Tempest data collectionn during low battery conditions

Thought I would share some data of what happens when Tempest experiences low battery conditions but still tries to collect weather data. It has been completely overcast in Bellevue, Nebr for over 2 days. There has no sun to charge the batteries via the solar panels, hence, the battery level has dropped to 2.39v. (This similar condition occurred for about 8 day back in May.) This puts Tempest Field Test (ST-7764) http://tempestwx.com/station/15413 into what is known as:

Power Save Mode 2
Voltage <= 2.39 or >= 2.375 up from Mode 3
Sensor sample interval set to every one minute

Lightning data is only collected once per minute instead of continuously as when batteries are charged to full capacity or >2.50v.

Now if we compare that to Tempest° WeatherSystem (ST-2827) http://tempestwx.com/station/21476 . . .in which the battery level has dropped to 2.43v (Power Save Mode 1; Sensor sample interval set to every 6 seconds)…you’ll notice that there are Many More lightning strikes than there are on ST-7764. The Tempests: ST-7764 and ST-2827 are < 3 feet apart, so they should both be receiving the same amount of lightning strikes. But. . .since the battery level of ST-7764 has drained more that the other one, the sampling is degraded because of lower power.

Also of interest. . .in Power Save Mode 2, the Average wind speed equals the Gust. . .with Lull = 0. i.e. there is not enough battery power to sample winds but once every minute. . .so the lull value is set to zero. Notice all the blue shading (with 0 Lull) under the ST-7764 wind graph as compared to the ST-2827 wind graph below.

Can only observe the weather. . .but can’t make it change!



are you sure? I don’t think lightning data is sampled. What I’ve read is that at low voltage the lightning sensor is completely disabled. That you still see strikes in your graph is because the graph not only shows data from your sensor, but also data from third parties. It gets that data from the server and it might well be that you only get that once a minute.

sunny. . .I don’t know how else to explain the fact the Test Tempest (ST-7764) has ~700 fewer strikes than ST-2827 [being only <3 feet / ~75cm apart]. . .unless. . .it because ST-7764 is the beta pre-production model and ST-2827 is the production model. . . but. . .both are on firmware v129 with the Hubs both being on firmware v143. Or perhaps that the lightning sensor chipware has improved greatly since January.

As I said, it probably is just disabled due to low voltage. Therefore you are missing out on all the strikes that the unit itself would have detected and you only get the data from the third party sources in your graph.

this topic explains when what is slowed or disabled


I assume the battery level has dropped very quickly because this is a pre-production model? Otherwise what is most concerning to me is that it only takes two days of low sun levels for the battery to be depleted sufficiently for the Tempest to enter power saving mode. That doesn’t bode well for winter, and is arguably much worse performance than the SPA attachment for the older SKY devices.

1 Like

For all of the reasons and replies listed above. . .that is why I am “Patiently waiting” to test the . . .

" small power booster (battery or solar or wall-powered) for the hopefully small number of users who live where there simply isn’t enough sun to keep the Tempest charged year-round."

“We’re close to a final design and hope to have prototypes by the first of July and, with luck, expect to start field testing by the end of July or early August.”

1 Like

How do I sign up to test this power booster?

How do I sign up to test the power booster too?

I guess WF will announce it (somewhere on one of these community forum pages) when they are ready. . .otherwise. . .just express a deep interest (like I have). . .to let them know you definitely want to participate in this when they call for field testers.

Hi all,

I’ve had my weather station going for just over a week. As is normal for our area the solar radiation level is really low this time of year (avg 50 W/m2). I’m going to relocate the unit on a taller mast but I’m not sure it will be enough to power the station. I’ve been running at 2.35 volts for the past couple of days and the haptic sensor is disabled.

I’d be very interested in field testing a battery unit for the Tempest if it becomes available.

Thank you


I don’t know how good my sky unit is calibrated, but last year the max solar irradiation measured was below 300W/m2 during November, December and January. This is in the Netherlands. If the tempest needs 300W/m2 before it starts charging, it won’t work well during these months. Days are getting short in the winter anyway, so there isn’t much change for it to charge at all.

With the vertical orientation of the solar panels on the Tempest, angle of incidence of the sun will improve during the winter. Does anyone have a table of SR measured value vs. angle of incidence? If not it might be worth finding out because a low SR value in winter might actually be fine for the solar panels.

1 Like

that graph by itself should be a cosine.

But we don’t know the effect of the lens on top, it may change that. Even then, I would expect that a cosine function would be a good first approximation. This would mean your max SR of 300W/m^2 would be equivalent to a higher SR in the summer.

1 Like

On a sunny day some one could take a reading and move their Sky or Tempest at angles to the sun to measure the effect. My place has clouds today so it is something for another day…
cheers Ian :slight_smile: