SOFAR 8kW PV Installation: Opinions, Durability, Failures & Experiences with Inverters

---Ryba--- 489729 1705

Best answers

How reliable are Sofar inverters for an around-8 kW PV installation, and what problems or failures do users report?

Users report mostly acceptable results: one 6.6KTL-X had no problems for over a year, and another 8.8KTL-X owner said a firmware update to v2.20 removed the old clipping around 8 kW and let the inverter reach full peak power [#18112697] [#18662870] [#19158337] The most common problems in the thread were not outright inverter defects but installation/configuration mistakes: too many panels in one string or the wrong input mode caused DC overvoltage errors such as PvOVP, BusOPV or permanent fault, and the cure was to wire the strings correctly and keep each MPPT within its limits [#18322870] [#18323108] [#18359020] [#18360610] [#18699492] Another recurring issue was grid overvoltage: users saw ID01/Grid OVP shutdowns when mains voltage rose to about 253 V, and one case reached 280 V at the customer side; the inverter was behaving as designed in those cases [#18624602] [#18665849] A few users also mentioned Wi‑Fi/software quirks and awkward warranty handling, with the installer, seller, distributor and Sofar service bouncing responsibility between each other [#18306997] [#18322870] [#18323064] [#18331261] [#18772451] There was at least one real hardware failure report, where an 8.8KTL-X tripped AC overcurrent after about a month and then would not restart, so the unit had to go through warranty [#18771167]

Generated by the language model.

Treść została przetłumaczona

Zobacz oryginalną wersję tematu

Report a violation of the law

Reply | New topic

#571 19393851 21 Apr 2021 23:22

toolpusher toolpusher

Level 23
» | Helpful post? (0)

Post #571
19393851 21 Apr 2021 23:22

I have recently enlarged the installation and gradually added panels - in our south there has not been a sunny day for several days. Here is a fragmented generator then had 5.44kWp. By the way, does downloading pictures from Solarman work? Because I have to do it via prtscreen. I don't know if chrome is blocking me or I can't download jpeg. Although the same on MS Edge.

Attachments:

theo.png (45.13 KB) You must be logged in to download this attachment.
ADVERTISEMENT
#572 19393861 21 Apr 2021 23:30

Janusz_kk Janusz_kk

Level 39

» | Helpful post? (0)

Post #572
19393861 21 Apr 2021 23:30

You can see the difference, so you've elegantly tuned the inverter
#573 19393888 21 Apr 2021 23:55

toolpusher toolpusher

Level 23

» | Helpful post? (0)

Post #573
19393888 21 Apr 2021 23:55

. 1 string on 2 MPPT in parallel mode. The interesting thing is that I was looking for the original MC4 Staubli tees but nowhere to find only the stiff ones. I have the original MC4 for the entire installation and I did not want to have Chinese at the inverter, so I was forced to do it myself. I clamp all the ends and solder additionally.
All in all, it is strange that there are no original MC4 flex tees and I have written to several stores.
#574 19393892 22 Apr 2021 00:03

theo33 theo33

Level 27

» | Helpful post? (+1)

Post #574
19393892 22 Apr 2021 00:03

Saving jpeg doesn't work either, I copy similarly.
I understand that you physically have the inputs connected in parallel and you also have the inputs set in parallel?
Upload the graphs of voltages and currents
ADVERTISEMENT
#575 19393904 22 Apr 2021 00:18

toolpusher toolpusher

Level 23
» | Helpful post? (0)

Post #575
19393904 22 Apr 2021 00:18

So I have switched in the inverter to parallel. I once corresponded with a serviceman from China regarding such a large oversizing of the installation and he told me to do it. Anyway, it was probably logical. I just wanted confirmation. For the readability of the chart, I only gave 1 string.
I turned on the inverter late because I changed the jumper last night and wanted to be turned on. The best part was that at night I had nightmares that when turned on, the inverter went up in smoke

Attachments:

Smn1.png (28.91 KB) You must be logged in to download this attachment.
#576 19393907 22 Apr 2021 00:24

theo33 theo33

Level 27

» | Helpful post? (0)

Post #576
19393907 22 Apr 2021 00:24

These nightmares are uninteresting

And you can turn off the DC power and voltage from this diagram and turn on the pv1 and pv2 currents, if the current from the modules is divided equally on both inputs
#577 19393913 22 Apr 2021 00:35

toolpusher toolpusher

Level 23
» | Helpful post? (0)

Post #577
19393913 22 Apr 2021 00:35

I also skimmed the report and so 99% is even, but sometimes there is a difference of 0.1A. There may also be rounding of values.
I also have the impression that with 2x MPPT the inverter has become cooler.

