Jump to content

SiliconKid

Members
  • Posts

    62
  • Joined

  • Last visited

  • Days Won

    2

SiliconKid last won the day on November 27 2022

SiliconKid had the most liked content!

About SiliconKid

  • Birthday 16/02/1974

Profile Information

  • Gender
    Male

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

SiliconKid's Achievements

  1. Correct, as stated in my original post above: "but unfortunately as the Grid draw increases and the CT accuracy improves the compensation you made by adjusting the CT ratio becomes an over compensation to a small extent, and there does not appear to be any way to combat that phenomenon." We actually need SunSynk to make some changes, treat the negative load values as zero, and stop using the negative values in their internal calcs.
  2. On the inverter itself go into the Advance menu and then go to the Others tab and you will find CT_Ratio there as an integer number you can edit. You can also change it via the inverter settings from the Equipment page in the mobile app, under the Advanced options.
  3. Problem Description Your SunSynk inverter and Solar Assistant are showing power consumption values that do not make sense and do not add up when using conventional logic and even the built-in display on the inverter itself is showing grid load numbers that do not make sense at face value. In addition, the latest version of the SunSynk mobile app that now has a detailed power flow view that splits the loads into Essential and Non-Essential loads will now sometimes show a negative value for the non-essential load, but at the same time the display on the inverter itself will show that value as 0 (ZERO). Explanation of observed behavior and values In order to explain this problem it is necessary to first show the most simple real world example of this actually happening on my inverter display. When this photo was taken the Solar panel trip switches on the Solar DB were off to prevent Solar from getting involved and simplify the troubleshooting process, and the battery was at 100% and the inverter was configured to not use the battery at all. In addition, all trip switches for all non-essential loads, on the non-essential (original) DB board were dropped, to ensure no very low latent load (e.g geyser sensors using 10 watts) from any non-essential loads either, again to simplify troubleshooting. A photograph of my inverter's built-in flow display was then taken: The above photograph demonstrates the problem perfectly: 674 Watts is a fake and invalid number and 632 Watts also does not make sense The 0 Watts indicated above the house icon (non-essential load) appears to be accurate at first glance because, as stated above, I turned off all non-essential loads completely on the old DB board, which means that there should be no draw at all from non-essential loads, and that seems to be the case in the photograph. Unfortunately things are not as they seem because the 674 Watts shown in the second blue circle is a completely invalid number that does not correlate to any real world draw in any way. And if you had looked at the detailed power flow diagram in the latest version of the mobile app you would have seen that it did not show 0 Watts at the same time that this photograph was taken, instead it actually showed -42 (negative 42) Watts as the non-essential load. -42 is a very interesting number in this case because if you subtract -42 (add 42, because subtraction of a negative value is actually addition) from 632 (the actual amount of Watts the grid was allegedly supplying at the time), you get 674, which is the incorrect grid input total being reflected in the photograph above. If you add together the amount of power that the Inverter itself was consuming at that point in time (39 watts) and the amount of power being supplied to the Essential loads at that point in time (551 Watts) you get 590 Watts, and 590 Watts is actually what the grid input total was at that point in time, not 674 watts. Note: 590 (Actual grid input to inverter) + 42 (The hidden non-essential load value) = 632 (The actual watts allegedly being drawn from the Grid at that time) Note: The 632 Watt "actual reported draw from CT" reflected in this example is also puzzling because the -42 Watt non-essential load was not actually a load, it was actually a CT calibration error, and in fact there was ZERO non-essential load at that point in time drawing 0 watts, which means that the reported current draw from the CT should have been 590 Watts, not 632 Watts. So what is actually going on there, how is the "Grid power supplied to the Inverter" value (674) inflated, and how can the non-essential load value be a negative number? From some experimentation I have isolated the problem to inaccuracies on the part of the CT (The CT is not calibrated to zero accurately), and to make matters worse the inaccuracy of the CT varies dependent on how much current is being drawn from the Grid. The more current you draw from the Grid through the CT, the more accurate the CT becomes, and this is because the standard CT used with a SunSynk inverter uses a CT ratio of 2000:1 (2000 amps are reflected as 1 amp for the purposes of monitoring) and consequently a very low primary current results in an extremely low fractional secondary current that is being monitored by the inverter, to the extent that the inverter can actually under read to the point where the calculated non-essential load value reflects as a negative value when it should be zero. You can adjust for this problem on the SunSynk inverters by