Controller Doesn't Turn Off

[Ur3nigma]

I bought the rear wheel version of this 48v 800w kit:

http://www.ebay.com/itm/48V800W-26-Front-Wheel-Electric-Bicycle-Motor-Kit-E-Bike-Cycling-Hub-Conversion-/290696803295?pt=LH_DefaultDomain_0&hash=item43aee133df#ht_1410wt_1396

When I connect the battery to the controller there is a rather vigorous spark. It does this even if the motor is disconnected from the controller so I know it's the controller drawing the current. I've been disconnecting the battery every time I'm leaving the bike for more than an hour or so because I'm afraid it has a slow drain on the battery. Also, the LED throttle never turns off. It has a button to turn off the power to the wheel, but the LED lights stay on 24/7.

Is the sparking indicative of a bad controller? Should I leave the controller plugged in despite the fact that the LED lights never turn off? I'm thinking the sharp draw of current upon plugging it in can't be good for the battery. Should I just get a new controller?

 

[dnmun]

when the controller is powered up by plugging in, the input capacitors suck up a lot of current rapidly and that causes the spark. totally normal.

some people use a 'precharge' resistor to reduce the current flow until the controller caps charge up. if you put the controller circuit current which goes through the small red wire on a switch, and leave the large red wire for the output FETs connected directly to the battery, then the capacitors stay charged up and the switch turns off the circuit current so it doesn't drain the battery.

the mosfets in the output on the big red wire do not leak current when off so when the switch is off the battery is isolated and you don't have to worry about the controller draining the battery dead.

 

[John in CR]

Dnmun summed it up nicely. I'd only add that connecting the way you have been is actually bad for your controller, because those brief but high current surges wear on the capacitors and other circuitry in the controller. My battery has been connected since October, and the previous battery hadn't been disconnected in a year and a half.

The higher in voltage you go the more "vigorous" that spark gets, and with a high power controller with big capacitors the spark can be like a small firecracker.

I like to use a keyswitch on the wire that actually turns the controller on for added safety and security. Plus the locals here for some reason a impressed that my ebike has a key.

John

 

[miuan]

the mosfets in the output on the big red wire do not leak current when off so when the switch is off the battery is isolated and you don't have to worry about the controller draining the battery dead.

be careful. many controllers employ a small SMD resistor that drains the charge off capacitors permanently. on some of them, it's marked R203, and it goes from VCC directly to GND. i always dump the sucker.

 

[Drunkskunk]

The switch on your controller doesn't actualy turn off the controller, it just disables it from being used. Its still drawing a small abount of power and will kill your battery if left plugged in.

 

[dnmun]

i disagree. show me where there is a current drain on a controller with the circuit current turned off. you have an ammeter in your digital volt meter, use it and prove your distortions. there are people who read these posts and may believe what you say. so prove it or retract it.

 

[Ur3nigma]

I thought it might be capacitors in the controller. There's simply no other explination as to why it could draw that much current so quickly. However, if the capacitors are always leaking current, most capacitors can't hold a charge all that long, wouldn't they also drain the battery?

 

[dnmun]

the capacitors do not leak current. once they are charged up then the current stops flowing. these previous comments were totally uninformed. you can put your ammeter in series with the big red wire after it has charged up the caps and you will see that no current flows at all when the capacitors have charged up. they could do the same. but they won't.

it is just too bad that they have now stuck you with an uneducated opinion.

 

[miuan]

dnmun, come on, have you ever traced the R203 in an infineon controller?
some don't populate this board position, but many do, and it WILL suck current from the battery unless you disconnect one of the BIG wires.

 

[Ur3nigma]

A capacitor is two opposite charges separated by an insulator. It's always been my understanding that there is no perfect insulator and the charge differential will eventually dissipate across the insulator. Furthermore, if the charge did not dissipate over time, the sparking I'm observing when I plug my battery into my controller would not occur. The capacitor would be at full charge at the time it's unplugged and would remain so until it's plugged back in. There would be no capacity to draw any current from the battery and no sparking. The sparking is occurring because the capacitors have been drained of current.

That is imho...

 

[wojtek]

some people use a 'precharge' resistor to reduce the current flow until the controller caps charge up. if you put the controller circuit current which goes through the small red wire on a switch, and leave the large red wire for the output FETs connected directly to the battery, then the capacitors stay charged up and the switch turns off the circuit current so it doesn't drain the battery.

i was thinking about Single Pole 125 Amp Switch Disconnector
http://www.expert-electrical.co.uk/Circuit-Breaker/Switch-Disconnectors-/-Isolators/Single-Pole-125-Amp-Switch-Disconnector/prod_871.html

but Neil also recommended me pre charge resistor..

I used switch disconnector with success for last 1.5 years. It eliminates the spark nicely and is easy to switch on off and adds extra security.

But i think it does draw some amps (???).. When i switch on Lyen's 24fet 4115 controller i can see 10w on the CA being drawn..

Could you kindly send me a link to such pre charged resistor that would be good for 100v100a application please? tried to google but no luck

many thanks!

