diy solar

diy solar

Deye / Solark 8K Setup, Some photos an diagrams!

I got to travel to where my inverter is stored and took a pic of the fan model number.
Since I now know they are 12vdc, I can use a pc supply to test the noise level and compare to possible replacements.
I can do this before I mount the unit to get an idea what approach to take. Also looks like an adapter (80mm to 120mm) could be used to allow120 mm fans. (should be much quieter))1651458613777.png

Kid
 
I finally got to test the possible replacement fan. (Noctua NF-A8 ULN) The Noctua is silent compared to the ambient noise while the stock fan produces a very high pitched loud whine. Note that the 3 pin connectors have different pin outs, so that has to be changed should anyone do the fan swap. I am choosing not to do the swap, due to the lower cfm output of the Noctua fan. I will try to make a duct on the exhaust side of the cooling chamber and use some 120mm fans (quiet and good cfm output) to pull air through constantly to try to keep the stock fans from running except in the most extreme power draw conditions.

Kid
 
I think the Deye, if similar to Sol-Ark, has a high voltage DC rail. PV feeds high voltage rail from which AC is made (rather than conversion first to battery voltage), battery is boosted to same high voltage rail, or charged from it. For models which AC couple, AC also goes to high voltage DC before charging battery. Likely AC line input does too.
I've been told before this is the case, the high voltage (or really medium voltage in the larger scheme of things) DC rail seems to replace the first logical block of stepping up DC voltage before chopping it up with PWM to make the waveform in high frequency type inverters. It eliminates a conversion step, and the battery would need to be stepped up anyways. I would suspect that even though Deye/Sol-Ark has the high-voltage DC rail in the center of their marketing material as one of the sources of its efficiency - that most of the all-in-one inverters probably follow a similar route. Why step down to 12/24/48V internally just to step it back up again? I suppose it makes sense with a low frequency inverter though.

I would expect a separate PV connection with MPPT circuit charging that rail, and not backfed, but maybe not?
As you note, RSD boxes between panels can't do anything about voltage on wires from inverter. Although, if RSD is activated by inverter, that may discharge those wires.

Working on my GT PV inverters, after disconnecting AC and DC I have to wait a while for capacitors to discharge on DC side. That is much slower than RSD requirement, so would require additional circuitry or operation of inverter into a load.

One Deye I looked up specs for accepts 100 to 500V input. The circuit would be buck/boost to work around 244VDC.
Although there is a diode in buck-boost topology, it might leak to charge up capacitor.


Now you know to check all voltages with meter before working on it.

My friend got bit by the PV inputs while working on a Deye and measured the same standing voltages on the PV inputs. I always assume there could be some sort of voltage potential and check everything twice before working on it, and when working on it I always make sure I'm focused on exactly what I'm doing, where my hands are, and don't allow anything to distract me until I'm done with the individual step. It's generally not a good idea to stick both hands into something electrical (I'm always nervous about making a direct path across my chest/heart). I would rather get another person to lend a hand, but 99% of the time I can do everything with one hand at a time when making electrical connections.

Watch YT videos where people have to point and touch everything makes me nervous as hell. It's those moments when you aren't quite thinking about what you are doing, or when you get slightly distracted during performing a routine step that an accident happens.

I was really into electronics as a kid until I got shocked (and arced across) really good a couple of times, and didn't come back to it until I was almost 40. At first I would become uncontrollably shaky when dealing with it and took me a couple of years to get over it.
 
They are very loud and spin very fast, around 5000rpm i believe, so its kind of high pitched noise. I havent had time but it should be relatively simple to connect the fans to a PC fan controller and make then spin all the time to 3000rpm or something similar, or even attach a temp sensor in there.... Just not sure if the inverter will try error codes for that, but im guessing you can place a resistor or something instead of the fans to fool it.

FWIW I was told that Deye (or Sol-Ark engineers, I dunno) was going for what SMA found: To keep the inverter internals in a consistent temperature dust-free environment, as it makes for the longest life of components to not be heat cycled. I've noticed it keeps its internal "AC Temperature" and "DC Temperature" readings in between 50-60C. It turns off the fans below 50C and they ramp up to full power at 60C. So unless you figure out a quieter solution that also keeps it in a tight temperature range, I would try to limit the noise by putting it someplace where you can't hear it. I found when I put a heat pump in my garage (after insulating the doors and one of the walls) and set it to keep the garage at 75F, the fans almost never kick on in that temperature anyways.
 
