Lynx Shunt failed and now the current measurement is no longer correct

DIY
, , , ,
1

Hello community,

First, a bit about my setup. I have three MultiPlus II/48/8000 units, managed via a Cerbo GX. They are connected exclusively on the AC-IN side and are connected to my house network. On the DC side, six battery banks, each consisting of 16 EVE cells (around 96 kWh in total), are connected via Lynx Distributors and a Lynx Shunt.

The house network includes a Fronius Gen24 (which is also connected to the grid), two Symos, an Ohmpilot and a Wattpilot. The Cerbo GX is controlled via Modbus in mode 3. The Cerbo does not have its own AC-side energy meter. I wrote the energy management system myself. I control the Fronius BYD charging, Ohmpilot, Wattpilot and Victron systems every second. The system runs brilliantly, and since mid-July I have been almost completely self-sufficient.

On the Cerbo, I can see both the charging/discharging setpoint that I specify via Modbus, what is currently flowing from AC to DC or vice versa across the three phases, and the charging current measured by the Lynx. Until yesterday, this was always around 92%. In other words, when I had the Victrons charging at 10 kW, the DC charging power towards the batteries was around 9,200 W.

Yesterday, the Lynx Shunt failed (on the CAN side). I checked the wiring several times, etc., but nothing worked any more. The Cerbo could no longer detect the shunt. On the dealer’s advice, I then removed the small cables from the negative busbar at the bottom of the shunt; the LED went out. I then put everything back as it was, and the Cerbo detected the shunt again. However, the SOC and charging current were incorrect. The charging current was showing absurd values such as -20 kW. So I was told to switch off the MultiPlus units, after which the dealer calibrated the charging current.

The system is running again now… BUT: The previous 92% has now become around 50%! If I charge at 6 kW and the Cerbo also shows 3*2 kW on the AC-IN side, the Lynx Shunt shows less than 4 kW. Was the calibration carried out incorrectly? Nothing at all has changed in the DC wiring, MultiPlus units, etc.

The dealer has already provided a great deal of support. However, given the certainly unusual setup, it is understandably difficult for them to deal with this Lynx Shunt problem with yet more work free of charge as part of the usual warranty and support services. So I wanted to ask here before incurring additional costs elsewhere. Can anyone give me a tip?

How/when will the SOC calibrate itself again?

Best regards
JP

2

Could you post the screenshot/settings of the shunt here?

If the battery is safely charged, you can also synchronise it manually if you have access.

Have the updates been installed?

3

As far as I know, the dealer carried out an update of the Lynx Shunt hardware.

image
image4070Ă—1884 303 KB

image
image1920Ă—1133 181 KB

image
image4068Ă—2286 478 KB

4

One more piece of information:

Until a week ago, the Lynx Shunt was connected to CAN2 on the Cerbo, and the Boostech BMS to CAN1 on the Cerbo. Due to the Boostech firmware update (it can now obtain the SOC directly from the shunt), I had to move the Lynx Shunt from CAN2 to CAN1. As a result, it was presumably (!) recognised as a new device, and the information above may no longer match the original values. The 1800 Ah value is correct. However, I’m not sure about the other values, such as tail current, Peukert exponent, etc.

5

Looks like a default setting for lead-acid batteries (Peukert 1.25).

What do your battery datasheets say?? Did the dealer back up the previous settings before the update and then correct them manually afterwards??

6

image
image2534Ă—1338 209 KB

But the Peukert effect would only distort the SOC profile, not the charging current, would it?

7

According to the manual, you need to adjust “tail current”, the Peukert exponent", “charge efficiency” and “lower discharge limit” when using Li batteries.

If you have a “managed BMS” (i.e. a master–slave combination) and it can be connected via CAN, do that.

8

My dealer says, “you can leave the values as they are for now”, which is also confusing me somewhat regarding the Peukert value. But incorrect settings would not explain the drastically changed purported charging values from the Lynx either. The problem must lie elsewhere. Does anyone have any other ideas? I have checked all the cables again and tightened everything.

Could resetting the Lynx have somehow damaged something in it? The fact that it suddenly stopped responding on the CAN bus is already “strange”…

I would also like to clarify that I am not assuming that anything went wrong on the dealer’s part in particular. The dealer has always proved to be very knowledgeable and has supported me within the scope of the agreed services, and certainly beyond that as well.

Neither the dealer nor I would want a Lynx Shunt to disconnect all of a sudden, and presumably neither of us is to blame. ■■■■ happens. I can’t get my head around the fact that a system that had previously been running properly with very plausible values is now, after this hang-up and the subsequent reboots of the shunt, suddenly reporting completely different (implausible) values. If the problem were related to the basic setup or the basic wiring, it would have had to have been present in this form over the past few months. The fact that the problem is unambiguously connected in time with a shunt that had hung and the reset that was subsequently required points to a problem with the shunt, or simply to something I have not thought of yet. Hence my question to the collective wisdom here: does anyone have any other ideas?

I am in the process of getting suitable current clamps so that I can measure the actual currents and see whether the Multiplus/Cerbo are right (they claim, for example, that 9KW is flowing towards the DC/battery) or the shunt is right (it claims that it is only seeing 50–60% of this power).

Thanks to everyone who reads, offers support and contributes their thoughts—and of course to the dealer, who has supported the project up to this point.

9

According to the manual, the default lead setting was reactivated after the update.

Find your battery data and adjust it if necessary.

10

Right. I’ve bought a clamp meter. It confirms the measurements. I tell the Cerbo “draw 9 kW from AC-IN towards the batteries”. The three MultiPlus units each draw around 3 kW AC from the mains. On the DC side, each outputs 53.5 A according to the clamp meter. Each of the six battery strings is carrying around 27 A. Both figures are within the measurement tolerances for around 161 A. The MultiPlus Modbus values also report around 160 A!

Only the shunt reports 98.5 A.

Since I’ve been logging the AC and DC values almost every second for months, I can pinpoint the discrepancy precisely to the moment when the shunt was reset. And the error factor is linear across the entire measurement range, from charging through discharging, and at least from -9 kW to +9 kW.

So, unless there’s some way to magically set parameters such as nominal voltage/current in the shunt, and those parameters were corrupted during the reset, I’m afraid the shunt must be faulty…

11

The DC current clamp is the best option for a plausibility check. Given these symptoms, there is a suspicion that someone accidentally pressed the zero calibration button while current was flowing. To check this roughly, all loads and chargers must be switched off. The actual current should then be approximately 0. If a current is still displayed, it can be compared with the measurement from the current clamp. If the clamp shows almost no current but the system displays a much higher value, the shunt must be completely de-energised by disconnecting it, and zero calibration must then be pressed.

12

See my la

See my last post. The clamp meter confirms that the shunt is displaying/measuring incorrectly.

The calibration was performed with zero current. All inverters off. No current. Calibration. Everything reads 0. As soon as the inverter operates (regardless of the direction), I get a measurement error factor of 0.6. If the zero calibration were incorrect, everything would be offset, but it would not be consistently off by a factor of 0.6 across the entire measurement range.

13

Calibration at zero amps could still be the problem. At zero amps, only the zero offset is calibrated; to establish a characteristic curve, you need at least one additional point at a higher current.

Calibrate it again at a constant current of 100 A, and the factor should be correct again. Assuming this can be done with the Lynx Shunt. I know this from the JK BMS units, which have a corresponding function.

14

I wouldn’t know how. The Cerbo only offers zero-current calibration. And the other parameters should actually be stored permanently in the shunt (after all, it knows that it’s a 1,000 A shunt). Essentially, I’m looking for exactly this option: to specify the parameters permanently or teach them in, and you’re right, that would require two measurements. Unfortunately, those two measurements simply aren’t available (or I haven’t found them anywhere).

15

The only thing I can think of is compensating for this error within your own programmed control system. If it’s always a factor of 0.6, that should be easy enough to do. The only thing that would then be incorrect is the reading directly on the Cerbo or in VRM. Alternatively, replace the Lynx Shunt.

16

It really does look as though the shunt is defective. Otherwise, this would only be possible if it had accidentally been bypassed. I’ve just checked my shunt: I could only find the “Reset to default settings” option via a Bluetooth connection. Have you tried that?

17

But that doesn’t help me, though

I had already thought of that too. But that still doesn’t help me with the SOC, which would then surely be calculated incorrectly by the shunt…

18

I definitely haven’t bypassed anything. Between “everything was working before” and “now it’s measuring incorrectly”, the only thing that happened was disconnecting the power via the bottom three wires and then reconnecting it. Nothing else has changed.

Do you mean Bluetooth via the Victron VRM app? I haven’t tried that yet because I’m worried about overwriting important settings. I’ll gladly give it a try tomorrow with the support team (who are following this thread). Otherwise, the shunt is only three months old, and I’d be “grateful” if Victron replaced it.

Thank you all for your thoughts and help!

19

No, via the “VictronConnect” app / local. Tap the gear icon and then the three dots. Using your smartphone.

You can also find the option there to export or import the settings.

20

That’s the one I meant. Sorry. I expressed myself badly.

I probably first need to enable “local Bluetooth” on the Cerbo. I generally only use VRM, when I use it at all! And then do I see the Shunt via Bluetooth or through the Cerbo? Sorry for asking such a silly question.