Strange behavior (bug?) of one MPPT in a Multi RS

DIY
, ,
1

Hello developers,

As per title, at some point, when the sun was on its way down, one of the Multi RS MPPTs has shown a strange behavior.
It continuously scanned for the maximum power point over and over.

Both MPPTs have strings with the same sun exposure, the same tilt and the same orientation.
Both have the same type and number of panels.
So, in other words, identical strings.

In the video below, it’s the second row/MPPT on the Solar section.
You can see how it’s continuously scanning and, even if it can settle on a much bigger MPP than the other tracker, no, it doesn’t do it and continues to scan without stopping.

First string stays at about 50W (200V x 0.25A), but the second tracker keeps scanning between 60V - 230V and you can clearly see that at about 70V the current is about 1.5A, so a possibility of about 100W production, double that of the first MPPT.

Also because the PVs on both MPPTs have the same characteristics, for sure the first tracker can also produce 100W.

Is it possible that you lowered the bottom limit to 65V, but somehow you forgot some condition somewhere and you only validate maximum power points with voltages only above the old 80V, therefore disconsidering the points with voltages below 80V ?

This behavior is also consistent with the fact that I didn’t see ever maximum power points below 80V, even if you lowered it to 65V…

Thank you!

2

I think you are approaching this as if the top row is working correctly, and the bottom row is acting strangely?
The fact that the top mppt is steady at 50w, but the bottom mppt hits 100w sometimes makes me think that your arrays are not very identical, and maybe both are doing something unexpected.

If this were my system, i would re-commission and compare to your original commissioning docs. Re-commissioning is the only way to prove that the array is not the source of the problem.
If you can’t do the commissioning tests for some reason, make a video like you have, quickly swap the inputs over so that array A is now on mppt channel B and vice versa, and make another video, hopefully in similar solar conditions. Comparing A to B is not as good as testing the Voc and Isc, because if both suffer from the same fault, they both appear good, but it is better than presuming that both are good.

You seem to know a lot about mppt, but i really don’t know so please don’t take this as mansplaining or such, i’m just trying to be clear for you and others reading.
For clarity the critical commissioning tests are;

  • physical inspection - we can prob skip that if these other tests come out fine.
  • electrical test; Voc and Isc, comparing to the theoretical and the past tests.

Voc is easy - just turn off your isolator, check the voltage on both A and B arrays and document the result.
Isc is harder - you need to short out the array and check the current. You can’t do this by looking at the mppt data because this is not Isc, its a constantly moving number around Imp. Be careful of opening a contact on a shorted array - you can start an arc.

I’m sure you have your expected Voc/Isc/Vmp/Imp numbers and expected production, but if not, post your panel specs and location (city is close enough) and i’ll build you a table.

I think you will find that both arrays probably have the same Voc, but array A doesn’t perform the same when shorted as array B - this tells us that something (like a bad MC4) is different between them, and the comparison between A and B and the theoretical values tells us if both arrays are bad, or one is, or both are fine.

As an aside;
I feel like too many DIYers, but also installers, are saying “well, it seems to work, so i’ll say its done”.
In a situation where you have an XsYp array, and Y >1 (ie any level of parallel panels), you have a huge potential for the array to “work” but not perform properly.
For example: In a 6s3p array, break the string on ONE sub array, and you will still have a working array, still have the right Voc, but only 2/3 of the power.

3 Likes
3

Yes. Physical string commissioning. And checks.

Nice explanation.

4

I’ve done all you’ve asked:

  • during the phenomena, I’ve changed between each other the strings: A->B and B->A. After max a minute the same MPPT exhibit the same behavior. MPPT2.
  • string Voc when not connected, is close to 2-3V between strings. I.e. A=234V, B=235V. So less than 0.5%.
  • when producing, because didn’t made a short circuit test, I admit, is about max 20W difference, on the whole day. For example, A=1365W, B=1380W. During this: A=204.7V, B=206.8V. You can compute the current is also awfully close: 6.66A vs. 6.67A

So, you see, as I am pretty thorough and I want peak efficiency, they behave pretty the same and all values are well withing tolerance.
All the above data, if Victron is really interested, can be taken from the VRM database.

5

What comes to my mind is that another things can be involved here.

A classic MPPT is connected on the DC battery side and must deliver on 48V some energy that comes from a much greater voltage. So it’s easy to stepdown.

The Multi RS Solar’s MPPTs must deliver that energy into a DC bus at 400V. So, for a marginal condition, where the max MPP can be at, let’s say, 70V, it must boost that to 400V, which is an almost 6 times greater.

Maybe some instability there in the conditions, or some hardware limits will make the MPPT to not settle on such low value and try a bigger V like on the MPPT1, where at the time of the video it was about 197V, so for 400V it means a boost of 2.

6

Hi Alex, can you post a picture of the two arrays please, I have an idea but would need to see a picture first.
Thanks.

7

Please say it, as I have all details in my head and up until I will reach that location will take some time. :slight_smile:

8

Are they side by side or one Infront of another.