Well, inverter ended up in new utility room. I took note of what copec said regarding the unit trying to maintain a certain temperature range to lengthen component life. I am choosing not to try to add any fans for the time being. The rockwool insulation keeps the room very quiet and you can barely hear the fans if you are on the carport next to the utility room. (so quiet enough for me)
InverterWall.jpg
I will cover the rest of the walls when I return later this winter. Really impressed with the Deye 8K !!

Kid
 
Well, inverter ended up in new utility room. I took note of what copec said regarding the unit trying to maintain a certain temperature range to lengthen component life. I am choosing not to try to add any fans for the time being. The rockwool insulation keeps the room very quiet and you can barely hear the fans if you are on the carport next to the utility room. (so quiet enough for me)
View attachment 109145
I will cover the rest of the walls when I return later this winter. Really impressed with the Deye 8K !!

Kid
That looks awesome! Do you have communication going with those BMS’s?
 
Using Solar Assistant. I can only see total battery voltage. The individual cell voltages never showed on the bms page in the Deye menu.
Only total voltage and temperature. But while I was there I checked the battery states with the Seplos displays and everything was staying in sync.

Dashboard › Battery​

State of charge
68 %
Power
1838 W
Current
34.1 A
Voltage
53.9 V
Temperature

Kid
 
Well, inverter ended up in new utility room. I took note of what copec said regarding the unit trying to maintain a certain temperature range to lengthen component life. I am choosing not to try to add any fans for the time being. The rockwool insulation keeps the room very quiet and you can barely hear the fans if you are on the carport next to the utility room. (so quiet enough for me)
View attachment 109145
I will cover the rest of the walls when I return later this winter. Really impressed with the Deye 8K !!

Kid


Nicely done. Couple questions. Why not use the breakers in the main panel and the critical loads panel instead of the separate breakers? Was it simply to have shutoffs right at the inverter? Have you seen any swelling without compression? I know they say below a certain c rate that compression isn’t needed, but I’m just curious if you see any. Also where/who did you get your cells from?
 
Nicely done. Couple questions. Why not use the breakers in the main panel and the critical loads panel instead of the separate breakers? Was it simply to have shutoffs right at the inverter? Have you seen any swelling without compression? I know they say below a certain c rate that compression isn’t needed, but I’m just curious if you see any. Also where/who did you get your cells from?

I can talk by experience that you get swelling with minimal C rates, only by the process of charge & discharge (At least with my prismatic's) . Posted something about this last year (With photos of the swelling), and thats why i went ahead and builded my compression case.

It affects the cells?, by how much?, not sure. But just note that the manufacturer gives a longer lifespan with compressed cells.
 
Last edited:
I originally came out of the homemade powerwall/second life type community, so I watch my cells and plan to use them until they are totally dead — ignoring calendar age. So when I start using LFP I would want to use compression, that seems to be more an OCD thing for me though, I totally get the argument with not dealing with it.
 
Well, inverter ended up in new utility room. I took note of what copec said regarding the unit trying to maintain a certain temperature range to lengthen component life. I am choosing not to try to add any fans for the time being. The rockwool insulation keeps the room very quiet and you can barely hear the fans if you are on the carport next to the utility room. (so quiet enough for me)
View attachment 109145
I will cover the rest of the walls when I return later this winter. Really impressed with the Deye 8K !!

Kid
Good looking!, wich BMS's are you using?, are they good?
 
Using Seplos 200amp. They seem to work well and the displays are nice. Mine are from a mfg run last year that had some bad diodes that could fail prematurely. But each bms rarely sees more than 80 amps. I am hoping they will be fine.

Kid
 
The breakers are for safety/convenience. The main house panel and critical loads panel are 65 feet away. I have not noticed any swelling.
My cells are Eve 304 amp\hr from Amy at Docan. Shipped from the Houston warehouse last year.

Kid
 
Back
Top