Battery Pack ON/OFF switch with Relay Failure Mode...

Joined
Apr 24, 2008
Messages
812
Location
Victoria, British Columbia
I have a Conhis pack...

http://www.conhismotor.com/ProductShow.asp?id=456

The externally facing keyswitch fails to turn off the main power leads from the pack to the ebike. They are locked in the "on" position, even in "off" on the keyswitch, producing power whether the key is turned or not.

The keyswitch is connected to a relay which cuts the main (+) output...you can see the two thicker red wires going to the relay.

Here is the picture...

RelayCoil.jpg

After email with Conhis, they told me to open the relay cover and "see if the switch moves" when the key is turned. It does! So it wasn't the keyswitch I don't think.

Suddenly now it works - the system is back to normal - the keyswitch turns off the power in the Off position.

I'd like to understand why the relay wasn't working, and now is, after the cover is removed. It was not burned or welded shut, the inside of the relay.

I could just wire around the relay, leaving the pack "always on" powerwise which is ok, although it is nice to turn off the controller this way.

Do I need a replacement relay? Will an alternative automotive relay aquired here do the job if it can handle 40 amps?

The switch wouldn't be affecting the BMS operation I imagine, meaning there would be quiescent drain of the BMS, even when the pack is normally "off".

So, other than removing quiescent drain from the controller, the on/off switch doesn't do much, other than inactivating the throttle - should be ok is bike is left on charger when away.

Defaulting to "on" works the same. The bike works normally.

Is this just a "jammed" relay or weak spring?

Thanks, Chris.
 
It could be that the relay was physically stuck, which could indicate future failure if it was not spurred by a drop. Did the pack experience and physical trauma recently?
 
I can't tell if there was any physical shocks or dropping of the bike, as it was my son using the bike. Conhis indicated that the relay was probably jammed as well. I suppose I could get another relay, but its only been in use on the bike for a couple of months. If this is going to re-occur every now and then, I had better just wire out the relay and make a direct connection from the pack to the main output connector. The switch mainly seems useful only to remove power from the controller, not to remove quiescent drain within the pack itself, from what I have observed with slow discharge over a week without being charged. I don't like the charger connector on the pack either - it can be shorted if you don't push in the plug carefully. It would be better to just bring some Andersons out through that plug hole I think. All this concerns me as I am maintaining this bike for "a bunch of hippies"....

Chris
 
chvidgov.bc.ca said:
I can't tell if there was any physical shocks or dropping of the bike, as it was my son using the bike. Conhis indicated that the relay was probably jammed as well. I suppose I could get another relay, but its only been in use on the bike for a couple of months. If this is going to re-occur every now and then, I had better just wire out the relay and make a direct connection from the pack to the main output connector. The switch mainly seems useful only to remove power from the controller, not to remove quiescent drain within the pack itself, from what I have observed with slow discharge over a week without being charged. I don't like the charger connector on the pack either - it can be shorted if you don't push in the plug carefully. It would be better to just bring some Andersons out through that plug hole I think. All this concerns me as I am maintaining this bike for "a bunch of hippies"....

Chris

How is the pack wired? Is the switch getting full pack voltage? If that's the case, you can remove the relay. But I'm under the impression that the relay is there to spare the switch from seeing all that voltage and eventually failing.
 
Sounds to me like your relay was welded shut for a while, but has now come unstuck. Can you take the cover off the relay and get a look at the contacts? If they're pitted up, this was probably the issue. If the relay can mechanically jam, it should be replaced. It should be replaced if it was welded too. Is there a precharge resistor?
 
wbk9...yes...I took the cover off, and it was operating ok at that point, physically moving with the on/off switch being turned. I had better take another look at it in closer detail with a magnifying glass maybe. I imagine it will be susceptible to the same problem in future, if I don't replace it. I don't have a precharge resistor, the pack is perma-connected to the controller as installed. I imagine there is considerable inrush in the relay, as the controller caps suck the amps. This is possibly another example of poor Chinese design choices, working for a while but not in real world. It "works for a while" then fails, on average.
 
chvidgov.bc.ca said:
wbk9...yes...I took the cover off, and it was operating ok at that point, physically moving with the on/off switch being turned. I had better take another look at it in closer detail with a magnifying glass maybe. I imagine it will be susceptible to the same problem in future, if I don't replace it. I don't have a precharge resistor, the pack is perma-connected to the controller as installed. I imagine there is considerable inrush in the relay, as the controller caps suck the amps. This is possibly another example of poor Chinese design choices, working for a while but not in real world. It "works for a while" then fails, on average.

Yes, I would definitely take a closer look at the contacts. The weld probably broke as you were removing the relay cover. That would explain the normal movement you saw and the restored operation after the fact of checking the relay. No precharge strongly suggests this is your problem. You may be able to smooth the contacts with emery paper and save the relay for the time being, but I would get a precharge resistor in there soon.
 
