question

The daily voltage graph. What causes the two step downs and reduction of power?

15 hours of bulk charge on a BLUE summer day without reaching the cut off voltage of the battery.

It almost looks as if the charger tried to increase the voltage every now and then but then measures a drop in current.

Before the upgrade to 1.64, the voltage would go up in the morning and stay up during the day yielding up to 198 W. Now with the reduction to about a third in voltage, the yield seems to max out at 67 W at about 5A (on the battery side).

To me this indeed looks like a faulty power tracking algorithm in the product that finds a local optimum and never leaves it.

The big question: what could be causing this?

please be so kind and post some pictures how it looked before and how it looks after the update.

please consider also that simply a full battery can cause lower yield.

what kind of battery do you use?

how are and how were your voltage settings before?

Hi @Niklas Schauberg

Thanks for your responses. The drop in voltage actually puzzles me because, me too, I would expect it to stay around the same value once there is light shining on the panels with only the current varying during the day.

Can the MPPT charger even control the input voltage to such a degree?

I am using an AGM battery with the recommended settings for the absorption and float voltages (14.8 and 13.7 V respectively).

I have watched the system operate over the years but I never did take screenshots. I remember not having noticed a drop in voltage as I observe it now.

On a day such as today, with a clear blue sky and the roof of the van tilted towards the sun, I would have expected the battery current to go up to close to 15A with the battery voltage I observed at noon (which would be about 200W). Yet production seems to have been capped at (strangely) precisely a third of the previous capacity and the active cooling of the controller with a fan would not have been required (I added the fan three years ago because I noticed that the device got hot at full production rate). The settings say that max battery current is 15A - not 5A as observed.

My impression (based on past observations) is that the solar voltage is the limiting factor (i.e. once production starts, it drops twice by a third - for unknown reasons) when the solar current is within the usual range.

Any idea what could cause this?

Can it be the charger or is this some weird coincidence where something else broke at the time of the update?

Daily curve BEFORE the issue developped. Note that the van was parked on the west side of a building and remained in the shade until noon.

Another daily curve BEFORE the issue developped. Note that absorption voltage was reached in the morning on this day.

In both cases, the voltage was around 40V as long as there was light.

Yesterday's curve showing the pattern of reduction in voltage from ~40 to ~30 to ~20 V.

Factory reset in the afternoon didn't make a difference.

Any hint regarding how ro troubleshoot this and bring back production at full capacity will be greatly appreciated.

@minthral

I think one of your panels may be faulty. Try them individually as a test.

thanks for the details, I see what your concern is. I am still not decided if it might be normal and caused by weather conditions. some arguments supporting that is, that on the 3rd picture the voltage is quite constant while delivering power and the current follows the expected radiation curve.

on the other hand the voltage drop between idle and load is quite heavy compared to pic 1 and 2 which is an argument against this.

so you did nothing else to the system than upgrading the firmware right? no cabling or anything physical?

if yes then please send me an aditional screenshot of your battery voltages for the same moments you have posted above and please a screenshot of your charging settings.

just want to make sure that the battery is not the cause of the behaviour.

you your know if there were any loads to the battery while the moments of your screenshots? especially the fridge. was ist allways on, was the compressor permanently running etc.

after interpretating this, I would propose to downgrade the firmware again to observe if old behaviour returns.

@Niklas Schauberg thank you again for your response.

Regarding the weather, this is what my array looked like yesterday at 12:11 hours:

As you can see, this is as close to a perfect exposure to the sun at this time of the year at this latitude (almost 90 degrees). So no, it is not the weather conditions. The voltage curves of yesterday were as:

The drop to zero was when I disconnected the battery to see whether this has an effect on the solar voltage - it did not.

When current flows into the battery, the battery voltage goes up. Yes, the fridge was on and that can be seen by the periodic drop in battery voltage.

The fridge has been in use for years, so no, it's neither the fridge nor the cellphone charging during the day.

Today, we have a clouded sky. Interesting to see that the voltage fluctuates between 25 and 40 V. I doubt a broken panel or a faulty connection would lead to this behaviour when there is no load (3 to 5 W equals an idle state).

So to me, this looks like MPPT doing something strange. At least this answers the question, whether the charger can adjust the voltage to the extent observed: yes it obviously can and does.

@JohnC Yes, the firmware upgrade was the only change. Everything else remained the same. No, changing the cabling of the panel from serial to parallel is not an option (especially not as this array has been working perfectly for years and especially not because that would lead to a higher current, a higher loss and thus less yield).

The charging settings. Doing a factory reset and selecting the battery type AGM results in the same settings.

thank you this is insightful information! and I agree in you setup with no other things on the roof partial shadowing might not be the issue at all.

so that one panel died or has a cold solder joint might be an issue combined with thermal expansion due to the fact it is symmetrical around noon. coincidences happen, but let's check software side first.

do you know which firmware version you had prior to the update?

do you know how to downgrade the firmware?

in fact 1.64 included changes to you specific model:

v6.06 (Released officially)
2024-06-06

Changes

Fix firmware update notification is not cleared after updating the firmware on remote connections.
Correct the issue where the selection dialog occasionally displays large gaps between options.
MPPT Solar Chargers:

Add firmware v1.64:
Internal improvements for 100/30-12V model.
Improve robustness of SmartSolar 75/10 and 75/15 during shutdown and startup.
Improve robustness of SmartSolar 100/20 startup under high solar potential.
Make firmware v3.15 update mandatory for SmartSolar 250/100 rev2

Hi @Niklas Schauberg

No, I don't know what firmware version I had other than it had been the most recent before 1.64 was released.

