Solis Hybrid inverter with DIY battery

daveharnett

I'm waiting on installation of a PV system with a Solis Hybrid inverter (RHI-6k-48ES-5G). I've independantly bought 16x Eve 280ah LiFePo4 cells and a JK Bms.

I'll use homeassistant to monitor not the bms and the inverter and occasionally change settings based on solar forecast - this is well covered on other threads.

My question is around the best way to get the inverter to charge the battery. I see two options:

1. Place a pi/esp32 between the inverter and the bms, and use it to translate between the BMS modbus protocol and Pylontech Canbus protocol. The BMS/modbus side has got good libraries, but I've not seen one for the propriatary canbus protocol. Has anybody seen a library for this (or even good documentation for the protocol?)
2. Tell the inverter that it's connected to a lead-acid battery, and tweak the parameters below into a profile that's suitable. Does anybody understand how the inverter behaves based on these parameters?

Thanks

graememk

Ah. I have stuff for you!

https://github.com/ASMotionLab/Daly2Sofar

CAN-Bus-protocol-PYLON-low-voltage-V1.2-20180408.pdf

SMA CAN protocol (2).pdf

Also look at the code for the zeva bms. That might help too for CAN packet building.

The pylontech is just a extension of the SMA protocol

daveharnett

Outstanding, thanks Graeme! I dare say that'll keep me out of trouble for a couple of weekends.

For posterity, I also found some images of somebody else's settings - turns out there are some more settings available for lead-acid which aren't apparent in the manual, which make the operation much more comprehensible:

Big Lar

https://www.boards.ie/discussion/comment/119059421#Comment_119059421

@daveharnett You could be my new friend here, I am having a bit of a time with the DIY battery myself and Daly BMS. Presently I am working on a solution and it was my intention to start a new thread with what I had learned.

1. The first Issue I had was that the inverter would shut down as it received an Over voltage error when the batteries were almost charged, from my investigations here and on FB groups it seems that under Control Parameter settings that the Battery Overcharge voltage should be set a little bit below 58V as there seems to be a bug issue on the Solis inverter, currently I am set to 57.5 but I really haven't got the weather to hit the 10Kw storage capacity yet. Now I could charge the battery by day for the experiment of it - But I am just too stingy for that as I would never get the full 10Kw back from the battery. Here is my post on FB https://www.facebook.com/groups/288045168816481/posts/1039820603638930/
2. I find that with Lead Acid settings that the max charge I can get is 40A no matter what inverter settings I use on the battery, 80A BMS. I have the setting on the battery in the SOLIS inverter set to 70A and lately I am charging from grid from 6:00am to 9:00am and the max I can get there is 40A also. Doing this for the craic really at the moment as I am on Day/Night meter but it is in intention that during the winter I would use this option.
3. Building on from above my new issue will be for the winter be; How will I wake the inverter to start charging at the night rate as the inverter is powered by DC and from my experiments of late when the battery drains down then the BMS turns off the battery thus no feed for the inverter in turn leading to the inverter switching off and not charging from midnight - If you can follow that logic then this is why the inverter only starts charging @ 6:00am as thats when the light hits the panels these days. So I will need something to wake up the inverter for the winter months to start charging.

I am of the opinion that while using the BMS to manage the batteries is a good thing but the downside is that the BMS also controls the power supply to the inverter, may be talking outta my ass here but Like in the end of the day all I really want the BMS for is to balance the resistance of the cells so that they do not overcharge and I was half thinking of just running a series of 16 X "Lets Say" 1M ohm resistors where each resistor would parallel with each battery and balance the resistance then ditch the BMS and let the inverter manage the battery from there - Love some feedback from somebody for this one.

