Station 1888 near Bear Creek NC Online

[Nekota]

Greetings and I'm pleased to have a station 19.3 online and reporting data. It's still in setup and testing while I try to learn the behaviour of the system and the local noise.  I have 2 of 3 ferrites (stock models) attached in horizontal planar at 90 degrees (A & C with B disabled). On each of the ferrites one of the leads has a loop knot suggesting a polarity so I am wiring all the ferrites with respect to that polarity.  The E field channel is using the probe PCB antenna and is disabled due to large signal pulse every second. I suspect the local electric fence charger is responsible for that noise. I have installed the board into the aluminum case and have not wired up the external switch so power is applied with plug. The LEDs show some testing on power up and after about a minute or two the GPS LED begins to flash and signal LED with audio clicks can be heard. 

Plans : I will attach the third ferrite perpendicular to the planar to form a box xyz corner or Y configuration. I plan to remote this box and antennas about 1000 feet from the setup location and need to have a very stable system.

About me : I retired from hard drive R&D a couple of years ago where I designed testing methods for asperities on the disk surface. One of the problem areas is ESD so there are some sensors that attempt to measure voltage and H fields but the E field signal is small and H field signals are huge so seeing these signals from distant lightning strikes brings back those struggles.  I also built much of my test equipment and develop software (MATLAB) and python to make measurements. Now I am back  to NC where  I grew up and still want to measure stuff with Raspberry Pi and this blitzortung blue is interesting tool for LF noise analysis.

Issues : I do have a problem that I have searched for and either my search keys or the subject hasn't been discussed.  After I have made changes to the preamp settings (Normal/Disable/Don't send) the system will startup and run for about 5 minutes and then the buzzer goes on solid for a couple of seconds and the system http page freezes. The power, mode, network and fault LEDs are steady but GPS, alarm, and signal are off and the system seems to be locked up. Since it's in a case, the push button switches are not accessible so reset is done with power cable cycle.   After reset with power, the system operates for hours doesn't show any error messages on the debug screen.  I do see the local control changing the gains and some remote messages about overall gain of 6400 and thresholds of 100. Any suggestions from others experience would be helpful.

Hi Nekota:

Welcome aboard from another fairly new guy here. I've been running a system blue since April and have learned a few things already--primarily due to Cutty's expert advice.

The standard for usage of two H-Plane ferrites is to first use channels A & B and disable C:


Put the controller in manual mode.  After making any changes you'll want to be sure to click "apply" then "save now" as seen at the top of the controller web interface page.
Then set the channel C gain to 1x1 and disable it.  Apply and save the settings again. After doing this you may go back to automatic mode if you like. I also recommend setting the HP (high-pass filters) to off on all amps since they are actually another gain stage and can introduce additional noise. 

RE, the resets, I've found that if I have the amp gains set too high, or if the controller goes into interference mode for too long, it will reboot itself. I've also found that the presence of moderately strong nearby RF fields even though many octaves away frequency-wise causes the system to begin sending nonsense data. I can cause this to happen by transmitting 100-watts on the 30-meter (10 mHz) amateur band. The only fix I know of is a reboot. 

A stable power supply delivering somewhere between 4.5 & 5.5 VDC is also important. A linear ps is best but you can find switching-mode supplies that are OK, too. The easiest way to tell if your ps is RF quiet is to put one of your antennas near it while watching the Signals page in the web interface.  When you get close to the PS does your displayed noise level go way up? If so, that's a noisy PS and you may want to try something else. When you disconnect all of your antennas (probes) do you still have noise on the signals page? If so the source is likely the power supply. I have a Ubiquiti PoE injector that generates a great deal of LF RF hash that was showing up on the signals graph. Being a cheap sort, I wrapped the entire injector in aluminum foil and greatly reduced the noise using that simple method of shielding.  I also put a ferrite bead RF choke on the CAT 5 near the injector.

To narrow down where your external noise is coming from try going to the signals page and set it to display only signals from one of ferrite loops. Check the Noise floor box.

Like so:



Slowly rotate the ferrite in the horizontal plane while watching the noise floor level. When you are pointing directly at the noise source you should see a null point or at least a point where the noise is significantly reduced.  The null point will be off the ends of the ferrite. So if you have the ferrite oriented north-south the strongest sferic signals will be received from east-west or broadside.  If the noise source is north or south is will be nulled or attenuated when the ferrite is pointed directly at it.  The null point can be quite narrow. If you have multiple noise sources then things become more difficult.   

There is a USN submarine communications station, NAA, in Cutler, Maine 80 miles away from my location.  It transmits 1-MILLION (!!) watts on 24 kHz.  There seems to be little I can do when it is transmitting.  You can clearly see a big spike at 24-kHz in all of my signals graphs when NAA is transmitting. Lately it hasn't been transmitting 24x7, so I am enjoying the quiet while I can. Since I am retired Navy I've considered calling and asking them to turn it off...  Of course the point is, you can't always get rid of all noise.

You can confirm that your H-plane probes are in phase by using the knots in the leads as you have done. If you ever cut those off or lose them and want to know if your ferrites are still in phase place them directly next to each other so that both ends match--like two crayons in a box of crayons- and the leads are both on the same end. Look at the signals page. You'll have two sine waves. If one of them is leading the other by 180 degrees you'll know they are out of phase. Reverse the leads on one of the ferrites to bring them back into phase. Lots of folks say this doesn't really matter and they are probably right--but I like to keep mine in phase.

I hope my attempt to explain a few things hasn't been too elementary for you and I wish you best of luck with your system blue. It's a very interesting hobby and a lot of fun.

--Kurt

Thanks Kurt for your detailed and clear comments.  

What I've learned in a week about this station is the system goes into interference and the unit appears to be locked up.   It turns out there is around 150 sec timeout for the interference mode and the system begins to operate again.  I have looked for noise sources and often the system shows no interference for hours.  I am only using the H amp with 3 ferrites in XYZ (ABC) and checked the polarity of the antennas by placing all in parallel as you suggested. I installed the most recent firmware which has a 60 sec timeout for interference and some notes about help with debugging interference signals which I have to discover.  My H gains range from 500 to 4000 and the noise floor ranges from 10mV to 60mV - again gain dependent.  I was hoping the noise could be expressed in root Hz so the gain and bandwidth are accounted for. The signals show 5 to 40 per second with 15 a common value and stroke values of 40 to 60 percent. When I get interference the signals are 400 to 999 in less than 2 seconds and I cannot see what they look like. Maybe they are available but I haven't found them yet. 
I would love to have a webpage feature to pipe the waveforms to a local storage device or maybe use wireshark to capture the UDP packets but that seems like overkill and lots of work.

On the construction I used the photo which shows the LED to be R G G Y B R R but the board chip LEDs are R G G R B R Y which makes more sense.  I don't think I'll swap the Fault and Signal LEDs at this point.