Attachments:

slmn2.png (21.83 KB) You must be logged in to download this attachment.
#578 19393914 22 Apr 2021 00:47

theo33 theo33

Level 27

» | Helpful post? (0)

Post #578
19393914 22 Apr 2021 00:47

As I thought, the module's current splits into two, here it comes out after about 4.5, so in my opinion the efficiency of the inverter will drop
Why did you connect in parallel, since you have one string and most modules will not exceed 11A, unless you have modules with some higher current?
ADVERTISEMENT
#579 19393926 22 Apr 2021 01:10

toolpusher toolpusher

Level 23

» | Helpful post? (0)

Post #579
19393926 22 Apr 2021 01:10

Because I have a 6.5kWp installation on a 4.4 inverter. The currents and voltages are normal, but the power is a bit too high, although it does not matter, because whether it cuts from 6.5kW or 5.0kW is probably the same.
On the other hand, in this manual for installers, such connection is recommended - page 27 - Link

What may the loss of fitness be? Looking at the graph, for example, the 600 and 800V curve is practically horizontal from 50% of the power.
#580 19394350 22 Apr 2021 11:43

prose prose

Level 36

» | Helpful post? (0)

Post #580
19394350 22 Apr 2021 11:43

@toolpusher And the info menu should now be 5 kW
#581 19394413 22 Apr 2021 12:12

theo33 theo33

Level 27

» | Helpful post? (0)

Post #581
19394413 22 Apr 2021 12:12

toolpusher wrote:
Because I have a 6.5kWp installation on a 4.4 inverter. The currents and voltages are normal, but the power is a bit too high, although it does not matter, because whether it cuts from 6.5kW or 5.0kW is probably the same.
On the other hand, in this manual for installers, such connection is recommended - page 27 - Link

What may the loss of fitness be? Looking at the graph, for example, the 600 and 800V curve is practically horizontal from 50% of the power.

The decrease in efficiency on weakly loaded inputs may be some, but probably negligible and if the graph relates to the AC output power, it is flat from 35%, but if it was also related to the power of the inputs, the load up to 3A at the input, i.e. up to 30%, would be losses. to be.
In my opinion, this parallel connection applies to strings connected in parallel when the summed current exceeds the capacity of the inverter input, while in your case it does not change anything, the previous limitation to 4.4 results from the output power after AC.
Once upon a time I made such attempts that I physically connected the inputs in parallel, but I left the settings independent and the current was distributed randomly, and in my opinion it was a result of the mppt algorithm because they will never work identically. And it looked like that with low powers one input worked, and with higher powers both. Unfortunately, I can't track these charts right now
#582 19394439 22 Apr 2021 12:35

Cytro12 Cytro12

Level 26

» | Helpful post? (0)

Post #582
19394439 22 Apr 2021 12:35

The graph grows up to 50% and again decreases towards 100% load, I note that the given graph is for a high power inverter there is probably 12kW, the smaller the inverter, the lower the efficiency, but generally the characteristics match (similar components and algorithms). Fronius has a little different again for an example of a different solution. We take into account the point at which the inverter has been operating the longest, not the extreme conditions.

The independent work of MPPT here depends on the speed of the processor and the quality of the algorithm.
ADVERTISEMENT
#583 19394526 22 Apr 2021 13:22

toolpusher toolpusher

Level 23
» | Helpful post? (+1)

Post #583
19394526 22 Apr 2021 13:22

prose wrote:
@toolpusher And the info menu should now be 5 kW

And is.

Attachments:

20210421_111728_HDR.jpg (879.2 KB) You must be logged in to download this attachment.
#584 19394750 22 Apr 2021 15:40

prose prose

Level 36

» | Helpful post? (0)

Post #584
19394750 22 Apr 2021 15:40

Now just check that the temperature doesn't go too high.
#585 19394769 22 Apr 2021 15:53

toolpusher toolpusher

Level 23

» | Helpful post? (+1)

Post #585
19394769 22 Apr 2021 15:53

Well, it seems to me that it has become cooler, but it is probably because 2 MPPT work but with less load. Besides, if it is the same inverter up to 6.6, I think it will be fine at 5.5.
#586 19394802 22 Apr 2021 16:12

prose prose

Level 36
» | Helpful post? (+1)

Post #586
19394802 22 Apr 2021 16:12

toolpusher wrote:
Well, it seems to me that it has become cooler, but it is probably because 2 MPPT work but with less load. Besides, if it is the same inverter up to 6.6, I think it will be fine at 5.5.

It got colder because you split the load equally on the MPPT inputs, 5.5 ktlx reaches the temperature at full power 50 degrees.

