In mine case the Victron lynx shunt 1000 calculates the SOC and I did see a 10% error after several weeks (JK BMS was better only 5% off).
I did hear the simple smart shunt is more accurate?
From my perspective, a BMS has many very important tasks, but collecting and providing data (State of Charge) is not one of them. These devices are not built or designed for that. A BMS is designed and built to monitor the fundamental parameters of individual cells in a battery system and to disconnect the system from the energy flow in case of deviations. These devices do exactly that very well and reliably. The fact that many people think they can tap into this data and thereby push their systems to the maximum of what is tolerable is a completely different story. But be that as it may. Measuring the State of Charge is certainly not one of these tasks; otherwise, all of these devices would be flawed designs. The mere fact that small charging and discharging currents are not measured at all proves this. A shunt is the preferred solution here. These devices are built precisely for that purpose. With batteries that don’t have separate power consumers, this works almost perfectly. The fact that more and more technologies like BMS, balancers, heaters, or numerous flashing and illuminated displays have been incorporated over time is a relatively recent development. In my opinion, one should react to this development by taking this consumption into account during configuration or by creating a way to take it into account.
Because JK BMS SOC accuracy is bad I installed a lynx shunt 1000, but the soc is 10% off after several weeks.
I did hear the normal smart shunt is far more accurate, why is this the case?
And what about lynx smart bms 500, has this device the same accuracy than the smart shunt?
This has nothing to do with the manufacturer or the size of the device. The shunt measures the current flowing through it, taking efficiency and other parameters into account. HOWEVER, the power consumption of all devices beyond the shunt, such as a BMS, a balancer, DC battery heating, or any DC displays, is not considered. These seemingly minimal power consumptions add up during 24/7 operation. That’s the problem.
I can only attest to the accuracy of the BMS Smartshunt 712 being at least an order better at coulomb counting than any JK BMS I have ever worked with. That is an anecdotal indication it must be based on at least a 16bit but probably 20bit or higher ADC for current measurements, able to track down to currents at or below 10mA. The old JK’s I have worked with would fail to even measure currents around 1A or below, newer ones are likely to have improved on that.
Initially, I worked without a SmartShunt, using VE-Configure to derive the SOC from it. Victron offers this functionality but unfortunately has a built-in bug. After several attempts to fix this bug here in the forum, numerous users told me that a SmartShunt is essential for an accurate SOC reading and therefore for the DESS to function correctly. Now I see the result: the SmartShunt cannot accurately calculate the SOC in winter because it doesn’t initialize.
Now I’m supposed to write my own program in Node-RED to compensate for that. What happens after that? Will everything work then? Who guarantees that to me? Had I known all this beforehand, I would have continued working with VE-Configure; it is accurate enough for my LIION batteries, and it can’t calculate the SOC correctly any more than the Smart shunt.
It would be simple to implement a fixed variable to integrate a constant quiescent current instead of passing that responsibility on to users through the use of Node-RED.
If I sell a device, it should deliver on its promises; that’s my opinion. Therefore, this whole discussion is pointless.
1 Like
The logical fallacy here is that, even though this might improve your SoC tracking significantly, it does not address the real issue for all situations. All you can hold Victron accountable for is the BMS SoC tracking performance of their own brand batteries, and to some extend their lack of transparency with respect to the experimental nature of DESS or the other way around overly simplified marketing of it’s future potential.
If Victron products don’t work properly (Ve-Configure) and I’m then advised, including by Victron employees here in the forum, to buy a SmartShunt, and now, since things still aren’t working correctly, I’m advised to learn Node-RED to get everything working, then I’m simply angry, you should just accept that.
I don’t think they advise you to learn Node-RED, do they? I do but that’s just me being pragmatic as an engineer based on my personal view on what can be expected, or not, from Victron. But yes, I’d love to see Victron shift focus from feature bloat to core technology innovation but that’s unlikely to happen and possibly not a viable business model for them. After a period of lowering my expectations I reached a relatively stable level where I either not expect to much to get frustrated anymore and build solutions myself where I really need them.
PS, rewiring the Smartshunt directly in between the battery negative and the BMS is how I solved the quiescent current issue.