BigC

@Powerforum Store.. thanks for knowledgeable response. And @frivan for your simple suggestion. Since the generator is a manual start, there is no hardship in operating a bypass switch from grid to generator, so I will implement that this week. The budget is a concern, so the suggestion to live with the Solis on the gen port and add the CT makes sense and is easy to do, since most of the work has been done already. If the weather permits enough sun this week I should have my answer soon.

Then you lose the contribution of the 3rd string during load shedding, which is kinda when you need it most. It's a tossup .. it's the cheaper solution, but not elegant at all.

Hi everyone .. it has been a while! I successfully got the Solis to work on the Gen port, using the frequency shift feature of the two to shut down the Solis when the batteries approached 100% SOC. It was not without problems since the Deye would occasionally "leak" power back to the grid because the Solis was too slow to respond and the grid smart meter would disconnect the grid. However, a bigger problem, after a lengthly grid outage (many days) in the Overberg, was how to add a standby generator to the system? Connecting the generator to the Deye grid port worked, but not a long term option. What I am going to try next is to connect an el-cheapo/used hybrid inverter (like a Kodak) to the Gen port and the generator to the Kodak grid port. Theoretically this should work well, with hopefully no over-feed problems. Question is .. can the Kodak work battery-less in this work-mode. If not a Kodak, does anyone have experience or suggestions on what inverter would work here? I so look forward to your responses .. love this stuff 😄

Thanks for your comments. The Deye does make this distinction with a tick box. Earlier today the Deye FW was updated, so I am hoping to see a change tomorrow. Additionally I will be assisted tomorrow by Solis to verify it's settings as well . . . 😀

Illness, weather, installation of 9 x 460W panels, and the addition of another Blade battery have delayed the commissioning of the system somewhat. However, it is all working well .. I got rather excited .. up to the point of the batteries reaching 100% SOC. Sadly the inverter does not react to this situation by increasing the frequency on the Gen port. Hence, at some point the inverter turns off the supply to the Solis with alarms sounding and all manner of consternation. The faults logged by the Solis are consistently "disconnected from the grid" It would seem that this Deye has older FW .. Ver 0-5388-1516, which does not implement this means of regulating the Gen port output. Additionally, I have no experience of the Solis inverter and the configuration seems a nightmare. I therefore have no idea whether its present configuration is suitable. First off I suppose I can lean on Deye support to download the latest FW .. is this a no risk scenario?

@HennieL : The garage roof is fully utilised and additionally not optimised for angle, etc. Additionally, the cables and panels were installed in clusters (don't understand why). The new panels will go on the house roof. I would have preferred to mount the Solis in the house and run a mains cable to the garage, but the configuration suggested below works much better .. will have to run PV cable to the garage instead . . . Thankyou @Scorp007 and @GreenFields .. "If this is all correct, then the logical thing for me would be to connect the single string off from the second Deye MPPT, and connect it back to your Solis 4G again (this time with the Solis inverter connected via the AUX input). This will free up the MPPT on the Deye that can take a 5200W string at least". That is the conclusion I came to. I spent some time on ChatGPT exploring all this. The result is 4.1kW (9 panels at 460W) + 2.6kW on the Deye and 2.9kW on the Solis giving me 9.6kW total. This is close to the 190A max charging current. I think this is as good as it gets. ChatGPT suggests keeping the Gen port input substantially lower than the onboard MPPT input as, it seems, the Solis responds more slowly to the frequency shift generated by the Deye and this can cause hunting. Thankyou all for your wonderful and knowledgeable contribution to this thread. I have confidence now that this will work .. can't wait to implement this upgrade. Ciao . . .

I love this solution .. now to commence the install. Thank you all for your valuable contribution to my enquiry. All done and dusted in < 12 hours. I just enjoy this forum . . . 🙂 PS . . . I may need help with the configuration, but will cross that bridge when I get there. Ciao . . .

@GreenFields . . . Just to clarify, the existing panels are old and were installed on 3 different occasions, as 3 strings, on a roof that is impossible for me to access. The cables all ended up in the garage, where I was able to measure OC voltage and SC current, and then carefully connected the panel sets that were identical in current, at least, into 2 strings. It is not possible to identify them, or change the configuration. @frivan and @hoohloc have identified a possible solution. I have 3 x Solis mini 2500 inverters (removed from the old installation)! However, I don't see how they can connect to the Aux/Gen port. Surely the grid-tie inverter needs to see a grid? Would the Gen port work like that?

If I opt to fit a second inverter rather, say a 3.6k or 5k unit, which is a bit more expensive, then you lose nothing .. just as long as they are compatible when connected in parallel. Can one mix inverters, as long as they are from the same stable?

2 dissimilar strings totaling 27 mixed panels, cobbled to give me 2 nice strings of about 4500W total. They were inherited so I did the best I could. They are working well. As @Kalahari Meerkat suggested, I may have to forgo the BMS connection, configure the battery mode as "use battery v" which means I lose some of the battery metrics. Was hoping someone tried this before I spend the bucks . . . 😆

I have a problem. I have 4k5W of panels in 2 strings feeding an 8kVA Deye inverter and 7.2kWh of Blade battery. I wish to add a further 4kW of panels (and another battery), but do not really require a bigger inverter. the solution is to add a Victron 250/60, say, charger to handle the extra string. Can this work, or will the Deye MPPT throw up it's arms in despair? Is there another solution that the esteemed members of the forum may have up their sleevies?

Item: 2 off Solis mini-2500-4G Solis S6-GR1P2.5K Solis EPM3-2G Export Power Manager Age: Bought early 2023 before load shedding really kicked in, and subsequently found to be unsuitable since only grid-tie. 3ph supply converted to 1ph and inverters being upgraded to a single large hybrid inverter, with the intention to move off-grid completely. Price: Solis minis .. R4500 ea. Solis S6 .. R5500 Solis EPM3-2G .. R 2500 Payment Method Accepted: EFT Warranty: Expired, but sold as working units. Packaging: Condition: Excellent Location: Stanford, Overstrand Reason: Upgraded to 1ph hybrid Shipping: Courier Collection: Link:

No .. just twisted pair. Seems that I should have used a shielded cable, or maybe even cat5 with the 4 pairs all in parallel (sort of) which would have been a better 'twisted pair'. Given the effort of burying the cable (in conduit) and pulling in more expensive cable, I am hoping the client will live with problem.

Hi @Sc00bs .. Apologies for the long delay .. The client went for the 45m of cat5 cable, so sorry .. no feedback. However, what did happen is that the 40m extension for the CT for the Deye inverter has proven to be problematical. Readings of load power on the inverter and on Solar Assistant are negative by a small amount .. once loaded the readings look normal. To test the connection I did the following: Removed the CT from the incoming grid cable .. problem remains Disconnected one leg of the CT at the inverter .. problem remains Reversed the polarity of the CT .. problem remains, although readings of load power are now negative. Disconnected both legs .. problem solved Despite this problem the inverter continues to export power back to the CT normally. I expect there is some low level interference in the long CT circuit.

Thanks @Sc00bs Looks like the answer to a maidens prayer. I will report back once I've tried this . . . Ciao