9

A combination of the above.
Because of the space constrains, they are like this:

  • string 1: 1 panel in front of 4 panels in a row
  • string 2: 2 panels in a row in front of 3 panels in a row
  • string 1 in front of string 2
  • spacing between rows enough for as less shadow as possible
  • panels in horizontal position, so when a row is shading the other only a section/diode is bypassed/conducting
10

Thank you Alex please bear with me on the questions.

So are they ground mounted or on a structure of some kind.
If mounted on a structure what is the distance off the ground.
What angle and direction do they face.
What location are the in longitude/latitude.
The spacing could be critical here so panels size and space between them to the nearest foot too please.
Panel specifications (do you have a pdf spec sheet?)

Sorry for all the questions but it helps me build a picture of your install.

FYI
I have six arrays on three RS450/100.

11

Thank you for the patience and willing to help.
So…

  • on a structure at 6 meter above ground
  • 25 degrees tilt angle, pointing at SSW
  • 44.2N, 28.6E
  • rows at 2 meters apart, CS3W-460MS, horizontal mounted

What could all this influence the MPPT behavior?
That never ending cycling on the entire voltage range?
It’s clear that it’s a full sweep, and one full sweep is enough to gather all the info needed.
So why not settling on the real MPP and keep cycling?

12

Hi Alex.
Without a picture this would be me best guess based on what i have seen on mine and what you have said.

Two issues come to mind.

First i suspect some of your panels are casting a shadow on the others, and sure you aware the impercations of this.
The height off the ground will probably only have a small effect so probably not worth worrying about.

Secondly you say as the sun is going down so you have a constantly moving target (decreasing in power) causing the MPPT to hunt all the time, something i have watched closely on mine.

My winter panels are set at 65 degrees to maximise winter sun, this time of the year it will peak (for me) 11am to 12pm then drop off quickly…

Hope this helps to explain what you are seeing, if not sorry not much else i can suggest

1 Like
13

Thank you for your advice.

Yes, shadows are bad, but this is why I’ve installed the panels on horizontal, to invalidate only one/two of three sections at a time, as the three sections on a panel are along the vertical.

I am thinking the sun movement is not so big/quick during that 5 seconds cycle of exploring the entire IV curve.
And once it finds a maximum, shouldn’t it stick to it, at least for a while, at least for those 5 minutes between scans?
Why it keeps scanning again and again?

Anyway, as I’ve said, is in the marginal harvesting area, so maybe 50-100W up and down is not so much.

Thanks again!

1 Like
14

Your welcome Alex and good luck :+1:

15

The Multi RS Solar’s MPPTs must deliver that energy into a DC bus at 400V. So, for a marginal condition, where the max MPP can be at, let’s say, 70V, it must boost that to 400V, which is an almost 6 times greater.

I don’t think you are correct on this - the Multi RS mppt controllers are converting your ~200v into 48v. The DC bus is at battery voltage, I don’t think there is a 400v DC bus anywhere on the Multi RS. AFAIK there are only a few Victron devices that have boost or buck/boost converters - almost all are buck converters.

16

Ok - i think I see the issue.

These arrays are producing around 1.3kw during the day, at ~200v.
I’m guessing that each array is therefore a string of 5 panels at 460w?
If so, this is about 10m^2 of panel per array, so the array in the vid producing 50w is producing about 5w/m^2, so your sunlight at that time is probably about 25w/m^2
So are we talking about the very start or end of the day, or extreme cloud cover?

If we are, that hunting for mpp is normal.
At 1000w/m^2 (STC), some dust, slight alignment issues, a few degrees in temperature difference, makes almost no difference to the energy captured.
At the absolute margins of production, tiny things have large effects.
I have many different SCC in the factory, and i can plug in a 100/20 and get double the power compared to an Epever 50A a few seconds before. I can double my production when i change from dirty panels to clean panels (regardless of which controller), or triple the production late in the day between panel flat on the factory roof (~5 degrees) vs a tilted one.
So I think this will come down to one of the mppts being ever so slightly more sensitive to something, causing it to hunt. I see that all the time on my controllers that are connected to panel in the factory where the only sunlight the panels are getting is through the (very dirty!) opaque roofing sections.

Also, a common thing you see is the voltage gets pulled down, current seems to shoot up (in your video to 100w, double the other string), but just like slamming your thumb on a hose to get a sudden jet of water, it doesn’t last.

Edit: Ohhhhhhhhh - i just saw the “sun is going down” part - my whole comment is moot :frowning:

17

It’s a high frequency All-In-One topology system and it does have a 400V bus… :slight_smile:
The same as all devices in the RS range of products.
Just take a look a the Victron training videos about MPPT RS.
You can see there the 400V capacitors and high frequency transformer that galvanically separates the MPPTs and 400 high voltage DC bus from the battery and 48V.
Also you can take a look here: Potential between earth and neutral - #10 by alexpescaru

18

Thanks for that Alex - I’ve not done an RS system yet, and i was presuming that because the DC battery side was 48, that the MPPT was as well.

More videos to watch!

19

Also thank you for the other thoughts about the marginal energy harvesting of the MPPTs.
It gives me more ideas to digest. :+1: