This once again supports my point that Victron prioritises Global MPP rather than chasing peak values and cooking the cells.
@Diessel
With “Partial shading detection”, Victron is exactly doing this… “chasing peak values”, by exploring the entire IV curve at 5 minutes.
Without it, it just tries to keep onto one point found at an earlier exploration at large time intervals.
@doro
Thanks for the graphs.
But here it’s another thing…
Chinese derived MPPTs software, tends to lock and look for a better MPP on the “left” side of the found MPP, decreasing the voltage by increasing the current. Remember, in the earlier days, the software had a glitch where one could find the MPPT stuck at voltages as low as 60V, up until you reset them or disconnect for a small period of time the strings.
Victron is doing the opposite, trying to look on the “right” side of the MPP, reducing the current and look for better MPP.
Both of them solved the problem in the same way, by scanning the IV curve at selected intervals.
But like you said, Deye found a way to “handle this problem better”.
I know that. What I can’t work out is why the Victron fails to pick up the lower peak when scanning in both modes. Meanwhile, the Deye finds the global peak at a reduced voltage (185.9 V), and maintains an operating current of around 9.0 A.
UPD: And above all, I’m really interested to find out which string has the bypass diodes engaged, and which doesn’t. That will answer all the questions. Once we’ve got the thermal images, we’ll see who’s cocked up.
@doro
Please try to take thermal images of the diode box area on both shaded panels.
This will help clarify things considerably.
These same two different approaches to finding the point also occurred to me.
That was the case before, and it still is. The point is, if there are shaded cells on the panel, with Imp at the upper limit and Vmp dropping, and the bypass diode hasn’t been activated and there are hotspots. And that’s definitely not good.
I’ve had another careful look at the latest graph, and I don’t understand where the downward trend is coming from when the upper string is relatively stable. I take it that looking at a wider sample won’t really explain anything either, but will only raise even more questions.
I look forward to seeing your thermal imaging pictures.
Thanks, that’s very informative.
To be honest, I’m rather baffled by the way the Victron algorithm works. If both strings are operating under identical solar irradiation conditions and the 450/100 is sticking to the upper voltage limit, I’d definitely like to hear what the developers have to say about this.
Please correct me if I’m interpreting this graph incorrectly, but you manually restarted the tracker here, didn’t you? After that, it picked up the string and locked onto the correct MPP?
Yes. I turned off the circuit breaker and switched it back on manually.
Yes. I turned off the circuit breaker and switched it back on manually.
I’m baffled.
It’s quite clear that the bypass diodes have done their job in both strings. What I don’t understand is why Victron is choosing the less efficient operating point.
It’s a mystery to me.
Could the Victron representatives present here please respond?
I am ready to carry out any necessary additional measurements.
Do you have any DC clamp meter and together with a voltmeter to verify the parameters reported by the MPPTs?
Maybe there are some discrepancies or you can clearly eliminate any doubts this way…
LE:
For example my Multi RS is reporting with about 0.5A more on the trackers than it’s in the real life.
And for a 200V string, it means 100W more…
Hi, Eduart, your post was shared with the team. It will be analised and we/they will get back to you.
Thank you!
By the way, yes, I’m now noticing the tracker sticking to the upper voltage.
I hadn’t noticed this before. And these are perfect strings with no shading.
UPD: After 10 min
The controller likes everything, but I don’t anymore
.
UPD2: After 25 min
This is already really sad.
UPD3:
I’ve checked several installations, and this isn’t an isolated case.
Similar behaviour can also be observed in installations with 250/70 and 250/100.
I’ll summarise my observations.
An hour later, with sunset approaching, this is already looking absurd.
I’ll explain why:
01-01 east-facing string, currently under diffuse light. The MPPT is doing its job here and isn’t chasing high voltage.
01-02 south-facing string, currently under diffuse light + some partial angled rays from the setting sun. Based on the PV module layout, it’s in a more favourable position and, in principle, cannot be performing worse than the east-facing tracker. Yet we’re seeing exactly what this topic is about. It’s simply gone to sleep, while the trend keeps creeping downwards. Its voltage is the best of all, no argument there, but there’s hardly any current.
02-03 west-facing string, currently under diffuse light + the last direct rays of the setting sun. Its module layout puts it in a less favourable position, and its output is higher than 02-04.
02-04 west-facing string, currently under diffuse light + the last direct rays of the setting sun, and once again the controller seems to favour the highest voltage point.
And this confirms that every time the controller scans the curve, it selects a higher voltage and a lower current, resulting in a loss of efficiency.
Привіт, Едуарте, твою публікацію поділилися з командою. Її буде проаналізовано, і ми/вони зв’яжемося з тобою.
Thank you very much! I am ready to provide any data I have that might help isolate and resolve the issue. I am also willing to grant access to my monitoring system and test alpha versions of new firmware on my own equipment. Please reach out if any further assistance is needed to address this problem; I am keen to see the device’s performance improved.
Measurement equipment is available. However, in this case, verification measurements seem unnecessary. The observed behavior manifests simultaneously across other sensors as well—for instance, those related to the battery. While some margin of error inevitably exists, it has little impact on the overall behavior.
And this confirms that every time the controller scans the curve, it selects a higher voltage and a lower current, resulting in a loss of efficiency.
I’ve tried to point out a similar or at least maybe related situation here, two years ago…
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 an…
A very interesting topic. I’m looking forward to seeing how things develop and, ultimately, to the solution. I think the guys at Victron will give this careful thought and then put a lot of brainpower into finding a solution.