Would you just put a resister inline on the "battery side" of the main power input to the relay? There's gonna be a helluva lot of people with this problem, as those types of pack cases are everywhere nowadays, stuffed with Samsung cells.
 
The relay contacts are rated at 40A 14vdc with a 48V coil. The contacts are undersized for this 48V application and will switch only about 10A at 48V. It's working more or less because it closes on no current (ZERO throttle) and is never opened under full load and so manages the 40A 'carry current' okay.

It's probably getting zapped by the cap surge. Almost certainly, the contacts were stuck, and almost certainly, this will re-occur if you have matching titty/pitting on the two contacts (characteristic of DC).

You might consider wiring a 1K 1W resistor across the relay contacts. This will leave the battery in perpetual pre-charge mode when it's installed on the bike which will virtually eliminate the troublesome current spike when the contacts close. Since it's always on pre-charge, a larger 1K value will reduce the worst case power requirements for the resistor. The large resistance will charge the cap slower, but unless you slap the pack on and race to turn the key, the delay from the large value should still allow full cap charge in a few seconds. This goes after the probable cause directly - I would try this first before eliminating the relay.

A small 1/2W resistor would be fine, but a 2W value could be connected across the battery all day long. The larger wattage components have more robust leads that I like better for vehicle applications. It's unclear how the wiring is attached to the relay (socket, fastons?) but you can just place the resistor next to the relay body and run the leads to the relay spades up snug to the relay bottom. Tack them with a bit of solder so the resistor is now part of the relay. This will be tidier/easier than soldering large gauge pack wiring.
 
Thanks Amber...I believe the reason they used a relay in the architecture was to be able to turn the pack on and off with a keyswitch mostly a selling point - "but I need this to work for a bunch of hippies", and I thought the pack architecture would work. The usual problem of burned contacts with the amp surge has been relocated from the usual output Andersons, which get burned out, to the contacts of the relay, inside the pack which burn out instead. The precharge needs to be "in situ" so the keyswitch activates the relay <with precharge always ready to go, before the relay>. I understand the usual precharge is to manually connect the resistor, charge up the caps, and then connect the main line. The end users of this can't do this, because it would be inside the case of the battery that you would have to manually intervene all the time to connect the precharge first every time. Is there a way to have a "passive precharge" that takes care of the problem without manually doing anything? I may be forced to just scrap the keyswitch and the relay entirely, and leave the massive Anderson (the biggest type > 40 amp type) that the pack has always connected to the controller. With the instruction to the hippies to "ALWAYS LEAVE THE BIKE ON THE CHARGER WHEN NOT OPERATING THE BIKE", so the additional quiescent drain of the on controller doesn't bring the pack to LVC in less than a few weeks.

So there are really three problems with these packs...the relay will burn out in normal use, and the off switch doesn't disconnect the BMS from the cells - so it isn't really off for the purposes of long term storage of the packs - the BMS quiescent drain continues in the off position. And the charger plug can short where the cord from the charger plugs in, if it is not perfectly centered.
 
Thanks Tek...got your suggestion just as I answered Amber...lots to ponder there. I may have to try your suggestions, if I can't find a better relay of the same form factor. I didn't understand your point about the relay "closing under load"...the on switch closes the relay and keeps it closed. Its that it won't unclose the relay when I turn the key to off, because the contacts are welded a bit. I think the problem with your idea, is if they try to run the bike, with the key switch off position, it will run! Through the resistor across the relay only. Wouldn't that zap the resistor, if they ran the bike in the key-off? I think the best solution may be just to make the battery like a Ping, with the controller always connected to a live battery.
The LVC would still work to protect the <more likely to run down in long-term storage> pack. Making sure I've got ebrake cutoffs for throttle safety.

Its a big dilemma for supporting packs for novice users.
 
chvidgov.bc.ca said:
I think the problem with your idea, is if they try to run the bike, with the key switch off position, it will run! Through the resistor across the relay only. Wouldn't that zap the resistor, if they ran the bike in the key-off?
With a 48V battery, a 1K resistor will supply 48V/1K = 48ma. 48V x 48ma = 2.3W. The bike is not going to move on 48ma and the 1W value seems workable - it will survive failed throttle tries without difficulty when switched off - likely the controller LVC will keep it off so no real power will be drawn. As I mentioned, if you want it to be bulletproof for dummies who hold the bike on WOT for a few minutes with the bike turned off and a controller with a really low (or no) LVC, then use a 2W or larger resistor. Even a 5W ceramic is workably small (see eBay).

This is the same strategy recommended by Kelly for their switch/contactor setup - nothing new here...
 
