Smooth trace, smooth spectrum - strange

[hp3ak2]

I am a new user for 3 days now.  I notice that a few of my reporting waveforms and spectrum plots are what I would call "very smooth" compared to most reports from my station and all other stations that I have viewed.  The system seems to be receiving and using these observations, as well as my more "normal" waveforms for calculations.  Can anyone explain these beautiful, but odd waveforms?  They only seem to show up at a range of 3,000 km to maybe 3,300 km ... very strange.

I have attached a screen shot of a report.  -Or- they show up often enough that you should be able to view them:  Station 1838 in Volcan, Panama.  Thanks for any information that you can give me. 

   

[BobW]

So I'm going to start with "I don't have all the answers"....
It appears that the really smooth waveforms are something relatively local to you, and are not due to lightning. The system appears to "think" that some of the pulses are related to lightning strikes, but it's really just coincidence. Look at the time offset on some of your detected strokes. It's many 1000's of microseconds. The pulse fits a stroke that occurred somewhere, but isn't related. Alot of this really is guessing that a timed pulse that's within a window is related. Look at some other stations and yours with some closer strokes. The time offset is much lower. The matches probably occur during heavy lightning periods like tonight in the US. That way, one of the observed pulses has a bigger chance of fitting a window.

Another way to make sense of this. You're in Panama, so I don't know if you do miles or meters, so you may have to convert....
Say there's a single lightning strike exactly 2 miles away. 10 seconds later you hear thunder. (speed of sound is 1100 feet per second, 5280 feet in a mile, so 5 seconds = 5500 feet, or pretty close to a mile)
So now say there's two equally bright lightning strikes that happen about 5 seconds apart. You don't know it, but the first one was 4 miles away, and the second is 1 mile away. 10 seconds after first flash, you hear thunder. 10 seconds later you hear another thunder - pretend it's loud also. So, Do you assume that the first thunder you hear is from the first flash, and the second is from the second flash and calculate the distances accordingly? Two miles for the first flash, because it was 10 seconds flash to bang, and the second flash at 3 miles because it's 15 seconds flash to bang for second flash?

As you can see, if there's occasional lightning with enough gaps between them, you can accurately distance - but not locate - them based on time of arrival of thunder. Put lots together, and it gets difficult. Now think in terms of the zap to pulse received in terms of the microseconds, or low milliseconds that this system has to deal with.  

So, how to locate the source of your noise? You can rotate your antenna so you get peak on one channel, and null on the other, The source is then in either of the two directions of the peak loop. It might be interesting to look directly at the waveform page on the controller with the "squelch" disabled. There may be a pattern to the pulses not otherwise visible.

And, I'm sure that there's others that can explain this much better.

[allsorts]

I like that variable distance from events resulting in the observations of the events getting jumbled. Something I hadn't twigged about a real world TOA system...

They are a strange signal but the spectrum plot reflects the impulse so I'd say they are real. As a check do they stop if you short the head amp loop inputs? How is the H-field head amp connected to the loops?

The other thing to note is that they are *BIG* signals, I've seen the peaks clipped off at something over 2.4 V peak to peak. The Y scale on many plots automatically scales itself to the size of the impulse, so what appears to be a similar sized signal can actually be very different. The fourth Y marker away zero on your plots varies from 0.08 V to 1.2 V. These "smooth" signals vary in amplitude from only occasionally 0.6 V P to P to overloading the amplifier at > 2.4 V P to P. Most "normal" signals are less than half that minimum, 0.16 to 0.3 V P to P.

As BobW says watch the "noise floor" display on the receivers Signals page and see if there is any sort of pattern or regular repetition. If they aren't very frequent getting a bearing from a loops null might be rather time consuming. On my 30 cm 20 turn loops the null is very deep but also very narrow. A tiny rotation, just a degree may be two, can take you from one side of the null to the other and with the amplitude varying anyway...

They look like switching impulses but can't think of anything that switches a fair amount of power (due to the signals size, unless it's very close) at what appears to be tens of seconds or a minute or two intervals.

[BobW]

"Something I hadn't twigged about"
Ok... Had to get the Brit to Yank translator out for that one!

So , watching the page on LMO, those pulses happen pretty frequently, so loop peak / null should be able to do it. It looks like they center around 10khz. Dog collar fence, electric fence, lawn mower loop? Power line comms?

[kriu]

I am a new user for 3 days now.  I notice that a few of my reporting waveforms and spectrum plots are what I would call "very smooth" compared to most reports from my station and all other stations that I have viewed....

