For my JK v19 BMS battery on CAN-bus , on Ekrano GX I used to periodically check Details > Lowest cell voltage and Highest cell voltage. Both values used to be shown with a 1mV precision - as provided by the BMS. After the recent OS upgrade to Venus 3.70 both values are now with two digits precision only. Looks tidy, but I am missing milivolts that used to tell me how well my balancer works. Is there a way to increase the Voltage precision back to 1mV (3 decimal places)?
Hi @Oldi,
you sure this was different before? I am running 3.66 (started with 3.60 6 months ago) and I had always 2 decimal places.
Only the in the remote console it shows 3 digits for me.. maybe you are referring to that?
Yes, i think he means this remote console. For me it also went from 3 to 2 digits after upgrading 3.67 to 3.70. I did a revert to 3.67 and it was back to 3 digits.
It’s not the only reason for the rollback. The JK BMS is complaining about overcurrent alarms at around the time when the battery reaches 100%. I never had that issue in the past 6 months. Let’s see what it does with 3.67 again..
1 Like
1 Like
I was referring to Ekrano GX screen as well as Ekrano access via LAN. Same picture as @Martijn provided.
While i do agree that the UI should be consistent across displays, console and VRM, i doubt that any BMS out there offers millivolt accuracy, so comparing mV readings between cells is pointless.
@chirgu, I agree absolute accuracy is not needed, but relative accuracy can be very helpful. I have 32 lifepo4 cells from three different batches, and my BMS is set to trigger balancing at 5mV difference. I need to monitor how those cells behave beyond the flat portion of the charging curve, as they are quite new. For years, Venus OS displayed those values with 3-digit accuracy, and I found it unfortunate to degrade this feature e.g. to improve esthetic (my subjective assumption).
At the moment my cell’s imbalance is at 7mV - which is super good, but I need to go out to my garage with a Bluetooth app to see this.
Doubt is good, it keeps you sharp. But I can assure you you are mistaken. On our batteries we flag issues as soon as there is a voltage difference > 5mV over all cells over the full operational range.
1 Like
I will get back to this, a 12 bit ADC (or above) with accurate / calibrated components should be able to measure this, or at least with the noise part being in the mV range.
In a recent discussion we established that the BMV Smartshunt has a 20 bit ADC, a search should surface that discussion. 12bit is by todays standards, prehistoric 
EDIT: 24bit:
That is something else, yes a smart shunt is accurate. This is about 3rd party BMS-es.
1 Like
For a mV reading we should expect the tool that measures it to be accurate to 0.5mV, so that the inaccuracy disappears in the rounding. And thats what i doubt. I dont doubt that people or systems can and do act on the mV readings they get, but that the actual measurement itself is less accurate than what the decimal places suggest. What good is a measured value of 3.405V when the cell actually has 3.4 or 3.419V?
1 Like
I don´t know about Victron BMS, but I have a JK balancer and Seplos BMS 3.0. Readings of all 48 cells match up within the range of 2 mV’s. That could be a coincidence, but I doubt it.
But in general I do agree that two decimals is good enough.
1 Like
Fair enough. My position on that topic is that most 3rdparty BMS-es, in the technical domain of our interests, are horribly outdated and underperforming.
On battery voltage and SoC%: Yes, on cell voltage level: No. Anyways, lets shut down that computer and start weekend. Cheers
1 Like
I mean good enough for the UI. Not good enough for BMS and control.
1 Like
Wat is een ui? Ik doe alleen command line. En nou gaat de computer uit.
1 Like
This thread is about the UI Internally still floats.
Using the assumption that a third party BMS can’t handle millivolts - which they can perfectly fine - and then develop (read: change) a GUI that is inconsistent across the board, that is simply a design failure.