Why run main power thru a BMS?

[John in CR]

Hi power BMS's are expensive, but why not just run only the power for the controller itself through the BMS along with charge current? A cell gets down to cutoff and the BMS turns the controller off. It doesn't need to do any more does it, except when charging does it?

John

 

[el_walto]

The BMS will cut out if the battery is shorted or a damaging amount of current is pulled from it. A fuse can do this also.

The BMS can not turn off the output leads if the output leads are not going through the BMS.

Do people need both these features? Many probably don't and prefer to use a throttle cutoff instead. Leaving your lights on, or anything else would still destroy your battery as the output leads would never turn off.

I think a few people have killed their packs by leaving auxiliary stuff on, or even a controller that is pulling current.

 

[agniusm]

I just wandered the same and was talking about this with my mate. Why use discharge fets at all? I made myself A123 starter replacement battery for my car, for winter and there is no BMS that can do 450-600A discharge while cranking. Ok for charging but BMS could have just an output to switch off controller for that purpose if cells approach LVC and people would program they controller to pull currents only capable by the battery if they want their pack to last. In most cases controllers are limited to small currents anyway.
So how to go about this to avoid going through discharge fets? I have no electric background what so ever.

 

[John in CR]

Exactly. I want the controller controlling the current, not a BMS, aka Battery Murdering System. Fuses are the correct protection in the event of shorts, and my keyswitch turns everything off except the BMS. I'd even add a by-pass switch for my controller brain power, so when the BMS cuts it off, and I check the battery with my multimeter and it did so in error, then it doesn't leave me stranded.

My thought is that if I'm not pulling main current through the BMS, then it shouldn't fail. How is less than 100mA going to make it fail except maybe during charging?

John

 

[heathyoung]

Yep - this a good way to use a small BMS on BIG cells - use its discharge FETS to run a contactor/relay, and not worry about trying to pass oodles of current through sub-standard FET's and shunts.

 

[agniusm]

I need to get myself cheap lifepo4 cells for test bench and start playing with electronics. Looks like discharge fets are there to make excuse for higher BMS price. None of distributed BMS system has current limiting or it is programmable option.
DNMUN has a solution i think.

Yep - this a good way to use a small BMS on BIG cells - use its discharge FETS to run a contactor/relay, and not worry about trying to pass oodles of current through sub-standard FET's and shunts.
OK, so you say connect negative from the battery to the controller avoiding BMS and connect some sort of relay on BMS negative output to controller? Sounds simple enough??!!

 

[heathyoung]

You still need to connect the negative of the battery to the BMS (for charging purposes) - BUT you use the output from the discharge FET (-ve) to run either an appropriatly rated relay off the pack voltage, or a DCC converter (or a resistor if you swing that way) to a lower volatge coil contactor.

OR

Use a small, appropriatly rated relay running from the BMS negative output (ie. coil and contacts rated for pack voltage) to switch the controller 'ignition' signal - no high current work is done, no voltage drop, no muss no fuss.

Either method means you can use a cheapie BMS (15/20A rated) on the biggest, baddest cells you like, and still have charge control, and individual LVC.

 

[agniusm]

OK, so basically i would tend to avoid all that complexity and rather just watch my voltmeter so the pack does not SAG below safe voltage. On few tests i done cells do not tend to go out of balance while discharging and as i was told it tends happens at the HVC and BMS will deal with that.
Same method as with ICE. If you have red light, you go to gas station, here, when voltmeter shows low voltage, you have to search for a socket, right?

A lot of folks here dont use BMS at all and HVC protection for high current freaks would suffice i think.

 

[John in CR]

Why should I go through the relay/contactor complication? I want to just run the BMS output to my keyswitch input that turns my controller on and supplies the controller itself the less than 100mA it requires. If the BMS trips, it turns the controller off just as if I turned my key off.

 

[heathyoung]

Yeah, thats the second method I suggested - use a small relay to do that.

 

[John in CR]

Yeah, thats the second method I suggested - use a small relay to do that.

I don't understand the need for the relay. Positive primary goes to battery positive instead of to the BMS like it normally would using a BMS, and then the smaller pack voltage supply wire that powers the controller brain goes to the BMS positive output with my keyswitch in between. If that's not correct then I'm missing something.

 

[bigmoose]

I have been mulling this over for a bit on the unused neurons in my noggin...

For first order I think the BMS should:

• signal the controller to stop during a LVC event
• signal the charger to stop during a HVC event

In other words the controller and the charger should be able to accept an "arm" command from the BMS. Without the "arm" command, the "fire" command (twisting the throttle) or (providing charge) should be disabled. Simple, low voltage and current signal path.

A fuse should be integral to the battery pack for safety.

Now we have to think of second order safety effects. For a brushed motor, if the controller FETs fuse (short) we can have a run away at full power. Hence I think a contactor is required in the battery pack to provide propulsive power cut off. It is signaled the same way, with an "arm" command from the controller.