This signal certainly does not come from lightning. This is a disturbance from a nearby electrical appliance. Do these types of signals appear in series a dozen or several dozen signals per seconds?
Besides, your stations capture extremely a lot of bad signals and only a small percentage of them really are from lightning.
I suggest to significantly reduce the gain - 2-3 times. The amount of sending signals to the server should not exceed 20,000 per hour during a strong storm and 3,000 per hour during a period of low storm activity. You will see that the amount of correct detected signals from the lightning does not decrease significantly.

[allsorts]

"Something I hadn't twigged about"
Ok... Had to get the Brit to Yank translator out for that one!

Whoops, sorry. "Something that hadn't occured to me before".

So , watching the page on LMO, those pulses happen pretty frequently, so loop peak / null should be able to do it. It looks like they center around 10khz. Dog collar fence, electric fence, lawn mower loop? Power line comms?

My turn. "LMO"?  

<strike that- LMO the onlilne version of the PHP MyBlitzortung that you can run on your own server>

Doesn't that just cycle through the last dozen or so signals showing each one for a few seconds. Without looking at the time stamps you don't know if those dozen or so signals cover an hour or a minute...

The Archives > List might be a better place to look at the signals. Being static you get a chance to look at them properly. Just spotted some that have that auto scaled Y axis at 2 V! Another feature is many signals have an "echo" approx 250 us after the main signal. 250 us is about 75 km at light speed in free space or 60 km in a cable. I can't imagine what in the world could cause a real, measureable, signal with path length distance difference of 75 km. Unless VLF is reflected by one of the ionosphere layers? But the "echo" is fairly consistent in being 250 to 300 us and it shows on small (100 mV PtP) and the massive 2 V PtP signals.

I doubt the head amp to receiver cable is 60 km long but I'd have a good look at the termination of the screened plugs to screened cable. The cable/plugs I used needs the cables screen drain wire soldering to the plugs screen. There is also something you do (or maybe don't do!) with the input side of each channel of the H-field head amp when using loops. Bridge (or not) a couple of solder pads.

[BobW]

This is updated every couple of seconds. Gives some indication of what's being observed and how frequently. They don't repeat, and the time stamps only move forward, and should be close to current time if station is being triggered frequently.
https://www.lightningmaps.org/blitzortun...o_sid=1838

[nerd65]

Has it been established that the source of the signals is actually external to the unit?  Do they disappear when the antennas are disconnected?  It isn't beyond the realms of possibility that they are actually internally generated signals caused by a fault condition.

[kriu]

To hp3ak2:
Take Laptop and connect your station to the USB (power) and LAN (network) in laptop.
Set up IP addresses on your laptop and station from the same domain.
Disconnect your all entire network of power energy at home and nearby.
Observe then the signals received by your station.

[hp3ak2]

To hp3ak2:
Take Laptop and connect your station to the USB (power) and LAN (network) in laptop.
Set up IP addresses on your laptop and station from the same domain.
Disconnect your all entire network of power energy at home and nearby.
Observe then the signals received by your station.

  Very good and detailed advice.  I will try that.  It will be a good way to get isolation from the house power & other potential noise.

Thanks to all who have made suggestions and offered advice.  ** I believe that I have determined the following at this point with my system Red:

1. The signals occur at approximately one second intervals, with some gaps.  This is identical in behavior to the electric cattle fences nearby in my shortwave receivers.
   
2.  They are very strong compared to other signals being received, 1.2v / .23v (Equivalent of kilometer long transmitting antennas, 1/2 km distant.)
   
3.  Slowly rotating my 1m crossed loops will null the signal in one channel, while maximizing it in the other channel.  I can make either channel very strong while making the other quite weak, or I can make both about equal, but very strong.  It does not appear that moving the antennas, within this neighborhood, will do any good at all.
   
4. The blitzortung computers usually ignore these when sent; however, some of them are close enough in time stamp to be processed.
   
5.  I have tried turning on the filters and setting the sensitivity back in each channel.  However, the signals are just too strong.
   
6.  Since most of the energy is from 5 to 40 kHz, I don't believe that there is a practical way in which to filter it out.  Suggestions?
   

I have been unable to locate a good set of instructions on manually setting up the gains, triggers, etc. in the forums.  ??

[allsorts]

Slowly rotating my 1m crossed loops will null the signal in one channel, while maximizing it in the other channel.  I can make either channel very strong while making the other quite weak, or I can make both about equal, but very strong.  It does not appear that moving the antennas, within this neighborhood, will do any good at all.

