This project has moved and is read-only. For the latest updates, please go here.

Incorrect temperatures being reported?

Topics: 4. Bugs
Sep 3, 2016 at 2:21 PM
First, a big THUMBS UP for a great application! I am pretty new to Raspberry pi's and all, and was looking for a more reliable way of uploading PV data to PVOutput from my Sunnyboy 4000TL-20, than a rather flaky Android app. This one works, straight out of the box! I followed the instructions CAREFULLY, and all is fine - thanks again.

As I learn more about python and sqlite, I intend to write some routines to display live data on an LCD panel, remote to the inverter.

However, I notice that the inverter temperature readings are suspect... figures start at around 50C at first light (ambient is around 22C at this time), and rise/fall gently over the solar day. The Andoid app I had showed the same error. I am pretty sure that the inverter itself is OK, it only fires up the variable speed fan when it needs to. So maybe it's not a problem with the application, but a data definition error? I see some similar and other wierd figures reported in others' data on PVOutput. My system there is WILKINS_WATTS

Any ideas?
Sep 5, 2016 at 12:28 PM
I have the exact same inverter, temperature is starting around 35C going up to 65C max. I never heard the fan running until now (except when I turned it on manually)
SBFspot gets the internal temperature of the inverter, that's all I can say about it. Unfortunately we can't compare it with SMA's Sunny Explorer as it lacks the temperature.
Sep 5, 2016 at 5:21 PM
Hi, thanks for your reply. Hmm, even your initial temperature of 35C seems too high! (unless your inverter is in a VERY warm room?). Though I can believe a high temperature of around 65C. Looking at the other temps in the PVOutput SBFSpot group, they are all over the place, even for the same inverter. Maybe the data simply isn't that reliable?
Sep 7, 2016 at 10:15 PM
Edited Sep 9, 2016 at 12:51 PM
hej laurence and sbf,

to get hands on the reality of the meaning of the value of the "temperature"
may i ask you to evaluate your figures from the perspective that SMA-inverters might report
the temperature with a multiplication factor OR with an offset

then we need to know:
-1-?- if the scale is not 1/1 -> what is the multiplication factor and
-2-?- if the scale has an offset -> what value does the scale start on

kr wim - - - -

can you == the users of sbfspot take the time to keep some history of temperature info,
e.g. -> an easy reference point -> early day the ambient === inverter temperature

-a- for sunny boy the Operating temperature range is -25°C to +60°C
-b- inverters will reduce output (when temperature raises) and finally stop output at >60°C -> see SMA-doc: Temp-Derating-TI-en-13 | Version 1.3
for this -> there is no use in a value that reports from 0°C <--> up to 255°C (scale at 1/1)
?? AND what value will an SMA inverter use to indicate -25°C ??
Sep 10, 2016 at 3:53 PM
Hello Wim, thanks for your interest in this issue! I can add a few comments:

I have only just started using sbfspot to upload data to PVOutput, and so I only have a few days' worth of temperature data (although five years worth of generation data!). You can see those records on PVOutput for yourself - my system is called "Wilkins_Watts" I have compared my results with other users in the "SBFspot" group, and the results are all over the place!

First thing in the morning when the inverter starts up, the temperature is reported as 50C, when in reality the ambient temperature is nearer 22C.

The reported temperature does rise and fall gently throughout the day, roughly in line with power generation. The highest recorded figure is around 64C

When the inverter shuts down at the end of the day, the reported temperature is around 57C. I estiamte the ambient to be around 25C

I am in the UK, and our peak power days are now behind us for this summer. My SMA 4000TL inverter has a fan, but this rarely comes on (and only then, and minimum revs). I have never seen the inverter "flat-line" suggesting any kind of thermal limits being applied. So I am pretty sure that theinverter is in control of its own temperature management, but it is not necessarily reporting its true value

It is not so easy to get at the rear heat sink, but I might be able to add a temperature monitor to the heatsink, and record/compare those values. However, the internal sensor may be located closer to the power transistors, with the heatsink temperatures running a few degrees lower.

I imagine that there is both a scaling and offset factor as you mention, yes.
Sep 12, 2016 at 8:26 AM
I don't see a reason why SMA should apply scale and/or offset for temperature measurements. The value reported by the inverter is of type int32 with a resolution of 0.01 degree.
Temperature derating is activated when the inverter can't dissipate the heat due to ambient temperature being too high. Me too, I never seen temp derating activated, having internal temperture of >60°C
for sunny boy the Operating temperature range is -25°C to +60°C
This is ambient temperature, not the internal as reported by SBFspot.
When I have the time, I'll open my inverter to see if it's possible (safe) to measure the internal temperature so we can compare both readings.
Inaccurate temperature reading might be a reason why it's not reported by Sunny Explorer, who knows. We can't ask SMA I'm afraid ;-)
Sep 12, 2016 at 8:54 PM

That a good idea. There are some stonking (=beefy!) power transistors bolted to the heatsink, and of course they'll get the hotest. I suspect the temepratature sensor (there may even be more than one) is close to them.

Looking at the PVOutput SBF group in a bit more detail, I can see some inverters that appear to report perfectly reasonable temperature swings, and others that do the same as mine - starting from a most unlikely value, even when just starting up or shutting down. It's a mystery!

On another matter - i wonder how I might go about recording other data (such as ambient/panel temps, net house consumption, etc, and uploading this to PVOutput as one of the extended (V7 > V12) values - any ideas? I will take the reading at get it onto the Raspberry pi, but it's how to best integrate with your application. I've not gone anywhere near your source code! (too scared...)
Sep 13, 2016 at 8:15 AM
Edited Sep 13, 2016 at 8:15 AM
Looking at the PVOutput SBF group in a bit more detail, I can see some inverters that appear to report perfectly reasonable temperature swings, and others that do the same as mine - starting from a most unlikely value, even when just starting up or shutting down. It's a mystery!
Keep in mind some users don't upload inverter temperature. Instead they use the ambient temperature feature of PVoutput. See this FAQ.

Yesterday I mounted a temp sensor to the heat sink (external, without opening the device) and I saw a difference of 9-10°C
This morning, inverter still off, the heat sink temp was 27.0 and the ambient 25.9
First temp reported by SBFspot was 37.3 (wasn't at home so I couldn't verify the heat sink temp, but it seems there is an average difference of 10°)

The temp sensor sits in the middle of the heat sink, pressed between two "ribs" with some flexible tube to get the best thermal contact.
Sep 15, 2016 at 7:54 AM
Example of ambient temperature instead of internal inverter temperature: Sundown
Have a look at "Weather Underground" logo at the bottom of the page
Sep 29, 2016 at 11:40 AM
Edited Sep 29, 2016 at 8:52 PM
hej laurence and sbf,

note: since i own a transformer type inverter -> the "heat production" in my inverter is higher (~2*) then yours
note: till today we could not find a field in the datastrings (for my inverter) that corresponds to:
  • fan-speed or -fan-active
from the limitted set of values available:
  • one can conclude that the values are NOT expressed in fahrenheit
  • depending on the inverter type OR firmware version there is an offset of 0°C to 25°C
    not yet known - which parameter (type, FWversion, ?? has influence
to measure the temp i do use a digital temperature tool from my wife's tool set (kitchen tool to measure temp. frozen foot <-> treated food in furnace) :-)
specific of that type of temp-tool ( it costs ~12Euro)
  • temperature range from -30°c to 275°c
  • granularity 0,1°c
  • has a hold button :-)
observations on my inverter (SB3300) in normal conditions:
  • the delta in temp between air temp below <--> air temp above the inverter is less then 10°c in low production
  • temp on the heat sink is slightly higher then air-temp flowing out of the inverter (delta, 1,2°c)
    measured by unmounting the airfilter/cover -> i can access directly the heat sink of the inverter
the fan of my inverter starts when the production is ~2600Pac Watt,
  • this is independent from ambient temperature (might be that fan is activated with high ambient temp)
  • fan speed depends on the Pac value - eg when Pac increases to 3600Watt then fan does step-step increase its speed
in my setup the ambient temp never gets above 35°c
  • so i can not judge whether there is an influence on fan by ambient temp.
kr wim