modifying the CT Ratio in the advanced menu to "calibrate for zero" when you actually have zero non-essential load, but unfortunately as the Grid draw increases and the CT accuracy improves the compensation you made by adjusting the CT ratio becomes an over compensation to a small extent, and there does not appear to be any way to combat that phenomenon. My current CT ratio is set to "2155:1" and that gives me a dead zero non-essential load reading on the built-in display AND in the mobile app detailed flow view now, if my non-essential loads are actually zero, and remains dead accurate for any very low non-essential load below around 250 watts, but any load higher than that starts to affect accuracy very slightly again. The following photograph shows the values reflected on my inverter after adjusting CT ratio and turning non-essential loads back on (incurring 30 watts of latent non-essential power draw): As you can see, the numbers all make perfect sense in that photo and so do the flow indicators. The amount of power we are talking about here, as reflected in the slightly inaccurate non-essential load numbers, is very small and not really worth getting upset about BUT SunSynk do need to update their UI to be consistent and also make sure that any graphs and reports are not incorrectly adding those negative numbers to derive inferred load values because otherwise reports over time will show totals that very inaccurate. As a simple example, if the CT is under reading by 50 watts on average and the "Grid supply to Inverter value" is therefore inflated by 50 Watts on average, that would equate to (24 hours x 50Wh = 1200 Wh = 1.2 Kwh per day) of additional grid draw being reflected that is a complete fabrication. Over 30 days that equates to (30 x 1.2Kwh = 36Kwh) of non existent grid draw being reflected. Conclusion and Suggestions In my opinion SunSynk need to make some adjustments to their firmware and app to address this issue. I have no problem with them ignoring negative values, hiding them from end users, and reflecting a value of zero instead, because most end users will not be able to grasp the technical reasons for those negative values and how CTs work. But if SunSynk are going to show 0 (ZERO) when the number is negative they must also treat the value as ZERO in any calculations, they cannot substract the actual negative number and thereby add it to totals that it should not be added to. On the mobile app they currently do show the negative number but they should probably change that to do that same thing they do on the built-in display and show Zero rather and then also make sure that the app does not subtract that negative number from any totals anywhere. Whatever they decide, it should also be consistent between their app and the on-device display. It also worth noting that 3rd part monitoring solutions like Solar Assistant (which I also have running) are also impacted by negative non-essential load values and the calculations made by those 3rd party systems are also affected and also produce seemingly impossible total values that make no sense because of the negative value being reported by the inverter and being used in calculations where a subtraction is expected to reduce the total and not inflate it. If SunSynk fix this by zeroing negative non-essential load values in memory on the inverter, that would theoretically fix the 3rd party systems automatically, but if SunSynk choose to retain the negative value in memory and simply display it as zero and consider it zero in internal calculations then the 3rd party software will also need to be updated to compensate for this phenomenon.
  4. 500.12 was definitely a problem and Growatt themselves instructed me to revert to the previous version (500.11 as I recall) to stop the fan fail warnings. As soon as I reverted, that nonsense stopped.
  5. Also NOTE something else that I learnt recently: The data is logged every 5 minutes, and that is controlled by the comms board inside the inverter. The setting in the Wifi Dongle settings that implies that you can control the frequency of logging is nonsense, it does nothing. You can change it to 1 minute or 2 minutes and it will still log every 5 minutes. NB! The comms board STORES data that hasn't been logged during the day if there is a comms or internet problem and data cannot be delivered to the Growatt server. And when comms is re-established all of that stored data is then batch uploaded to the server to back fill missing data! This is a critical thing to understand because that backfilling takes TIME. Quite a long time from what I saw. AND, from what I saw it blocks active current logging while it is busy playing catchup and uploading all that missing data, which then has the net effect of making it look like logging is not working again, but it actually is, it's just paused while a massive backlog of data is being dumped to the server. I think that's why some people say something like "I just gave up eventually and walked away in disgust and then it somehow sorted itself out and started logging again". In reality it was busy backfilling missed data when they gave up assumed it was a lost cause.
  6. Hi I actually managed to get my system into the exact same state you've described here when I recently decided to do another round of firmware updates. On that weekend I upgraded main fw to 500.12 and I also flashed the comms board to sort out missing data issues. At some point after all that flashing I found myself in the same state you are describing here. I tried deleting the datalogger multiple times and registering it again, logging in and out of my account multiples times, etc. etc. Nothing worked. Eventually I also hit a wall where there was no datalogger visibly registered but if I tried to register it, it told me it already existed. Ultimately what I did is I deleted my entire account. Full delete. Then I registered a new account, with a different email address. Setup the plant again from scratch and added the datalogger again. Then I had the datalogger back, but I was still not getting ANY data actually logged, despite the phone app confirming the datalogger was checking in and connection was excellent. At that point I went back into the garage, shut everything down again, flashed the comms board AGAIN, powered everything back on again methodically, reset the inverter and set every setting manually again, then plugged the Wifi dongle back in, and ... bingo! Data started flowing in again, every 5 minutes, and hasn't stopped since. And that was over a week ago. Note however that it took time for the datalogger to start sucking data in again because apparently it was backfilling data that had been stored on the comms board (see my other post below this one). If I look through the power graphs etc. there are no gaps going back days and days at this point. NOTE however, that I was then also forced to rollback the main firmware from 500.12 to 500.11, because 500.12 started to cause constant fan fail warnings and faults. Amos told me to roll back to 500.11 to stop the fan fail issues, so I did, and it did. Fortunately, after that rollback, the datalogger started working again immediately after plugging it back in, and didn't give me any hassles.
  7. Hi all Last year in February I installed an SPF5000TL HVM-P Inverter with 2 x US3000 Pylontech batteries, and ever since then it has been quite the journey to get everything working exactly the way I wanted it, without issues. I have only just recently achieved a state of operation that I consider to be optimal, where everything works to my satisfaction, including the Wifi data logger and remote monitoring. To achieve that state of Inverter nirvana required many hours (days) of fighting with hardware, communicating with Growatt reps in China, and flashing firmware in my garage myself like a mad scientist. Finding the right firmware has always been a challenge because there is no official repository of Firmware for Growatt Inverters and Growatt do not want us to maintain our own equipment, so we (the PowerForum community) have had to source the Firmware piece by piece through direct communication with various Growatt reps, over time. I have decided that it is now time to release my full curated and organised library of firmware and information to everyone. You will find the entire library here: https://drive.google.com/drive/folders/1rySPfRIlk29Ij3wTQlYmuqM0dSTr_UHd?usp=sharing PLEASE NOTE: Everything you find at that link is SPECIFICALLY for the Growatt SPF5000TL HVM-P Inverter. Do NOT flash any other Inverter with those ROMs. Not ever. For any reason. Ever. I only use Pylontech batteries and some of the information might refer to Pylontech specifically. Growatt DO NOT approve of us flashing firmware to our Inverters over USB ourselves, or at all for that matter, and I have also been told that I should only be flashing firmware through some kind of USB emulator device. I have personally never had a problem flashing over USB from a Windows 10 notebook without an emulator device, but I know what I'm doing and I'm methodical and careful. You flash at your own risk. Both firmware, and in public parks. I do not have any Solar panels attached to my setup. It is configured as a pure battery backup / UPS solution. I have seen that other users of the same Inverter have had issues specific to Solar and I have very little knowledge on that sub topic. It is very possible that certain Firmware updates included in the link above may improve your experience with Solar, but I make no specific claims with regards to Solar and you should consult with other PowerForum users who have hands on experience with Solar connectivity with the SPF5000TL before you do anything crazy. DISCLAIMER: I am providing this link, and these ROMs and tools, as a public service because I know how frustrating it is to not have this freely and easily available, and how hard it is to find this information and data, BUT I DO NOT TAKE ANY RESPONSIBILITY for what you choose to do with it. I am not responsible for firmware adventures that go bad and result in the loss of life of Inverters because they were bricked because you forgot to charge your notebook battery before you started flashing or your cat jumped on the USB cable and pulled it out half way through a ROM flash. I take NO responsibility for the consequences of your chosen actions, I am only the provider of information and knowledge. What you choose to do with it is on you. DISCLAIMER ENDS With great power (and a lot of ROMs from Growatt) comes great responsibility. Good luck, and happy flashing SiliconKid
  8. Unless the latest factory firmware on brand new units has finally been updated, of course. Then you might be ok with factory, but from what I've seen that's not the case.
  9. @Strauss If your firmware is factory your inverter definitely isn't working properly. At minimum you need the firmware I've been providing to everyone.
  10. In that case, I'm going to need that firmware update please, if it actually solves that problem. That's been driving me up the wall too. I had actually given up on it.
  11. Growatt can't get their firmware right so now they're blaming the batteries. It's not the batteries, other better inverters work fine with the same batteries based on anecdotal evidence. The firmware version numbers from Growatt are all over the place. I gave up trying to figure out which is version is the "most recent" or "correct". I have a firmware update pack that I've been giving to anybody who asks, that sorts out the original 4/20 BMS comms failures, as detailed by me in the very first post in this forum thread. That update will get BMS comms working properly over an RS485 cable but there are still issues related to Solar and input being turned on and off inappropriately which the SA Growatt community have been fighting for months. For a straight forward battery backup solution using the grid to charge the batteries and with NO solar connected, the firmware I offer is perfect. If you have Solar it doesn't look like there is any hope. Growatt can't get it right and it looks their inverters actually just aren't all that good. They work with solar, but the implementation is basic and non optimal. For Solar I would suggest getting rid of the Growatt and buying a Sunsynk or maybe Deye instead. The Sunsynk are very good. If you want the firmware I have, PM me.
  12. It's been discussed at length but it seems that Li-Ion batteries don't need a float charge like the older types do from what I've seen mentioned in these forums. The 100% down to 95% and then back up, in a constant cycle, is by design according to Growatt, so it's not considered a problem. If Solar is involved the Solar will get cut off at 100% and the inverter will again wait until the SOC drops to around 95% before turning on the Solar input again. And again, that is by design according to Growatt, and lots of discussion about that on these forums too. Long story short, the Growatt is good for pure battery backup (no solar) and serviceable and ok for Solar, but not great it would seem. When I add Solar I will probably sell my Growatt and buy a more expensive Sunsynk.
  13. I've been following this discussion with interest. It seems to me that the fundamental problem is that the Growatt Inverters always deliver Solar power VIA the battery. I don't know if that is what's going on, but it certainly sounds like it. The power it needs to keep itself running is also definitely taken from the battery. That's been confirmed by Growatt, so that's a constant draw on the battery, and if you are running without Solar and on Mains bypass you can clearly see your batteries slowly deplete to around 95% SOC over a consistent period of time (6 hours in my case with 2 x US3000) and then get charged backup to 100% in short period (20 mins or so in my case) and then the cycle repeats. And that's purely because of what the inverter itself is drawing from the batteries. The inverter seems to have no bypass that allows Solar input to get to the load without it first going via the battery? I don't have Solar hooked up yet, but from what I'm reading the problem there seems to be directly related to the fact that Solar always seems to goes via the battery. In other words, when you are "powering your load from Solar" I don't think you actually ARE powering your load directly from Solar, you are actually powering your load from the battery, which is immediately being replenished by Solar? And that's why when the battery gets to 100% it cuts the solar off, to prevent overcharging that battery, because effectively you are in a position at that point where the Solar being generated is MORE than you need to compensate for what the load is drawing, and there is nowhere else to put that excess Solar generated energy because the inverter isn't capable of sending it "directly" to the load without the batteries involved. Is that accurate? Or is it actually splitting the Solar input to: 1. Feed the load directly via a battery bypass circuit. 2. Charge the battery at the same time on the side using any excess? If I'm right, then the whole issue with Solar really just comes down to: "The inverter isn't expensive enough to include what it would need to convert the incoming Solar DC to the required clean AC on-the-fly via a bypass circuit, and also divert excess to the batteries to charge them, and to also stop charging them when they are sufficiently charged and simply allow the excess Solar generation that's available to be lost and wasted." I just get the distinct impression that the Solar implementation on these Growatt's is very basic and simple and will get the job done but is not using any intelligence to be as efficient as possible and protect the batteries as much as possible. And maybe that's ok if we consider the price point and what you are getting for that price. They really are not very expensive given what they CAN do, in one device, even if what they can do isn't necessarily what we would all like them to be doing. The Growatt 5000's also can't blend power sources, so that's another limitation that something like a more expensive SunSynk can do effectively (costs about 10 to 12K more than the Growatt). All very interesting though. I will probably go Solar fairly soon and right now I'm considering selling this Growatt and buying a SunSynk instead just to avoid the Solar related headaches and also because their UI and software is better.
  14. The inverter will be drawing some power constantly. I'm running battery backup only and no solar and I can clearly see how the inverter draws from the battery until it hits around 95% and then charges it back up to 100% again.
×
×
  • Create New...