Overiszing a Sunny Boy 4000TL

[JamesPa]

Does anyone have any experience of 'oversizing' a SMA SunnyBoy 4000TL (or one of the same series).  By oversizing I mean more panel peak power on the DC side than the inverter is specified for, without exceeding the max voltage or current.  I am looking at upgrading to 5.5kWp panels on this 4kW inverter.   

 

The manual doesn't say how it behaves and I can imagine three possible behaviours

 

1. Damage to the inverter

2. Clipping of the output at the max

3. Shutting down altogether

 

(1) seems most unlikely; I had hoped it would be (2), but an installer told me today that he has experienced another brand of inverter which does (3).  SMA so far have not answered the question although the sunny designer tool seems content.

[SteamyTea]

4 minutes ago, JamesPa said:

SMA so far have not answered the question although the sunny designer tool seems content.

Trust the tool, SMA are not brilliant at customer service.

I once stuck 3.5 kW_p on a single string 3 kW SMA inverter (can't remember the model number, it was the very light one).

Worked a treat and apart from the fan on full chat, would take 3.8 kW in occasions.

Make sure the inverter can draw in cool air, external air preferred. They can chuck out a lot of thermal energy when clipping.

Edited February 6, 2023 by SteamyTea

[Dillsue]

Bear in mind that if the panels are facing different directions the peak install will never be peak generation. We have a 3.6kw inverter with a max specified DC input of 5.7kw but its got 8.5kw of panels hooked up that are spread over 180 degrees.

[JamesPa]

1 hour ago, Dillsue said:

Bear in mind that if the panels are facing different directions the peak install will never be peak generation. We have a 3.6kw inverter with a max specified DC input of 5.7kw but its got 8.5kw of panels hooked up that are spread over 180 degrees.

Thanks and noted, however the panels are such that they will exceed the max power specified by the inverter during the middle of the day in the height of the summer.  If the behaviour is (2) - clipping, then the lost generation is negligible (Ive run it through PVGIS) but if its (3) - shut down, then its material.

 

Installer reluctant because he says he has experienced another brand that exhibits behaviour (3) although everything I can find about SMA would suggest this is unlikely.

[billt]

I've done this. At a previous house I had an SB5000 with 5.6kW of panels and an SB4000 with 4.4kW of panels, all facing the same direction nearly south. There was no issue with the inverters, they just restricted the output.

 

Here I've got an SB5.0 with 6kW of panels, but they face in opposite directions so they'll never produce 5kW.

[SteamyTea]

One should be careful about unbalancing each string (DC input).

Depending on roof angles and sun angles, you may get a higher than expected incidental input.

 

I would think that if the system was too powerful for the inverter, eventually it will shut down for safety reasons.

So while it may not be optimal, it should be safe.

[PhilT]

There is also cloud lensing to be aware of, which can briefly increase output to as much as 50% above the pure direct sun/blue sky output

https://nmsolargroup.com/effects-clouds-solar-panel/#:~:text=They will see complete direct,than its direct-sun output.

 

[billt]

Here's an example of what actually happens to the output. It limits at 5kW in the middle of the day.

 

The SMA Sunny Design web site lets you try various combinations of inverter and panel and will tell you if there is a significant mismatch.

 

https://www.sunnydesignweb.com/sdweb/#/

 

[JamesPa]

7 minutes ago, billt said:

Here's an example of what actually happens to the output. It limits at 5kW in the middle of the day.

 

The SMA Sunny Design web site lets you try various combinations of inverter and panel and will tell you if there is a significant mismatch.

 

https://www.sunnydesignweb.com/sdweb/#/

 

Thanks, is this an actual measurement or a prediction from SunnyDesign?

 

If the former can you tell me precisely the inverter model please?  Installer very sceptical although all the evidence seems to support the clipping you show.

[billt]

They are actual measurements. It's the reported output from the inverter sent to PVoutput.org at about 5 minute intervals.

 

That was an SB5000TL-20, no longer current, it was installed in 2013.

 

Here's another example. This has the SB5000TL-20, an SB4000TL-20 and an SB3800. The SB3800 is a transformer inverter and was installed 2 years earlier. The nominal array is 3870 watts, so only clips very occasionally. This is measurements of the AC output and much more frequently.

 

You can also see that the SB5000 has brief overshoots sometimes.

 

 

I've found an instance of the current inverters limiting. They are an SB5.0AV and an SB4.0AV. It doesn't happen very often because of the panel orientation, and lack of oversizing on the SB4.0.

 

 

I take it that your installer hasn't been doing this for long. It used to be very common to undersize inverters for the array.

Edited February 6, 2023 by billt
Added image.

[SteamyTea]

39 minutes ago, billt said:

I take it that your installer hasn't been doing this for long. It used to be very common to undersize inverters for the array.

Generally about 10% under.

But then it was easier to add an extra 10% to the capacity.

[JamesPa]

3 hours ago, billt said:

You can also see that the SB5000 has brief overshoots sometimes.

 

 

 

Can you clarify what's going on here with the oscillating output.  Is this clouds passing or the inverter cutting in and out?

[Christiano]

To add to what’s been said above, we have 11kW of panels with SB5000TL (23 x 260W on 2 strings) and SB3600TL (17 x 250W on 2 strings).

 

We can peak at up to about 9kW between the two and when it does it’s just capped off, no oscillating, no nasty noises. Been in since 2015, monitored by emonCMS since 2016.

[billt]

9 hours ago, JamesPa said:

Can you clarify what's going on here with the oscillating output.  Is this clouds passing or the inverter cutting in and out?

 

Yes, that's cloud, i.e. a normal days output. The first image was very unusual, a day with unbroken sunshine.

[JamesPa]

Thanks to everyone who has commented on this.  Amazingly SMA have replied to my email and have also confirmed the behaviour is clipping not shutting down.  Hopefully the combo of practical experience from this forum and the written confirmation from SMA will be enough to convince an installer or two to quote

[SomeGoodBits]

I guess JamesPa has moved on, and i am sure he has seen the document from SMA that i am uploading.  But i had to upload it when i saw this topic, it guided my choices.

7 reasons to undersize.pdf