help please reduced output on hot days

[billybee]

first of all hello to everyone.

i need some advice on reduced inverter output on really sunny days.

ive had my system for about 7 weeks, its a 3.0kw 12 a-sun 250w panels the inverter is a power one aurora 3.0 and the system is on 2 strings. my probelm is on mixed cloudy and sunny days the sytems peak out put can hit 3300w (great) and in the day the systems produced 18kws. now on hot days like the end of may the system only peaks at 2400w and produces around 18kws as well. ive done a lot of reading on it and it seems to be the inverter and some safty feature when it his 8 amps inputfrom the panels. is this normal ive watching a few systems and most seem to be doing this. some deal with it better than others. any advice on why this happens would be good.

thanks all

 

[SolarCity]

Hot temperatures will mean that the panels heat up - this actually reduces the efficiency of the panel and this will result in a reduced yield. You will find that you're more likely to reach peak power on bright yet cooler days.

 

[Gary K]

I think you will find its the panels that are affected by the heat, rather than the inverter, on partly cloudy days, the panels cool in the shade and then briefly produce strongly before they warm up again. On the constantly sunny days the panels are constantly hot and the output is reduced as a result. The panels are rated to produce the stated output at 25'c but up on the roof they may be at 60'c or more, for every degree above 25 you may be loosing .5% of the output.

 

[billybee]

i spoke to a-sun and they say its not that. they say the panels are good upto 80 and at .5 for every degree that means the panels must be hitting 80+ and when this happens you can see the change from as early as 10.30 am (i'm a nerd compaired daily outputs) no on a normal mixed day by about 11.30am the panels hit about 1800w by 12.30 3000w which is the way it went yesterday but i checked yesterday and neather of the 2 lines i have got to 8 amps. today started simler to yesterday and by 12.10 id it 2700w then by 1.15pm sun shinning lovely it was back down to 2450w it stayed at this (give or take a few cloudy spells till 4pm. but i noticed both lines stayed on 8 amps consantley. now output is similar (bit more today) but when you compair the two days the main diffrence is between 12 and 4 where it seems to lose upto 700watts and hour of the top of the graph.

the reason i'm asking is i'm compairing to some very simlar local systems (i know you can never do exact comparisions) and i'm getting about 10% less perfromance than them and its hot days that are affecting me the most.

 

[Gary K]

I can not find anything on a-sun panels, are you sure that is the manufacturer? Is this system on any site, EG PV output

 

[billybee]

yes
sunshine please 3.000kW

panels are ASUN250-MB_EN

thanks the reason i blame the inverter is i'm watching a sunnyboy inverter and its doing better on hot days. i have a solotion underway but not sure if it will work.

sorry sytem isnt live feed another problem with the inverter lol

 

[Gary K]

yes
sunshine please 3.000kW

panels are ASUN250-MB_EN

thanks the reason i blame the inverter is i'm watching a sunnyboy inverter and its doing better on hot days. i have a solotion underway but not sure if it will work.

sorry sytem isnt live feed another problem with the inverter lol

What system are you comparing with - if it as being run with say Sanyo panels, that may well cause the 10% differance

 

[Earthstore]

Almost definitively the panels heating up, some panels work better when hot than others, you need to remember the cells are black, and behind glass, so the temperature will increase when in constant sunlight.

An easy way to test is when performance is down, spray the panels with a hosepipe to cool them down, and watch the inverter at the same time.

But if it cracks the panels do not blame me...

 

[billybee]

im compairing with a system 10% bigger than mine that produces 20% more on average its 10kms away ive got all the info of him and we are the same orintation 2 degrees diffrence in roof pitch he has 14 235w ET panels. why would sanyo cause as much as 10% diffrence i read up on a lot of panels and went for the asun on cost and performance. also ive looked at a few systems and not many of the panels can hit 10% above there rateings and mine have peaked at above 3100 quite a few times. its manley hot days the problems happens. i read something a while ago about aurora inverters on the net with a similar probem and the solution (carnt find it again).

 

[billybee]

Almost definitively the panels heating up, some panels work better when hot than others, you need to remember the cells are black, and behind glass, so the temperature will increase when in constant sunlight.

An easy way to test is when performance is down, spray the panels with a hosepipe to cool them down, and watch the inverter at the same time.

But if it cracks the panels do not blame me...

sounds a bid dodgy lol. may be better just to put a temomiter on roof and check temps on days that it happens. or find another system with asun panels (ive tryed) but a diffrent inverter and see if sililer things happen.

 

[sambotc]

The inverter will max out at 16A so 8A per string sounds right to me?

 

[Earthstore]

There are so many things that affect panel temperature, roof type, dark slate would be hotter than say a fibre cement roof, type of mounting frame can make a huge difference, air flow around panels, a roof may have an obstacle in the way that reduces air air flow.

It is not just as simple to compare one system to another.

I am only trying to help by the way.

Added
It is the cell temperature that you will need, not just the roof temperature.

 

[billybee]

this is what i think, and something i read (but carnt find now) had a similer problem, it was something to do with the inverter cutting down to about 80% output, there was a solotion that aurora tech told him to do but i phoned and they wanted the installers electrition there to check things out in the service area. he came crashed the inverter waited an hour for there tech to phone. after a hour he found the way to put right what he'd done (googled on net) waited another hour for power one then had to go. he did look at my figures and agreed with the 10% below par.

now when i decided on what system i wanted i did always want a bigger inverter than the panels. i now everyone says the inverters perform best at around 96% but from say 50% to 105% theres less than a % diffrence in efficancy. so my soloution is going to be up the inverter to a 3.6 i'm hoping when i do this that when i hit 16 amps (8 each string) and the inverter cuts down i should then get about 2900w instead of 2400w. may work.

 

[cbw]

I was going with the 8 amps a string as above, am I missing something here?

Sounds GREAT to me, one I installed last week has been maxing out merrily from 08:30 till 18:00 for the last 2 days.

Graph looks great, one great big block with a flat line at 16A!

 

[billybee]

i know compairing is hard. thats the main reason i went for the one i did, i have a fiberboard roof and noticed when they installed with monts that had to be drilled though the boards that the panels are quite high of the roof, i also live on a hill so should be getting good air flow. the panels a copir with are on a terreced house same as mine chimley either end same as mine. i do know the only true way to compair is to have the house next door. with the same size system.

 

[Earthstore]

Chimney's.....
Do you get any shading at all?

 

[billybee]

yes thats mine nice clime till miday hits 8 amps 2450w's then an almost flat line till it hits where it would normally start to drop again. clouldy mixed day graph is a montine peak maybe a 2 hour period frome 1 till 3 where it can run around 3300w's. i dont no if a bigger inverter will help but i will certanly post any results.

 

[billybee]

tinny bit on the top string early morning and the top sting oftern takes about two hours to get level with the bottom string. but i spoke to the guy i'm compairing with and in his reply without even asking he gets the same on a morning. were both south at about 5 degrees west. were both terreced simler roof pitchs and have chimleys either end. the big diffrence is he runs and eversolar 4000. but a diffrence the other way is his panels 3.290kw have only ever gone above this 3 times and only by 2% mine and thats in 192 days. mine have gone above 100% panel output 10 times in 50 days.

 

[Gavin A]

The inverter will max out at 16A so 8A per string sounds right to me?

that's AC output and not related to the max DC input which is 10A per MPPT, so it's definitely not that.

If run in 2 strings the string voltage at presumably 6 panels per string will be right at the bottom of the voltage range for the inverter, which will account for a few percent of the performance loss. Run it in single string and the voltage would be around optimum.

 

[Gary K]

im compairing with a system 10% bigger than mine that produces 20% more on average its 10kms away ive got all the info of him and we are the same orintation 2 degrees diffrence in roof pitch he has 14 235w ET panels. why would sanyo cause as much as 10% diffrence i read up on a lot of panels and went for the asun on cost and performance. also ive looked at a few systems and not many of the panels can hit 10% above there rateings and mine have peaked at above 3100 quite a few times. its manley hot days the problems happens. i read something a while ago about aurora inverters on the net with a similar probem and the solution (carnt find it again).

Your panels loose .53% for every degree above 25'c compared with a 235w ET panel which looses .43% so thats 1% differance for every 10 degrees above 25'c so if the panels are at 75'c the ET panels are 5% more efficient if all other factors are the same.

Any one confirm this is right.

 

[Earthstore]

Go to the top of the class..
Your maths is spot on.

In the real world, the difference may even be higher.

 

[Gavin A]

If those figures for the percentages are correct, then the rest is right, although I doubt the panels will get to 75, more like 55-60 max most of the time in full sun, so around a 3-4% loss.

add that loss to the reduced inverter efficiency though and that's probably 6-8% reduction accounted for. Add in the -2% power tolerance, and you've pretty much got the reduction mentioned.

