July 28, 20223 yr Author On 2022/07/25 at 12:32 PM, Razeen said: What battery settings is best to put on the inveter? ie. Float V, Absorption V, Equalisation, Shutdown, Low etc... I basically want the settings so my battery remains full and only in use once there is loadshedding. then only use 70-80% of the battery(to try and extend the life). Does the battery communicate with the Deye..I am seeing alot of posts on this with mixed answers? What will my best solution be to let them comminicate in Sync... I have noticed that the Inverter shows a lower % SOC vc the Actual Battery LCD screen when running. I have Float set to 52.5V and Absorption set to 53.5V. I used to have them on 51.5V and the battery never fully charged. This is for a Narada 48NPFC100. It looks like the Deye inverter can communicate with the Naradata battery uses a Shinwa BMS however the Tian Power BMS (which is what I have) is not supported (yet). You can check a few pages back on which wires to crimp if your battery uses a Shinwa BMS. The SOC on the inverter will definitely vary from the true SOC reported by the BMS if there is no BMS comms - mine has varied by up to 30%. On 2022/07/25 at 4:33 PM, mrboss said: do you mr solarconvert got your issue solved or not yet?? Not yet. The last contact I had with Deye support was that they would consider adding support for Tian Power BMS for Narada batteries if there was enough demand. I gave them the info that they asked for and have not heard anything since. 8 hours ago, Kevin26 said: Hi. Has anyone had luck in communication with the Narada bms? I'm sitting withe the same problem with 2 batteries that u purchased from what now appears to be a suspect website and I can get my Sunsynk to see the batteries.... There have been several people on this thread that have got the BMS comms to work, but if you have a Tian Power BMS on your Narada batteries you're probably out of luck for now. You can try to find out what BMS you have by either crimping the inverter-to-battery cable according to the posts on the previous pages to see if it works or by trying out the two types of BMS software attached on the previous pages and see which works for you.
August 4, 20223 yr As most of you will know from this thread... Once / if you have the Narada working it takes up the BMS485 port which is what Solar Assistant and ICM use to communicate with the inverter if you dont want to use RS232 and / or give up the Data Logger. I have 2 Sunsynk installs... My one with the Narada batteries and my parents with a Hubble on CAN bus. Theirs has both ICM and the Data logger and it is really nice to have both options. So after some googling and failed attempts to split the RS485 port due to different comm protocols required for the batteries vs ICM... I came across a RS232 splitter and am hoping that this will solve my "problem" - http://www.mt-viki.net/product/33-en.html Has anyone had any experience with these type of splitter? Everything in theory says it should work and I have ordered it from Takealot, so if it does not.. Back it goes 😁. Sadly it is on a 5 to 7 working day lead time, so I will have to wait a little while to provide feedback
August 9, 20223 yr Hi Am having some similar trouble just trying to read data from my Narada 48npfc100 and could use some help / insight I have made up a cable for testing purposes that uses a usb -> serial then rs485 -> rj45. Incidentally a homemade rs232 cable seems to provide zero response at all. I have obtained the Narada software mentioned previously in this thread, but it cannot connect, it gives comms errors. Using cutecom I am able to send commands and see responses, but they dont make sense, so could use some eyes if I am missing something obvious. I have also noted I get different responses if I set to baud 1200 vs baud 9600 .. so am a bit at a loss. below see some examples, and hopefully someone can give an opinion, (dip switch 1 is up, all others down, and it is a single battery) baud request response 1200 27040FFF000105E8 00 6f 77 ff ff ff eb df 9600 27040FFF000105E8 fc e0 f0 1200 7e010100fe0d fd fd ff fb 7d 9600 7e010100fe0d f0 e0 None of those look close to what people where getting from either BMS ? or am I just not seeing it ? I have also noticed on the supplied battery rj45 cable, the end labelled battery has wires only on pin 3 and 5? Any direction would be helpful.
August 9, 20223 yr On 2022/07/10 at 9:15 PM, SolarConvert said: Ok I have done some more serial port sniffing. I should have done this long ago - I would have saved myself countless hours and wasted RJ45 cable crimps. On the Deye inverter, if I set lithium mode to 16, it emits the following with a 1 second gap in between each line: 27 04 0f ff 00 09 04 2e 28 04 0f ff 00 09 04 d1 29 04 0f ff 00 09 05 00 2a 04 0f ff 00 09 05 33 2b 04 0f ff 00 09 04 e2 2c 04 0f ff 00 09 05 55 2d 04 0f ff 00 09 04 84 2e 04 0f ff 00 09 04 b7 2f 04 0f ff 00 09 05 66 30 04 0f ff 00 09 07 09 31 04 0f ff 00 09 06 d8 32 04 0f ff 00 09 06 eb 33 04 0f ff 00 09 07 3a 34 04 0f ff 00 09 06 8d And then this is repeated from the start. Knowing this now, I can see how the inverter would have never communicated with my Narada battery with a Tian Power BMS with lithium mode 16. I take it the above commands are Shinwa BMS ones, not Tian Power ones. The inverter appears to be scanning for a Shinwa BMS. I still intend on creating the separate thread about my Narada with the Tian Power BMS, but suffice is to say, the BMS responds to the following totally different commands: 7e 01 01 00 fe 0d 7e 01 06 00 fc 0d 7e 01 33 00 fe 0d 7e 01 42 00 fc 0d 7e 01 45 00 fe 0d 7e 01 dc 03 06 00 00 c2 0d Now if I switch the inverter to lithium mode 9, the scanning commands more closely resemble the ones that my battery's BMS responds to: 7e 01 43 00 fe 0d 7e 32 32 30 32 34 41 34 32 45 30 30 32 30 31 46 44 32 37 0d 7e 02 43 00 fc 0d 7e 32 32 30 33 34 41 34 32 45 30 30 32 30 31 46 44 32 36 0d 7e 03 43 00 fa 0d 7e 32 32 30 34 34 41 34 32 45 30 30 32 30 31 46 44 32 35 0d 7e 04 43 00 fc 0d 7e 32 32 30 35 34 41 34 32 45 30 30 32 30 31 46 44 32 34 0d 7e 05 43 00 fe 0d 7e 32 32 30 36 34 41 34 32 45 30 30 32 30 31 46 44 32 33 0d 7e 06 43 00 fc 0d 7e 32 32 30 37 34 41 34 32 45 30 30 32 30 31 46 44 32 32 0d 7e 07 43 00 f2 0d 7e 32 32 30 38 34 41 34 32 45 30 30 32 30 31 46 44 32 31 0d 7e 08 43 00 fc 0d 7e 32 32 30 39 34 41 34 32 45 30 30 32 30 31 46 44 32 30 0d 7e 09 43 00 fe 0d 7e 32 32 31 30 34 41 34 32 45 30 30 32 30 31 46 44 32 38 0d 7e 0a 43 00 fc 0d 7e 32 32 31 31 34 41 34 32 45 30 30 32 30 31 46 44 32 37 0d 7e 0b 43 00 fa 0d 7e 32 32 31 32 34 41 34 32 45 30 30 32 30 31 46 44 32 36 0d 7e 0c 43 00 fc 0d 7e 32 32 31 33 34 41 34 32 45 30 30 32 30 31 46 44 32 35 0d 7e 0d 43 00 fe 0d 7e 32 32 31 34 34 41 34 32 45 30 30 32 30 31 46 44 32 34 0d 7e 0e 43 00 fc 0d 7e 32 32 31 35 34 41 34 32 45 30 30 32 30 31 46 44 32 33 0d 7e 0f 43 00 e2 0d 7e 32 32 30 31 34 41 34 32 45 30 30 32 30 31 46 44 32 38 0d The above repeat from the start until a battery is found. I will go into more detail around what the Tian Power commands do once I have a solid understanding, but at least I can now prove that Deye/Sunsynk lithium mode 16 is not for Narada batteries with a Tian Power BMS. I don't know if I should be sad or happy at this point. Just in case you didnt see it, the post by ziva on June 4th, includes a zip file. In that zip is a doc file that has commands similar to yours. It also includes the software, so if you want to reverse engineer, I'm sure that would be useful (If you havent seen it already obviously), and looks like it will work with your battery
August 19, 20223 yr Author On 2022/08/09 at 6:06 PM, jetlee said: Hi Am having some similar trouble just trying to read data from my Narada 48npfc100 and could use some help / insight I have made up a cable for testing purposes that uses a usb -> serial then rs485 -> rj45. Incidentally a homemade rs232 cable seems to provide zero response at all. I have obtained the Narada software mentioned previously in this thread, but it cannot connect, it gives comms errors. Using cutecom I am able to send commands and see responses, but they dont make sense, so could use some eyes if I am missing something obvious. I have also noted I get different responses if I set to baud 1200 vs baud 9600 .. so am a bit at a loss. below see some examples, and hopefully someone can give an opinion, (dip switch 1 is up, all others down, and it is a single battery) baud request response 1200 27040FFF000105E8 00 6f 77 ff ff ff eb df 9600 27040FFF000105E8 fc e0 f0 1200 7e010100fe0d fd fd ff fb 7d 9600 7e010100fe0d f0 e0 None of those look close to what people where getting from either BMS ? or am I just not seeing it ? I have also noticed on the supplied battery rj45 cable, the end labelled battery has wires only on pin 3 and 5? Any direction would be helpful. Hi @jetlee, Did you try both software archives from this thread - the one from zivva and the one I posted? The one from zivva did not work for me but the one I posted did. I'd keep the baud at 9600 and try out all of these requests because with Lithium mode 16 the inverter tries all of them out: 27 04 0f ff 00 09 04 2e 28 04 0f ff 00 09 04 d1 29 04 0f ff 00 09 05 00 2a 04 0f ff 00 09 05 33 2b 04 0f ff 00 09 04 e2 2c 04 0f ff 00 09 05 55 2d 04 0f ff 00 09 04 84 2e 04 0f ff 00 09 04 b7 2f 04 0f ff 00 09 05 66 30 04 0f ff 00 09 07 09 31 04 0f ff 00 09 06 d8 32 04 0f ff 00 09 06 eb 33 04 0f ff 00 09 07 3a 34 04 0f ff 00 09 06 8d Your battery seems to respond better to the Lithium mode 16 commands than to the Revov/Tian Power ones. Do you have a Deye/Sunsynk inverter? If so it would probably be easiest to crimp a cable as per the first few pages of this thread and try out Lithium mode 16 whilst the inverter is connected to the battery. My battery uses the Tian Power BMS so the Shinwa software doesn't work for me.
August 19, 20223 yr Author On 2022/08/04 at 1:35 PM, Jay-Dee said: As most of you will know from this thread... Once / if you have the Narada working it takes up the BMS485 port which is what Solar Assistant and ICM use to communicate with the inverter if you dont want to use RS232 and / or give up the Data Logger. I have 2 Sunsynk installs... My one with the Narada batteries and my parents with a Hubble on CAN bus. Theirs has both ICM and the Data logger and it is really nice to have both options. So after some googling and failed attempts to split the RS485 port due to different comm protocols required for the batteries vs ICM... I came across a RS232 splitter and am hoping that this will solve my "problem" - http://www.mt-viki.net/product/33-en.html Has anyone had any experience with these type of splitter? Everything in theory says it should work and I have ordered it from Takealot, so if it does not.. Back it goes 😁. Sadly it is on a 5 to 7 working day lead time, so I will have to wait a little while to provide feedback Did you have any success with this? I am contemplating whether to embark on trying to convert CANBUS to RS485 with my PI if it is at all possible. Basically trying what out what this thing does https://powerforum-store.co.za/products/canbus-to-rs485-converter What I have also found, unfortunately, is if my battery does not get fully charged for a long time, the BMS loses track of the charge state of the battery. It has happened several times where it gets so bad that the battery is essentially flat (running at 44V only) whilst the LED lights still show full or above 75% (4 full lights). This is a really bad spot to be in, the BMS should not do this, however it is not unique to this BMS. Pylontech actually notes it in their CAN protocol document as well: Quote **Request full charge: Reason: when battery is not full charged for long time, the accumulative error of SOC calculation will be too high and may not able to be charged or discharged as expected capacity. Logic: if SOC never higher than 97% in 30 days, will set this flag to 1. And when the SOC is ≥ 97%, the flag will be 0. How to: we suggest inverter to charge the battery by grid when this flag is 1. So the only reliable way for me to know whether the battery is flat or full is to check the battery voltage - 44V is flat, 54V or thereabouts is fully charged. So much for the emphasis on the BMS whereas it can actually be quite inaccurate if the battery is not fully charged for a long time.
August 19, 20223 yr 5 hours ago, SolarConvert said: Did you have any success with this? I am contemplating whether to embark on trying to convert CANBUS to RS485 with my PI if it is at all possible. Basically trying what out what this thing does https://powerforum-store.co.za/products/canbus-to-rs485-converter Sadly not, the splitter did not work at all with either the Pi or Data logger or both. It also did not power up the logger, so it went back. I am going to try and test with with a windows machine, I just need to get a few more USB to Com cables. If that works, then I am going to dig properly to see if I can get a Raspberry Pi based virtual serial port splitter I thought of going CAN to RS but it comes to the same problem as splitting the RS485 or tapping into the 2nd pair on the port. ICM and Solar Assist both only work when the LiBMS mode is set to 0 where as Narada's need the mode to be 16. The moment you change this more, they both receive CRC errors. If you are going to convert CAN, it is still going to be on mode 16 which will be a problem. Hence my theory of splitting RS232. My next option was to use the 2nd pair on the RS485 port and try and convince Mannie to get ICM to talk to the inverters on mode 16
August 19, 20223 yr 6 hours ago, SolarConvert said: What I have also found, unfortunately, is if my battery does not get fully charged for a long time, the BMS loses track of the charge state of the battery. It has happened several times where it gets so bad that the battery is essentially flat (running at 44V only) whilst the LED lights still show full or above 75% (4 full lights). This is a really bad spot to be in, the BMS should not do this, however it is not unique to this BMS. Pylontech actually notes it in their CAN protocol document as well: So the only reliable way for me to know whether the battery is flat or full is to check the battery voltage - 44V is flat, 54V or thereabouts is fully charged. So much for the emphasis on the BMS whereas it can actually be quite inaccurate if the battery is not fully charged for a long time. So funnily enough, about a week ago I noticed the same issue. I used to have this problem on my Axpert Inverters and to solve it, I needed to fully charge the batteries to max voltage. I tried this over the last few days on the Sunsynk but it stops when the BMS reports 100%.. In my case, this is currently at 50v. From reading on this forum, the way around is to set the inverter to AGM and do a full charge via voltage and then set back. This is going to be a very annoying issue in winter months. Any ideas?
August 20, 20223 yr Author 13 hours ago, Jay-Dee said: I thought of going CAN to RS but it comes to the same problem as splitting the RS485 or tapping into the 2nd pair on the port. ICM and Solar Assist both only work when the LiBMS mode is set to 0 where as Narada's need the mode to be 16. So my idea would be to use the CAN port of the inverter to communicate with the RS485 port of the battery, using Lithium mode 0, with the PI doing the translation. The RS485 port on the inverter therefore frees up for monitoring the inverter. This would however mean in your case that you’d need two PIs - one for the translation and one for monitoring, because Solar Assistant does not allow the PI to be used for any other purpose. In my case however, I use Home Assistant for all of my automation and reporting needs, so one PI would suffice for all of the above. The above is all based on the assumption that CAN to RS485 translation can be done on the PI, but I don’t see why it couldn’t be done. As mentioned though, I am not sure whether the juice is worth the squeeze when there are times when the BMS itself is inaccurate when it comes to SOC. 13 hours ago, Jay-Dee said: This is going to be a very annoying issue in winter months. Any ideas? Yes this is exactly when the issue occurred for me, multiple times - in winter - when the amount of PV production from my panels is insufficient to cover my needs. The battery ends up not getting charged fully for several days and the BMS SOC becomes less and less accurate. This is not a problem in the summer months when there is almost always excess PV - the extra “wasted” PV ensures that the battery is actually fully charged and the BMS calibrated. My only solution to this is to monitor the voltage pattern of the battery. I can definitevely know when my battery is fully charged when the voltage shoots up rapidly to around 54V and then the battery stops accepting more current, then the voltage drops due to the lower charge current. This is very reliable. In fact, before load shedding I either charge from the grid until I see this pattern, or I set the priority mode on the inverter to “battery first” rather than “load first” to ensure that all PV is used to charge the battery rather than reduce my load costs. So then the question every day becomes “is my battery (really) charged to 100%” rather than “what is the exact SOC of the battery” and then one of the main reasons for wanting proper comms from the inverter to the BMS in my case actually starts to fall away. If the battery is really 100% charged, then BMS SOC calibration or “estimated SOC” on the inverter is not really an issue the next day. When the inverter is set to use Battery % mode (not Lithium mode), even when the inverter thinks that the battery is at 100% SOC, it can carry on charging (sometimes even for hours) the battery until it is really charged. I have also found that even when the BMS thinks that the battery SOC is at 100% when really it is not, it continues to accept charge current, until the correct voltage of all of the cells is reached. So to answer your question, I think it is fine to continue charging the battery when the inverter and/or BMS thinks the SOC is at 100% because eventually the battery will be fully charged and automatically the charge current is reduced to a trickle charge (I am not sure whether the BMS or inverter is responsible for reducing the charge current).
August 20, 20223 yr Author Just to add to my previous post, this is the pattern I'm talking about. Notice that just after 13:00 the inverter thought the battery was full, but actually it was not. The battery continued to be charged as can be seen in the power chart as the PV was gradually reducing. Just before 17:00 I decided to top up the battery from the grid because of imminent load shedding, so I could tell that the battery was fully charged by 17:00. On this particular day, the discrepancy between the inverter SOC and the real charge state was not big, but there have been days where it is huge, even between the BMS reported SOC (which I monitor separately) and the voltage.
August 20, 20223 yr 9 hours ago, SolarConvert said: So my idea would be to use the CAN port of the inverter to communicate with the RS485 port of the battery, using Lithium mode 0, with the PI doing the translation. The RS485 port on the inverter therefore frees up for monitoring the inverter. This would however mean in your case that you’d need two PIs - one for the translation and one for monitoring, because Solar Assistant does not allow the PI to be used for any other purpose. In my case however, I use Home Assistant for all of my automation and reporting needs, so one PI would suffice for all of the above. ............... So, I am running ICM but had an install of Solar Assistant as a means of comparison. ICM, like ICC allows for other things to be done on the Pi at the same time, hence my theory of the virtual serial port software as then it could run on the same Pi. However with the battery issue, I may be forced to revert to running on voltage and not SOC from the BMS, very annoyingly so as this was one of the primary reasons for changing to the Sunsynk (better battery control). ICM can read the Narada BMS directly and manage that but I have not tested it yet (I do still have the cable from my Axpert days). My thinking is that I run the inverter on voltage to allow full charge as well is proper cut off / cut off at the correct voltage not SOC and I can then still read the battery details via ICM. ICM how ever will still not have a true SOC reading. This way, my RS 485 port frees up and I can then use the logger without issue The other alternative is limit how much the battery can run down during winter and prioritise the battery charging over load, to ensure it charges fully on anything but a rainy day. It will be almost as cost effective / savings effective. While typing this, I think I am going to give the last option a try tomorrow and Monday after manually charing the batteries and see if it keeps the SOC real. Failing that, I will try moving to voltage only. Once those are resolved, I will then move back to the serial splitter if needed (unless you get the CAN translation working)
August 21, 20223 yr Author I wouldn’t necessarily switch over to voltage battery mode on the inverter. That’s probably the worst mode for lithium batteries. The voltages help with determining whether flat or full but not so much in between. The time of use programs (the 6 time windows with battery percentages) would also need to be managed with volts if going this route. My current setup on the inverter is to use percentage mode on the battery, not voltage mode and not lithium mode. With percentage mode you can still set the float and absorption voltages, which I think is the key to allowing the battery to charge fully based on voltages but still get the “estimated” SOC from the inverter in %. This allows me to manage time of use with SOC % and is arguably much better than trying to get an estimated SOC % just from the battery voltage. I think the inverter uses the energy flow in and out of the battery to estimate the SOC in battery percentage mode. However, I have not had a chance to try out lithium mode on the inverter due to my BMS not working with lithium mode 16, so the battery percentage mode is all I have to work with. If you are able to fully charge your battery with lithium mode (you can tell by monitoring the voltage), then lithium mode is probably the best setting. It all depends on whether the inverter stops charging the battery when the BMS indicates that the SOC is at 100%, or not - I currently do not know the behavior. If the inverter stops charging the battery when the BMS indicates 100% SOC, then you can run into the issue where your battery is not really fully charged if you haven’t fully charged your battery in a while because in those instances the BMS SOC can be inaccurate.
August 21, 20223 yr Author The charge and discharge curves seem to confirm my 54V and 44V observations of the battery being fully charged and flat, respectively. This is my interpretation - please correct me if I am wrong. The voltages above need to be multiplied by 15 since it’s a 15 cell battery. So, when charging, the cells need to reach around 3.65V to be fully charged, so 3.65 x 15 = 54.75V for the battery. Notice the sharp voltage increase at the end of each charge curve, which corresponds to the graphs from the voltage monitoring I posted earlier. Looking at the discharge curves, my assertion of 44V being flat for the battery seems to correspond to the sharp decrease in the discharge curves (44V / 15 cells = 2.93V). I believe the BMS protects the battery from the voltage dropping lower than 44V - this I have seen in practice in winter when the battery hadn’t been fully charged for many days. The 4 battery SOC lights were on, the inverter thought that the SOC was above 75%, but the battery simply would not discharge further. I thought the behaviour was so odd, until I realized that the battery was flat. Then load shedding started and there was an actual outage at my house because of this scenario. Since that experience, I look at the battery voltage a lot closer and try to ensure the battery is fully charged at least once every 3 days. And yes, diverting all PV to charging the battery rather than reducing the load costs makes no financial difference if there is simply not enough PV in winter. Of course if there is excess PV, then it is far less likely for the above scenario to present itself since the battery will be fully charged at least once on most days, except for the rainy days.
August 21, 20223 yr 4 hours ago, SolarConvert said: I wouldn’t necessarily switch over to voltage battery mode on the inverter. That’s probably the worst mode for lithium batteries. The voltages help with determining whether flat or full but not so much in between. The time of use programs (the 6 time windows with battery percentages) would also need to be managed with volts if going this route. My current setup on the inverter is to use percentage mode on the battery, not voltage mode and not lithium mode. With percentage mode you can still set the float and absorption voltages, which I think is the key to allowing the battery to charge fully based on voltages but still get the “estimated” SOC from the inverter in %. This allows me to manage time of use with SOC % and is arguably much better than trying to get an estimated SOC % just from the battery voltage. I think the inverter uses the energy flow in and out of the battery to estimate the SOC in battery percentage mode. However, I have not had a chance to try out lithium mode on the inverter due to my BMS not working with lithium mode 16, so the battery percentage mode is all I have to work with. If you are able to fully charge your battery with lithium mode (you can tell by monitoring the voltage), then lithium mode is probably the best setting. It all depends on whether the inverter stops charging the battery when the BMS indicates that the SOC is at 100%, or not - I currently do not know the behavior. If the inverter stops charging the battery when the BMS indicates 100% SOC, then you can run into the issue where your battery is not really fully charged if you haven’t fully charged your battery in a while because in those instances the BMS SOC can be inaccurate. So today, I have charged on voltage and have noticed some interesting things... When I stated on Lithium mode, the battery (4 lights) & inverter reading the battery were showing 100% but at 50.1v. When putting it onto AGM V, the battery charge to "full" 54v float... It dropped the estimated SOC to 42% @ 50v and then look about 15 to 25 mins after reaching 54v to float and stop absorbing charge. I have now put the inverter back onto lithium mode and it is floating the battery around 55v and charging it at 57.6v as this is what it is reading off the BMS. It charged for a minute or two more and then floated again. I am going to leave it in lithium mode but only allow it to drop the battery to 50% with leaving the priority on battery charging to see that it maintains a true SOC If you look at the below image (taken while floating)... The 55.49 is the live battery voltage. The 57.6v never changes and from what I believe is the charging voltage as can be verified by the 2nd image which is a volt meter connected to my batteries and image taken while charging just before it floated. The 0.00A is the live current in or out of the battery. The 43v is the cut off from reported from the BMS. 0A, I am not sure.. 200A is presumably the discharge capacity. 100% is the SOC and 224Ah is the the total live available Ah. So the odd thing is, I assumed it was a 100Ah per battery and pleasantly it is more. The voltages that we have always been told to use are wrong and the internal BMS does not report Ah or SOC correctly when not fully charged over a period of time 🤔