With a precharge connected all the time the bike might move, but barely. Biggest risk there is they try to ride it like that and burn up the resistor, but even this shouldn't happen because the BMS will see the voltage sink like a stone when throttle is applied, sending the unit into LVC. So this is one viable option.

Were it me, I would add a precharge resistor and use the keyswitch to switch that in and out, and then add a toggle switch to turn on the relay manually after key on. Might be too complex for the hippies though. Another option would be to add a RC time delay to the relay coil so that the key turns on the precharge instantly, and the RC circuit delays the relay powering on until precharge has taken place, but you would probably need two sets of contacts on the keyswitch to make that hippie-friendly.
 
Thanks guys...tips were greatly appreciated...btw I love those hippies...as they are relatives and a lot of fun to chill with...just that their minds are on......other <virtual> things......probably about like me in 15 years. So all I have to do is solder a resistor across the relay stubs for the two stubs that are the main contacts inside the relay, leaving the bike in "very lame" powered up mode in key off, changing to "full power" on with the key-on position.
 
chvidgov.bc.ca said:
Thanks guys...tips were greatly appreciated...btw I love those hippies...as they are relatives and a lot of fun to chill with...just that their minds are on......other <virtual> things......probably about like me in 15 years. So all I have to do is solder a resistor across the relay stubs for the two stubs that are the main contacts inside the relay, leaving the bike in "very lame" powered up mode in key off, changing to "full power" on with the key-on position.

That should work. Only caveat is watch that the battery doesn't get overdischarged this way. Presumably the controller has very low standby current and when LVC kicks in, no more current can flow from the battery. If so, you can avoid the worst case scenario of a ruined battery. Good luck!
 
Right...gotcha...the LVC will kick in. I will tell them to "keep the bike charged up at all times" over and over again. Sort of like a Lead Acid. I could also put a simple off switch on the controller on/off (red small wire that goes to + input), on at least on the two controllers on bikes with the same packs described.
 
chvidgov.bc.ca said:
I will tell them to "keep the bike charged up at all times" over and over again.
Whoa - "them?". This is not your bike? A less informed user puts a little different spin on this...

Also - the business of "keep it charged all the time" made me go back and look at the link you posted. I confess to have only looked at the first photo using a tablet so the pic was small - I thought this was a removable battery, which made this strategy okay for a day's use or so. The battery could always be pulled overnight for charging or storage. But I see the battery is not a slide-in type and will be connected pretty much all the time. My Bad. This is not a good situation for this high drain always-on pre-charge setup.

So - apologies - but I recant the recommendation. The combination of 'other folks' and a permanently installed battery make this approach look ill-advised. You need something more bullet-proof - like a toaster.

chvidgov.bc.ca said:
I could also put a simple off switch on the controller on/off (red small wire that goes to + input), on at least on the two controllers on bikes with the same packs described.
Switching the controller 'ignition wire' is a good solution although I would like to see a keyswitch - which adds some mounting inconvenience. But - if you can mount up another wire to the battery (either as a connector on a short whip or perhaps using another PowerPole bulkhead fitting - see Powerwerx) you could take the switched +48V that drives the relay coil and run it (also) to the controller ignition wire. This would preserve the single switch design - it would simultaneously switch on the relay and power up the controller. In this situation the pre-charge current would be negligible because the controller would normally be disabled and there would be no quiescent idle current. Pretty close to only 'toaster' level care required....
 
The battery is on a slide plate, but that is only for very occasional removal if at all in this user environment, as it is padded with yellow camping foam between the slide plate and the battery, to prevent rattling of the battery on its plate on the back of the bike it is relatively difficult to pull it off the sandwiched foam and/or line up the removal pin from the keyswitch with the slide plate hole. I'm thinking I should just go back to old school "disconnect or connect the sacrificial Andersons (the really big ones) to turn the battery off and on. The bike operates in a shared environment of novice twenty-somethings, some of the time, otherwise one user most of the time. It is a big pack capable of 40 amps no problem on a spike. Wouldn't your idea work for a week or more of no charge before LVC? It will certainly hit LVC a lot if I allow it to happen in this human environment.
 
When the ignition wire is turned on the controller will start to suck amps. The pack relay must be in the on state then to prevent overcurrent. So all you can do is never turn the relay off. So why bother with it at all - its just a gimmick. Just get rid of it, and use the ignition wire for on/off functionality, like Ping batteries that are left on a bike with just very big Andersons directly from the BMS, with an ignition switch.

Instructions would be "please disconnect the existing main Anderson connectors (bigger than the for long term storage of the pack or pack removal. Use the switch (controller ignition switch) to turn the bike on or off in daily use", whatever kind of switch it is on the ignition wire. That switch couldn't be the repurposed pack keyswitch without moving it, as that would prevent pack removal. And of course the mantra "always keep the bike on the charger when not in use". So it seems the ignition wire switch is the way to go to minimize quiescent controller drain. And the relay would be disconnected entirely, obviating that issue. Just the giant Andersons, for pack removal.

