question

Wow, it's all coming thick and fast now.

First I installed the new VictronConnect 5.98 beta 1 today, which contained the new 1.15 for the RS, and now a new update is coming straight away?

Ok, it's already installed and the sun is even shining at times and there's still some shade on the module field. We will see...

Me too updated all three will see what happens to the single tracker tomorrow morning.

Thanks guys very much appreciated.

Dave.

Also updated both of my 200's half an hour ago and its working quite fine. Can tell more about the shadow management in a couple of hours when one string is shaded and the other is not.

Thanks for providing the beta @Guy Stewart (Victron Community Manager) !

Also updated my 450/200 now. I'm curious how it will perform...

I just updated my shaded RS450/100 and the results are incredible! Right now I harvest around 4 times more than before!
(Red Line is Point of Firmware Update)
Thank You for your great work Victron!

Many Thanks for this new and better beta! :-)

This time the timing was perfect for tests, as today is the only day of the week with full sunshine, and I could update both my 450/100 with 2 trackers each fully loaded around 10am, and I took the time to check nearer.

First of all, v1.16 is generally better than v1.14 albeit still far from perfect.

General remarks:

- Full scans are done all 10 minutes, ... sometimes. Which is better than none, but

- from what I saw, while tracker 1 does a full scan every 10 minutes almost each time, tracker 2 (on both units) doesn't do them each 10 minutes time, or just up to VOC but not down to 60'is volts, even with very low output of around 100W or less (see graph 1 below), and sometimes sweeps only up or does a very limited scan (not going down less than 200 volts, and certainly almost never the 60'ish volts, even with low power output at higher voltage (especially at low output and even if the other tracker has 1000+W of output). But sometimes none does it (graph 2)

- And 10 minutes is still too long, it's less good than every 5 minutes (like e.g. Fronius does), as far shading can substantially change within a few minutes (see graph3 below where just 2 panels get shaded one after the other within 10 minutes). With far shading, 4 or 5 minutes seems like a good interval for full scans.

- I see some permanent local tracking around the local MPP, which is good, but may I suggest to "auto-learn" the number of panels, and do more aggressive local scans for +-1 panel voltage every minute, or at least when there is a steady and significant decrease or increase of power, like in Graph 3 ?

- One of the MPPT 450/100 is more proactive in 10 minutes periodic scanning than the other, e.g. at one stage there was full sun on the pannels, but for over 20 minutes the string produced 100W, then finally it made a full scan, and produced 2000+W. But there was a 20 minutes loss of production.

Thus imho, as full scans take only less than 5 seconds (I see more like 2 seconds per tracker): Here a few suggestions of improvement:

- full scans could be performed every 4 minutes, maybe not one just after the other, but spaced.

- local scans (+- 1-2 panels / 3-6 diode voltages) could be performed every 2 minutes.

- full scans should be systematic, on both trackers. Make sure to go down to minimum voltage (especially if current tracker production is not "great") and up to Voc.

- If power on a single tracker goes steadily down (or up) more than 20% within 90 secondes while other trackers remain stable, no need to wait for interval for full scan, do it immediately.

- Imho, only once these basics are in place and well tested in field, start trying to be more smart.

- Additional suggestion: Open-source your MPPT software, as it's not yet great, it can only benefit from more eyes imho, and there are already nice open-source MPPT software and hardware on GitHub, so imho nothing to loose or risk ;-) .

Will post more if I find more ;-)

Looking forward to an even more improved beta 2 :-)

Graph 1: Tracker 1 does full scans all 10 minutes, but tracker 2 does only upwards (while having a poor power output because of partial shading):

Graph 2: No full scans during over 20 minutes:

Graph 3: 10 minutes are too loooong: here we have horizontal panels on tracker 1, and you see how power goes down and down again from 2250W to 300W because of shading (while tracker 2 remains stable, which is an added sign of shading moving, no need to wait 10 minutes for full scan, or at least do them every around 4 minutes):

My MPTT one- 1 has shading

updated to 116 early this morning, so v115 in center and v116 to the right

if that helps

Bonjour à tous,

Merci pour cette mise à jour qui semble effectivement efficace. Malgré tout j'ai une chose étrange. suite à la mise à jour en 1.16 sur un MPPT RS 450/200. J'ai eu une erreur que je n'ai jamais eu depuis 1 an d'installation... Votre avis ?

Je ne vois aucune valeur excessive sur aucun graphique.

Hi all, but most specifically @beat,

I have been reading and enjoying all of your feedback and suggestions, but I think it's time to set some expectations. At least in the short term.

The RS engineers who are now focused on shade performance are also involved in many other important improvements to the RS range which are still yet to launch (see the Limitations chapter in the Multi RS manual for clues).

While this issue has been immediately prioritised, once this narrow aspect of shade performance and associated bugs are resolved, that is likely to be the end of it for a while.

The other nice to have things, such as preparing the codebase for open source, more detailed UI, logging and analysis tools are not possible now. There is too much else for them to do.

The big advantage of how Victron operates is that a post on this community can reach the CEO and the R&D engineers immediately and directly, and they can begin pushing out firmware changes that week. However that speed means that we have to focus very carefully.

Sorry for the cold water, I am enjoying reading and thinking about the possibilities, and 'levelling up' the range, you're welcome to continue to propose them, and you never know, but I would keep the above in mind for expectations.

Ok, @Guy Stewart (Victron Community Manager) back, razor-sharp focussed on RS 450/100 shading bugs and v1.16 beta 1 feedback :-) :

Here another bug of v1.16 beta 1 that I can report now that we have a partly sunny partly cloudy weather:

When a cloud goes away and leaves place to sun, it takes most of the time a lot of time to RS 450/100 to "pick up " the extra power, e.g. going from 200W to 2500W.

(as comparison reference, the MPPT 100/20 on same roof pane picks power up almost immediately).

But when another similar cloud comes, the production reaction is (of course) immediate. But interestingly, it doesn't go down immediately to 200W, but first goes to a higher value e.g. 800W and then gradually goes down to 200W.

I could catch and graph in Victron Connect such a situation here, which shows the problem nicely:

In the begin there was sun. Then a cloud comes for around 30 seconds. You see the power going down from around 2200W to around 1000W (sharp decline of blue W line, small decline of orange V line, showing that the MPPT1 did a local MPP adjustment).

But then, the sun is back (when there is the first small increase of the orange V line, but power continues to decrease. The sun coming back didn't increase at all the power.

The sun goes even stronger, and suddenly the MPPT1 lets Volts going up to almost VoC, then another light cloud comes over and there are slight changes, aloways downwards of power.

Finally, the MPPT RS 450 v1.16b1 decides to de a periodic full scan (but it's not full, MPPT2 did go further down this time), and picks up a much better MPP:

Then again, it lets the V go up and looses track of MPP.

During all this time, there is enough room in the battery, so RS is in MPPT mode 2, to maximize MPPT without having to limit charging current.

So it looks like the local shading MPPT adjustments (letting voltage go slightly up or down, or not trying to draw more current?) outside of the periodic scans are buggy or vastly suboptimal too...

I just stumbled on this thread, was lamenting that the sun wasn't out totally today and the panels were at 3000 approx, updated to this and almost kissing 5000 with cloudy sun, whatever is in this firmware should not be changed forever lollllll

hopefully tomorrow, i get more data and compare if the tracking adjustment is actually improved on my end

Hi all,

hi @Guy Stewart (Victron Community Manager) :

Sorry! My analysis below was total rubbish.

The power drops shown on the graphs were real, but NOT a result of bad MPPT behaviour! MPPT is working great!

Here the BMS reduced the DVCC MaxChargeCurrent as the battery was close to be fully charged.

Hi @Guy Stewart (Victron Community Manager) :

FINALLY some sun around here and a very nice start of the day. Absolutely no cloud and up to now no shadow on my panels as trees will show up later.

However, some things your devs might want to look into.

450/100. Both trackers to identical 2x7 330W (JAM60S-17-330-MR) panel arrays each. T0 directed NNE. T1, SSW.

Analysis: only T1 (purple and blue), as T0 has no relevance during that time of the year.

Nice ramp-up between 8:25 and 10:16. Then, at 10:16:30 Voltage jump -> Power drops.

Interesting is the zoom. Voltage sweeps down and returns to the good MPP. Then voltage rises to the lower MPP:

Saw similiar behaviour a lot during the last days. MPPT is doing a great job limiting the Power as the Battery Management requested so.

Topic #34 Input excessive current: I can see this every 3-5 days.

Thanks!

If it's not a too sensitive question, can I ask you which kind of MPPT algorithm, you guys at Victron, are using in your products?

Particularly in these 450/100 and 450/200 versions ?

Because on smaller, low voltage models, like 100/20, 150/35, etc, those problems doesn't seem to be present and I am trying to put some things together seeing the graphs the users so kindly shared...

And as @beat said on Nov.8, I totally agree with the idea below, although Victron could have a different oppinion... :-)

Open-source your MPPT software, as it's not yet great, it can only benefit from more eyes imho, and there are already nice open-source MPPT software and hardware on GitHub, so imho nothing to loose or risk ;-) .

Thank you!

Hi

Looking forward to checking this on my Multi RS 6000 Dual Tracker
My 2 fields are 90-degree angle on the house facade and shading happens. So definitely will be a good beta tester for you @Guy Stewart (Victron Community Manager)

Good morning everyone !

Apologies as this is a thread regarding the 1.16 firmware.

My 450/100 developped the dreaded #200 error and was replaced by a new unit.

However production from this unit has been quite low even after changing the firmware from 1.15 to 1.16 beta.

I have 2 strings of 7x410w canadian solar connected to the 450/100.

However today by chance I tried downgrading and intalled the 1.14 firmware and on a cloudy morning my production jumped to over 300w ,same sky cloudy conditions!

and still climbing..