Now what safety shuts offs are required for 3P BLDC motors? We should discuss this and the known controller failure modes. Like twist throttle failures resulting in "full forward" commands. Could be handled with a contactor, or could have another layer of low voltage "arm" signals in the controller from say an accessible key switch to enable the PWM circuitry.

We should think this all through and evolve to a 21st century "system topology" for a high powered eBike.

Also a topological recommendation from now defunct A123. They required that if you had a microprocessor in the mix for a "smart" BMS, that there needed to be another, hardware only, redline BMS that was redundant and could shut things down if the micro hung/failed. Both TI and Maxim make chips that do this function in hardware.

This discussion would center around the "Failure Tolerance" of the BMS. In the A123 example above, they are recommending single failure tolerance. What is prudent on a high powered eBike?

 

[jonescg]

I've not had much experience with BMS's for ebikes, but I was using the EV-Power BMS on Voltron and it worked like a charm. All it did was sound an alarm when LVC was hit and then charging, HVC resulted in the charge enable circuit of the Elcon/TC charger to get turned off. I could have wired the LVC into the throttle, or used it to disconnect the contactor or whatever, but at speed that wouldn't be wise.

So most BMS's I have had anything to do with never manage high currents, they simply warn you of a low or high cell. This is normal for most larger EVs. The EV-Power ones also bleed high cells down to 3.6 V, a nice feature, but not necessary if you keep your cells well balanced.

 

[John in CR]

Now what safety shuts offs are required for 3P BLDC motors? We should discuss this and the known controller failure modes. Like twist throttle failures resulting in "full forward" commands. Could be handled with a contactor, or could have another layer of low voltage "arm" signals in the controller from say an accessible key switch to enable the PWM circuitry.

We should think this all through and evolve to a 21st century "system topology" for a high powered eBike.

Also a topological recommendation from now defunct A123. They required that if you had a microprocessor in the mix for a "smart" BMS, that there needed to be another, hardware only, redline BMS that was redundant and could shut things down if the micro hung/failed. Both TI and Maxim make chips that do this function in hardware.

This discussion would center around the "Failure Tolerance" of the BMS. In the A123 example above, they are recommending single failure tolerance. What is prudent on a high powered eBike?

Ebrake first, and then an easy to reach keyswitch that supplies the MCU are my BLDC safeties for a runaway condition along with a fuse at the battery.

Regarding BMS, I have some low power BMS's and I thought ebike BMS's tried to be idiot proof and when a cell hits cutoff the BMS cuts main battery power which flows through the BMS. I never hit the low cutoff, but the overcurrent cutoff (which is why main power goes thru BMS to protect battery from high current), turned the battery off when it tripped. If I connect my keyswitched controller brain supply wire to the BMS instead of to battery positive, then if I hit a low cell, it turns my controller off by cutting power to the brain. I don't understand why I need a relay or anything else.

John

 

[dnmun]

what is happening there is that the drop in voltage as the cell is discharged in combination with the extra voltage drop across the mosfets causes cell #1 to see the LVC prematurely, before the cells above the ground hit LVC? and when you check, they all read 2.7V after they bounce back up from getting turned off.

the problem i see damaging the most batteries by BMS failure is when the output mosfets have burned up and shorted, so when the BMS cuts off the mosfets at LVC they don't shut off because they are shorted. seen that a buncha times now. almost killed one of my headway packs that way.

and like i said on the other thread, to discharge without using the BMS output mosfets, you can wire up an led on the dashboard between your dashboard power + and the BMS P- with a current limiting resistor on the led that allows you to monitor the LVC status of the pack and shut the equipment down manually . the big risk of running without a BMS to cut off the power at LVC is leaving a load on it unattended. that has happened to a buncha people. manufacturers have to include it in their equipment too or the customer will destroy it day one.

 

[heathyoung]

Yeah, thats the second method I suggested - use a small relay to do that.

I don't understand the need for the relay. Positive primary goes to battery positive instead of to the BMS like it normally would using a BMS, and then the smaller pack voltage supply wire that powers the controller brain goes to the BMS positive output with my keyswitch in between. If that's not correct then I'm missing something.

Easy. Due to the use of N channel FETs, the negative lead is switched. High current and low resistance P channel fets are very expensive compared to N channel.

Have a look at the design of the BMS - you will find that your positive lead is NOT switched. Relay serves to allow you to switch the positive lead to your controller's 'ignition' lead.

 

[John in CR]

Damn, you guys are going to make me have to learn some electronics beyond the most elementary stuff after all.

 

[dogman dan]

It seems to me like a generally sound approach John.

It's just the details of how to wire it that gets me too.

For smarter people who pay attention, perhaps just a large indicator light on the bars connected to the bms? Or maybe run the stereo through the bms? Light goes out, or music stops, stop riding. If all you run though the bms is accessories like lights, turn signals and horn, shouldn't the output fets tend to last a very long time?

Then the controller just bypasses the bms.

You 'd of course want to put a switch in line with that light, so it doesn't burn all night or for weeks. Double switch so one throw turns off both controller and bms light?


Another thing I would sort of like to have, 16 led's on the dasboard, one for each cell. Glow green when each lifepo4 cell is at 3.5v, glow red when approaching bms cutoff. The idea being to know when a 48v pack reaches about 46 or 47v, but also know if one cell is going down early. For sure, you know that once below 48v on the CA you are nearing the end.

Cellog of course will work, but you won't be reading those on the fly at my age. Maybe I just need a Mr Magoo version of a cellog 8

 

[t3sla]

Get a relay which opens when the BMS hits LVC and wire that from the BMS-ed side to the ignition switch on the controller.

You can get 10S lipo bms for cheap, I have a bucket of them if anyone wants to tinker with using them with rc lipo.

It's something I've been meaning to try for ages, an area which I really want to easily document so that noobs can get into things easier.

Thinking about it, you could have a 9V battery with a piezo buzzer that is triggered by the relay.....


EDIT: Anyone know where to get large JST connectors, 10-12s size?

 

[John in CR]

What is being switched off at the negative end of the BMS, between the BMS and battery negative? Is there a voltage differential there or just continuity to ground? Isn't it pack voltage, and in the BMS between the positive and negative end is just a series of switches, one for each cell that opens when the cell gets too low? In that case why can't I simply supply my controller brain from the negative end, putting the controller brain in series with the BMS? Then BMS effectively connects to battery negative inside the controller with my keyswitch in the circuit to turn everything off? Actually my DC/DC converter for lights is supplied off the keyswitch too, so the route to ground is through the controller and through the DC/DC converter too?

John

 

[IBScootn]

Bump

 

[snowranger]

If you put a keyswitch in between the negative bms and negative battery, you can manually shut everything off, but the system is still powered through the bms when you switch on. Plus, there will be a lot of current going through that key switch.

A relay that breaks the controller + signal when the bms opens doesn't need to handle much current.

 

[Arlo1]

I have been mulling this over for a bit on the unused neurons in my noggin...

For first order I think the BMS should:

• signal the controller to stop during a LVC event
• signal the charger to stop during a HVC event

In other words the controller and the charger should be able to accept an "arm" command from the BMS. Without the "arm" command, the "fire" command (twisting the throttle) or (providing charge) should be disabled. Simple, low voltage and current signal path.

A fuse should be integral to the battery pack for safety.

Now we have to think of second order safety effects. For a brushed motor, if the controller FETs fuse (short) we can have a run away at full power. Hence I think a contactor is required in the battery pack to provide propulsive power cut off. It is signaled the same way, with an "arm" command from the controller.

Now what safety shuts offs are required for 3P BLDC motors? We should discuss this and the known controller failure modes. Like twist throttle failures resulting in "full forward" commands. Could be handled with a contactor, or could have another layer of low voltage "arm" signals in the controller from say an accessible key switch to enable the PWM circuitry.

We should think this all through and evolve to a 21st century "system topology" for a high powered eBike.

Also a topological recommendation from now defunct A123. They required that if you had a microprocessor in the mix for a "smart" BMS, that there needed to be another, hardware only, redline BMS that was redundant and could shut things down if the micro hung/failed. Both TI and Maxim make chips that do this function in hardware.

This discussion would center around the "Failure Tolerance" of the BMS. In the A123 example above, they are recommending single failure tolerance. What is prudent on a high powered eBike?

This is what I think it shoud do as well. Basicly Methods HVC LVC boards but it would be nice to have auto ballancing every charge...

 

[John in CR]

If you put a keyswitch in between the negative bms and negative battery, you can manually shut everything off, but the system is still powered through the bms when you switch on. Plus, there will be a lot of current going through that key switch.

A relay that breaks the controller + signal when the bms opens doesn't need to handle much current.

I only want the BMS to supply the current needed by the controller brain, so when the BMS trips the controller turns off. There's no way I'm paying for a BMS that can handle 300-400A. I like simple, and you guys say I need a relay, but I wonder why the complication and I don't have a relay that handles over 80V. So what if the BMS switches at the negative end. That means I just put things in a different order, and the circuit becomes battery positive - BMS - keyswitch - controller brain - battery negative. If somehow that leaves the BMS's brain ON which would drain power, then I'll just use the second circuit on my keyswitch that switches with the same turn of the key. I never understood why my switches had 2 independent circuits closed with the same key position, but I think I may need it now.

John

 

[t3sla]

What controller are we talking?
A Kelly type with separate power for control logic?

I'll try video a few examples his weekend if free