SRNE 12kW IP65 HES and 10kW ASP

[OgreMustCrush]

I think I have a possible theory why the voltage spiked. There are 2 different registers available over RS-485 that appear to refer to the overvoltage protection voltage, though the names/descriptions are ambiguous as to which is the actual overvoltage protection limit that shuts down charging. Anyway, the BatChgLimitVolt was at 57.6 while the BatOverVolt one was set to 60V. Think it might have gotten set that way when I was debugging the RS-485 communication and trying to figure out which values actually work or do anything, that register might have gotten sent invalid data and defaulted to that value. Given its name, maybe it isn't actually an overvoltage protection voltage but rather an "overcharge" voltage. Anyway, I dropped it both it and the other one down to 56V just to be extra certain it doesn't happen again. Neither of those are exposed in the UI, and as far as I can tell the BMS doesn't affect them either.

Did you try it in Inverter Logic with the BMS set to Disable to see how it performed with some sun in that case?

That's a good suggestion; I haven't actually tried changing that charging current logic setting yet. It presumably should default to something when no BMS is connected, I wonder what scheme it uses it that case. The Inverter one seems like the one that is the most ambiguous of the 3 options though, at least the HMI logic tells you it uses the maximum charge current in the Battery Manage screen, and I can see exactly what the BMS is requesting on a second to second basis, but I don't know what algorithm the "machine's derating logic" uses.

I do not know your system, something I learned after 3+ years with Growatt AiO, I stopped using BMS closed circuit with the AiO, went to BMS to the SA (solar assistant), changed the AiO to US2 Lithium no BMS and closed circuit of the AiO to SA. it stopped all the fighting between the AiO and BMS batteries.. letting SA be the man in the middle (hub) so to speak. Everyone is happy including myself.. I see better management and control of the batteries and operation of the AiO..
Now it may not be for you.. but just wanted to share something, you may consider.. wish you the best.

Yeah, I get that, I did something similar with my old inverter for a while before I finally found the magic settings where BMS comms didn't mess anything up. But Solar Assistant's support for this inverter is really poor right now, most of the inverter settings are missing and half the ones that are there don't work, so I can't really rely on it to automate anything. I'm probably going to eventually move to doing my automation through Home Assistant once I have a robust enough RS-485 monitoring/control solution setup for this inverter, but until I get to that point, I'm limping by with the inverters built in features which requires it to know SoC to do anything useful.

I'm not asking anyone to solve an issue with my setup, I'm working through that myself just fine, more providing feedback about what I am seeing and asking if anyone else has seen anything similar. Haven't seen anyone else post any end-of-charge voltage/current plots w/ from an HESP since I brought this up yet. Would love to see a nice clean one with no cycling so I can find out what at least one set of known good settings are.

 

[OffGridForGood]

It presumably should default to something when no BMS is connected, I wonder what scheme it uses it that case.

It defaults to the battery-set up settings you input - or the default factory ones based on which battery chemistry/cell-count you input.
This is also user-settable - "USE" - ie user-battery settings and input your own parameter values - the default tail current is 2Amp

Were you able to check if any cells in the battery packs may be hitting OVP at the cell voltage level? ie cutting off charging, then the cell voltage drops and charging resumes? - you noted the cells are not well balanced 'yet' and I suspect they can't actually balance under the circumstances, ie not holding high enough pack voltage long enough for balancing to take place.
The bench source will likely be tiny current compared with your PV array, so in that case the runner cell is less likely to take off on you and you don't see the jigsaw charge effect with the bench source. - just guessing at this point - can you confirm the pack cell voltages during the erratic charging behaviour?

 

[OgreMustCrush]

It defaults to the battery-set up settings you input - or the default factory ones based on which battery chemistry/cell-count you input.
This is also user-settable - "USE" - ie user-battery settings and input your own parameter values - the default tail current is 2Amp

Were you able to check if any cells in the battery packs may be hitting OVP at the cell voltage level? ie cutting off charging, then the cell voltage drops and charging resumes? - you noted the cells are not well balanced 'yet' and I suspect they can't actually balance under the circumstances, ie not holding high enough pack voltage long enough for balancing to take place.
The bench source will likely be tiny current compared with your PV array, so in that case the runner cell is less likely to take off on you and you don't see the jigsaw charge effect with the bench source. - just guessing at this point - can you confirm the pack cell voltages during the erratic charging behaviour?

The packs were both initially balanced, the level of balance degraded with the cycling since I switched to this inverter but has been improving after I tweaked the voltage so it keeps the pack voltage above 55V during cycling. I also did some manual balancing when I had the bench supply hooked up to each pack (brought pack to 55.2V then charge individual low cells to 3.45V with seperate supply until current drops below 1A.) . Haven't been hitting cell OVP as far as I can tell other than when the inverter took off to nearly 60V earlier today.

 

[OffGridForGood]

The packs were both initially balanced, the level of balance degraded with the cycling since I switched to this inverter

what inverter did you have before this one?

but has been improving after I tweaked the voltage so it keeps the pack voltage above 55V during cycling. I also did some manual balancing when I had the bench supply hooked up to each pack (brought pack to 55.2V then charge individual low cells to 3.45V with seperate supply until current drops below 1A.)

what cell voltage did you set in the JK's to start balancing ?
typically we recommend a start balancing cell voltage of 3.42 (for 16s this is 54.72 pack voltage) because below 3.42 maybe 3.40 balancing can cause un-balancing - the cells need to be high enough voltage to be out of the flat linear part of the charge curve.

I have my balance start set to 3.42 v/cell in the JK's and the max pack voltage during charging set to 56.8
After months of typical Winter - poor charging seldom hitting 100% SOC- I have two packs I pull and manually balance 'close enough' to allow the BMS to do the rest. The recent return of the sunny days, the BMS's take care of the packs on their own again now.

 

[OgreMustCrush]

what inverter did you have before this one?

what cell voltage did you set in the JK's to start balancing ?
typically we recommend a start balancing cell voltage of 3.42 (for 16s this is 54.72 pack voltage) because below 3.42 maybe 3.40 balancing can cause un-balancing - the cells need to be high enough voltage to be out of the flat linear part of the charge curve.

I have my balance start set to 3.42 v/cell in the JK's and the max pack voltage during charging set to 56.8
After months of typical Winter - poor charging seldom hitting 100% SOC- I have two packs I pull and manually balance 'close enough' to allow the BMS to do the rest. The recent return of the sunny days, the BMS's take care of the packs on their own again now.

Old inverter was a Voltronic Axpert Max Ii (basically a LV6548, which is an Axpert Max I)

I've moved the balance voltage a few times, but currently at 3.42V. I prefer to run my charge voltage near 55.2 (3.45V/cell) to minimize voltage related degradation at higher cell voltages and that should still be well into knee, just takes maybe a bit longer to absorb. But it is low enough that if the inverter is cycling it cells drop below the balance threshold pretty quickly.

In theory I should be able to get to 100% everyday (at least until I get the additional 32kWh of battery I am in the process of ordering) unless it is super cloudy with my current setup, and even then I fill in with grid charging to make sure I have battery capacity during peak power rates. Still tweaking my settings on that, I have 7.6kwh of grid-tied solar AC coupled off the grid connection of the inverter, but the HESP doesn't seem to use it to charge the battery automatically based on the CT readings, I have to force grid charging during peak solar hours and I inevitably either import or export some grid power depending on the production that day. it might be better putting the grid-tied system on then gen input w/ the microinverter input setting, going to try that soon.

 

[OffGridForGood]

I've moved the balance voltage a few times, but currently at 3.42V. I prefer to run my charge voltage near 55.2 (3.45V/cell) to minimize voltage related degradation at higher cell voltages and that should still be well into knee, just takes maybe a bit longer to absorb. But it is low enough that if the inverter is cycling it cells drop below the balance threshold pretty quickly.

not aware of any damage/degrdation that will occur below 3.60v per cell. As far as I can see from battery research the cells will age out first.
I started with 3.45v myself - based on the comments from Andy and off-grid-garage, but the cells didn't balance at these low numbers, and they drift farther and farther with time. As I built more and more ESS capacity the problems only compounded. I am happy with 56.8v (3.55/cell) and how the packs/cells hold balance now. I wouldn't worry about pushing them a bit higher if I saw a need, but this setting seems to be sufficient. - I suggest you keep an eye on your own cells/packs see how they do over a period of time, and maintain an open mind.

 

[OgreMustCrush]

Anyone seeing this sort of charge cycling near 100% SOC on the HESP? Seems to ramp up for some amount of time, then suddenly switch to discharge mode, and then the cycle repeats. Seems to do it on both grid or solar charge, cycling gets more frequent the more fully charged the battery is. Doesn't seem to be influenced by whether using BMS communication or running open loop or the actual set battery charge voltage though. It doesn't appear to be a data acquisition issue as I can see the same cycling on the inverters UI with the charge state toggling between quick charge, constant voltage charge, and charge off. It's making it hard for my battery to successfully balance since the voltage keeps dropping below the threshold every time it drops back into discharge mode. It isn't the BMS(s) as far as I can tell, they keep both charge and discharge enabled mode continually and not seeing any cells get high enough to trigger over voltage protection.

(Note the drop in SOC is from me disabling BMS communication and raising the charge limit settings in the inverter, otherwise would have stayed at 98% like it did for the first 35 minutes shown. Also note that negative current in this case is charging)

View attachment 276468
Previous hour (not as close to full so less frequent cycling):
View attachment 276470

I just want to put in an update that I'm now having good results on the HESP with running closed loop with a slightly increased bulk voltage (to 56V, which really seems to result in the inverter charging up to about 55.8V) from what I was using previously, and using Inverter Logic selected for the Battery Charge Current Limit setting. What I've discovered is that the behavior of the Inverter logic is to follow the BMS/HMI current limit until the voltage/SOC reaches certain points, at which point it progressively decreases the allowable current as you get closer to 100%. This pretty much eliminates any cycling, as when the battery is near 100% it will only charge at 10A max since the voltage/current doesn't fluctuate anywhere as much during the absorption phase (which was resulting in the inverter dropping into "Not Charging" state and cutting off charge.) Since the cycling goes away, the cell voltages stay consistently high enough in the voltage knee to balance effectively.

The Inverter Logic setting appears to limit current to 100A at 85% SOC, 50A at 90%, 20A at 95%, and 10A at 98%, with likely some equivalent voltages approximations based on the bulk voltage setting when running open loop. One downside is it does make that that last 15% take quite a while to charge up (from less than an hour to over 3h on my pack,) and if you don't either a useful load or grid sell to pump any excess solar to during that period, it might not be the best strategy. It also might be a bit too aggressive of limit with larger battery banks, even on my ~30kWH it seems like a bit much dropping down all the way below 0.1C at 90%.

 

[TK1979]

I just want to put in an update that I'm now having good results on the HESP with running closed loop with a slightly increased bulk voltage (to 56V, which really seems to result in the inverter charging up to about 55.8V) from what I was using previously, and using Inverter Logic selected for the Battery Charge Current Limit setting. What I've discovered is that the behavior of the Inverter logic is to follow the BMS/HMI current limit until the voltage/SOC reaches certain points, at which point it progressively decreases the allowable current as you get closer to 100%. This pretty much eliminates any cycling, as when the battery is near 100% it will only charge at 10A max since the voltage/current doesn't fluctuate anywhere as much during the absorption phase (which was resulting in the inverter dropping into "Not Charging" state and cutting off charge.) Since the cycling goes away, the cell voltages stay consistently high enough in the voltage knee to balance effectively. The Inverter Logic appears to limit current to 100A at 85% SOC, 50A at 90%, 20A at 95%, and 10A at 98%, with likely some equivalent voltages approximations based on the bulk voltage setting when running open loop.

View attachment 283103View attachment 283104

Very interesting and good to know!

 

[42OhmsPA]

I just want to put in an update that I'm now having good results on the HESP with running closed loop with a slightly increased bulk voltage (to 56V, which really seems to result in the inverter charging up to about 55.8V) from what I was using previously, and using Inverter Logic selected for the Battery Charge Current Limit setting. What I've discovered is that the behavior of the Inverter logic is to follow the BMS/HMI current limit until the voltage/SOC reaches certain points, at which point it progressively decreases the allowable current as you get closer to 100%. This pretty much eliminates any cycling, as when the battery is near 100% it will only charge at 10A max since the voltage/current doesn't fluctuate anywhere as much during the absorption phase (which was resulting in the inverter dropping into "Not Charging" state and cutting off charge.) Since the cycling goes away, the cell voltages stay consistently high enough in the voltage knee to balance effectively.

The Inverter Logic setting appears to limit current to 100A at 85% SOC, 50A at 90%, 20A at 95%, and 10A at 98%, with likely some equivalent voltages approximations based on the bulk voltage setting when running open loop. One downside is it does make that that last 15% take quite a while to charge up (from less than an hour to over 3h on my pack,) and if you don't either a useful load or grid sell to pump any excess solar to during that period, it might not be the best strategy. It also might be a bit too aggressive of limit with larger battery banks, even on my ~30kWH it seems like a bit much dropping down all the way below 0.1C at 90%.

View attachment 283103View attachment 283104

I use an automation in Solar Assistant to do similar, taper current as voltage rises, with my ASPs; I'm running open loop.
I wonder if you could do the same thing with the HESP or if closed loop would override?

All in all are you happy with it? Are you selling any power back?

 

[OgreMustCrush]

I use an automation in Solar Assistant to do similar, taper current as voltage rises, with my ASPs; I'm running open loop.
I wonder if you could do the same thing with the HESP or if closed loop would override?

All in all are you happy with it? Are you selling any power back?

Not sure if there is enough functionality on the HESP in Solar Assistant to do that or not, there are several settings that don't work on it the same as they do other SRNE inverters and a lot of the HESP specific settings haven't been implemented in Solar Assistant. I stopped using it a while ago and have instead been contributing to https://github.com/timbit123/srne-modbus/ for connecting to Home Assistant over MQTT. I believe if you set the Charge Current Limit to HMI logic it will use the current limit set in the inverter over what the BMS sends. In my case, I'm actually overriding the native BMS logic anyway with https://github.com/Sleeper85/esphome-yambms combining my separate BMS, so I could programatially change what the "BMS" sends as well (it actually has a feature to ramp down current in YamBMS already, but it isn't aggressive enough to make much difference.) Might do that eventually, at least once I get more batteries.

I'm fairly happy with the HESP, been working through its quirks, which are mostly just poorly documented features where I have to find out exactly how they work since it isn't spelled out in the manual. Like that charge current mode, or how the inverter handles AC coupled grid tied inverters. My original assumption was it would use the CTs to detect other inveters on the grid side and could pull power from them to charge the battery to limit the export to whatever the sell limit is, but it doesn't seem to charge automatically from grid-tied when it is on grid side behind CT. I rerouted the grid-tied into the generator input using the "Micro inverter input" setting and that will charge from that power automatically, but only when the grid connection is down, it seems to otherwise perfer to let the ac-coupled inverter export to the grid rather than consume the power into the battery with no way to change that option. Basically have to instead program the inverter to charge "from grid" when the grid-tied is producing power, but obviously can't perfectly match the current production with the grid charge limit at all times. In my case since I normally have some sell back anyway (from both the grid-tied system but also the panels directly connected to the HESP) once my battery fills up so it's not critical to get perfect, just having to tune it so it doesn't inadvertently charge from the actual grid too much (and even some of that is fine during off-peak on cloudy days since I'm trying to make sure the battery has enough charge to cover on-peak periods) and/or sell back too much when I'd rather it charge my battery since the current sell-back rates are about half what I pay to import.

 

[42OhmsPA]

Thanks for the detailed follow up.

 

[uslsml]

I needed to adjust my programmed charge/discharge schedule last night and decided to use the opportunity to test the daylight savings time transition. I adjusted my schedule to create a transition at 2:30 AM. Thankfully the ASP did perform the transition when the time jumped from 2 AM to 3 AM. The ASP passed this single-transition test case. I did not try creating multiple transitions between 2 AM and 3 AM to see if the ASP correctly chooses the correct state at 3 AM.

The fall time change would present different test cases. If there is programmed transition between 1 AM and 2 AM, after the time falls back from 2 AM to 1 AM, would the ASP repeat the transitions between 1 AM and 2 AM, or would the ASP avoid repeating the transitions a second time? I think I'd be OK with either choice.

 

[TK1979]

ASP and SPI owners - what would you do with hooking up a 22kw generac (for when needing to charge up the batteries)?

A - wire the generator to a loads panel, then a few plugs for chargeverters (which would charge up the batteries when low voltage hits?)
B - wire the generator directly into the ASPs them selves and have startup controlled by the drycontacts (when low voltage hits)?

Has anyone tried either of these options? Or something else I am not thinking of?

 

[Zwy]

I would use the Chargeverter method. The Chargeverters will have clean power and ramp up instead of instant on. The ASP would have to go to bypass to charge which means any PV would go to the batteries instead of loads. Which means either you have to change settings in the inverter or somehow put it in bypass. Using the Chargeverters allows you to charge any time you want and at any charge rate easily.

 

[42OhmsPA]

I didn't even connect the grid to mine.
Rectifier(s) charge when the sun is crap.

 

[TK1979]

Ok - my ATS is 200A service entrance rated - I'll be able to add some breakers into this panel and wire in plugs for the chargeverters (what is good about this is I'll be able to use grid or genny as the breakers are after the switch)

 

[mrmanny]

Hi everyone I have a question on my SPI-10K-UP:
How can I get full 120v output from grid when only connecting L1 + N + G ?
Is that even possible , as of now I cannot even change 01: UTI to SOL
Please help.

 

[42OhmsPA]

Hi everyone I have a question on my SPI-10K-UP:
How can I get full 120v output from grid when only connecting L1 + N + G ?
Is that even possible , as of now I cannot even change 01: UTI to SOL
Please help.

So you want to feed 120V and get 10kW 120V out. No split phase 120V/240V?

You'll need to change setting 68 to 0.

You'll need to parallel the input cable to L1 and L2. You'll also need to parallel the output cable.

 

[Zwy]

Hi everyone I have a question on my SPI-10K-UP:
How can I get full 120v output from grid when only connecting L1 + N + G ?
Is that even possible , as of now I cannot even change 01: UTI to SOL
Please help.

I have to ask why?

 

[mrmanny]

@420Heysus - thanks for the response , think Im still somewhat lost.
1. You'll need to change setting 68 to 0.
- Inverter doesn't let me do that.
2. You'll need to parallel the input cable to L1 and L2. You'll also need to parallel the output cable.
-When you say parallel the inputs - do you mean like do a jumper cable from L1 to L2 ?
-Same for the output ?

@Zwy -
All I need to run is 110v Devices.
Don't need anything 240v , was thinking it would be easier then running another cable for the 2 legs.

My current setup btw.
https://drive.google.com/file/d/1y_qQS12hdfcUdtzGmLLWJzcuLd3saYg3/view?usp=drive_link, https://drive.google.com/file/d/1yXODauNTK97qWkfOJQ8Cq4cnNgLZOC45/view?usp=drive_link,
https://drive.google.com/file/d/1yVMrEGQ_-o2HOGAcB1fv7iR646lr81zP/view?usp=drive_link,

20250401_153249.mp4

drive.google.com

 

[42OhmsPA]

thanks for the response , think Im still somewhat lost.
1. You'll need to change setting 68 to 0.
- Inverter doesn't let me do that.

Is it in standby? Rocker switch off, enter menu before turning off.

2. You'll need to parallel the input cable to L1 and L2. You'll also need to parallel the output cable.
-When you say parallel the inputs - do you mean like do a jumper cable from L1 to L2 ?
-Same for the output ?

Correct.

@Zwy -
All I need to run is 110v Devices.
Don't need anything 240v , was thinking it would be easier then running another cable for the 2 legs.

My current setup btw.
https://drive.google.com/file/d/1y_qQS12hdfcUdtzGmLLWJzcuLd3saYg3/view?usp=drive_link, https://drive.google.com/file/d/1yXODauNTK97qWkfOJQ8Cq4cnNgLZOC45/view?usp=drive_link,
https://drive.google.com/file/d/1yVMrEGQ_-o2HOGAcB1fv7iR646lr81zP/view?usp=drive_link,

20250401_153249.mp4

drive.google.com

I would advise to run a traditional 120/240V panel, balance the loads between the legs and call it a day.

Your AC cables look way undersized, what gauge are they?

 

[GuyG]

Your AC cables look way undersized, what gauge are they?

I noticed that too. 12 AWG is my guess.

 

[Zwy]

@Zwy -
All I need to run is 110v Devices.
Don't need anything 240v , was thinking it would be easier then running another cable for the 2 legs.

Running 120V only requires larger wires than running 240V. Running 240V for a 40A load means you can use 12AWG wire on each leg and a neutral. Using 120V for a 40A load means using 8AWG for both hot and neutral wires. That gets expensive and is a waste of copper.

Your photos show some very small wire on the AC output. The minimum needed is 6AWG which is the largest that will fit in the terminals. You have to run 6AWG to a 60A breaker in a panel, then you can downsize wire size with the proper circuit protection such as a 20A breaker for a 20A outlet with 12AWG wire. At a minimum, there needs to be a breaker on the inverter AC output with 6AWG wire, then any circuit off that needs proper over current protection for the ampacity of the wire.

If you decide to stick with 120V only, and parallel the inputs and outputs, you will still need to run 6AWG to a breaker panel with a pair of wires from inverter L1 and L2 and combine in that box. Each "leg" can only have 6AWG to fit the terminals and backfeed into a pair of 60A single pole breakers in a 125A box. You can't just combine at the inverter as each leg is actually a seperate inverter inside and requires it's own circuit protection. Same thing on the AC input, you have to split the power into each inverter leg. This would require a 125A box with a pair of 60A single pole breakers and can only use 1/2 of the panel.

You won't save anything as you still have to run basically the same amount of wire and install the same number of breakers as 240V.



This needs to be corrected also. Get the cables centered in the hole.

 

[mrmanny]

Is it in standby? Rocker switch off, enter menu before turning off.

Right now its on Solar/Battery Mode according to Solar Assist , and Grid is showing 62.1V but when I click in Inverter it shows whats in the pic bellow:

I would advise to run a traditional 120/240V panel, balance the loads between the legs and call it a day.

I think this is what I will end up doing.
Your AC cables look way undersized, what gauge are they?
Currently running 10 gauge in 12 out, again , only 110v devices will change to 6 as adviced.

Just to double check ,
Attached all my connections.

@Zwy - will do what I can from my side, might have to get an electrician out here...

 

[Zwy]

@Zwy - will do what I can from my side, might have to get an electrician out here...

Don't worry about those panels yet, you need to install a 125A panel at the inverters. If you haven't seen my shop build thread videos, I strongly suggest checking out the entire build.

Here is my installation of the combiner panel, be certain L1 of each inverter goes to the same phase in the panel and L2 the same phase in the panel. Originally I hung a 100A panel but decided to install the 125A shown in this part of the first video.

I see many things that make me cringe a little. The wires run thru the bare holes in this panel is a no no. It could start a fire or make the box hot, you come along, touch the box and next thing you know you are being loaded in an ambulance or worse. Don't think things like that don't happen. I ran some 12AWG NMB in my work trailer build I've been building the last 3 weeks and one Romex clamp at a junction box just happened to push the ground wire into the hot wire in one tiny little spot. Plugged the trailer in and the GFCI did exactly what is should do, it tripped.

It's no fun getting that slight burning sensation. I did 2 years ago when I was moving wires from a live box to another box. Slightly touched the neutral wire after disconnecting it from the neutral busbar. The problem was it was still hot as I had turned off and disconnected the wrong breaker. Probably only about the 3rd time in over 40 years I ever got the buzz. The time before was this incident.