General Weather/Earth Sciences Topics > Lightning

Microsferics TOA network 2022

<< < (2/9) > >>

DaleReid:
Thanks, Joe:

The problem with distant stations trying to triangulate the location is the azimuth error, even if tiny, becomes a huge factor at distance.

I've not seen nor read about how this system's designer has approached the math to try to filter out signals of less value. The ideal is to have equally scattered stations (not there yet, but with this new batch of equipment available the areas will start to fill in), and have the math pick the highest value, best condition to maximize accuracy.

To rule out a local station seeing a strike 3 miles away may have some value.  It certainly doesn't when the strike is 100 miles away and with a good separation of another station so that the angle is maximized.

Flying for many decades using VORs at a distance was proof enough that close is good, and angles of wider separation is good.  The factors I don't know are how the GPS timing signals can remove some of the observation error, with the timing available from GPS making the absolute best possible case something in the range of a kilometer anyway.

Lots that we don't know.  I'm hoping for a ton of stations to come on line and we can compare.  But your 20 miles apart stations (good for you!)  would be far different from Lorrick's and mine being about 4 miles apart, when there are few stations within states of us.

I don't recall if you are a Blitz participant also, but I looked at the sferics site last afternoon, and at the TOA (used to be IBM associated) and to the Blitz sites.  The facts for the time being (likely due to the lower number of stations) sferics was showing 1-3 strikes/minute, while Blitz with its hundreds of stations showing an almost blinding number of strikes, circles and reporting station lines to the storm that moved through your area (and was hitting the launch complex).

Until the spread and coverage of stations is more equal, the logic and rationale of the implementation of station filtering and more sophisticated code to pick the best of those stations reporting the same strike won't be known for sure, even if there IS a best approach.  More than one way to skin a cat, you know.

I'd love to hear more (either official or word of mouth) how these places approach the processing and selection  of signals to use.  Knowing just enough to be dangerous, I believe long distance propagation of the strike gets mussed up by the various paths it can take.  I think that the sweet spot is a couple hundred miles from the strike.

Dale

JupiterJoe:
Iím sure Relko will explain the technicalities of this far better than I can, but for longer range strikes, Iím hitting efficiencies of over 80% in most cases.
I am Blitzortung contributor as well. As soon as Microsferics gets more members up and running and filling in dead spots, the accuracy will be similar. The one really nice thing about this system is that it truly is plug and play(with some set up, but no soldering and such) and there are no waiting lists. Also Relko is very good at support whereas you donít hear from the developers on Blitzortung very often.

relko:
Dale: Both Microsferics and Blitzortung use Time (Difference) Of Arrival (ToA) for locating lightning. There are no angles involved hence no azimuth errors to account for, although the ToA method has of course its own set of problems. With ToA, each lightning signal is timestamped on the detecting site and the data processing server can then compute a geographical position by analyzing a batch of different timestamps and site locations.

The simplified schematic below illustrates the ToA principle where T1-4 lines/circles show "time of flight" from the point source (a lightning discharge) to the receiving sensor. The ToA solution is the intersection of 4 or more circles and when the position is known, an exact time of event can be determined as well.

 [ You are not allowed to view attachments ]

Uncertainty of a ToA solution (a.k.a. error) is determined by performing a reverse check; Somewhat simplified, for each site that was involved in deriving a ToA solution an expected time of detection is computed. The result is an ellipse that will vary in shape and size depending on how much the accumulated expected times are off. A small ellipse signifies good accuracy/high confidence and vice versa. Plotted on a map it would look like this:

 [ You are not allowed to view attachments ]

Generally two sites in close proximity is not a problem in most cases. If the error is large, the second site can be used to try to refine the solution. If the error is within some predetermined acceptable limits, both sites are counted as providers of the ToA solution.

The main issue, as JupiterJoe touched upon, is when lightning is very close or overhead two nearby sites. There are a lot of weak cloud discharges in an active thunderstorm which are only detectable when the lightning activity is within a few miles/kms. This "natural noise", while very useful for single point lightning detection (used in e.g. NexStorm as Noise Ranging Assist to improve close storm distance estimations) is harmful for ToA purposes. It will overload the sensors with lightning data that sites farther away cannot detect. This will render the sites useless for a period until the storm moves away. We have some methods of partially dealing with the problem such as auto-adjustable sensitivity that reduces unwanted weak signal detections. It's an area of ongoing research.

Moreover, the other site in Eau Claire, which has been online since the inception of the North American network, is currently a V4 variant which taps its signal data from the Boltek sensor. The new V5 system comes with its own custom E-field sensor. The V5 is generally more sensitive and will detect more lightning farther away (2x in some cases) compared to Boltek.

cjohlandt:
Hello all.  I'm thinking about purchasing one of the Microsferics TOA kits, but had some questions about antenna location and interference.  Where can I find more details about optimum siting?  The product page states the sensor "must be mounted some distance away from potential noise generating sources".  How can I determine what might be a noise generating source?  Would my solar panels be a noise generating source?  Also, I need some help with "some distance".  If I can get a good GPS signal, do I need to make sure I'm high enough to clear nearby houses and trees?  Is there a website or some other resource available?

Also, I have attempted to register for the astrogenics forum.  I received the initial welcome email but never got the follow-up email confirming acess.

Thank you!
-Chris

DaleReid:
Chris, I don't have a unit in hand yet, although it is on order.

I can perhaps give some general info to a couple of your site planning questions.  If you have a portable car GPS you might take it to the area you are considering to see what sort of GPS constellation you geet after it is on for awhile and registering satellites.  Mine will tell me how many, how 'good' the signal is from the satellite, and how many it calculates should be visible.

To refine it a bit, I have a Blitz unit running with the local unit's data visible on a little program provided by Blitz to do just that.  A nice pattern of satellites (for two units when I had a Red, also) showed no time that a good solid constellation for determining position and of course, time, wasn't locked in.  So if you can see most of the sky from say 30 degrees up, you are likely fine.  Do a GPS (I have Garmins in the cars).  I have a TOA unit that is on top of my shed and never got complaints from them unless I loose power and the UPS goes dead overnight, but they don't have as easily accessible health-o-meter as I call it to see what is going on. 

Trees seem to make little difference for my units.

As far as noise, I'm speaking from the old Boltek ADF type system and the Blitz systems, and things like dimmer switches in the house, certain kinds of motors (mostly those I have in the shed like Skil Saw and all but my lights are now all LED in the garage and shed and so far, knock on wood, they aren't showering the area with noise.  A neighbor 500' away sometimes welds and that really is a bad time.

Trying to use a noise sniffer wasn't easy for me to do and frankly I've forgotten what frequencies the other guys used to walk around looking for quite spots. 

I can't comment on the forum access, since I've not had too much time to do much other than look at the US web page display for Microsferics right now.

Maybe some of the other guys (and gals?) with a station already on line could add more detail or email you directly.  Hope this helps and is of some encouragement.  Dale

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version