Strange behavior involved with shielding

[diorcety]

Hello everyone,
At work I faced a problem that I can't really explain even after many days to experiment.

The situation: EV traction with electrical motor, inverter, some distribution boxes and a battery stack of 600V (>60kw). All these devices are "grounded" to the vehicle including the +/- cable shields at each sides. +/- are fully isolated from the chassis (bigger than 5000ohm/v)

New vehicle in conception, including a more powerful motor (3x3 phases, the previous one was 2x3phases). We noticed lot of noice on the min/max voltages on the cells.(-200mv noise on some 3.2V cells), when the inverter start working (just pressing a bit the accelerator pedal, so not lot of current). After few days of investigation we arrived to this conclusion: this is not a measurement issue on our system, we directly measured the voltage on the cells, across multiple cells, the noise is really present on the cells. If we keep the shielding braid to the battery enclosure, but removing the shielding of +/- cable on few centimeters, the noise disappear completely (200mv of noise to 1mv).

So it doesn't look like a voltage potential issue or a loop issue (we tried with the battery outside the vehicle without shielding braid, the result is the same), but more like the shielding of the cables conduct the issue. Strange that the shielding which is isolated from +/- introduce that issue, and that a better current path(using shielding braids) doesn't change something in this behavior. Resonance issue on the system?

 

[andre_luis]

The situation: EV traction with electrical motor, inverter, some distribution boxes and a battery stack of 600V (>60kw). All these devices are "grounded" to the vehicle including the +/- cable shields at each sides. +/- are fully isolated from the chassis (bigger than 5000ohm/v)

Although intuitively grounding both sides of a cable may seem like a good idea, depending on the type of application, particularly where high-intensity magnetic fields are generated by powerfull motors, this loop can rather act as an real antenna, picking up noise at one point in the cable and retransmitting by induction to the circuit at another part of the cable.

 

[diorcety]

Thanks for the reply. Yes I also think that, but I can't really understand how a "noise" can produce such voltage variation on the cell voltage itself, which is very low impedance(the cell can deliver up to 200A). My second through is that there is an isolation issue at one moment: during the switching of the inverter with sparks and a very short high current call, which may be undetected by the vehicle isolation test system and the inverter one, for example on one phase. I will try to investigate this idea.

 

[D.A.(Tony)Stewart]

If the cable is not matched impedance at both ends the signal coupling to ground at both ends will cause circulating noise like ground loops. This may also cause shoot thru spikes which definitely will cause voltage drop.

Solution: use ground shield on source end only.

 

[andre_luis]

...but I can't really understand how a "noise" can produce such voltage variation on the cell voltage itself, which is very low impedance(the cell can deliver up to 200A....

Although you stated this, "...this is not a measurement issue on our system..." I wonder how you actually came to the conclusion because in these vehicles the dimensions are large and the way we measure matters a lot. Were these values of 1mV and 200mV measured by a multimeter or oscilloscope? Unlike the 2nd case where a coaxial cable is used, in the 1st case, the cables if not twisted can act as an antenna, the larger the free area inbetween.

 

[D.A.(Tony)Stewart]

I think he scoped both readings but failed to report the pulse width and rep. rate of the battery drop voltage from small acceleration.

He was getting signal integrity issues with a shield ground loop from mismatched , I believe from mismatched impedances that caused shoot-thru of the FET bridge.

The ESR of the battery if you measure it will tell me how much current caused those battery pulse drop voltages.

The ESR is valid, if SoC is good, BMS is working and batteries are all in equally good condition. We can assume nothing but if a cell is only 3.2V and is Lithium Ion it may have much higher ESR than if it were 3.6V.