When a current is flowing via a reactive load, say a 1-2kW inverter, or indeed, a fault current through the inductance of a few metres of cabling, when the circuit is interrupted you can get very high inductive transient voltages across the thing doing the interrupting (usually some sort of "switch" - a fuse in this case). Here's an example circuit lifted from incompliancemag.com:
What is a VDC here is the battery, the switch is our fuse, the LCR components are the load. The shunt isn't shown.
The battery is fairly low impedance (for the purposes of this discussion, let's call it a perfect battery at 0Ω), so anything on the battery side of the "switch" (fuse) will not see much in the way of any transient - you just won't budge the battery-side terminal of the switch, V1 referenced to ground will stay at more or less 24V in this example regardless of what spikes occur across the switch (V2-V1).
The load side of the switch, the voltage at V2 referenced to ground, will see pretty much the full spike. Any sensitive components - say, a semiconductor - connected to that node will cop a hiding.
One fix for this is to use additional components (flyback diodes across the load, MOV across the switch) to try and suppress these transient voltages. Given these parts of the circuit are outside the control of the BMS designer however, a more reliable - and simpler - arrangement is to install any sensitive stuff that you want to avoid these transients either before the switch (i.e. on the battery-side of the switch), or on the low-side of the circuit (i.e. on the negative terminal). The SBMS0 is designed for high-side sensing, so we're left with the one option of "before the switch".
IMHO, this is a perfectly reasonable and sensible approach - I don't see any drawback to putting the shunts on the battery-side of any disconnects (breakers, fuses, BPs, etc) so long as you avoid overly long cabling between the battery and shunts (i.e. same as you would/should between the battery to fuse if you weren't using a shunt).
What is a VDC here is the battery, the switch is our fuse, the LCR components are the load. The shunt isn't shown.
The battery is fairly low impedance (for the purposes of this discussion, let's call it a perfect battery at 0Ω), so anything on the battery side of the "switch" (fuse) will not see much in the way of any transient - you just won't budge the battery-side terminal of the switch, V1 referenced to ground will stay at more or less 24V in this example regardless of what spikes occur across the switch (V2-V1).
The load side of the switch, the voltage at V2 referenced to ground, will see pretty much the full spike. Any sensitive components - say, a semiconductor - connected to that node will cop a hiding.
One fix for this is to use additional components (flyback diodes across the load, MOV across the switch) to try and suppress these transient voltages. Given these parts of the circuit are outside the control of the BMS designer however, a more reliable - and simpler - arrangement is to install any sensitive stuff that you want to avoid these transients either before the switch (i.e. on the battery-side of the switch), or on the low-side of the circuit (i.e. on the negative terminal). The SBMS0 is designed for high-side sensing, so we're left with the one option of "before the switch".
IMHO, this is a perfectly reasonable and sensible approach - I don't see any drawback to putting the shunts on the battery-side of any disconnects (breakers, fuses, BPs, etc) so long as you avoid overly long cabling between the battery and shunts (i.e. same as you would/should between the battery to fuse if you weren't using a shunt).
Last edited: