With lower resistance, it would be relatively easy to increase the number of windings and/or the polarization current. The graph shows that the present operating point (1.2 A, 447 turns, about 2" ID and 1.8" long coil, 1.8 seconds) is viable, yet more sample polarization could be had with longer polarization times. On the other hand, once reliable operation is attained, is there any advantage to more sample polarization? Also, I do not believe there is any advantage to fast cycling or a higher data rate for an amateur geomagnetic observatory. Since we are not relying on moving averages, but rather taking one high quality point (through auto-retry using FOM filtering), there does not seem to be any reason for oversampling in a basic magnetic measurment interval of about two minutes (a different question than oversampling of the precession waveform). There might be a field of study of micro changes of the geomagnetic field over very short periods of time <<2 minutes, however, I think this very specialized area is probably well beyond the scope of magnetic storm monitoring that most amateur observatories will be looking for. Note that the relatively slow measurements of an amateur magnetic observatory are also very different from some amateur archeological applications, where an increased measurement rate might allow for better ground coverage in a shorter time. Some good news, zero current switching with the FET-small signal relay hybrid appears to be working well. Here is the latest autopsy photograph of a relay in service from about April 18 to May 12 (no burning or arc marks): JPG. Friday, May 14, 2010: Overnight: PDF, TXT, USGS PDF, pretty uneventful; the car going out is about it (the +18 nT offsets). Waiting for the #18 wire and new Xtreme shielded cable for the sensor run (about 100 feet). Saturday, May 15, 2010: Last couple of days the downturn in the diurnal cycle started earlier (was around 10 am EST): PDF. Cable arrived today JPG. Hope to get it installed in the next few days. Sunday, May 16, 2010: The new cable is in. It does appear at first glance that the filtered envelope waveform is cleaner and that the number of auto-retries is down. We'll see what the overnight run looks like. The shield is open at the sensor end and connected to the center-tap common of the counterwound coils at the FDM magnetometer connection. With warmer weather, went back to distilled water and a 2.2 second polarization time: PDF. Evening PDF. I am thinking when the #18 wire gets here next week, it might be worth using a higher polarization field around 100 Gauss (was 200 Gauss in error) or more with the new coils. I might try about 600 turns, staying with a relatively low polarization current around 1A to 1.5A. I suppose more turns should give slightly larger precession voltage as well. Interesting to note that since the inductance goes up as N squared, with 600 turns, the inductance will be pushing 20 mH from the present 10 mH per coil. That means a review of the discharge time. Since series coils add in inductance, as opposed to N squared, I wonder if inductance is what was really bothering Wadsworth when he wound his coils with breaks, rather than the concern about self capacitance that he spoke of in his article? I still don't see the effects of self resonance of his coils down in the 1.5 kHz to 2.5 kHz frequncy ranges needed for Earth's field magnetometry. Monday, May 17, 2010 Overnight: PDF, TXT, USGS PDF, sample spectra this am: s1, s2. Spikes are from vehicles (newspaper delivery, garbage trucks, shool bus, etc.) Running on the new Alpha cable. Still convinced that a far more inexpensive cable will do, however, no question that some additional noise has been suppressed. The surplus cable I was running on was two twisted pair with one pair having an additional foil shield. That was less than ideal, since the ADI balanced amplifier input achieves best noise rejection with a matched input impedance for each coil feed. The capacitance was imbalanced with the old shield arrangement. Also, there were noisier spectra yesterday afternoon with the new cable. Presently running on distilled water with a relatively long polarization cycle (2.5s). Time between measurements is 2 minutes, with an auto-retry cycling at about 5 seconds for a figure of merit (FOM) > 2e-6 (~0.1 nT). If the FOM is too high, the auto-retry feature continues to take data until a "good" measurement is achieved, then a new 2 minute interval begins. Among other advantages, the FOM filter (part of the FDM executable software module) rejects passing vehicles since the field is changing during a digitization interval (1.1 seconds) causing a relatively high FOM. Very slow moving vehicles or momentarily stopped vehicles are not rejected (e.g. garbage trucks). Waiting for the #18 wire to re-wire the coils (presently #24 wire). Planning to increase the number of turns on the counter-wound coils from about 450 turns to about 600 turns, while maintaining a polarization current in the range of 1 to 1.5 Amperes. Wednesday, May 19, 2010 - COILS! Wound the new counter-wound coils with 600 turns of #18 AWG magnet wire yesterday and and balanced and installed them today. I used 5" long 2" PVC sections (about 2.3" diameter) with threaded cleanouts on either end. The threaded cleanout provides both an end stop for my 2" long solenoid coils, as well as a mount to affix the coil form to the coil winder. To make a crude coil bobbin, I cut a 2.5" diameter hole in one cover of a 3M electrical tape package with a hole saw in the drill press. The hole saw generally did not finish the cut, so I cleaned them up a bit with a razor blade. Next, I would vinyl foam insulation around one end of the threaded PVC cleanout to support the 3M electrical tape cover. Then I placed the tape cover over the insulation on the theaded cleanout. Here are a couple of more pics: pic1, pic2. Then I