 

[d8veh]

I always use watt-meters and can firm that with both the KU93 and KU123 controllers from BMSBattery, the watt-meters show 4w when the controller is plugged into them, and zero watts when the contoller is un-plugged. I have my switch on the thin red wire configured so that the throttle LEDs go off when the controller is in standby mode.

 

[dnmun]

if your controller has the small red wire, put the switch in that wire, and permanently connect the battery directly to the large red wire that goes to the mosfets, that is the line that sucks up the big initial charge. the small red wire has about 60mA for the circuit current, but that will drain the battery if left permanently on. use a small toggle switch since it is just small current. no breakers needed if it has the BMS since the current surge will shut off the BMS if it shorts. all jmho.

 

[wojtek]

Installed a switch on the thin red +5V wire and intend to switch it on and off during shorter periods of not using the bike.

i have no BMS in my 100v lipo pack.. So i could either use 125A breaker or pre- charged resistor... [looking for one]

the spark on 100v battery disguist me more than old grandmas on the nudist beach.

 

[Ur3nigma]

I hardwired the battery to the controller the other day and just unplugged the LED indicator lights. Since then I've had the BMS trip twice. The battery has to be unplugged for it to be reset so I'm going to have to put a breaker or something on the wires leaving the battery. Where would I get a breaker with such a weird (48v 20ah battery) voltage/amp?

I've also since checked, and the controller draws 0.01ah when everything is unplugged. I figured I'd just leave the battery plugged into the charger when I'm not using it unless it's going to be for several days or weeks. Any issues with that? I have a LiFePO...

 

[dnmun]

there is an issue with the controller if it is drawing current with the circuit current turned off. one of these guys finally posted a picture of a 20k resistor, surface mount, across one of the caps, which he unsoldered to stop the leakage. that is the only controller i have seen so far with a draindown resistor on the input caps. so if you have that controller you will need to unsolder and remove that resistor like he did.

not sure why your BMS is shutting down.

 

[Ur3nigma]

Do you have a link to that pic?

I'm not sure why the BMS is shutting down either. I'm assuming it's acting like a circuit breaker and shutting down due to too prevent over drawing the battery. That would be weird though since it's a 1000W battery and the motor is 800W. Also, the supplier sent me a new wheel/controller due to several previous issues. The BMS issue didn't happen until I installed the new motor and controller. Another weird thing was that the BMS reset itself the first time after I plugged it in and started charging the battery without unplugging it from the controller. I'd only ridden it about 6 miles so it should of had at least 3/4 charge. I'm not sure what that was about...

 

[cor]

i disagree. show me where there is a current drain on a controller with the circuit current turned off. you have an ammeter in your digital volt meter, use it and prove your distortions. there are people who read these posts and may believe what you say. so prove it or retract it.

Wow, rather harsh, don't you think?
At any rate, my (stock EVG 24V) controller has the exact problem that he described:
- when the battery pack is inserted, it connects to the contacts with a fat spark.
(It does not help that in both the controller and the battery pack a capacitor is embedded that make the internal resistance lower, giving a bigger spark)
- the controller is always on as long as there is charge in the battery pack, the key switch is simply one of the digital inputs to the controller.
- Since the controller is running always, its power regulator is also always-on. It takes a moderate amount of power, only about 20mA of continuous drain, but that means that even if I have a fully charged and very large 20Ah battery pack that it will be completely drained in 1000 hours, which is just over a month. Then the battery will be toast (unusable). The only option is to remove the battery from the bike to stop the drain through the controller's supply.
I have little experience with the drain on other controllers - some may have only their caps connected when the key is off, others may also have continuous drain from the controller running. It depends on the design of the controller and how it is switched on/off.
Cor.

 

[NeilP]

I have to agree that almost all controllers left plugged in will eventually drain the battery even if it does have an on/off switch.
They virtually all have some drain, be it by design, ( drain down resistor on the cap) or not.

But it is worth checking your particular controller if you want to be sure...but only possible if you have an accurate and reliable low current ammeter, that will read milliAmps.
Do as suggested and connect it in series with the main battery wires...After you have precharged the controller first, otherwise the initial charge current when the controller is plugged in will likely blow the ammeter.

I have two controllers here, both the same board(EB218) one built by Lyen and the the Xlyte. The Xlyte one holds it cap charge for days/weeks, while the Lyen one with the discharge resistor drains the caps within 5 seconds.

So all controllers are different.

Answer..

Disconnect each time from battery, either with switch or just unplug.
When you re plug the battery in, Connect with a pre charge resistor first. 1Watt 100 Ohm will do, although a wide range is suitable. If you have a load of old resistors, try what you have.

 

[John in CR]

Wojtek,

A precharge resistor that will work fine for you is something in the 200 ohm range. The connection is made first through the resistor allowing the caps to be charged much more slowly, eliminating the spark. There are numerous threads on the subject, and how often you disconnect your battery will determine whether you need a precharge circuit or just a stand alone resistor like I use.