Lotz of lads here on boards recommend the Seplos BMS but from what I can see there is some beater of a thread on that with fault finding issues so it seems that BMS is not all its cracked up to be either.

https://www.boards.ie/discussion/comment/119056655#Comment_119056655

I have all the stuff for this on order from Aliexpress and hope to get it built as soon as it arrives, should have it up and running in about 4 - 6 weeks as I have the bits and pieces ordered with a few weeks now.

daveharnett

https://www.boards.ie/discussion/comment/119070559#Comment_119070559

If i've understood correctly, I think the solution to point 3 would be to set the BMS's protections to be slightly more "permissive" than the inverter's protections settings. Eg. if you set the inverter's overdischarge_protection to 10% SoC, then set the BMS's overdischarge_protection to 5% SoC. This would allow plenty of energy in reserve to keep the inverter ticking over until it's time to charge.

This would let you retain most/all of the benefits of the BMS:

• Acting as a failsafe for the inverter's over/undervoltage protection.
• Some amount level of smart balancing (and imbalance protection).
• Thermal protection

You've obviously researched the passive balancer idea much more than me, but I'm sceptical. Wouldn't the current bleed from a 1Mohm resistor at 3.65v would be miniscule? It might help to counteract tiny imbalances from accumulating long term I suppose, but I'd want to actively monitor it for a long time before i trusted it.

40A sustained from an 80A bms is pretty shocking, even by Chinese standards. Maybe it's a "40A (80A Max for 20 seconds)" model that got mis-sold?

graememk

If one cell is running high/low double check for a loose connection.

a 15s is possible as well

There is active balancers too that can balance at higher power levels but anything they balance has to be dumped as heat.

You'll never get full 10kwh, due to inverter losses etc, also your not going to 0 either. 8kwh out is fairly reasonable

Big Lar

https://www.boards.ie/discussion/comment/119071334#Comment_119071334

Yeah, I had that thought also about setting the BMS's protections to be slightly more "permissive" than the inverter's protections settings

The balancer idea does seem over simplified and if that worked then nobody would be buying a BMS I guess. Plus I guess if I did go that route then thats the end of Canbus to the inverter

As for the 40A, it seems to be a setting in the control parameters option that I have tried to change but cannot, I can change the Charge Limitation to 40A+ but it just seems to revert to 40A so its not a BMS thing at all. BMS will hit the discharge limitation no problem.

I am hoping that the Daly2Sofar Arduino BMS is going to fix everything 😶

graememk

https://www.boards.ie/discussion/comment/119072642#Comment_119072642

*Not an Arduino, it's an esp32!

reklamos

https://www.boards.ie/discussion/comment/119072642#Comment_119072642

Instead of Lead-acid use user-definied as a battery and there you shoudl not be limted to 40A.

With battery the inverter should not power off during night time when properly configured. As @daveharnett already wrote. Set your dicharge limits properly. On inverter make it higher than on battery. Also you can configure FOC to keep battery at minimum state of charge all the time.

Big Lar

https://www.boards.ie/discussion/comment/119074293#Comment_119074293

user-defined comes up with "Can Communication Failure " so that's not a runner at the moment unless there is another setting to disable that alarm

meelick

Hi all, I'm in a similar situation myself. I'm using a 6kW Solis RHI series inverter, and I built a 12kWh DIY battery with Daly BMS, but no go on the "User Defined" settings (CAN Fail), so going to build the canbus bridge mentioned above to try and get it looking like a supported battery.

In the meantime, Can anyone point me in the right direction to get the settings for the "Lead Acid" settings? The inverter sees the 57.4V from the battery, and it's not powering off at night as it used to before I added the battery (used to be no modbus metrics for my grafana dashboard when no solar generation, with the battery it's pulling enough from it to keep the inverter and wifi-stick working through the night). My problem is that I'm not seeing the battery discharge at all, even if I put a load on the system greater than the solar generation, it still pulls from the grid rather than the battery. Even when there's zero solar, it still pulls from the grid. Nothing from the battery. Is there some "mode" I need to change to tell the inverter that it should pull from the battery?

reklamos

https://www.boards.ie/discussion/comment/119158467#Comment_119158467

It should work without any special settings. Check what is discharge limitation you have under control parameters if it is 0 it will not discharge. As for other settings we would need more details about your battery first.

