Tempest lightning ground clutter 🌩

I have my tempest 30 feet above the ground station on a chimney mount filled with sand and properly grounded. I’m on my 2nd tempest so I know it’s not the unit. I seem to be getting a lot of electrical interference from pools Jacuzzis and air conditioning units. When tstorms are in the area the circuit for lightning opens and allows all electrical activity in. Even if storms are 20 miles away I see ground interference from nearby electrical units.

Anyone else have this problem. I attached 2 photos. One is my unit and another is a nearby tempest. You can see the other nearby station is how the graph shoud look. My graph has a solid blue line from 0-3 miles. That must be the interference. Station 49666 Massapequa Shores tempest.

indeed that looks like nearby interference. Get out your AM radio, tune it to some empty frequency and see if you can pick up on it and do something about it. Worth a try.

If you can’t find the source, contact support and ask if they can tune sensitivity a bit down to help filter out. But as Sunny said, try first to find the source if possible. Always better.

Just curious… what firmware are your units on? I’m having the same issue and never had any issues prior to this past week when my firmware was updated to 156.

My hub is version 170
My tempest is version 153

So you have one of the later firmware versions then. Wonder what the one you are comparing to is? I’m thinking something is up with these newer versions. I’ve never had any issues until the firmware update.

Doing it now station 47575

Look at stations North and South of me they don’t get electrical ground clutter.

Latest Tempest FW has indeed opened the ears from the Franklin sensor a bit more. This ‘could’ explain some have more clutter but this means also that near you something is spamming pretty hard. As Sunny said already, get a little portable am radio, tune it to a free frequency and walk around. Often you can find the source and eliminate the problem (cheap transformer, a loose electrical wire …)

If you can’t find any source, only option is to contact support and ask to tune it down again, but please try first to locate the source. It is always better to eliminate the problem versus hiding it (hence you will also hide real lightning further away)

yesterday they turned it down and the 0-3 distance still came through hard. i asked them to reverse it as that didn’t help at all. I honestly think all my neighbors in-ground pools, jacuzzi, heater , central air conditioners. just too much going on. Its probably not even in my yard my i’m suffering from my neighbors electrical arcs.

could easily be the case. But if you can pin point it to one source, they might be willing to do something about it. If it is a device that isn’t supposed to be sparking this much, it usually isn’t a good sign and fixing it might be wise. If it isn’t fixable, you can put a metal case around it to prevent the signal from going very far.

I had the case with my lightning detector (not the Tempest) and I went to the neighbour and talked it over. Starting by explaining what was the detector etc and finally we searched both the source and found it. In the end now he knows the webpage to look for when he wants to know where the thunderstorm is coming from … he isn’t a geek like me but he was happy I shared the info with him. Win-Win :wink:


The current Tempest detector system using magnet EM pickup coils only, needs to be filtered.

On Project NOSL, NASA, 1980 we used optical detectors to determine the strike counts from the Space Shuttle.
Otha H. Vaughn, Bernard Vonnegut and I devised a simple ‘Filter’ to extract the exact lightning data without EMI ‘false’ man made noise… Today no ground station array has done this.

Using the top sensor of the Tempest/Air, and a simple low noise op amp, one can extract the optical impulse signal from the lightning. Even the large solar panels can be used, as this was what we used in our early [1977] device we built. Later at ASRC, I developed a simple IR filter using an off the shelf IR Emitter, connected as a ‘Receiver’. That narrow [optical] band further improved distance and dynamic range by 60 dB.
Experiments at SUNNY, Albany and at ASRC, we detected lightning at a distance of several miles very accurately… Combining with the EMI pulse, then filtered with the Optical pulse, gave direction and confirmation that the EMI pulse was indeed lightning. This totally eliminated 'Ground Clutter" as you speak.

NOSL, ‘Skeet’ Vaughn, B Vonnegut, R Orville, L. Weed, J. Justo c 1977 - 1985.
Google Books: “A Personal Storm Warning Service”


Please realize that the Air and Tempest use an off-the-shelf lightning sensor (Franklin AS3935) chip, which is intended for use in consumer electronics devices to alert soccer moms of an approaching storm so that they can call their kids off the pitch. AFAIK, only limited adjustments can be made by software, and any attempt to treat the thing like an actual scientific instrument will end in disappointment and frustration.

If you want to dabble in lightning detection, check out the Blitzortung lightning ToF network. The current waiting list is 2-3 years to buy a receiver kit…lots of soldering required…


  1. Night-Time/Daytime Optical Survey of Lightning Experiment (NOSL)
  2. Otha H. Vaughn

Compared to the sensor in the tempest, which detects up to 40km away, this optical method that detects “several miles away” seems limited. But nice work.

Yes. It was used in conjuration with the EM Pulse data to filter out the close man-made noise pulses. Therefore yes again, the limitations of optical detection worked within this 5 mile radius perfectly.
It simply used an AND Gate; if both signals are present within that 5 mile radius, then the output measured TRUE and passed the data to the program. If it was only one, then it passed a NOT to the registers. Simple. [of course if it was only one optical pulse, then it is probably an atomic bomb, so who cares at this point :grins]
Modern Radar data collection algorithms use the same technique to extract and delete the ‘ground clutter’ near the dish.

This information on my posts is really for the Weatherflow designers, and is posted here for general comments. If you have constructive ideas, then lets have it!
WeatherFlow system is fantastic and yes, it does need a tweak here and there.

BTW, My 1980 experiments detected by the IR Diode alone had confirmed lightning pulses over 30 miles away, and I could not even see them. Believe it or not.

To accomplish this I used a Logarithmic Amplifier so as to gain well over a 110dB range to prevent clipping.


Those who want to duplicate my original 1977 design, I used an off the shelf [then] IR Emitter, connected to a LM358, 1.2mhz low noise phono pre amp. The diode was connected to the [-] Neg Inverting input with a 1meg ohm feedback. Gain therefore was very high.
Original Lab notes with Dr. Vonnegut are still available at SUNNY, Atmospheric Research Center, Albany, NY. SEE: L. E. Weed 1977-1980 Lab notes. VONNEGUT - WEED
Using the second part of the LM358, was a electro-magnetic coil. Both outputs went to a AND Gate, and each had its own output for comparison on a storage oscilloscope.

Reviewing the Franklin unit, it can also be used for optical, by simply substituting the coil with an IR diode. Have not tried it, but in theory it should work. Go for it!
Note: This one simple IR detector runs on a single supply rail, and is still the best lightning detector-tracker ever made. NASA uses it today as well as for all surveillance satellites used for enforcement of nuclear events from orbits.


Additional and Precise Details of my CG - LF to IR Lightning Discriminator Designs.

The illustrations below are additional documents of a simple loop antenna designed with a Coax [75 ohm] Loop. Experiments with a Hula-Hoop 8 Turn 30mil magnetic wire [transformer wire] design with similar results. The Magnetic Induction Loop was substituted for the ‘Reference Diode’. Therefore all the ‘AND’ comparisons [point of processor input] were done at the point of input to the Optical Infrared [nano meter wave length] The Infra-Red can be any IR Emitter with a 10 degree aperture facing upward. This detects CG and CC as well from about 50 miles in any weather condition [cuts through atmospheric moisture]

Both amplifiers need to be adjusted for ‘Gain’ to optimize the Distance of the noise filter.

Unit tracked thunderstorms from 100 miles [magnetic] and precisely located them within 0 .1 miles from about 30 miles to direct overhead, and produced or plotted vs time a perfect ‘V’ shape plot of the CG cells as they moved in and out of my station [similar to Tempest or Air plots]

The TI LOG114 [Burr Brown] was used as a logarithmic [160dB] amplifier processor. Other DC to 1 MHz amps can also be used such as Analog Devices Log Amps or your own design using a LM358 with a PNP transistor in a feed forward [feedback] instead of a 1 meg ohm fixed resistor. [see TI-LOG114, or AD8304]

The LF frequency was Zero to 10Khz, or B-Dot [current]. Do not use the 500Khz band [see graph]
The IR detector detects SPRITES [see Vonnegut Weed 1980] from over the horizon ~[100-200 miles] as well as CC and CG discharges.

This design overcomes all the previous ‘noise clutter’ near your station

Below are my experimental loop antennas, comparisons of details vs frequency

lightning bands
loop antenna

1 Like

upped your profile, you should be able to post more images (edit your post above with the pencil hidden under the tree dots

Thanks, updated and corrected . .