Attachments:

8.PNG (40.6 KB) You must be logged in to download this attachment.
#587 19394818 22 Apr 2021 16:21

toolpusher toolpusher

Level 23

» | Helpful post? (0)

Post #587
19394818 22 Apr 2021 16:21

I wonder if this module temperature is not a detector on the heat sink. Because in my case, the heat sink at 50st the inverter is slightly warm and now it's even summer.
#588 19394911 22 Apr 2021 17:14

prose prose

Level 36

» | Helpful post? (0)

Post #588
19394911 22 Apr 2021 17:14

From what I was able to find out the module is the IGBT AC temperature on the heat sink.
#589 19395585 22 Apr 2021 23:05

toolpusher toolpusher

Level 23

» | Helpful post? (0)

Post #589
19395585 22 Apr 2021 23:05

theo33 wrote:

toolpusher wrote:
Because I have a 6.5kWp installation on a 4.4 inverter. The currents and voltages are normal, but the power is a bit too high, although it does not matter, because whether it cuts from 6.5kW or 5.0kW is probably the same.
On the other hand, in this manual for installers, such connection is recommended - page 27 - Link

What may the loss of fitness be? Looking at the graph, for example, the 600 and 800V curve is practically horizontal from 50% of the power.

The decrease in efficiency on weakly loaded inputs may be some, but probably negligible and if the graph relates to the AC output power, it is flat from 35%, but if it was also related to the power of the inputs, the load up to 3A at the input, i.e. up to 30%, would be losses. to be.
In my opinion, this parallel connection applies to strings connected in parallel when the summed current exceeds the capacity of the inverter input, while in your case it does not change anything, the previous limitation to 4.4 results from the output power after AC.
Once upon a time I made such attempts that I physically connected the inputs in parallel, but I left the settings independent and the current was distributed randomly, and in my opinion it was a result of the mppt algorithm because they will never work identically. And it looked like that with low powers one input worked, and with higher powers both. Unfortunately, I can't track these charts right now

Ok then I'll go back to 1 MPPT. If it is actually one inverter for the 3.3-6.6 range, then it should support power up to 6250Wp per MPPT according to the card without any problem. As a colleague Cytro12 wrote, most of the time the inverter does not work at full power, hence there may be large drops in efficiency for 2 x MPPT.

Added after 2 [minutes]:

Cytro12 wrote:
The graph grows up to 50% and again decreases towards 100% load, I note that the given graph is for a high power inverter there is probably 12kW, the smaller the inverter, the lower the efficiency, but generally the characteristics match (similar components and algorithms). Fronius has a little different again for an example of a different solution. We take into account the point at which the inverter has been operating the longest, not the extreme conditions.

The independent work of MPPT here depends on the speed of the processor and the quality of the algorithm.

I have not seen another chart, it is only the one for the 11KTL-X. Additionally, I found that in some manuals for KTL-X three-phase inverters there was a graph for single-phase TL series where the last curve was for 500V.
#590 19396742 23 Apr 2021 18:31

adamvfr adamvfr

Level 11
» | Helpful post? (0)

Post #590
19396742 23 Apr 2021 18:31

Is it normal that the current is always lower in the PV2 string by about 0.5A
Hello, Sofar 3.68KTLM Panels 2 strings, 8 pcs each, the same.
For a few days I had only 7 panels plugged in the string PV2 and looking at the statistics I had better production with seven than when I plugged in the eighth one. Maybe this is a margin of error, but I certainly did not notice an increase in prod. rather a reduction (I have a comparison to an adjacent installation). Mathematically, I got about 4% less total production after plugging in the eighth module.

Attachments:

napiecie.jpg (66.53 KB) You must be logged in to download this attachment.

roznica.jpg (24.89 KB) You must be logged in to download this attachment.

prad.jpg (82.89 KB) You must be logged in to download this attachment.
#591 19396797 23 Apr 2021 19:05

Janusz_kk Janusz_kk

Level 39

» | Helpful post? (0)

Post #591
19396797 23 Apr 2021 19:05

You may have a partially damaged panel and therefore have more production without it, or it is shaded which gives the same effect.
#592 19396800 23 Apr 2021 19:07

toolpusher toolpusher

Level 23

» | Helpful post? (0)

Post #592
19396800 23 Apr 2021 19:07

What version of Solarman are these charts from?
#593 19396817 23 Apr 2021 19:13

theo33 theo33

Level 27

» | Helpful post? (0)

Post #593
19396817 23 Apr 2021 19:13

These charts show that you already have 8 modules here, show the charts from 7 where the production was higher and compare which parameter made it worse.
Also, swap the inputs because the input algorithms are not the same, and if the situation changes, it could be considered that the strings differ, maybe in length, lighting, I think that the parameters of the modules do not differ?
#594 19397503 24 Apr 2021 09:09

adamvfr adamvfr

Level 11
» | Helpful post? (0)

Post #594
19397503 24 Apr 2021 09:09

Charts from Solarman Smart.
The thongs are identical, no shade.
I have analyzed these charts in more detail, I can also give the login and password to someone who is insightful, because I would like to understand it better myself.
So you can see the difference in Amps even when there were 7 panels, but in the peaks it is much smaller then.
At least that's how I see it.
On the third day of operation of this installation, the allegedly low V grid and frequency was turned off (see diagram below), but the inverter did not turn on by itself, so I turned off DC and AC and fired and started normally.
Two days later (also the graph) the inverter recorded zero DC and AC voltages too! but it did not turn off, why?

Attachments:

inst grzes.png (632.96 KB) You must be logged in to download this attachment.

7 pan.jpg (101.83 KB) You must be logged in to download this attachment.

7 pan piki.jpg (113.97 KB) You must be logged in to download this attachment.

wpiecie osmego.jpg (97.99 KB) You must be logged in to download this attachment.

wylaczenie falownika.jpg (137.06 KB) You must be logged in to download this attachment.

same zera.jpg (119.09 KB) You must be logged in to download this attachment.
#595 19397511 24 Apr 2021 09:20

adamvfr adamvfr

Level 11
» | Helpful post? (0)

Post #595
19397511 24 Apr 2021 09:20

Today's data.

Attachments:

dzis.jpg (43.8 KB) You must be logged in to download this attachment.

dzis wykres.jpg (64.4 KB) You must be logged in to download this attachment.
#596 19397541 24 Apr 2021 09:44

theo33 theo33

Level 27

» | Helpful post? (0)

Post #596
19397541 24 Apr 2021 09:44

Drop on private. this access and enter the page, I can browse.
And as I wrote before, swap the inputs
#597 19397698 24 Apr 2021 11:14

Cytro12 Cytro12

Level 26

» | Helpful post? (0)

Post #597
19397698 24 Apr 2021 11:14

If you have a clamp meter, check the actual values on the cables, because these measurements of these inverters are of different quality. If Sofar has different algorithms for MPPT, this equipment is weak because they say that there are two independent MPPTs and it turns out that there is one and the other as a divisor because the inputs are too weak.
I wonder if this equipment will last at least until the end of the warranty :)

It would be a pity because they put it en masse into every cheap installation.
#598 19397809 24 Apr 2021 12:46

Janusz_kk Janusz_kk

Level 39

» | Helpful post? (0)

Post #598
19397809 24 Apr 2021 12:46

Cytro12 wrote:
If Sofar has different algorithms for MPPT, this equipment is weak because they say that there are two independent MPPTs and it turns out that there is one and the other as a divisor because the inputs are too weak.

Can you write something closer?
#599 19397831 24 Apr 2021 13:03

adamvfr adamvfr

Level 11

» | Helpful post? (0)

Post #599
19397831 24 Apr 2021 13:03

I would also read. So every 2 thongs but basically one? artificially divided?
#600 19397890 24 Apr 2021 13:45

prose prose

Level 36

» | Helpful post? (0)

Post #600
19397890 24 Apr 2021 13:45

ZdzislawDyrmaZasadni wrote:
Hello, I have a general attention to SOFAR inverters. According to in the current manual, these inverters must not be installed in basements or any other low-lying rooms, because min. the required distance from the ceiling must be 80 cm.
If someone has to buy Chinese, better choose SOLAX.

And solax does not need ventilation?
Create an account, log in here. You will receive points by participating in discussions.
Join this discussion.

Install Elektroda application

Reply | New topic

Report a violation of the law

Topic summary

✨ The discussion centers on experiences with Sofar inverters in approximately 8kW photovoltaic (PV) installations, focusing on technical issues, durability, and configuration challenges. Users report generally stable operation with models like Sofar 6.6KTL-X and 5.5KTL-X over extended periods, though some face connectivity problems, especially with Wi-Fi and software availability. A recurring technical problem involves input voltage exceeding the maximum allowed 600V DC for single-phase Sofar inverters (notably the 4KTLM-G2), causing errors such as PVOVP and BusOPV and inverter shutdowns. This is often due to too many PV panels connected in series on a single string, with 15 Q.CELLS Q.PEAK G4.1 305W panels sometimes exceeding voltage limits, especially in cold conditions where voltage rises. Solutions include splitting panels into two strings (e.g., 8 and 7 panels) connected to separate MPPT inputs, though some users report this does not fully resolve the issue. The maximum recommended panels per MPPT input is around 9 to avoid surpassing power and voltage limits. Users also discuss the need for three-phase inverters for larger installations above 7kW on a single phase due to grid operator restrictions. Software and firmware updates are sought after but not easily accessible, with some users lacking Polish language support. Comparisons with other brands like SMA and Fronius highlight Sofar's cost advantage but raise concerns about long-term reliability and service support, which is still developing in some regions. Network-related errors such as Grid OFP (over-frequency) have been reported, with troubleshooting involving network frequency checks and inverter restarts. Overall, Sofar inverters are considered a cost-effective option with some technical caveats related to system design, installation quality, and support infrastructure.
Generated by the language model.

FAQ

TL;DR: Field owners report 0–20 % daily energy loss when Sofar inverters trip on 253 V grid peaks; “set Country Code 12 for Poland and the problem goes away” [Elektroda, 3301, post #18624621] Why it matters: one menu setting often fixes most unexpected shutdowns without replacing hardware.

Quick Facts

• Max. DC input voltage 1000 V & MPPT window 240-850 V (Sofar 8.8KTL-X manual [Uri, #18921613]) • Recommended DC/AC oversize ratio ≤ 1.2 : 1 for KTL-X models [Elektroda, BikeBarian, post #18360391] • Country Code list: PL = 12, DE = 07, NL = 05 [Elektroda, Defence & 3301, post #18851696] • Typical firmware: V2.30 (increases AC limit from 4.0 kW to 4.4 kW on 4.4KTL-X) [Elektroda, prose, post #19034268] • Service e-mail: service.pl@sofarsolar.com; hotline +48 22 123 98 58 [Elektroda, AT PRO, post #18772451]

How reliable are Sofar inverters in everyday use?

Most 6–9 kW owners report trouble-free work for 1–3 years [Elektroda, iborkim, #18112697; Elekt...

Which menu code should I select for Polish grid parameters?

Use Country Code 12 (PL). It lifts the over-voltage trip to 253 V/10 min and 264 V/1 s, matching Polish NC-RFG limits [Elektroda, 3301, post #18851696] Selecting 07 (DE) raises the limit to 264 V but violates declaration of conformity.

My inverter shows ID01 ‘GridOFP’ and stops at noon—what now?

Measure L-N voltage; if it rises above 253 V, ask the DSO to lower tap settings or move PV to another phase. Until then, temporarily change Country Code to DE (07) or install a 230 → 220 V autotransformer [Elektroda, MariuszFEG, post #18623709]

What causes PvOVP / BusOPV errors on 4KTLM-G2 units?

Voc exceeds 600 V in cold sun. Split the 15-panel string into two of 8 + 7 or use a 3-phase inverter that accepts 1000 V DC [Elektroda, noja102, post #18323557]

How many Q-Cells 305 W modules can a Sofar 4KTLM-G2 safely handle?

Fourteen—seven per MPPT keeps each string below 2600 W and 600 V [Elektroda, kosmos99, post #18531958]

Is the MPPT Scan function worth enabling?

Enable only when partial shading occurs; it rescans the IV curve at user-set intervals. Service says it "improves harvest under shade" but reduces output 1–2 % on clean arrays [Elektroda, 3301, post #18622124]

How do I update Sofar firmware myself?

Copy the /firmware folder to the SD card.
Insert card, choose Menu → 5.Software Update, enter password 0715.
Wait until ‘Update DSP1 OK’, ‘DSP2 OK’, ‘FUSE OK’, then reboot [Elektroda, Grucha383, post #19029218]

The display boots for one second at dawn—is that harmful?

No. The capacitors charge when DC rises above 180 V, then discharge if current is < 20 mA; this soft-start repeats without affecting lifespan [Elektroda, rydzyk673, post #19035241]

Why does Solarman show 0.9 A per phase even at 0 W?

Firmware ≤ V2.30 reports reactive standby current as active. Real clamp-meter reading is ~0.2 A; ignore the dash-board figure until a patch arrives [Elektroda, wilk125, post #19038447]

Edge case: can the inverter fail after a breaker trip?

Yes—one 8.8KTL-X died after a 16 A AC breaker opened; display stayed dark despite correct DC voltage, requiring warranty swap [Elektroda, gotycki1, post #18771167]

How to auto-run a 1.5 kW boiler only on surplus PV?

Install a PR-energised relay with the inverter’s CT loop; set closure above 4.5 kW export so the heater engages only in excess production [Elektroda, 3301, post #19053070]

Generated by the language model.