Also it could be that your BMS is not allowing discharge

meelick

https://www.boards.ie/discussion/comment/119158660#Comment_119158660

Thanks for that. I've tested BMS with an external load (DC-DC converter and 12v battery charger), and can see charger switch on and off when I enable/disable discharge in the BMS app. BMS also correctly shows the voltages of the 14 packs. Pics attached of Inverter settings. Really appreciate the help. Oh, and pic attached of the DIY battery also. Storage Mode is "Self Use" and battery type is "Lead Acid". Battery is at 57.8v, which might be a little high, but my thinking is that it should at least pull something from it 'till it gets within the nominal charging range.

irishchris

https://www.boards.ie/discussion/comment/119160677#Comment_119160677

Am I correct in only counting 14 cells in that setup? If so therein lies one of your problems for a start

meelick

https://www.boards.ie/discussion/comment/119160853#Comment_119160853

Hi Chris, please explain?

irishchris

I'm open to being wrong on this but if your cells are rated at 3.2vx14 cells it's only 44.8v. generally batteries are made up in 16s so 16 cells. Pylontechs are 15 but think your pack at 14 cells and you are coming in way too low on battery pack.

Did you only buy 14 cells?

irishchris

Jeez just realised you said your pack was at 57.8v. surely I'm misreading that as that would mean each cell is 4.12v each. Surely I'm missing something as cells would be getting destroyed at that level and don't see any bloating in your cells. What did you top balance each cell to?

meelick

https://www.boards.ie/discussion/comment/119160934#Comment_119160934

I'm using 18650 cells with a nominal voltage of 3.7v and a fully charged voltage of 4.2v. I'm using OPUS BT-C3100 charger to capacity test each cell before use, any dead cells, heaters, or cells that don't hold their charge for at least a week are discarded. 3.7V nominal, not 3.2V.

irishchris

https://www.boards.ie/discussion/comment/119160948#Comment_119160948

Ah ok, from looking at photo it was hard to distinguish as looked like calb cell cases in the mount. Well done on your battery mounting system though, nice and tidy 👍

meelick

https://www.boards.ie/discussion/comment/119160960#Comment_119160960

Ah. I see. That's 1400 individual 18650 cells soldered into 14 packs of 100. :)

10-10-20

https://www.boards.ie/discussion/comment/119160677#Comment_119160677

That battery pack should probably have at least a vapour barrier or at best a vapour & thermally insulative (& fireproof) barrier between it and the wall. That would be especially true if the wall is exterior on the opposite side. Otherwise the wall will hold moisture and cause localised corrosion on the cells.

reklamos

https://www.boards.ie/discussion/comment/119160948#Comment_119160948

The Control parameters screen is actually showing that the battery is discharging at 20Amps? This is ~1kW where is it going(you said it is not dicharging)? According to Lead battery settings on the screen 20Amps is that max tha it can charge/discharge at. Does Daly BMS show that it is charging/discharging.

Since it is lead acid battery profile the currents are limited by set capacity. You can set higher currents if your battery is capable of more than 20Amps: see more https://solis-service.solisinverters.com/support/solutions/articles/44002330432-lead-acid-battery-parameter-settings-for-rhi-and-rai-inverters

I am not familiar with 18650 batteries but you need to verify your min/max values. I think 42V is too low for undervoltage as that would mean all cells would be at 3V which is the miminimum for 18650(If I rememeber correctly) The inverter does not see the exact battery voltage, they may cause cells to drop below 3V.

meelick

https://www.boards.ie/discussion/comment/119162005#Comment_119162005

The 20 amps is the max that I set for testing. It's not actually drawing 20 amps, the BMS app shows 0. Main screen with icon shows 0.01kW. Here's a pic I took at the same time as the other three, at the same time the control settings screen was showing Discharge Current at 20A (should be called Max Discharge Current, I think).

18650's can go as low as 2.8, 3.0 is a reasonable cut-of, but I'd agree that it's a bit low maybe 3.2 might be a good value. Still though the battery in the app shows a discharge of only 10w, enough to keep the metrics going to my grafana overnight via modbus over IP. The daily battery discharge for today shows 0.2kWh, so it's pulling about 10W from the battery, I'm just flummoxed why I cant get it to pull a decent load. :) I'll ping installer, and see if he has any insight. He was very interested in the DIY battery, sent me a mail recently to see if I had it working yet.

graememk

https://www.boards.ie/discussion/comment/119162433#Comment_119162433

Oh Ive ran across what your seeing, I noticed the value it shows going to and from the battery in that picture, in the latest firmware is from the BMS data.

How did I find that out? I had my shunt wired backwards, so bms was reading charging when it actually was discharging. - could explain why you are not seeing any readouts there.

reklamos

https://www.boards.ie/discussion/comment/119162433#Comment_119162433

in picture below battery current is the actual current at that time and then current direction is Charge/Discharge. At least this is what it is showing on my inverter. Every time I get to that screen I see diferent value. Now it could be that that this value is provided by BMS and since there is no BMS in your case, it is displaying set value as you stated.

I see that use have 'Self timed of use' mode configured. Can you check your timers and stop them. Set from Run to stop and set all values(time and Amps to 0) maybe the settings there prevent the battery from discharging. Wrong config there could explain what we are seeing.

If you can, read Solis registers 33133-33152 and drop them to grafana for 12 hours, as this is where most of battery metrics are stored . Registers 33134/ 33135 are battery voltage/current the way inverter sees them. 33142/33143 the way BMS sees them and provide to inverter. This could give better picture what is going on.

meelick

https://www.boards.ie/discussion/comment/119166299#Comment_119166299

In the Lead Acid battery settings there's a min ambient and max ambient set to 5deg and 45deg. Ambient temp currently reading at -000.0degC, so it's outside the range of the battery settings. So that may be a factor. I can't reduce the min temp below +5degC. I'm currently looking to see if I need an additional sensor to update the ambient temp.

Edit: Seemingly there is an optional temp sensor for RHI series inverters when using lead-acid battery systems. So I can try and get my hands on one of those, or try and get the canbus bridge working :)

meelick

So I borrowed a Dyness 4.8kWh battery, and it works perfectly. Charges, discharges, etc. Used a standard network cable for the canbus, Solis inverter and Dyness battery both have rj45 connectors for their canbus port. At least I've proven that the Inverter is OK.

So the research continues to get a Daly BMS working with a Solis RHI hybrid inverter.

graememk

https://www.boards.ie/discussion/comment/119179468#Comment_119179468

Are you planning on building the esp32 device? The translator from daly to CAN?

meelick

https://www.boards.ie/discussion/comment/119180843#Comment_119180843

I have most of the bits here already, I'm just waiting on the CAN module. But there could be a ton of work matching the uart output of the Daly with the CANBUS requirements of the inverter, i.e. emulating a supported battery. But, to be honest, I don't fancy a bunch of electronics between the battery and the inverter, so I've just pulled the trigger on a Seplos BMS from Alibaba, which emulates Pylontech batteries (including canbus connection), and is compatible with Solis inverters. If I have tome to play with the ESP32 before it arrives, maybe I'll get it working. But I can't see me doing anything when the Seplos arrives. :)

Valuable learning, though, pick a compatible BMS to start with.

reklamos

https://www.boards.ie/discussion/comment/119184516#Comment_119184516

Well you should have read Seplos thread because it does not emulate Pylontech(or does it poorly) hence none of us using Pylontech protocol. Also is Seplos going to work with 14S setup?

I got my CAN modules and will try to make CAN device when find time for it.

meelick

https://www.boards.ie/discussion/comment/119184549#Comment_119184549

O sweet jeebus. :) Should I cancel so and wait to try and get the bridge working? I just want the fekking thing to work :)

reklamos

https://www.boards.ie/discussion/comment/119184573#Comment_119184573

Check or ask if you can use 14s configuration. Also verify that it can charge cells to higher voltage as none of using 18650 cell. I do not think it should be a problem but the firmware might restrict that. Pylontech protocol is not a show stopper as it does work with diferent protocol.

meelick

https://www.boards.ie/discussion/comment/119184602#Comment_119184602

Ah Ok. I had ordered the one with 14S and 3.7 nominal cell instead of 3.2 nominal for LFP. I'll read up on that thread first before committing, see what the issues are.

meelick

So, phenomenal progress this evening. I got my wee UART to CAN bus adapter, (about £5 on amazon), and I had a few ESP32 boards lying around. So I used the git repo in the second post on this thread and built a variant of what's in the repo diagrams, no display, no digital isolator (that the impatience in me). A few hours to get the sketch to compile, there's a particular version of the Daly UART lib that's needed. Guru Meditations (crashes) all over the place, but now I'm finally getting places.

This is with the battery set to User Defined, and it's successfully reading the SoC of the BMS though the CAN bus interface and displaying it on the inverter! It's also showing "CAN OK".

It's not doing much else at this stage, but I think this is a huge milestone for me, and I think I'm on the right road to getting my battery working with the Solis inverter, finally. :)

graememk

https://www.boards.ie/discussion/comment/119198750#Comment_119198750

Try the aobo battery setting. That's what the rest of us is using

meelick

https://www.boards.ie/discussion/comment/119198942#Comment_119198942

At this stage I've got my canbus bridge emulating user-defined, aobo (using SMA protocol), and Pylontech (I did the changes to give the correct name and the metrics in the right places according to the pylontech doc above), and I'm still not getting current pulled from the battery. I've verified the battery with a resistive load and can easy pull 10A from the BMS, app shows info fine. Just not working with the Solis no matter what I try.

meelick

https://www.boards.ie/discussion/comment/119205355#Comment_119205355

Got it working. I'm not entirely sure what changes I made to the ESP32 sketch, but I'll analyse that over the coming days. I'm using User Defined, and Inverter now starts charging when SoC is <80%, and discharging when the load requires grid pull, at which time the battery kicks in after a few seconds and supplements the solar (if any). Here's a pic with a 2kW heater, causing draw from the battery.

And a pic of the battery info screen:

The sketch is using v1.0 of the daly uart library, the latest version uses significantly different API, and would not build. I haven't got the MQTT side working yet, but will do soon. And I might update the sketch to the latest daly uart lib api, as it might be able to pull more info from the BMS.

I like the fact that User Defined works, as I can tweak the voltages to suit my 14S100P Li-ion setup.

Thanks to everyone on this thread that helped!

meelick

Here's a quick update now that I have the battery working. I was able to get the battery charging at night between 2-4am (1/4 price night rate), as well as topping up the immersion without interrupting each other. All pulled from the grid, then at 4am switched back to powering the house from the battery. :)

reklamos

Why your battery charge rate is slopping down after about an hour? Is it intentional or is there something that starts limiting it? To make most of your money I would move water heating closer to use time. Having hot water sitting and cooling down is a waste of energy. What doe you use for impersion Eddi/iBoost?

meelick

https://www.boards.ie/discussion/comment/119218958#Comment_119218958

On the charge slope, I believe that's the charge curve of the battery, CVCC, so once it gets to the target voltage, it ramps down the current till it gets to 0.5c rate, at which point it's charged. On my tariff, 2am to 4am is the 1/4 price rate (7.5c/kWh). Anything outside that is 1/2 price rate (15c/kWh till 8am). You might be right on moving the Eddi boost time. I'll try boosting immersion closer to use time, but each unit later in the night costs double, so I'd only want to put it on for half the time, and see if it ends up at the same temperature. I'll give it a go. Yes, Eddi for the immersion.