The loops don't *have* to be at 90 degrees. 90 degrees just produces a nice omni directional polar pattern. I have two VLF transimitters within 30 miles one at 285 degrees the other at 260 degrees, these produce huge signals. I have one loop orientated to pickup as little as possible of both transmitters, they aren't equal strength so the null of that loop is not on bearing 272.5 degrees but more towards the stronger. This channel has the LPF set to 54.7 khz (did I say there is also a 60 kHz time signal transmitter co-located with one of the VLF transmitters?)

The other loop has its LPF set to 17.3 kHz (the Tx's are on 19.6 and 22.1 kHz) and rotated away from the first loop such that the levels of the interfering carriers are the same for both channels. This means the angle between the two loops is only about 30 degrees. Yes the polar pattern is severly compromised but if I had the loops at 90 degrees the one "looking" at the transmitters wouldn't see anything other than those transmitters.

Perhaps you could do something similar? How well do you get on with the farmer with fencer? Would they llet you try supressing it? How well maintained is the actual fence? Are all the energised wires well insulated from ground and vegetation kept clear?

As for a guide for gains, thresholds etc, every stations enviroment is different so what works for one probably won't be particulary applicable to another. My station is in "automatic". There is a recomendation to have the first gain figure higher than the second on the settings page, ie use 10 x 8 rather than 8 x 10.

[BobW]

how is E-field vs. H-field? More or less? Also, your gain looks pretty high. Mine is broken right now... Lawn mower got the cable to my loops. Will try to fix this week. But, my gain is usually 4x5 or so with a 1 meter 4 turn loops and I routinely participate in strikes in excess of 2000 Km away. Try knocking your gain down a bit and see what the pulses look like and see how much participation you get.
I looked at the satellite pics of where you are, and it looks rural. You are probably on to something with the fencer. Not sure what can be done about that as the output of those is a big spike of short duration. They may easily be 10Kv or higher spikes, and you want to make a lasting impression on whatever gets near the fence without doing anything permanent. As the other poster said, see if you can confirm fencer, and maybe which one. If you can isolate it - verify who's by seeing if you can get it shut for a bit - offer to patrol the fence looking for arcing - hot spot on your radio - and possible vegetation contact. If that might help you and the fence owner.

[hp3ak2]

how is E-field vs. H-field? More or less? Also, your gain looks pretty high.

I looked at the satellite pics of where you are, and it looks rural. You are probably on to something with the fencer. Not sure what can be done about that as the output of those is a big spike of short duration. They may easily be 10Kv or higher spikes, .....

Some good advice, thanks!  System Red and Amplifier 12, so cannot look at the E-field.  There are dairy farms and electric fences all around me.  The farmers are generally uncooperative, especially with a Gringo like me.  The main computers are generally sorting out the false hits, except then the time stamp happens to be very close.  It sure hurts my statistics though.  Again, a FAQ on all manual settings and their effect, in one place, would be most helpful.  But, like Linux and all things geek, you're just supposed to know this stuff or look it up, piece-by-piece!

[allsorts]

Not having a comprehensive manual is part of the fun. B-) Encourages experimentation, though I must admit it would be nice to have a target range for the noise floor as presented to the ADC's. A lot of settings have a blue "i" that provides some float over help.

[hp3ak2]

Not having a comprehensive manual is part of the fun. B-)  Encourages experimentation, though I must admit it would be nice to have a target range for the noise floor as presented to the ADC's. A lot of settings have a blue "i" that provides some float over help.

I now see that I posted this thread in the wrong area of the forum.  My apologies.  I will not be reading any more replies here.  Thanks to everyone who tried to help!

*** "Welcome to British Experimental Airways.  As a passenger we wish to give you the most control of your flight possible.  My name is "Automatic", and I am your robotic Captain.  I have been programmed to give you an average flight, from an average airport, in average weather conditions, going to an average destination via an average route at an average altitude." 

"Since we know that you can probably do better yourself, please note the little steering wheel and various knobs and levers in front of you.  You may take control of most functions of the aircraft from your seat.  There are clear labels on each control, but if we told you what each one does and how they interact, then where would be the fun in that? It's a grand experiment!  If you are a regular passenger, then you probably have a good idea of what to do to get us all there more quickly and safely.  ... Or, you can just leave "Captain Automatic" in charge...trust me! " ***

[BobW]

"Wrong area of forum"?
So what? I think it's interesting reading for anyone following. I take the labels as general guidance, not law.

You can do e-field testing with your h- field amp. Un wire the loop on one channel - or both - and stick a wire in the inin grounded side of the nput. You may find that e-field holds some hope.