 

[moggy1968]

a sun are talking rubbish, all panels produce a curve type graph of temperature coefficient. As the panels get hotter, they lose volatge, thats physics, and ye canee change the laws a physics!
Temperature coefficient graphs are one of the things worth looking at when comparing panels and is one of the many differences between cheap imports and better quality panels.
A 3000 inverter is ideal for your system, a 3.6 will be inefficient and will become more so as the panels age and their efficiencies drop further.
As above, running 2 strings will reduce efficiency, all things being equal, because of the lower voltage, but given your slight shading may be the best for your roof, it depends how they have been configured.

 

[Gavin A]

Temperature coefficient graphs are one of the things worth looking at when comparing panels and is one of the many differences between cheap imports and better quality panels.

not necessarily - the cheapest panels we offer have one of the better temperature coefficients at 0.41% (obviously not in Panasonics range, but beats a lot of higher price panels)

A 3000 inverter is ideal for your system, a 3.6 will be inefficient and will become more so as the panels age and their efficiencies drop further.
As above, running 2 strings will reduce efficiency, all things being equal, because of the lower voltage, but given your slight shading may be the best for your roof, it depends how they have been configured.

ah yes, shading. Hard to say which is going to have most impact, but the lower voltage will certainly be causing the impact in peak sunlight levels.

 

[billybee]

can i ask where you got those figures just out of intrest. as i phoned asun and they recon its not that. also why would it be so presise the loose on my system. i could understand at 25'c geting 110% 3300w maybe at 30'c losing a bit and more as the temp goes up but it doesnt happen like that when its a mixed day and hiting over 3000w the only time it drops is with cloud (and ive checked this next comment) on a hot day like today when we have had some cloud cooling slightly is drops a little but still only goes back to 2450w. i asked asun is there a safty feature that says right panels are to hot run below max for safety and they say no. ive looked at the data sheet and it says above 45'c i can see the .53% but dont know where the 25'c is coming from. thanks

 

[Gavin A]

the panels are rated at STC (standard test conditions), which is 1000W/m2 of light at 25 degrees panel temperature.

the temperature coefficients are therefore related to the 25 degree point.

this is standard across all panels, manufacturers etc. 0.53% per degree though is the highest I've seen, so your panels will be impacted significantly more than others, as discussed.

eta - this isn't a safety feature, it's simply how the panels work.

 

[Gary K]

All manufacters rate their panels at 25 degrees its makes for easier comparison

 

[moggy1968]

Which panels are those Gavin?
It's a generalisation for sure, but certainly a factor, as with all things, some branded stuff is tosh while there is some good budget stuff about, I'm a big fan of Yingli for example

 

[billybee]

so can i ask what this NOTC 45°C +/-2% means. i do understand what you say about temps and take it on board but why would it cut down the out put on 6 days to neally exactly the same amount each day. wouldnt i expect to see some fluxuation in the peaks as 5'c change in temp would mean aprox 150w's change in peak output. there doesnt seem any middle ground its all or nothing so to speak. i'm getting the inverter changed next week and going off what your saying it shouldnt or wont make any diffrence. i'm a bit old fashoned and strongly believe that if you run something at 100% you shorten its life. so yes ive done the figures and i know i'll get a slight performance lose with the bigger inverter (could make things worse lol) maybe as much as 2% but it will confirm that its a panel problem if outputs stay the same. and if its down to amps i should see a higher peak on hot days.

i wish i could find some asun panels to compiar with on pvoutput.

 

[billybee]

i have the aurora portable device (not very good) it has to readings Vin1 and Vin2 which im guessing are my two strings and the on thing i can 100% conferm of that is that sometines when i'm hiting 3000w the amps on both are 8 and on hot days at 2450w the amps are 8 never goes above 8, hot days its constant cloudy mixed days it moves about.

 

[Fronius]

Your panels loose .53% for every degree above 25'c compared with a 235w ET panel which looses .43% so thats 1% differance for every 10 degrees above 25'c so if the panels are at 75'c the ET panels are 5% more efficient if all other factors are the same.

Any one confirm this is right.

I doubt that 75C has been hit on the average panel so far this year. Asphalt can melt at 60C. I have a temp sensor on the underside of my array ( 33deg S-facing ) and have not been past 47C.

 

[cbw]

Front faces were in excess of that (47) on testing this week

 

[billybee]

well going of 47'c thats a 2% diffrence in output. now going of pvoutput im seeing anywhere from 7.5% to 14% diffrence in performance. also no taking into account other factors to get an 8% diffrence due to temps if 10'c is 15 then we be looking at over 100'c to do that

 

[Gary K]

I doubt that 75C has been hit on the average panel so far this year. Asphalt can melt at 60C. I have a temp sensor on the underside of my array ( 33deg S-facing ) and have not been past 47C.

Fair enough. I just pulled some figures to work with. I have to get a 3 piece ladder out to get to my panels but will get the laser thermometer out at some point.

 

[Gavin A]

so can i ask what this NOTC 45°C +/-2% means. i do understand what you say about temps and take it on board but why would it cut down the out put on 6 days to neally exactly the same amount each day. wouldnt i expect to see some fluxuation in the peaks as 5'c change in temp would mean aprox 150w's change in peak output. there doesnt seem any middle ground its all or nothing so to speak. i'm getting the inverter changed next week and going off what your saying it shouldnt or wont make any diffrence. i'm a bit old fashoned and strongly believe that if you run something at 100% you shorten its life. so yes ive done the figures and i know i'll get a slight performance lose with the bigger inverter (could make things worse lol) maybe as much as 2% but it will confirm that its a panel problem if outputs stay the same. and if its down to amps i should see a higher peak on hot days.

i wish i could find some asun panels to compiar with on pvoutput.

you're rarely running it at 100% though.

its maximum rated output is 3300W, its nominal output is 3000W, and you're seeing it running at 240oW in long periods of full sun. If you get a bigger inverter the annual output will reduce for no benefit.

 

[billybee]

thanks for the help and i will report back after the inverter change. but taking in what you all have said that is possibly a couple of % on temps maybe a little on having 2 strings rather than 1, compairing 2 similer systems in the same area facing the same way with the smae pitch. maybe his 10% bigger system could out perform me by 13 to 14% (not factoring in that my panels do have out put peaks the same as his bigger system) but in 50 days the avarage diffrence is 20% today i got 17340w he got 21171 and i think thats around 21% diffrence.

 

[billybee]

you're rarely running it at 100% though.

its maximum rated output is 3300W, its nominal output is 3000W, and you're seeing it running at 240oW in long periods of full sun. If you get a bigger inverter the annual output will reduce for no benefit.

am i reading the technical data wrong the way i understand it is above 35% of rated Output Powerthe diffrence is 95.2 to 96.8 on perfromance anything below 35% i could pick a bit up as 3.0 starts inverting at 160 and the 3.6 at 120 and lets be hounest if thats all i'm getting were talking a few watts. if the bigger inverter increases peak out but on hot days (which im beginning to doubt now) i could pick up 400watts an hour. losing another 1% would be 180w on my best day and i could get some back of the slightly earlyer start up and shut down.

 

[SibertSolar]

Sounds to me like you have modules with shoddy junction boxes/bypass diodes fitted, either that or the cells in use are particularly low grade. Could it be possible that one or more bypass diodes are kicking into reverse bias because of individual sub-string cells getting too hot maybe? I've never come across a-sun modules and don't know what/whose cells they use, does anyone have any links/spec sheets?

NOCT = Normal Operating Cell Temperature and 45degreeC is within the typical range of most poly cells.

 

[billybee]

http://www.a-sun.co.uk/download/ASUN250-MB_EN.pdf

 

[SibertSolar]

Ok, thanks. Looks decent enough tbh. Well, if it's not something related to a defect diode then perhaps it is down to a combination of all the temperature related factors already outlined by others above. Would be interesting to see a thermal image of the array when the apparent losses occur!

 

[billybee]

so if its + or - 2% at 45'c should i expect to lose 10% thata 20x 0.53% ish over that temp range.

 

[billybee]

i carnt really call them losses just dont get above 2450w on hot days and because its already arrange to change the inverter, ive got my fingers crossed that its the panels kicking to much amp out for the inverter and by uping the inverter that it also ups the reduced peak output by the same %. i will post when i get some results either way. if nothing changes then its back to asun for an expernation of them on why ther claiming the temps we've had shouldnt effect the peak out put. as they told me there tested and used in italy with no reported power drops and its hotter there than here.

 

[sambotc]

They obviously know something that everybody else doesn't then, and bearing in mind they have temperature coefficients printed on their spec sheet, i'd say they are not telling the truth.

 

[FB.]

Today was a great day to see the temperature phenomenon.
It was about 20'C, no cooling benefit of any breeze and a mixture of sunny spells and cloud.