August 21, 20223 yr 4 hours ago, SolarConvert said: And yes, diverting all PV to charging the battery rather than reducing the load costs makes no financial difference if there is simply not enough PV in winter. Of course if there is excess PV, then it is far less likely for the above scenario to present itself since the battery will be fully charged at least once on most days, except for the rainy days. This is the plan... I know what I generate in winter on a non-rainy day and even on a poor day, I should do enough to charge the batteries.. So hopefully there will not be too much lost in charge and discharge the battery vs power directly from solar but what ever it is, it is better than having the system die 5 mins into load shedding thinking you have at least 40% SOC. My rational of going this route and keeping the inverter reading the wonky BMS is that when it works, it works and I feel that if I can keep it working and learn it's flaws that this is better in combination with the Inverter talking to the BMS than not and hoping for the best. The down side is that I still need to find a way to split the serial port successfully to have the logger, Pi and batteries all linked at the same time
August 21, 20223 yr Author 4 hours ago, Jay-Dee said: The voltages that we have always been told to use are wrong and the internal BMS does not report Ah or SOC correctly when not fully charged over a period of time 🤔 Yes this was a shocker for me. I always thought that getting proper BMS comms was the holy grail, the solution to it all, only to find out that these can also be incorrect. 4 hours ago, Jay-Dee said: When I stated on Lithium mode, the battery (4 lights) & inverter reading the battery were showing 100% but at 50.1v. Interesting, so you experienced the exact same thing I did. 4 hours ago, Jay-Dee said: The 57.6v never changes and from what I believe is the charging voltage as can be verified by the 2nd image which is a volt meter connected to my batteries and image taken while charging just before it floated. If this 57.6V is understood by the inverter, then I wonder why it did not try to charge to that voltage whilst on lithium mode. So the inverter had just stopped charging at 100% BMS SOC before you changed to voltage mode?
August 21, 20223 yr 1 hour ago, SolarConvert said: Yes this was a shocker for me. I always thought that getting proper BMS comms was the holy grail, the solution to it all, only to find out that these can also be incorrect. So, when I had an Axpert, I had this problem when ICC was reading the batteries' BMS and I hoped that it would be solved with the Sunsynk communicating directly. At first it seems to help but it just took more time to rear its head again. So I can confirm it is definitely the battery / BMS (and we have 2 different BMS which is even more odd) and not the inverter 1 hour ago, SolarConvert said: If this 57.6V is understood by the inverter, then I wonder why it did not try to charge to that voltage whilst on lithium mode. So the inverter had just stopped charging at 100% BMS SOC before you changed to voltage mode? Correct and yes.. This seems to be a Sunsynk issue where it stops charge as 100% SOC. I am going to log this with the team from Sunsynk and see if it is something they will look into but it also requires me having my Logger back in for them to apply an update. If they did fix this "issue" it would sort of solve the BMS problem in lithium mode when one has a good day
August 22, 20223 yr On 2022/08/19 at 4:07 PM, SolarConvert said: Hi @jetlee, Did you try both software archives from this thread - the one from zivva and the one I posted? The one from zivva did not work for me but the one I posted did. I'd keep the baud at 9600 and try out all of these requests because with Lithium mode 16 the inverter tries all of them out: 27 04 0f ff 00 09 04 2e 28 04 0f ff 00 09 04 d1 29 04 0f ff 00 09 05 00 2a 04 0f ff 00 09 05 33 2b 04 0f ff 00 09 04 e2 2c 04 0f ff 00 09 05 55 2d 04 0f ff 00 09 04 84 2e 04 0f ff 00 09 04 b7 2f 04 0f ff 00 09 05 66 30 04 0f ff 00 09 07 09 31 04 0f ff 00 09 06 d8 32 04 0f ff 00 09 06 eb 33 04 0f ff 00 09 07 3a 34 04 0f ff 00 09 06 8d Your battery seems to respond better to the Lithium mode 16 commands than to the Revov/Tian Power ones. Do you have a Deye/Sunsynk inverter? If so it would probably be easiest to crimp a cable as per the first few pages of this thread and try out Lithium mode 16 whilst the inverter is connected to the battery. My battery uses the Tian Power BMS so the Shinwa software doesn't work for me. So I voided my warranty and took it apart to find out. Mine is using Tian Power. I unfortunately have an older Axpert 5KW (RCT Brand) .. which every 2 or so days, completely shuts off due to some BMS condition. I monitor my inverter with a homebrew app on the PI, and it is definately not over current / voltage related, so the only way I can see whats actually going on, is to read the BMS to find out why it goes into protection. Switching my inverter back to my (very old) AGMS works a treat and also happens to force the Narada back to good health, until it shuts off again. Disconnecting it from the inverter instantly also results in it working .. so I have a wierd issue, I'm hoping the BMS will tell me what it is... Worst case, Ill buy another BMS and swap it .. but Id prefer to get the original to work ... Going to try another rs485 adapter and new cable before I completely lose my mind, but at least now I know I have the Tian Power
August 23, 20223 yr Author 14 hours ago, jetlee said: every 2 or so days, completely shuts off due to some BMS condition Maybe monitor the battery voltage closely, especially before shutdown, as it would giver more insight. Since you've confirmed that you have a Tian Power BMS, you can use the software I attached a few pages back to read it. Also make sure you have the RS485 A/B signals on the correct pins.
August 23, 20223 yr 1 hour ago, SolarConvert said: Maybe monitor the battery voltage closely, especially before shutdown, as it would giver more insight. Since you've confirmed that you have a Tian Power BMS, you can use the software I attached a few pages back to read it. Also make sure you have the RS485 A/B signals on the correct pins. Yes, I monitor the voltage every 10 seconds, and log it, and it seems like theres no pattern, it can go at 49 volts or 48.5 volts or 48 volts .. I am not getting anything looking like a sane response from the BMS, so the next step for me, is going to get a different rs485, in case mine is faulty or just not compatible. Then I'll run through the Tian Power commands in the hope that I can get something looking like a correct response, even if I have to run tests on every wire combination. Once I have that, as you suggest, I'll use the provided app to read the error... if it still doesnt help, then either I'm swapping the BMS or buying an inverter .. will have to see which ..
August 23, 20223 yr 22 hours ago, jetlee said: I unfortunately have an older Axpert 5KW (RCT Brand) Not sure if you've heard of the premature float bug which is rampant on these older inverters. It causes the batteries not to fully charge and from the info in this thread it seems it would make the BMS go haywire.
August 23, 20223 yr 3 hours ago, jumper said: Not sure if you've heard of the premature float bug which is rampant on these older inverters. It causes the batteries not to fully charge and from the info in this thread it seems it would make the BMS go haywire. No, but Ill have a google .. thanks for the heads up
August 25, 20223 yr OK, so after all that .. the rs485 just wasnt great, replaced it, and now the Tian Power software is working, next is to figure out why its shutting down / Alarming, and then if I can fix it, on to writing code to read the protocol now that I am getting valid responses
August 26, 20223 yr 23 hours ago, jetlee said: rs485 just wasnt great, replaced it, Hey @jetlee, do you mind letting us know which rs485 adapter worked for you? I've had 2 serial converters that don't work so want to be sure I buy the right one next time. Thanks.
August 26, 20223 yr 1 hour ago, jumper said: Hey @jetlee, do you mind letting us know which rs485 adapter worked for you? I've had 2 serial converters that don't work so want to be sure I buy the right one next time. Thanks. Sure .. bought this one https://www.robotics.org.za/RS485-MINI , all the others I found were like 300 rand plus, this one with shipping (95 rand in SA) was still cheaper even with ludicrously expensive (proportionally) shipping .. Hooked up to a homemade RJ45 .. Good luck
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.