Using solar W/m2 to determine solar panel efficiency

Hi -

I am looking at using the Tempest to get

Solar power in W/m2 and compare it against the W that my panels are producing. My panels are exactly 1m2 and 170W, or 170W/m2 per panel, of which I have (9). I’ve made sure my load will pull as much solar as possible (drawing from battery too) and noticed that the panels and the tempest are almost the same. I would think that if the Tempest is reporting 500W/m2 and I had (3) panels perfectly oriented I would be seeing 1.5Kw power, correct? Green line is Tempest, Yellow is solar panels.


well, that is probably a lot more complicated. I think the tempest tries to tell you how much energy would be received, where as your panels probably try to tell you the very maximum they could produce in the end…
So if you panels are oriented perfectly in the direction of the sun they might produce 170W each. However the incoming energy is a lot more. If you are lucky your installation overall (not only the panels) has an efficiency around 25%. So if tempest is reporting 500W/m2, the overall power of a single panel would be around 125W (lots of assumptions going on here, but you get the idea, I hope)

1 Like


Yeah, I understand this is a lot more complicated with time of year, shading, panel orientation, panel efficiency etc. My panels are 18% efficient and I have see 1.5kW near Death Valley in July (was running the A/C via solar)

My goal is to see if I could use the Tempest to say “your potential is 500W/m2” and I see 400Wm2 that I could say that I’m leaving 1.1kW on the table due to shading, bad orientation etc. Ultimately I want to use some ML to build a forecaster based on the APIs from Solcast to tell me something like “at the current consumption rate you will not have enough solar today to meet your demand, you will need shore power to offset demand based on historical usage” kind of thing… I am currently logging all my Victron data and Tempest data to rPI with influxDB.

1 Like

For example, I turned on a 2kW load and you can see my (9) panels were not able to produce more than what the Tempest was using and drawing from battery.

but it should be like “i see 500W/m2 and assume an overall efficiency of 15%, so I expect to generate 75W/m2”
(15% is less than 18 because you will have losses during conversion)

Understood. The efficiency is already baked in through, I can (and have) see 170W from my 170W panel. The wattage being read from the panels is directly from the panels (Victron MPPT) not after AC inverting (10% loss off the top). Thoughts?

ok, so what you basically see is that the tempest shows at some point 500W/m2 and your 9 panels together also do approximately 500W? (which is 55W/m2)
((why do the solar panels dip down to zero every minute?))


Yup, Tempest shows 500W/m2 at some point and the (9) panels show 500W. The MPPT dips down to 0 every minute for the MPPT tracking algorithm is my understanding.

what is the max solar radiation tempest recorded in the past? 1.5kW/m2? That would create 165W/m2 on the panels, close to the expected max.

I’ve see as high as 1kW/m2 but only had it logging for the last 2 days and didn’t have a load on at that time. I see, so 1500/9 = 166W/m2

only because your graph showed that 500W/m2 on tempest happens to create 500W on all 9 panels (which is kind of coincidence, having 18 panels would have generated 1000W). So each panel creates about 11% of what tempest records.

My highest solar radiation recorded was 900W/m2, but I’m in the Netherlands, not as sunny as Death Valley.
I’ll compare my tempest with my Solar panels. Takes a while.

Awesome, would love to understand this more. Note the slightly elevated panel production than Tempest. Now, the Tempest could be un-optimal orientation as well. Thanks for the conversation, this is great!

This is for one of my solar panels (I’ve micro-inverters, so I’ve data for each panel). In green the solar panel in blue the tempest. My solar panel is around 1.5m2. Note the different vertical scales used.
So for the point at 12.30, the tempest indicates a solar radiation of 475W/m2 (or so), the panel produced 175W which is 117W/m2 (175W / 1.5m2). That would imply an efficiency of 25% (117/475*100%), which would be amazing. More likely some measurement isn’t really that exact, or I’m doing something wrong.

As mentioned this is a coincidence having double the amount of panels would just double the value. Tempest measures W/m2, but your panels give W (not per m2)

Interesting. I am looking for rPI voltage/current sensors I can tee off each panel to do the I *V = W since the voltage is dynamic on the panel to get per-panel stats as well…

just for your entertainment, here is the data I have for a single panel. Each panel has its own inverter and uses its own Power Point Tracking. Temperature is the temperature of the panel (real outside temp was max 22 degrees… celsius of course, what else :wink: )


btw Don’t trust the values of solar radiation of the tempest too much. The sensor is a combined sensor that also measures UV values. I don’t have a good way to check solar radiation values, but the UV values were off by a big amount. UV values are automatically calibrated against long term expected values at your location, so they might be more or less correct after a while. I don’t think the Solar Radiation values are calibrated automatically.

there is another very important factor to take into account with solar panels, temperature


sure, that’s why my panels measure that as well :wink:

I picked a clear day for better comparison. What I learned today is that I should rotate my panels around 7 degrees :slight_smile: (Earning money from my weather station)