I have a problem that I would like help with if possible. I have an EasySolar-2 GX 48/5000. It has been working fine for 2 years. Everything was cool.
I had some DC loads (2 Network switches, a NAS and 2 security cameras - 35-60w constant according to smartshunt) running direct from my battery for the past couple of months. That was fine but I wanted the unmonitored consumption to be included in the consumption data in VRM for the system. I purchased a SmartShunt 300 and installed it. I set it to Energy Meter - DC System mode and all seemed well the DC panel was showing the load in watts.
When I looked at the VRM the following day something was wrong. It registered no consumption at all during the night and the consumption during the day was an exact mirror of the solar energy yield. When I turned off “Has DC system” and looked again the system was registering AC consumption properly. Weirdly this same behaviour occurred about 6 months earlier before I had the SmartShunt at all. I just turned on “Has DC system” to experiment and the same behaviour occurred. As soon as I turned it back off everything returned to normal.
Hopefully I am not leaving out any essential information. Does anybody have any idea what is happening here? Perhaps it is something obvious that I am doing wrong.
Do you have a main shunt that measures the battery? If not then i guess this would be the issue.
You can use your shunt to measure the battery, including what the EasySolar uses/produces. Then enable “Has DC system” to let the system calculate what the DC side uses. It will subtract/add what the EasySolar uses/produces from what the shunt reports, the rest is the DC side load.
To get that “rest” to show correctly/more precise, you can use an additional shunt as a DC meter. Looks like the main shunt is necessary for that to work though.
Thanks for reply. The battery has it’s own shunt and BMS that feeds all that info(SOC, Voltages, Current, etc.) to the Victron over CAN bus. The SmartShunt as energy meter is only connected across the negative going to the DC loads. Have you seen or read about this same behaviour happening before? I can’t understand what it thinks it is doing; it doesn’t make sense to me but I am low on experience with Victron.
I have read around a bit since and phoned a friend of a friend who has done a few victron installs and I can’t find anything at all mentioning that a smartshunt as energy meter should be a second of 2 smartshunts. It doesn’t make sense either in my mind that the architecture should be that way. The installer said once your battery already has a shunt and BMS that is reporting properly to the GX then you should be fine. He has never seen or heard of the behaviour mine is displaying before.
Does anybody have any useful information or suggestions, I don’t fancy being the proud owner of a €75 paperweight.
While yes you have a smart battery with its own internal measuring, but its very well possible that the DC energy meter function only works correctly if theres a main battery shunt also.
Is the shunt configured as “DC system” in its details page? Or did you choose some other category?
Yes as I said originally it’s set up in Energy Meter - DC system mode and the DC display panel is displaying the loads correctly.
I knew what you meant and I have seen those diagrams but I presumed that the diagrams function was to illustrate the two separate uses of the smart shunt and how they can work together in one system. It doesn’t suggest anywhere that you cannot use one as a standalone energy meter. Yes of course it is entirely possible that it is the case but it seems very unliikely that (1) it wouldn’t be mentioned anywhere in documentation and (2) that it would make sense to engineer it that way. Surely the smart shunt in the diagram as battery monitor is only necessary when you are using raw batteries without that inbuilt functionality?
In everything I have read myself in trying to understand this (victron documentation and forum posts) I have not previously seen anyone suggest this to be the case. I have asked an installer (it’s possible he is incorrect of course) and he reckons once the GX is properly receiving information over can from the battery that this setup should work fine. I appreciate your guess and I know it is possible to be correct but I have no way of testing it without dropping another €75 + €15 for VE.DIRECT cable and altering wiring (possibly unnecesarily) so I was hoping somebody would have something better than a guess if you understand where I am coming from. I imagined someone might know for a fact what is happening or have seen or read about the behaviour mine is exhibiting.
Thanks and kind regards
For what it’s worth I have run my exact setup by ChatGPT and Gemini also just now and both stated that I do not need a second Victron smart shunt and that a single unit is capable of running as a standalone energy meter IF my battery is reporting SOC, volatage, current etc. properly over CAN to the GX (which it is). This seems to mirror what the installer advised me. I’m still confused about what is happening.
I would expect the same thing. But yes, its not clearly stated one way or the other.
My initial answer was due to not knowing that the battery itself reports data, so by not having a main shunt in the system, then the behaviour could be explained. To be fully honest i would still expect it to work bette than just counting MPPT and DC energy together.
