Should a metal battery box be connected to Negative?

[Simon]

If I am putting a 48V lifepo battery in a metal box. should the box be connected to the batteries negative terminal?

[floobydust]

I would leave it unconnected, as one less fault that needs to happen before disaster.
The box might end up unintentionally grounded if the box is laying say on a metal vehicle chassis. This can cause other problems.

I'm reminded of the Li-ion battery fire in the FedEx truck, following air transport. NTSNB Lithium-Ion Battery Truck Fire Following Aerial Transport
It was a custom built pack and the insulation has no standards, and chafed through due to vibration in shipping. At least it did not do it on the plane!
"...  was an electrical short circuit between the battery terminal bolt and the upper cells of the lithium-ion battery module..."

[ajb]

Three reasons you might do this come to mind:

- the battery box/chassis serves as a common electrical connection point (like in cars)
- the box needs to be bonded for shielding purposes, and battery negative is directly connected to the shielding node (eg circuit ground)
- there are hazardous voltages inside the box, and it needs to be bonded to prevent those voltages from appearing on the outside of the housing in case of a fault (like in metal-cased mains-powered appliances, but could apply to a device that generates high voltages itself from a battery, if there are other connections that could make this hazardous)

So if any of those or some other good reason apply, then yes, you should connect the box to battery box to the battery.  Otherwise there's not really anything inherent to batteries/battery boxes that requires them to be bonded.

[Siwastaja]

+1 for unconnected. You still should be very careful with designing the electrical insulation (including wear due to vibration etc.), but at least you have one layer of safety more when you don't purposely connect the box to any potential within the pack.

[Andy Chee]

If you choose to connect the metal battery box to negative, then you should have a terminal cover over the positive, and have a main fuse physically located at the positive terminal underneath the cover.  See automotive 12V engine bay installations for examples.

[Simon]

Three reasons you might do this come to mind:

- the battery box/chassis serves as a common electrical connection point (like in cars)
- the box needs to be bonded for shielding purposes, and battery negative is directly connected to the shielding node (eg circuit ground)
- there are hazardous voltages inside the box, and it needs to be bonded to prevent those voltages from appearing on the outside of the housing in case of a fault (like in metal-cased mains-powered appliances, but could apply to a device that generates high voltages itself from a battery, if there are other connections that could make this hazardous)

So if any of those or some other good reason apply, then yes, you should connect the box to battery box to the battery.  Otherwise there's not really anything inherent to batteries/battery boxes that requires them to be bonded.

The whole system is floating. We only used a metal box for durability. There is no electrical reason for the box being metal. One thing does occur to me. If someone decided to put two of them in series, the box will have potential, and if two boxes touch, bang! there is a fuse and the batteries should have a BMS that will cut out - these can fail. We don't expect to use these in series.

[Simon]

There is actually more positive wiring in the box than negative as there is a fuse, two contactors and two output sockets.

[floobydust]

A floating (ground) box will pickup RF and the BMS susceptibility you'd have to be solid about.

Another option is to "soft ground" the enclosure as is done for ESD and EMI.
A 1MEG or less (for 48V) and 10nF would stop ESD hits to the box from possibly arcing to the internals.

[coppercone2]

or a fuse

a indicator fuse would be nice to show if there was some problem. I think the RF ground would be alot better then some capacitor/resistor combination if its just a low current fuse

[Simon]

I am confused. Why are we assuming that the metal box is any form of ground. It's a metal box that sits on rubber feet. It could be plonked in a muddy field up to it's bottom in wetness just as much as it could be on hard tarmac on a dry day. If the box is hit with RF then it will shield the batty if unconnected. I expect that if connected it will just pass anything it picks up into the negative power rail.

[floobydust]

The metal battery box:
-Is a source of system ESD when moved or plonked on the ground. A person touching the box can discharge themselves and nail it, or any stray charge on the system will discharge to the box and the earth ground when it's set down.
-Is an antenna.

I would use the resistor+capacitor soft ground - it will not allow any high DC currents, also keep the box at low potential at high frequencies so the BMS does not see common-mode RF, and ESD hits will be kept to a low voltage.

[Manul]

I am confused. Why are we assuming that the metal box is any form of ground. It's a metal box that sits on rubber feet. It could be plonked in a muddy field up to it's bottom in wetness just as much as it could be on hard tarmac on a dry day. If the box is hit with RF then it will shield the batty if unconnected. I expect that if connected it will just pass anything it picks up into the negative power rail.

If the metal box is floating, then it is just a huge capacitor plate, another one being the battery internals. If the battery internals are not connected anywhere, then as you say, it's all just shielded. But the battery output will likely be connected to something. In that case any AC voltage appearing on the box is capacitively coupled to the battery internals and passes to the output. AC bypassing the box directly to the output (say negative terminal) creates a direct path, avoiding parasitic coupling through the internals (say BMS).

[Simon]

Yes the battery is connected to a machine with a servo motor in it. Soft linking sounds like a good idea, I am really nervous about just connecting the whole thing straight to the negative.

[Siwastaja]

If whatever connects to the battery has proper filtering, EMC will not be a problem. Having to ground a protective battery metal box for EMC sounds like a ugly hack to me. Think about EVs for example where hundreds of A are switched in a motor controller, battery charger and DC/DC converter. Battery DC is isolated from any of the external metal casing for safety.

[SteveThackery]

Hang on, I thought the box acts as a Faraday shield against the electromagnetic radiation from inside the box escaping. It does not need to be connected to anything to do that.  Faraday shields do not need to be grounded.

