Good day

First disclaimer, I am an enduser and not technically proficient on solar at all. So if what I am trying to explain makes no sense then please forgive me.

I have two 5kw Deye single phase inverters, setup as master and slave, with that I have 24 x 455w panels, and 5 x 5kw batteries

The system has been running for 1029 days and for the first 908 days the behavior described below did not occur

Each inverter has 12 panels linked to it, and standing in front of them looking at their live performance the load is (up to date 908) equally balanced - PV/Grid/Load/Batteries. So the one does not work harder or perform better than the other. Then on 10 April this year the PV produced, suddenly started to show a difference, with the Master producing less - on ave around 700-800w, at times

I did not do anything physically to the system, or to software, or to any setting - it just happened by itself as far as I can see. And also the all 24 panels are on the same level and angle, with zero shade on them during the day, so a shadow of a tree is not the issue

I am not sure if I am looking at the correct stat, but going through the different options on the software I saw that on DC Current PV I could see the difference. I have no idea how the physical setup was done, but as explained both inverters have 12 panels, and somehow the "stats" see it as DC Current PV1 and DC Current PV2 for both inverters (so 4 sets of data for the 24 panels)

So up to 10 April all PV DC stats were "perfectly aligned" on stats for both inverters - se below what I am trying to say

Master

Slave

Master on 10 April

Since then this has always been the case, with DC PV2 on the Master producing less (red line). Looking through all the days it seems to start in line with PV1, then drop off somewhere for a couple of hours and then catches up again by producing the same. Most days its not inline with PV1 - especially if you compare to how closely PV1 and 2 follows each other on the slave

Here for example are the stats for this week

Master

Slave

So the question is twofold. What is causing it to make me loose generating capacity of around 700-800w most days on Master for a couple of hours, and secondly how do I fix it?

Edited August 8, 20251 yr by Kalaharipro

48 minutes ago, Kalaharipro said:

Good day

First disclaimer, I am an enduser and not technically proficient on solar at all. So if what I am trying to explain makes no sense then please forgive me.

I have to 5kw Deye single phase inverters, setup as master and slave, with that I have 24 x 455w panels, and 5 x 5kw batteries

The system has been running for 1029 days and for the first 908 days the behavior described below did not occur

Each inverter has 12 panels linked to it, and standing in front of them looking at their live performance the load is (up to date 908) equally balanced - PV/Grid/Load/Batteries. So the one does not work harder or perform better than the other. Then on 10 April this year the PV produced, suddenly started to show a difference, with the Master producing less - on ave around 700-800w, at times

I did not do anything physically to the system, or to software, or to any setting - it just happened by itself as far as I can see. And also the all 24 panels are on the same level and angle, with zero shade on them during the day, so a shadow of a tree is not the issue

I am not sure if I am looking at the correct stat, but going through the different options on the software I saw that on DC Current PV I could see the difference. I have no idea how the physical setup was done, but as explained both inverters have 12 panels, and somehow the "stats" see it as DC Current PV1 and DC Current PV2 for both inverters (so 4 sets of data for the 24 panels)

So up to 10 April all PV DC stats were "perfectly aligned" on stats for both inverters - se below what I am trying to say

Master

Slave

Master on 10 April

Since then this has always been the case, with DC PV2 on the Master producing less (red line). Looking through all the days it seems to start in line with PV1, then drop off somewhere for a couple of hours and then catches up again by producing the same. Most days its not inline with PV1 - especially if you compare to how closely PV1 and 2 follows each other on the slave

Here for example are the stats for this week

Master

Slave

So the question is twofold. What is causing it to make me loose generating capacity of around 700-800w most days on Master for a couple of hours, and secondly how do I fix it?

I am making an assumption that you have 6panels in series on each string. 2 strings connected to each of the MPPTs. I would look when the master output is less like after 11h00 what is the PV value indicated on each inverter screen and provide the 4 readings.

If the delta was 11h30 to 14h00 I would have thought may be a bad connection in one of the six in series on the master inverter.

30 minutes ago, Scorp007 said:

a bad connection in one of the six in series on the master inverter.

That would be my first checkpoint too. There is of course the possibility of one panel that may have deteriorated.

1 hour ago, zsde said:

That would be my first checkpoint too. There is of course the possibility of one panel that may have deteriorated.

That was my 1st thought but at peak production after 12h such a panel was still producing the maximum current for over 90min.

Thank you for the feedback so far. At one stage I thought it might be heat related, ie when it gets to a certain surface temp it looses some efficiency (which is generally true but then all panels becomes less efficient), but I started to watch that closely and even in that there was no consistency. It seems as if PV2 on Master struggle to keep the "load" consistently, but then again on some days it does.

Could it be a loose connection that does it?

Yesterday for example it was freezing cold with rain till around 14:00, then it was mostly fun sunshine. See how it "keeps up" in the cloudy and rainy weather, then totally fall back when the sun comes out, but at around 15:20 it then suddenly starts to track the PV1 almost to the tee.

2 hours ago, Kalaharipro said:

Thank you for the feedback so far. At one stage I thought it might be heat related, ie when it gets to a certain surface temp it looses some efficiency (which is generally true but then all panels becomes less efficient), but I started to watch that closely and even in that there was no consistency. It seems as if PV2 on Master struggle to keep the "load" consistently, but then again on some days it does.

Could it be a loose connection that does it?

Yesterday for example it was freezing cold with rain till around 14:00, then it was mostly fun sunshine. See how it "keeps up" in the cloudy and rainy weather, then totally fall back when the sun comes out, but at around 15:20 it then suddenly starts to track the PV1 almost to the tee.

A loose connection is not ruled and that is why I asked to see each of the strings for PV value on the master. One would just expect a loose connection not to stay at highish current for a longish time like after 15h30 before it dropped in line with the slave due to lower irradiation.

can you plot the inverter's internal temperatures over the same time-frame as the pv anomaly ? .... maybe at around mid-day one inverter is much hotter (or thinks it is) and then throttles pv power to compensate

So since day 1 Slave has always run hotter than Master - so the slides below are inline with normal behavior as far as temperature goes. I don't see any correlation between the AC/DC temp and the drop in PV power

10 August
MASTER Peak temperature DC 66 AC 55

SLAVE Peak temperature DC 95 AC 56

Try getting sunsynk help on the phone - I have found them very helpful with deeper technical issues ... I assume you're in ZA 010 100 3589

Ok thanks will go that route - but its a Deye unit, so will call them

I managed to get two photos earlier today of the inverters in action. Master - PV2 shows the lower Watts created

MASTER

SLAVE

The above was taken around 09:40 - graph below

MASTER

SLAVE

Edited August 13, 20251 yr by Kalaharipro

@Kalaharipro ,

Can you please also plot the PV Voltages for both strings for Master & Slave.

I find it weird that if all your strings are equally composed & exposed, why Master PV-2's voltage is 26V higher than PV-1's voltage.

Pure speculation. Just for understanding. What does the load graph look like at the same time? Just trying to see whether maybe that one MPPT is tracking the AC load fluctuations while the others are tracking the DC battery charging. Considering that the slave's DC temp is higher, it doesn't look like an overheat throttling on the Master. Is this setup running on-grid or off the grid? Are you using battery priority or load priority? Could there be any extra dust or dirt on the panels in question that's not on the rest? Bird poop?

39 minutes ago, Sidewinder said:

@Kalaharipro ,

Can you please also plot the PV Voltages for both strings for Master & Slave.

I find it weird that if all your strings are equally composed & exposed, why Master PV-2's voltage is 26V higher than PV-1's voltage.

Are you referring to these?

19 minutes ago, GreenFields said:

Pure speculation. Just for understanding. What does the load graph look like at the same time? Just trying to see whether maybe that one MPPT is tracking the AC load fluctuations while the others are tracking the DC battery charging. Considering that the slave's DC temp is higher, it doesn't look like an overheat throttling on the Master. Is this setup running on-grid or off the grid? Are you using battery priority or load priority? Could there be any extra dust or dirt on the panels in question that's not on the rest? Bird poop?

It is grid tied - PV first, then battery, then Escom.

The panels are cleaned every 3 months or so - and in summer with rain most poop is also washed away. As I mentioned in the first post this scenario suddenly appeared in April after running perfectly for 908 days.

Browsing through all the days since then there doesn't seem to be a 100% fixed pattern. Most days it occurs, sometimes for an hour then lines up with Pv 1 and then maybe drop again, today for example the drop was until the batteries were fully charged - ie the PV production "need" dropped. It sits mostly at a 600-700W drop in production, but sometimes it's 200-300w. Its never been more than 700w

So its a bit random, when it does happen it is normally around 09:00 till batteries are full, and then after that it mostly tracks PV1, but not always - and not nearly as closely as what the Slave does. Slave is almost a carbon copy on PV1 and 2 - not at all the case on Master

And as mentioned, nothing changed that I am aware of - not software or hardware, or any settings. Not sure if Deye has remote access and could have updates software etc?

Edited August 13, 20251 yr by Kalaharipro

@Kalaharipro ,

Tough one...

My first port of call could be to inspect all connections, including disconnects, fuses, MC4's etc.

Updating Firmware...can be done - think one can request that one Deye's www. Not 100%, as I have SS. Your installer might know, if all fails.

Lastly, but a real shlepp, would be to bypass each panel in the string, one by one, till the problem disappears, so very tedious, got to get on the roof and reconnect to bypass. In doing so, you might just pick up the problem, as any "loss" normally manifests itself as heat, so feel the cable by hand, especially at the MC4's. You get MC4's and you get mc4's😂. You might just uncover a rats nets of installer boe-boe's like this.

If any on the above doesn't resolve, then maybe that MPPT got damaged somehow, but also unlikely.

Have you tried changing the strings on the inverter - also a bit of a schlep, but at lease you will see if the issue stays with the MPPT, then that is at fault. IF the issue migrates to the other MPPT with the string, then the issue is with the string of panels (somehow, to be determined), but at lease you can rule out either the inverter or the strings.

Adding to the method of finding if a panel is not within spec a easier way might be to draw up a table for current and voltage for each panel in the input to the string with low output.

Cover one panel at a time in FULL and record the V and A values. With these readings one might find a clue. I presume all the panels do have bypass diodes.

This might be less sweat than the great suggestings from @Sidewinder for a start.

Just do the test between panels at the same time like 11h to 11h30 as not to have irradiation change affecting readings.

Edited August 13, 20251 yr by Scorp007