Its possible that this is solely a VRM issue, since thats where the discrepancy shows up. On the GX it seems to work just fine
Yes maybe it is just a VRM issue but whatever it is means that the entire purpose of me buying the smartshunt is not being achieved. It’s registering the loads just fine in real time but badly messing up the graphing data completely as soon as it is activated. Also very strange that it exhibited this same behaviour even before the smart shunt was installed when “Has DC system” was turned on. At that stage I thought it was just because of the lack of smartshunt but apparently not. Very frustrating.
I can’t help but wonder is there just some setting change somewhere that could fix this glitch but I can’t seem to find it anywhere. I’ll keep trying anyhow.
It seems after more investigation of this that @chrigu is actually sort of correct in that while IT APPEARED TO ME that the Seplos BMS was providing all good data to the GX device over CAN it has in fact got some slight quirks with it’s Victron protocol implementation.
It sends almost all data correctly (all real time figures are good) but it is either not providing or providing incorrect historic total charge and discharge figures. The Seplos seems to record and present those figures in amp hours but the victron expects it in kW/h. I cannot be certain that it is passing those figures but it seems to make sense as the figures I’m seeing on the battery panel in remote console look to be out by a factor of roughly 50 (my aproximate voltage). Saw somebody else talking about using node red to set up a virtual battery, convert the figures from the seplos can and feed them into the gx via the virtual battery.
Seems like crazy hassle to fix this for me. I have to learn node red, make a CAN to TTL adapter and all sorts of crazy stuff and will it even be a stable setup in the end? Looks like I am out of luck unless I get a second smart shunt and use it to track the battery instead of the BMS. Very frustrating because it’s only a small bug but it messes the VRM graphing up completely and unless Seplos release a new firmware that addresses this I can’t fix it myself easily. If Smartshunts weren’t very expensive and it wasn’t such a pain to adjust my wiring and panelling to accomodate a second one then I wouldn’t hesitate but I’ll have to ponder what to do.
You don’t need 2 Smartshunts to do proper DC energy monitoring as long as your battery manufacturer has implemented the Victron protocol perfectly. It seems mine has not quite
*An afterthought was that I could try setting my battery to emulate a different supported manufacturer’s CAN protocol and setting GX to same. Maybe that might fix it if Seplos have done a better job on one of the other protocols? I might have a mess around when I get a chance.
No need for that. You read the Canbus battery into NodeRed, then change whatever needs changing, and send that out through a VirtualDevice. In VenusOS you then set the controlling BMS both in Batteries and Charge Control to Manual and the VirtualDevice.
You would need a compatible GX device, or raspberry pi with a suitable can usb adapter.
Then there are nodes to read CAN packets and you would need to add a translation layer to a virtual battery device.
It is easier if the battery is victron compatible to start with.
Well I wouldn’t use a PI because I already have a home lab running right adjacent to the solar system so I would use that. I have installed NodeRed in Docker on that today and started to read some tutorials. It actually looks pretty easy. I thought I would need an adaptor (CAN>TTL, CAN>USB, CAN>SERIAL, etc.) to get the data to the computer but @chrigu was saying there is no need for that. I was wondering how I would read the Canbus without an adaptor. All I need now is a way to get the data from the battery.
Yes it would be super easy if the battery was Victron compatible then we wouldn’t be having this converstaion. It claimed to be but now looks like their implementation of the Victron protocol is imperfect. Everything works great until “Has DC system” is enabled at which point the graphing completely glitches out but all real time data in VRM is displayed perfectly.
I have found three or four other people on here reporting the same symptom in VRM graphing as me. I just have to look back and see if they are all using batteries with emulated Victron protocols to confirm my hypothesis. Wish I could know the details of exactly where the VRM graphing acquires data and how it does it’s calculations exactly. Then I could figure out easily why the glitching is happening. That’s one of the troubles with proprietary, closed source systems I suppose.
You have a GX, in the EasySolar. Install NodeRed on it through the large OS image. Both the battery and the MP of the EasySolar are connected to it already.
Ahhh yes I see now. I was foolishly missing the fact that NodeRed could be installed directly on my GX device - I thought it had to be run elsewhere. I don’t think it’s a fully fledged GX like a Cerbo or whatever in my unit but hopefully it still has this much functionality. Thanks for the help mate I have a good path to look into now. Kind regards.
Nodered can be run separately, but it has a specific configuration to connect to a GX.
See manual setup instructions below, far easier to run directly on the GX.