[tszaboo]

Three reasons you might do this come to mind:

- the battery box/chassis serves as a common electrical connection point (like in cars)
- the box needs to be bonded for shielding purposes, and battery negative is directly connected to the shielding node (eg circuit ground)
- there are hazardous voltages inside the box, and it needs to be bonded to prevent those voltages from appearing on the outside of the housing in case of a fault (like in metal-cased mains-powered appliances, but could apply to a device that generates high voltages itself from a battery, if there are other connections that could make this hazardous)

So if any of those or some other good reason apply, then yes, you should connect the box to battery box to the battery.  Otherwise there's not really anything inherent to batteries/battery boxes that requires them to be bonded.

The whole system is floating. We only used a metal box for durability. There is no electrical reason for the box being metal. One thing does occur to me. If someone decided to put two of them in series, the box will have potential, and if two boxes touch, bang! there is a fuse and the batteries should have a BMS that will cut out - these can fail. We don't expect to use these in series.

You should identify which standards apply to your product, and read though that.
I think IEC-62133-2 is a good start.
If I were you, even if I would connect it, it might be with a fusible resistor or something like that.

[ejeffrey]

If the metal box is floating, then it is just a huge capacitor plate, another one being the battery internals. If the battery internals are not connected anywhere, then as you say, it's all just shielded. But the battery output will likely be connected to something. In that case any AC voltage appearing on the box is capacitively coupled to the battery internals and passes to the output.

Not really any differently than if the box isn't there.  If you aren't worried about capacitive coupling from outside the battery to its internal electrodes then a floating metal box around it doesn't make that appreciably better or worse.

I don't see any real need to connect the battery box to a given potential, and there are definitely some potential downsides (such as series connected batteries).  Instead I would say don't leave it to chance.  If there is an exposed metal frame that is connected to circuit ground (such as in a car) then either insulate the box so it doesn't make accidental contact or bond it intentionally to the circuit ground, regardless of which (if any) battery terminal that corresponds to.

[floobydust]

It's not a Faraday cage when the inside contents has conductors leading out. This allows an E-field to exist between the box and towards the batteries and BMS inside it.
We don't know a bunch of details, such as the machine the servos are mounted too, if the machine is grounded etc. But from experience I know the RC soft ground works well enough, and the RC can be depopulated or replaced with gas tube, polyfuse etc. whatever you think.

Leaving the box floating, the isolation voltage rating (insulation) between box and insides would likely not be enough for ESD so it will nail the contents... somewhere. This is bad.
If you impose an RF signal on the box, the capacitor's other plate is the BMS and batteries so common-mode currents will happen and I have no idea if the BMS can handle it. EE's get stuck thinking everything to do with a battery is simple DC and no RF bypass here or there can cause unexpected susceptibility.
Really this situation is no different than putting a PC board inside a metal enclosure, that has outside wiring coming in to the board. Floating the box does not work well.

A hard ground you have to do something with limiting DC fault currents so wiring could not melt and burn. It's the worst way to go I think. Even a fuse, you have no idea it blew so what was the point?

[Simon]

The machine has a metal frame. The motor driver does have a resistor and capacitor between the negative rail and the heatsink. The machine itself is is not earthed, but as it is lugged around there is no telling if it will end up grounded or not. It is intended to work as a floating system.

[PCB.Wiz]

The machine has a metal frame. The motor driver does have a resistor and capacitor between the negative rail and the heatsink. The machine itself is is not earthed, but as it is lugged around there is no telling if it will end up grounded or not. It is intended to work as a floating system.

Then I would float your metal box too.

You could add a capacitor to from case to negative and if the system has sensitive parts, you could also add a ESD part between a rail, and the case.
That gives a predictable discharge path for large ESD strikes, and avoids the ESD find it's own path thru something sensitive.
At normal battery voltage it appears open.

[floobydust]

Some machines generate static electricity hence the "machine discharge" ESD model. It seems to be when they are dealing with ungrounded machines dealing with machining plastics or rolls of plastic etc. it gets really nasty.

I still say float the box for DC, if it's on the ground then add a static dissipative resistance 100k-1MEG, and Y-cap to shunt RF and attenuate ESD I use 10-33nF.
A GDT gives a hard discharge path and the currents are high, it makes EMI.

[Randy222]

Is the metal frame/box an item that you ground other things too for current flows? If so then ground loops are to be evaluated. This also means the frame/box becomes current carrying metal.

If you leave it non-grounded then you risk having the metal box becoming +V, which could may become an issue.

With current protection on the +V I would just tie metal box to GND to hold the metal at the "zero" potential, and do not use the frame/box for ground points like it is done in automotive world.

[Simon]

The battery box is separate from the machines. They can be used independently with whatever the customer likes providing they put a link in the cable plug to activate the contactor when plugged in. Loads are always connected by means of a cable. The box is not attached to the equipment although contact cannot be ruled out.

[Simon]

You should identify which standards apply to your product, and read though that.
I think IEC-62133-2 is a good start.
If I were you, even if I would connect it, it might be with a fusible resistor or something like that.

That standard concerns the battery itself. We buy the battery off the shelf with the BMS built in and simply strap it into a box to protect it with connectors and contactors to access the battery power.

[Siwastaja]

If the system is built out of certified modules it has high chances of passing EMC as a whole, too. If it fails, then investigate. A metal box won't make it any worse for EMC. Tying the box into battery negative would be the last resort option IMHO; if it fails EMC try other means first (filtering on power and communication lines).