BMS "Semi" Bypass

[chuyskywalker]

I'm working on a build where I'm going to put a 14s5p 18650 Molicel p26a battery together in a Polly DP-6 case.



This case has just enough space for the cells, the necessary wiring, and a small 35A separate port, key switch enabled, bms. The pack, though, is capable of pushing up to 125-175A (5x25-35A each cell) with those cells, and I'd like to take advantage of that (really, only about 100A), yet I don't want to lose the BMS safety features. Problem is, a capable 100+A BMS will not fit in this case.

I believe I can maintain the same relative safety, though, using a feature that a few controllers come with: the "ignite" (or key switch) on the controller itself. This feature has the controller pass battery positive out a small gauge wire and expects the voltage to return on the "ignite" wire. This is typically used both as a way to signal the controller to turn on, but also to get one of those half-twist throttles with voltage display and key switch combos to work up at the handle bars.

Here's what I propose:


(Yeah, my diagram is decidedly NOT electrical engineer accurate, oh well!)

On the battery pack case, it has 5 pins. I'll use pins 1&2 for positive and 4&5 for direct negative. These will route to the controller directly. Internally, the BMS will be wired up as normal but the P-, instead of supplying the negative for the pins on the case, will connect to pin 3 only. Pin 3 from the cradle will then be routed, along with a split off the positive pins, to a DCDC convert to get me down to a workable relay range. This relay, along with a key switch, will interrupt the "ignite" wire circuit.

In this fashion:

• If the BMS goes into LVC/HVC for all or even a single group, the P- cuts off, the relay closes, and the controller turns off
• If the BMS switch is turned off, the P- cuts and the controller is offed
• If the key switch is disconnected, the controller is off
• The overall physical size of this setup is very small (the dcdc and relay are tiny, especially compared to a contactor of this size)

What this does not save me from, which the normal setup could, is if the BMS wants to be shut off, but the ignite line is being ignored for some reason and the controller just keeps pulling. I'm not super worried about this scenario, though. The other thing this doesn't prevent is a connection spark; ie: as I put the battery on/off the cradle, it'll still fast charge the capacitors on the controller. I plan to address that with an XT90 AntiSpark loop key on the positive line between the battery tray and controller.

Have I missed anything? Is something going to blow up in my face here?

 

[amberwolf]

Your block diagram should work, but I would put a switch between the BMS P- output and the DC-DC - input, or between the BMS B+ output and the DC-DC + input, so you can turn it off and prevent power drain. Alternately if you have a master B+ cutoff switch for the whole system at the main battery B+ terminal it would do the same thing when the system is not in use at all (or needs maintenance, or is parked as an antitheft measure, etc).

If you have a 12v system for lights, I'd just use the 12V (13.6-14.4v really) DC-DC that runs those to control a 12v-coil relay, instead of the 5v one. Then you can use a separate low voltage switch on the 12v side for turning the lights on and off if you don't just run them all the time the system is powered on.

The ignition (KSI, keyswitch / ignition, doorlock, etc) wire on controllers usually provides the battery power to the LVPS that runs the controller brain and all the other non-battery powered stuff run off the controller.

So cutting power to that disables the system by removing all power from those things.

Not disconnecting the pack completely from the system only leaves whatever load there is from the FETs themselves, any display, and any DC-DC or other devices powered from the unswitched battery voltage.

If you put all those devices on the relay with the KSI wire, then the only unswitched load is from the FETs, which under normal operation will be virtually zero when the KSI wire doesn't get power. The only time they should create a battery load is when the KSI has power and the controller is switching the FETs on to pass current to the motor, or if there is a FET failure allowing current flow (shorted FETs).

As long as you have an appropriate fuse or breaker, right at the output from the battery, sized to allow all the current you need but blow in the event of any wiring shorts or device faults that create currents exceeding wire capabilities, then shutting down only the controller brain this way should work fine.

 

[chuyskywalker]

Your block diagram should work, but I would put a switch between the BMS P- output and the DC-DC - input, or between the BMS B+ output and the DC-DC + input, so you can turn it off and prevent power drain. Alternately if you have a master B+ cutoff switch for the whole system at the main battery B+ terminal it would do the same thing when the system is not in use at all (or needs maintenance, or is parked as an antitheft measure, etc).

That's a good point, I think I'll just repurpose the AS loop key and have the DCDC positive come after that. Thus if I pull the loop key, nothing is connected to the battery at all.

If you have a 12v system for lights, I'd just use the 12V (13.6-14.4v really) DC-DC that runs those to control a 12v-coil relay, instead of the 5v one. Then you can use a separate low voltage switch on the 12v side for turning the lights on and off if you don't just run them all the time the system is powered on.

Won't be a lighting system on this one or any other 12v stuff -- it's a fair weather, putz-about fun ride only kinda thing. So I just went with the smallest dc/dc I could find that I've used in the past successfully. If I need to change that up in the future, the relay I went with has a range of 3-32v DC for the coil.