tapped (no PVC cement) the assembly together to make the 2" coil form bobbin. I used some rubber cement to better seat the covers (not shown). Then I screwed each bobbin onto the coil winder (winding could be done by hand without a coil winding machine). Note that I run the coil winder from the large hp supply. Also, I modified the winder motor by bringing out wires from the series connected motor to a reversing switch that I added to the winder (reversing polarity to a series connected DC motor does not reverse the direction because the polarity of both the stator and the rotor windings change). To reverse a series connected DC motor, you literally need to break the series connection and reverse either the stator or rotor winding with respect to the other. I only use the counter on the winder and stop it by reaching across and turning down the current limit knob on the hp supply. Note that counter-wound coils are not the same as like-wound coils with one coil flipped! I did not get this for some time. I wound the counter-wound coils, marking the direction of each wind. It was remarkably easy this time to balance the coils (match the inductance) using my like new, but very old, hp 4562B Impedance Bridge (a super nice piece of equipment!). I did the balancing on a wood floor in the middle of a room. It only took one turn to match the inductance. The counter-clockwise (CCW) red coil is 17.61 mH, Q 36, R 2.84 ohms, weight 1346 grams. The clockwise (CW) blue coil is 17.61 mH, Q 36, R 2.85 ohms, weight 1359 grams. I applied power and marked the leads for N-S field with a "+" label. The "C" marker designates what will be the center-tap back at the polarization PCB (both coils are brought back on their own twisted pair and later joined during switching at the polarization circuit (see April 17, 2010). The coils still fit in the storage bin on the coil test stand, pic1, pic2, pic3. (It is possible that better cancellation could be achieved with some more separation between the counter-wound coils, perhaps 1/2 to one radius, I hope to get back to this question.) The counter-wound coils are wired into the short cable that goes to the water-proof connector, pic1, pic2. The new pair resonated with about 131.5 nF. The 125 ml sample plastic bottle sits on a wood plug on top of a backwards mounted PVC threaded cleanout plug, which places the center of the sample volume about in the middle of the powered (red) coil. Note that as described earlier, only one coil is powered, while the other (blue) is only switched in for ambient noise cancellation during a digitization interval. It will take some study to optimize adjustments and settings, however, all seems fine with Prestone windshield washer fluid, 1.2 A polarization current for 2 seconds. Here is the afternoon record so far (the measurement scatter is because the geomagnetic field is a slightly active this afternoon): PDF. Too early to tell much about performance. The 2220 60 Hz harmonic has been relatively hight at times (even after cancellation): sample spectra1, spectra2, spectra3. I might have the resonating capacitor tuned a little low (there appears to be a little bias below the center frequency for our nominal field, presently around 2290 Hz). There seem to be a lot of FOMs in the e-7 and e-8 range, however, there are auto-retries as before too. Also, I lowered the amplitude threshold for a good measurement when I got a number of zero returns. The amplitude is higher than before as expected with 600 turns versus 447 for the original prototype coils. At least I no longer need to worry about overheating if the polarize system fails and leave the polarization current on steady state (such as when I terminated the program during the current on part of a polarization cycle some months back). Still looking good this pm: PDF. I would like to try raising the polarization current to about 1.5A to present a polarization field of around 100 Gauss (was 200 Gauss in error) or more, but that will have to wait for another power supply (this one maxes out at 1.2A). However, it is not clear to me a higher current or higher degree (percentage) of polarization of the sample fluid would give better results (e.g. less auto-retries). The reason is that if the gradient or rate of change of the field at the sensors is too high at any given moment, it is not clear that more polarization would help. It is possible, in those cases, that more polarzation would simply yield a higher magnitude signal with an unacceptable FOM. As an aside, I noted some days back that while I am not doing any serious data taking yet, I have noticed the total field fall off by some tens of nT during the months that I have been developing the experiment. While sorting through some old pages today, I remembered the NOAA geomagnetic field calculator. Turns out our local average total field is falling off at about -114 nT / year! Thursday, May 20, 2010 Overnight: PDF, TXT, USGS PDF. late morning: PDF. Evening: PDF, TXT, USGS PDF. Getting a few too many failed readings at times (below the FDM amplitude threshold), otherwise, honestly, I do not see the results getting much better than this. Resolution is below 0.1 nT and figure of merit (FOM) readings in the e-7 and even the e-8 range are not uncommon. It is probably time to step back and look where the project is. I do not think it is quite ready for laying out PC boards yet, but I do not think there is a lot of room for further improvement in the basic operation ... Maybe I will start a first draft of the construction article. Friday, May 21, 2010 Overnight: PDF, TXT, USGS PDF. Project Articles! Project Documentation (very early stages) Past Project Journal Notes QUESTIONS/COMMENTS/notice of typos, etc. send email to joegeller @ gellerlabs dot com COPYRIGHT © 2009, 2010 JOSEPH M. GELLER, All rights reserved.
Manuals
Tech Notes
About Geller Labs
Contacts
Links
Ordering