The pack switch would only be used for physical removal of the battery from the slide plate.

Probably a "hidden ignition wire switch"...small one, pushbutton or similar...I know at least one of the controllers has an ignition wire.

Thanks for talking this through with me. It's interesting how the thinking has changed as I went along.

In reflection, its stuff like this that prevents the ebike biz from flourishing.
 
chvidgov.bc.ca said:
The pack relay must be in the on state then to prevent overcurrent.
As explained above, there is no 'over-current' situation. Please ignore posted remarks about the resistor 'burning up' if the controller draws power - an electrical fantasy. As calculated, you can hook a 2W 1K resistor directly across your battery and all will be well - putting a controller in series with it cannot magically pump more current through it.

The business of always leaving the pack on the charger is a separate matter of discussion relating to battery life - but a switch design that forces you to do so is pretty poor.

chvidgov.bc.ca said:
...So why bother with it at all - its just a gimmick. Just get rid of it, and use the ignition wire for on/off functionality, like Ping batteries that are left on a bike with just very big Andersons directly from the BMS, with an ignition switch.

Ya - you PMed about using a breaker as well - both strategies are fine - you will just be using the battery in a conventional DIY ebike way and supplying your own switching for an essentially non-removable battery.

I have to admit that this kit/mounting plate/relay thing seems pretty hokey. I don't understand the thinking behind that system unless the key is three position and has UNLOCKED-OFF/LOCKED-OFF/LOCKED-ON states. If it has that, you could afford a third wire with Anderson for the ignition wire so the battery is still removable as mentioned in the previous post.

That said, I agree with your idea of removing/bypassing the relay entirely and using the Blue Seas breaker to also supplant the fuse. I use those units and they are fairly small - it might fit in the battery case, but if not, you could mount it up externally. An option here is to rig the ignition wire as a kill switch next to the throttle which will improve the safety and get you a controller on/off option as well. If you need to lock the bike, use a keyswitch in place of or in series with the kill switch. I like keyswitches for security systems (more weather proof) that allow the key to be removed in either ON or OFF positions. When I'm just pleasure riding, I leave the keyswitch ON and power the bike ON/OFF with the breaker - no key required, and when I park in a public place I use the key. Best of both worlds.
 
I think I will go with the Blue Sea..the breaker situation means that if the bike is left off the charger, and the breaker is turned off, the battery won't drain, and should be minimal inrush issue. That's the best you can hope for in this situation.

Instructions to hippies: "Always turn off the circuit breaker when the bike is not in use, especially when the bike is not on the charger. Always leave the bike on the charger as soon as possible after ride is over - you have been warned"


Risk to Pack - Blue Sea Circuit Breaker
On Charger Circuit Breaker On Small Risk if Charger fails (LVC will be hit)
On Charger Circuit Breaker Off No Risk
Off Charger Circuit Breaker On Risky (LVC will be hit)
Off Charger Circuit Breaker Off No Risk
No Inrush issue

--------------------

Risk to Pack - Relay with Precharge resistor (controller always on but lame with switch off)
On Charger KeySwitch On Small Risk if Charger fails (LVC will be hit)
On Charger Circuit KeySwitch Off Small Risk if Charger Fails
Off Charger KeySwitch On Risky (LVC will be hit)
Off Charger KeySwitch Off Risky (LVC will be hit) -------------don't like this line....circuit breaker is better.
Inrush solved, uses dinky Relay with resistor

-------------------

Using ignition wire with no relay is functionally equivalent to the circuit breaker scenario. The main thing is to ensure the controller doesn't get any juice at all in the switch-off state (either breaker or ignition wire), off the charger. Which isn't so with a precharge resistor. Yet prevent inrush issue.

Time to wrap up. Thanks again all.

Chris.
 
Today I got a Blue Sea 50 amp breaker installed/duct-taped to the outside of the pack, controlling the main current. The relay was disconnected and removed from the case - the keyswitch is only used to remove the pack. No issue with inrush current tripping the breaker off. Legible instructions on pack:

"Turn Breaker Off when bike is parked"
"Leave Bike on Charger when not riding"

That should alleviate my responsibility to the hippies.

This might one solution to the issue of human interface to these bikes. Lots of work with Andersons today.
 
I have a simple question, do I put the ignition in the off position if I am charging the battery or in the on position ?
 
if you need to disconnect and reconnect the battery then you can use the output mosfets on the BMS as a switch to turn the battery off when you connect it to the controller so the current surge does not damage the contacts of a switch or breaker.
 
Back
Top