Tempest charging with artificial light

I thought this one deserved a topic of its own, so it contains some info from previous posting (which I removed)

Due to lack of sunny days in mid winter, the tempest discharged to a point where it sampled only once every 5 minutes and disabled the rain and lightning sensor (power save mode 3). In the beginning I used a single 15W Led source up close to the tempest. I managed to charge it to mode 2, but charging went very slowly. I decided to use two 15W Led sources instead, one against the east solar panels and one against the west panels. The actual distance of the Leds was about 5 cm. Note that the amount of light captured decreases with the square of the distance, so putting the Leds at double the distance, would decrease the charging rate by a factor four (you can’t/shouldn’t put incandescent or halogen lights at 5 cm from the unit because of the heat they generate).
Here is the curve:

It takes only 2 days to charge from 2.37V to 2.80V, at which point the charging stops.

The power consumption of the unit increase when the sampling rate goes up. If anything, you might expect charging to go slower. But the slope in the red and orange part is less then the slope in the beginning of the blue part. There is a sharp increase in charging rate above 2.62V. Something I learned today: this increase in voltage when a battery gets close to 100% capacity is the normal behavior for the LTO batteries used in the Tempest (https://www.researchgate.net/figure/The-OCV-SOC-characteristic-of-the-LTO-based-battery-measured-at-different-temperatures_fig4_324117029).

Interesting enough, due to the fact that only 32 data points are seen in weatherflow’s curves ( https://tempestwx.com/station/33745/graph/105013/battery/3) it is pretty hard to see the increased charging rate in those curves. (I’m still hoping at some point we will get a better graphing system :wink: ). I created this graph by using the Rest api at https://weatherflow.github.io/Tempest/api/swagger/ and did a couple of requests for the last few days in csv format. Saved the output containing 2900 measurements, and imported it in Excel.

In general Leds are 10 times more efficient compared to incandescent lights. What’s more, many incandescent lights emit light in all directions, but Leds are always directional. I got mine from Amazon (https://www.amazon.de/gp/product/B07WD8Y92W), but I’m sure you could find even more powerful led lights)

For more fun, I’ll keep this setup running in order to see it discharge a bit and pick up charging again.

Note that a battery power booster is in development by Weatherflow, so hopefully this artificial charging isn’t needed anymore next winter.

10 Likes

Liked this statement. . .A Lot!

This is @storm_tracker patiently waiting for the battery booster in development. :joy:

4 Likes

Having problems in UK with lack of sunlight to charge so welcome this development

after an extra day of running this setup, it occasionally shows a value of 2.79 Volts, but goes straight back to 2.80 V. It even once showed 2.81 V. So it appears there is no automatic discharge to some value before charging again. During normal usage such a cycle would be generated by the sun setting :-).

Next up, discharge it again, and have a better look at the wind, because I don’t think wind is working good with this unit.

I have an old tempest, with a clearly broken wind sensor. I did put that in the same led setup as my unit mentioned above. This old tempest only charged from 2.51 to 2.56 in a single day. So it might be that 30W of Led isn’t generally enough to get it charging! However, I did noticed yesterday that one of its solar panels is a bit damaged on the surface, so it might not be charging optimally.

Mine went offline in a rather gloomy UK when the voltage dropped to 2.33. I’ve just got it back using a 30w led work light (a Meihua 3400LM) and a couple of mirrors to improve the light collection. It took 24 hours to go from 2.33 to 2.56 volts, and I’ve now reinstalled it.

Really looking forward to a battery booster as mine is dead now. At 70 degrees north we won’t see the sun until the end of the week so no help there.
Trying to charge with a 50W LED, fingers crossed!

1 Like

it will work just fine. put the light as close to the tempest as you can, unless it gets really warm.

Hi, because my tempest shut down as well 3 days ago (mostly clouds here in Berlin-Germany) I was redirected to this topic.
So I put 3x10Watt LEDs as close as possible (about 1-2 cm) and also a 12V fan (for cooling) to the solar panels and started charging the station. It took about 14 hours to charge from 2,35 to 2,79 Volt - so I’m happy again.

The only disappointment is that the voltage already dropped down after only 1 day to 2,6 Volt - so the power consumption seems to be quit high when the tempest is sampling a full speed.
If the discharge continues at this speed I guess I have to charge the station again already after a couple of day. But I have bought now a 100W LED and will be really curious how long it will take with this LED.

By the way does anybody know if there is a possibility to show the actual sampling rate of each probe in the app? Would be at least nice if it could be blend in.

Or is there a possibility to set the sampling rate manually for each mode? So that I can adjust them in wintertime to keep the tempest longer running with rates I prefer.

Christian

there just isn’t a lot of energy stored above 2.6V. As you can see in my curve, when it reaches 2.6V it starts to increase rapidly.
The same is true for a discharge, it discharges very rapidly, until it reaches something around 2.6V
Below you see the graph I took from some scientific article.

Be carefull with a 100W lamp. it might get too hot. Reducing the distance of a LED, quadratically reduces the power. But let me know how it goes. I’m curious.

No you cannot set the sampling rate from the app. I am currently running some experiment. But it takes a long time to get the discharge curves (which is kind of good :slight_smile: ). I will report later. Perhaps in a week or so.

good to know that the discharge rate somehow stabilizes after reaching 2.6 Volt.
I’ll keep you posted how it develops here at my place and when I have to charge again.
With the 100W LED I’ll probably have to add 1 or more fans to the setup to keep it cool.

this graph I just posted, makes me wonder why the unit shuts off way before it reaches 2.2V. Apparently there still is plenty of charge energy left. Perhaps they need the remaining energy for something else? Or perhaps the 2.2V is too low to drive the circuits inside the tempest. But from first looks it seems such a shame to stop working when there still is something like 40%-60% energy left.