There were cloudy spells with generation being just a trickle, and when the sun came out power soared to 3.5kW. Then gradually the power output declined (as the panels got hotter) and settled around 3.0kW. Then a cloud would block the sun again, power would drop to 700W again, the panels cooled, the sun came out again and the power jumped back up to 3.5kW.

So I reckon that I was seeing a 15% loss of power output due to temperature.

The power-derating temperature coefficient of my panels is 0.43% per degree Celsius. So a rise of 35 degrees in panel temperature would explain the 15% drop in power output.

With ambient temperature around 20'C, the panels would appear to have been reaching 55'C after being in strong direct sun for many minutes.

 

[billybee]

well i agree there, the only other thing is on cold days i offtern get the same diffrence out of the last 21 days from 4 to 10kw production ive been out performed on 17 days by upto 23% but in saying that i have beaten him 4 times so even though were 10km apart weather could play some part may be cloudy here and raining there. thanks for the help anyway

 

[billybee]

i'm not an expert like the members on here, but today here in durham has been simler, i dont no what size system you have so dont know if your 3.5 is the size off your system but mine 3.0kw and both the inverter and panel have maxed on occasion at 109% output and today i never got above 80% peak. i do fully understand the maths involved (i did ask for a-sun panels but am i little disappointed that this coefficent was not expained as a 0.09 diffrence per degree is a lot over 30 degrees) but its not enough for the overall diffrence. there has to be more than just that. and the biggest diffrence on the two systems is the size of the inverter.

sorry and he's also running a 4kw inverter with 3.3kws of panels witch i'm being advised against. lol

 

[billybee]

Today was a great day to see the temperature phenomenon.
It was about 20'C, no cooling benefit of any breeze and a mixture of sunny spells and cloud.

There were cloudy spells with generation being just a trickle, and when the sun came out power soared to 3.5kW. Then gradually the power output declined (as the panels got hotter) and settled around 3.0kW. Then a cloud would block the sun again, power would drop to 700W again, the panels cooled, the sun came out again and the power jumped back up to 3.5kW.

So I reckon that I was seeing a 15% loss of power output due to temperature.

The power-derating temperature coefficient of my panels is 0.43% per degree Celsius. So a rise of 35 degrees in panel temperature would explain the 15% drop in power output.

With ambient temperature around 20'C, the panels would appear to have been reaching 55'C after being in strong direct sun for many minutes.

the heat thing with what youve said above i can understand and agree with, cool down power up heat up power goes down and if that was happerning and mine was going up after some clould great, but mine never go back past the same output once it hits 2450 thats is best for that day (hot days) half an hour clould and it will drop but within minutes of clould clearing back to 2450 again and thats it.

 

[FB.]

well i agree there, the only other thing is on cold days i offtern get the same diffrence out of the last 21 days from 4 to 10kw production ive been out performed on 17 days by upto 23% but in saying that i have beaten him 4 times so even though were 10km apart weather could play some part may be cloudy here and raining there. thanks for the help anyway

Does your roof have the exact same slope, and the exact same facing?

My system is 3.75kWp (Kinve KV250-60M) and 3.6 inverter (Power One Aurora PVI3600 indoor model - in a cool downstairs utility room).
However, the 3.5kW output dropping (-15%) to 3.0kW as the panels got hot in strong sun is the main point I was making. The actual size of the system wasn't so important as the percentage drop-off in generation as the panels got hot.
System faces East-South-East on an unshaded 40-degree pitched roof.
Readings showing the above fluctuations and temperature degradation were around 10am-12noon.

 

[FB.]

the heat thing with what youve said above i can understand and agree with, cool down power up heat up power goes down and if that was happerning and mine was going up after some clould great, but mine never go back past the same output once it hits 2450 thats is best for that day (hot days) half an hour clould and it will drop but within minutes of clould clearing back to 2450 again and thats it.

Then it sounds as if a shadow is cast across one or more of your panels, and the whole panel goes offline until much later in the day, when the sun has gone behind the panels or is much weaker.
A whole panel can be shut down with just a few cells shaded (if they happen to be just in the right places). So if you have two panels with some shaded cells, your maximum generation could be down from 3.0kW to 2.5kW.
However, although a panel can shut down with just a few cells shaded, it does depend on which cells.
I previously posted a sketch of the typical layout of bypass diodes, which I will repost here - just one cell shaded in each third of the panel will effectively take that subsection of the panel out of production:

 

[moggy1968]

or a dodgy panel?