[roach374]

I'm doing an EV conversion on a motorcycle ('82 CB900F, for those wondering). I'm planning on a 32S LiFePO4 pack, so around 10kW. In previous (non-electric) motorcycle projects, I always use a chassis ground to simplify my wiring. Given the voltages involved here, and all the good arguments against using a chassis ground, I'm curious to know about people's personal experiences: Have you used a chassis ground at these voltage levels? Any horror stories, or did it turn out fine?

(Obviously, I plan to take all necessary precautions, cap my terminals, etc).

 

[cricketo]

You don't want to run large currents through the chassis for any purpose given the fact chassis is typically made out of materials that aren't great conductors (i.e. steel), plus connections between the components aren't great. For low current applications like blinker / stop lights it may be acceptable. The thing to watch out for is some battery cells have a positive terminal on their bodies, so if the chassis is intentionally interconnected with the negative, chances of a short circuit are increased in case of a collision. There may be other considerations with isolated DC-DC converters. Bottom line - if you're reusing existing wiring to 12V automotive accessory equipment, then chassis may be fine - but that's also where isolated DC-DC comes in, you may want to connect the negative output from DC-DC to the chassis, but leave the traction battery decoupled from the chassis.

 

[mpgMike]

Probably not your concern, but doesn't current flow through steel promote oxidation/rust? Any nicks in the paint would become major rust spots (if I remember my high school chemistry right).

 

[brian_]

Definitely use the chassis for a "ground" or negative return path for the 12 volt stuff that is always done that way, but don't connect either side of the traction (high voltage, EV, whatever you want to call it) system to the chassis.

 

[remy_martian]

In addition to what's already been said, ideally, you'd want to fault detect HV to chassis leakage current. Likely not within the grasp of many homebuilders, but, as I said..."ideally".

 

[roach374]

Thanks all for your help, this is super valuable info! My main concern is that my DC-DC converter only has one ground (basically, it's HV in, 12V out, and a single ground wire). While all the drive components will be wired directly to the battery negative (not the chassis), at some point, the chassis will become the common ground for the 12V and HV circuits (won't it?).

Since I also have direct, copper, 2AWG wiring to the battery negative for all the drive electrics, won't current take the path of least resistance?

...But the point about increased risk in the event of a collision is a good one. Need to think about that.

 

[Electric Land Cruiser]

For that reason I put my BMS inside the battery box the same as OEMs do it. I saw a video of a homebuilt EV conversion and his BMS was outside the battery box (or maybe there was none) and he was detecting full pack voltage against the chassis

 

[cricketo]

Thanks all for your help, this is super valuable info! My main concern is that my DC-DC converter only has one ground (basically, it's HV in, 12V out, and a single ground wire). While all the drive components will be wired directly to the battery negative (not the chassis), at some point, the chassis will become the common ground for the 12V and HV circuits (won't it?).

So you have a non-isolated DC-DC. Isolated would have separate negatives (grounds) on input and output. What kind of battery cells are you using ?

 

[remy_martian]

No. The DC-DC should be isolated from the HV side. You have two cables in for the HV side, a plus and minus, not one (I hope...).

The GND should be for the 12V side only.

If it's not, you have a safety issue and you should punt the converter to the curb and get another.

 

[cricketo]

It is a 100v system, isolation would be warranted. Sevcon makes decent industrial grade isolated DC-DC converters, they're used for instance on Zero motorcycles. Sevcon DC/dc convertor 13.5v 300 watt for 80v system | eBay (need to check the max voltage, but I think 80V is the one that goes above 100v for input max). There are likely cheaper options available as well.

Now I don't know about systems above 48V, but 36 and 48v systems are pretty common with negative traction on the chassis. I have two industrial EVs that have that from the factory.

 

[gunnarhs]

Now I don't know about systems above 48V, but 36 and 48v systems are pretty common with negative traction on the chassis. I have two industrial EVs that have that from the factory.

The (old) rule was that below 50V was allowed to be exposed, meaning negative HV was allowed to have common negative (ground with chassis.).
This was common in older applications with low-frequency PWM modulators or armature resistance controllers.
Today the "High Voltage" side uses "high" frequency (15-20 kHz) PWM even for low voltages ( 48V).
This high frequency modulators (which are both in the Motor-controller and DC/DC-converter) can produce EMI problems for electronics designed for common 12V DC.
So even if the regulations allow it it is not recommended to have common negative (also chassis ground in most cases) for the high voltage and low voltage side.

The other security aspect is that over 30V DC can create a quite descent "lightbow" if close distance is between "+" and "-" and some conducting material gets between.
Even if no fire is started this creates ugly damages on the cables and surroundings. This is the reason many auxiliary systems in cars are limited to 24 V.

 

[gunnarhs]

Now I don't know about systems above 48V, but 36 and 48v systems are pretty common with negative traction on the chassis. I have two industrial EVs that have that from the factory.

The (old) rule was that below 50V was allowed to be exposed, meaning negative HV was allowed to have common negative (ground with chassis.).
This was common in older applications with low-frequency PWM modulators or armature resistance controllers.
Today the "High Voltage" side uses "high" frequency (15-20 kHz) PWM even for low voltages ( 48V).
This high frequency modulators (which are both in the Motor-controller and DC/DC-converter) can produce EMI problems for electronics designed for common 12V DC.
So even if the regulations allow it it is not recommended to have common negative (also chassis ground in most cases) for the high voltage and low voltage side.

The other security aspect is that over 30V DC can create a quite descent "lightbow" if close distance is between "+" and "-" and some conducting material gets between.
Even if no fire is started this creates ugly damages on the cables and surroundings. This is the reason many auxiliary systems in cars are limited to 24 V.

 

[cricketo]

The (old) rule was that below 50V was allowed to be exposed, meaning negative HV was allowed to have common negative (ground with chassis.).
This was common in older applications with low-frequency PWM modulators or armature resistance controllers.
Today the "High Voltage" side uses "high" frequency (15-20 kHz) PWM even for low voltages ( 48V).
This high frequency modulators (which are both in the Motor-controller and DC/DC-converter) can produce EMI problems for electronics designed for common 12V DC.
So even if the regulations allow it it is not recommended to have common negative (also chassis ground in most cases) for the high voltage and low voltage side.

The other security aspect is that over 30V DC can create a quite descent "lightbow" if close distance is between "+" and "-" and some conducting material gets between.
Even if no fire is started this creates ugly damages on the cables and surroundings. This is the reason many auxiliary systems in cars are limited to 24 V.

One of the two machines I mentioned is indeed is the older low-frequency SCR-controlled machine. But the other one runs on Sevcon MOS90 - older, but still relatively modern MOSFET controller. It has a DC-DC onboard for 24VDC output, and it's non-isolated I believe.