Lightning Strikes Weewx Database

@vinceskahan I think you’re right–I was following the advice from @vreihen in this post above:

I’d be inclined to make this mapping:

lightning_ energy = energy.AR-00005820.evt_strike
lightning_strike_count = lightning_strike_count.AR-00005820.obs_air

What’s the consensus on that?

Looks reasonable to me.

So, that mapping seems to work on my AIR, but not on a tempest.

AIR mapping:

    lightning_energy = energy_AR-0005820.evt_strike
    lightning_distance = distance.AR-00005820.evt_strike
    lightning_strike_count = lightning_strike_count.AR-00005820.obs_air

TEMPEST mapping:

    lightning_distance = distance.ST-00012563.evt_strike
    lightning_strike_count = lightning_strike_count.ST-00012563.obs_st
    lightning_energy = energy.ST-00012563.evt_strike

AIR lightning graph:

air

temptest

Looks like the Tempest is putting out random distance and strike data constantly?

Looks like the ‘average distance’ is maxing out at 10 as well…

To be clear, there was no (real) lightning activity during the period in those graphs.

Couple things…

  • you can not equate Air lightning data to Tempest lightning data. They’re typically very different.
  • remember that the weewx driver simply listens to the UDP broadcasts from the Hub. You would need to capture the UDP broadcasts (probably on a different raspi or the like on your network) and line them up with events that the weewx driver recorded.

The weewx driver can log everything it hears if you turn up its debugging, but it will log a ‘lot’. You might get there with my separate UDP listener a little easier.

I don’t even bother mapping lightning on my station, as it’s hopelessly wrong by definition. WF creates what it reports for your station by aggregating your readings (if any) with some proprietary mechanism taking from multiple undisclosed sources of data. To me, that means don’t bother trying to use weewx to capture any lightning-related info at all, as it’s wrong guaranteed.

2 Likes

Wow, they should tell you that before selling us that not-cheap tempest thing.
So, essentially, if you are used to running weewx for your weather info display and archiving, and you see WF state that their tempest station is supported for use with weewx, it basically means:
No, it’s not, you’re on your own figuring out how to use the data we blurt out.

Not understanding your comment. WF supports ‘their’ app only. If you want to write alternate software, that software’s author needs to work within the WF APIs to make that happen, which is what the authors of the couple variants of the weewx driver have done.

2 Likes

if tempest is supported by weewx, it means that the authors of weewx build support for it, not the other way arround. Tempest makes the interface which weewx uses. That interface is supported by Weatherflow. Where did you get the impression that weatherflow does the support for weewx?

^^^^^ captain-coredump says the same thing about end users of third-party drivers… :wink: ^^^^^

2 Likes

The question is where did you read this. Enlighten us.

back to your original remark, you are not on your own to figure out how the API is working. It is pretty well documented and you can even try it on line by clicking the API link in the top right of the screen. If you need support for how to setup weewx together with weatherflow, I’m sure there are plenty of users that would help you out if you have a specific problem. Asking a question is much more effective than just complaining.

There’s a whole bunch of product logos on the Tempest literature going back to the original Kickstarter campaign, which could reasonably be construed as implying official support from those products when in fact the integration is via a third-party driver/application…

That is not the point. Pre-sales they are heavily advertising its lightning sensor feature, bragging about the API and compatibility, so you buy it, assuming you actually buy a sensor suite you can reasonably easily connect in weewx. Turns out you’re not. WF is simply using lightning-maps or blitzortung or whatever source and matches that with your stated station GPS, making you wonder if they even truly contain a lightning sensor at all. Looks more like a ‘virtual’ sensor to me.
I’ve been trying to get weewx to parse its electric discharge data for months now. This is just another thread with that same conclusion; It’s a mess. You should stop trying.

Rest assured it does contain a lightning sensor, and it is supposed to send lightning events out through UDP. Weewx could pick those up and process them.
Slightly more than a year ago they added lightning detected from third party sources (which are NOT send through UDP and weewx cannot see them through UDP). It is now really hard to see which lightning is detected by your own station and which comes from third party. Weatherflow already agreed to adding an option to select only you own data to be displayed (

[quote=“dsj, post:133, topic:6262”]
You have convinced us that we need to add that for lightning too, because some users will want to see the data coming off only their hardware
[/quote] ) but that isn’t yet implemented.

1 Like

I think it is well established that the WF lightning implementation is quite unique and outside the bounds of being able to capture things accurately in realtime via the UDP interface. I think it’s also very well established that WF is very good with refunding the money to people who aren’t satisified with their purchase.

So I’d suggest “you should stop trying” to complain here about it. Please go for the refund, save yourself the blood pressure, and shop elsewhere if another vendor meets your expectations better at your price point.

1 Like

There most certainly is a Franklin AS3935 lightning sensor in the Air…and two in the Tempest if I recall correctly. It is a common chip used in all kinds of consumer electronics to detect lightning, and far from a scientific instrument in terms of its accuracy in every manufacturers’ devices that utilize it…

Here’s what I don’t understand; Why would you not implement a truly scientific sensor suite?
It would be more expensive, but honestly, I’d still buy it because it’s a more accurate sensor.
There’s a huge market for high accuracy pro stations without the hassle of having to create your web-interface or app for it. Most accurate sensor devices out there have horrible support for plain displaying, archiving and navigating their output data, one first has to become or hire a coder for that to work. The least expensive lower quality sensors, like netatmo, ecowitt, etc, have that down to a tee. I don’t get those brands. People want the best of both worlds, no compromise on accuracy, and they’ll pay for it.

I think you have very unreasonable expectations at a under $350 price point.

  • Please feel free to buy a Davis VP2 Plus for $1155 plus a WeatherLinkLive for $250 and your total cost will be only $1405 plus tax and delivery charges. That should meet your expectations, except it has moving parts and a pretty ugly 1970s style console interface.

  • Or a Kestral cellular station for $1400 or so but they don’t have solar or uv or lightning for that cost.

  • Or an Orion LX for $6500 - oops, no solar or uv or lightning in that one either.

  • Or a WeatherHawk 916 for $4000 - that has moving parts but still no lightning sensor

  • Or a Vaisala for in the $3000 price range but still no light/lightning sensors there either.

Bottom line is every company has to figure out how much they want to invest in dev/test/regulatory/manufacturing/support/usability and what they think the target audience and price points they can target that will make them an acceptable profit. It’s not easy. It’s not cheap. Raise the price a hundred bucks and you lose a ‘lot’ of hobbyist market share.

I’d suggest you take further followups to the wxforum.net What Weather Station Should I Buy? forum where there are thousands of posts and almost 19,000 members who have been down this road before. Post your requirements there and see what they say there.

5 Likes

Funny you should mention that about ‘moving parts’. The vast majority of Davis VP2 owners have never had any issue with the moving parts, especially not with the latest model anemo- and rain-hardware. I have owned a Davis station before (I’m old), it most likely would outlast the Tempest in MTBF. (The battery in the tempest alone will probably die before one in a Davis would.)
You can buy Davis stations without the console or WL-Live, you simply hook it up to your meteobridge Pro and you’re done. If you want the lightning data, you connect an extra ecowitt sensor to the meteobridge. This is the setup I have done for my father in the South of France, works like a charm.

I’ve been a member of the wx forum ever since it exists, so you don’t have to act like I’m new here. The main reason I bought a Tempest was because they emailed me about it being available in Europe now, and its size footprint came in handy, but frankly, I was annoyed to find out its homebase has no ethernet port. I hate WiFi and all its inherent latencies and insecurities.

The Orion LX is more talk than truth. Its sensors are not that accurate, in fact, its temperature sensor is rated ± 0.3C (probably just another STH3x), and its RH sensor is rated ± 3%. Good grief, the money they ask for that is ridiculous.