reklamos

https://www.boards.ie/discussion/comment/119223871#Comment_119223871

Interesting. It starts sloping after 1 hour, so you only put have 3kWh before it started sloping. Since I do not see SOC I assume that the battery was above 50% otherwise it starts slopping too early.

yankinlk

https://www.boards.ie/discussion/comment/119223871#Comment_119223871

Wait. Is anyone complaining of cold showers? Interested for when I am forced to smart meter. ATM I'm day night so I boost the last hour and no one has issues with hot water.

But I'd try what you have setup to save max money if it stil provided comfortable showers

meelick

https://www.boards.ie/discussion/comment/119224042#Comment_119224042

Exactly, I'd estimate the battery is between 11-12kWh, but had only discharged about 40%. I've it set to depth of discharge of 80%, but even then, I'm not hitting anything close to that overnight (at this time of year).

meelick

https://www.boards.ie/discussion/comment/119224074#Comment_119224074

Not so far, decent showers in the morning, but I'd estimate I'm only getting about 3-4kWh of energy into the cylinder. I have 4 temp sensors attached to it, and the bottom never gets much about 20degC, where the top sensor shows about 63degC. Thinking of longer immersion element or getting a willis immersion fitted. I've been told that I could get much more energy into the cylinder that way, up around 6kWh, so a lot more hot water that way. Current setup is not circulating water at all, so it's all sitting at the top. The average of all the sensors drops from about 44devC to 42degC between 4am and 8am.

graememk

https://www.boards.ie/discussion/comment/119224211#Comment_119224211

Or a de-stratificion pump? The relay board for the eddi can control that.

But your right, only the top bit of your tank is getting hot and the rest is sitting cold.

Another thing to check is it wired into the bath element rather than the sink element

meelick

https://www.boards.ie/discussion/comment/119224344#Comment_119224344

Good suggestion, I'll look into a destratification pump as well. Eddi wired into both sink and bath elements, it's definitely set to bath, the sink setting heats far less. My problem in controlling the pump from the eddi is that it's downstairs in the utility, whereas the cylinder is upstairs, and I've no easy way of running a cable to the pump from the Eddi.

graememk

https://www.boards.ie/discussion/comment/119225554#Comment_119225554

Possibly if your looking to disregard the sink element you might be able to reuse the wire for that.

reklamos

https://www.boards.ie/discussion/comment/119224189#Comment_119224189

Why are you charging overnight then when there is plenty juice in the battery and we have sunny days? I only charge at night if battery is low and also based on solcast forecat for the day.

meelick

https://www.boards.ie/discussion/comment/119227999#Comment_119227999

I reckon that the cost of a kWh charge/discharge in/out of the battery costs about the same as the cheap rate between 2am and 4am (in battery lifetime reduction), so I'm using the grid rather than reducing the life of the battery.

bullit_dodger

https://www.boards.ie/discussion/comment/119384056#Comment_119384056

I think you've got the math wrong there meelick. Most batteries have a lifetime duration in the 6,000-10,000 range. Assuming you (fully) cycle the battery everyday that's 365 times a year. So you're looking at ~20 years for some batteries. Even at that even at the end of the published lifetime, the battery isn't dead, just that it will perform at (say) 70% of the initial performance.

Batteries are there to be used - that's why you bought it.

unkel

https://www.boards.ie/discussion/comment/119384142#Comment_119384142

You're mixing up chemistries. Most of us use LiFePO4 which has several thousand cycles, but he is using lithium ion 18650s, which typically have several hundred cycles. So he is very prudent not to cycle the bejaysus out of them! And apart from the cost of the battery, I'm sure he doesn't want to weld together again 1400 batteries every other year 😂