There sometimes are strange coincidences where the hardware happens to break when a firmware upgrade is performed. Some seem to always assume that something else is the cause. I see threads on this forum where an earlier upgrade bricked devices and some would still insist that the firmware upgrade did not brick those devices but rather something else.

It could also be that - in my case - one or both panels have characteristics that lets the (new) MPPT think the optimal setting were somewhere were it is not. So that would be the combination of charger and pannel array.

A faulty panel doesn't kind-of make sense to me because the drop in voltage is twice a third of the usual value - if one panel is faulty, shouldn't that be a fifty percent drop?

Also the number of individual cells on each panel - 40 in four by ten - adds up to the voltage (20 to 24 V per panel) but does not have numbers that would explain the observed one third. But that may just me not seeing something that doesn't exist anyway.

With my limited knowledge as a layman and taking into account the numbers and behaviours I see, I would not be surprised at all if the MPPT were the culprit.

Today is a cloudy day so I guess not the right day to further investigate. Also I don't feel confident disconnecting the array in sunlight to measure the individual panels and I habe nothing to cover the panels.

I wonder whether a faulty panel could also be diagnosed by looking at what happens when one or the other panel is covered while production is running. If a panel is faulty, my expectation would be that covering it has no effect on production. Or would it?

@Niklas Schauberg BTW: one week ago I was also thinking about thermal expansion. There was almost nil wind and the roof felt really hot to the touch. On the way back home, the roof got its cooling on the freeway for about an hour. The battery charger was disabled while driving. I watched the development of the solar voltage and yes, it went up by a third (which it also does under load at that time of the day). When I stopped and turned the battery charger on again, there was no change to the solar voltage beside a small decrease.

So the temperature did not seem to have an impact.

Regarding the firmware revision, I wouldn't even know which revision of 1.64 I have from this list:

Clearly, no older firmware release is provided for download.

There seem to be many customers looking for older revisions of the firmware because applying a firmware upgrade either led to specific issues or bricked the device altogether.

My impression is that Victron should take a look into this. Or maybe the percentage of upgrade issues is so low that it's not worth it.

In the future, they should support loading two firmware revisions onto the devices and then provide the users with a selection which one to use. That would help a lot troubleshooting issues with a suspected connection to firmware revisions.

Same behaviour on a cool day under a cloudy sky where people wear hoodies instead of swimming trunks and where the panels are cool to the touch.

Voltage drops twice by a third.

Note the regular drop in voltage that occurs precisely every ten minutes. It would be very strange if a faulty connector or panel caused this, wouldn't it? This looks like a periodical check whether the optimum could be somewhere else than at the current location.

Also, the maximum power harvest stalls at precisely a third of the maximum capacity.

The looks one to one the same as on a blue day. If one or both panels where faulty, shouldn't the power harvest be lower on a cloudy day? How can it be equal to the harvest on a blue day if the connectors or the panels were faulty?

This observation points to the algorithms of the charge controller, doesn't it?

The roof is no longer tilted towards the sun and the sky is covered. No shadows can be seen on the ground but the harvest maxes out at precisely the same level as on yesterday's blue day with the same reduction in voltage being observed.

The revision in my case is A053.

"Test flight program" - I seem to be mistaking myself with a paying customer who bought to product to do a job (and not help develop/ diagnose etc. it).

Hi @JohnC

Thank you for your response - very much appreciated.

The 'paying customer' part refers to the charger and its firmware only. There I am a bit sensitive due to past borderline experience with devices and firmware (NOT Victron and not solar energy) - so that prospect triggered me. My apologies for how that came over.

If I'd get redirected to the installer, that would be me (as this was/ is my own project from parts). I enjoyed doing it and I still am (interesting topic, puzzle to solve, a lot to learn).

I put my van into storage mode yesterday as I will not be using it for some weeks or maybe months, unfortunately due to surgery. Before I did that, I did some quick tests. I covered one panel. The voltage stayed the same at around 17 V but the current dropped. When I removed the cover, the current went up again. Then I covered the other panel and saw the same effect.

Interestingly, production went up on my way back yesterday - voltage was at around 30 V and production peaked at 120 W in direct sunlight. Then the voltage dropped again to around 17. Vibrations?

I now feel that this is an instance where something else happens to break at the same time as the firmware upgrade. My guess would be the panels or the connections or both.

Not sure whether I follow you regarding what you see on the pictures of the panels. I don't remember seeing anything out of the ordinary when looking at them.

My plan is to disconnect all, check the panels individually, check the connectors and connections and the re-assemble.

In the best case, I will be doing this in about four weeks. In the worst case, this is not going to happen before 2025.

I'll be fostering my knowledge in between.

@minthral

What's this thing in your pic?..

I called it a 'glob', but might be just a blow-in. Not a good look in a pic designed to show clear panels.

Also I wondered about bypass diodes on your panels, whether they might be contributing to this steppimg V?? If none, look for a dead short.

Good luck with your surgery. Speedy recovery..

Hi @JohnC

Thats a leaf from a tree. On another picture, its between the two panels and on yet another one it's not there at all.

The picture was supposed to show the orientation and give an indication of the light conditions.

The tree leaf is likely not the culprit.

Thank you for your wishes.

Hi @klim8skeptic

That is plant material from the trees.

I got myself a ladder the other day and took a closer look at the panels. There is no indication of any damage (which doesn't mean there is none).

From what I currently understand, the connectors are like the culprits.