Ignition switch (anti spark)

[mistercrash]

Here's the problem, I ride an electric scooter and it has an ignition switch to turn the power on. I'm pretty sure that the ignition switch is meant to work on a 12V system like on an ICE motorcycle. But the full voltage of the battery goes through it so when I turn that key to the ON position, there is a big spark that is generated inside the switch and in a short time, those sparks eat away at the copper inside the ignition switch and one day the power just doesn't go on because there's no more copper to make contact. There's only two wires going in the switch, the voltage of these two wires is the full voltage of the battery.

Is it possible to rig up an anti spark system to this ignition switch?

Thanks
Ray

 

[dnmun]

what type of battery? if it is a lithium battery with a BMS then you can turn the battery off with a small DIP or rocker switch in the BMS circuit current.

 

[mistercrash]

The battery is 16S made of cells from Makita tool packs. dmun, I'm going to tell you a secret, don't tell anyone ok, I don't have a BMS. After I wrote the post above, I did some more research and found something about rewiring the switch so that it only turns on the 12V system, and ad a start button to the 12V system that triggers a latching relay circuit that turns on the higher voltage of the battery. It sounds nice but I'm just not exactly sure of how to wire everything.

Here's a wiring diagram of my scooter.

 

[teklektik]

Nice detailed wiring diagram.

I would say with almost certainty that your problem does not lie with the controller but with the 12V DC/DC converter. It's a running quite a lot of 12V stuff and so has a pretty high capacity. It has big moose caps on the front end - just like the controller. Your switch is not switching main controller power, it's only switching a few tens of mA of 48V on the 'ignition' wire. Harmless. However, the switch is handling the converter caps directly with an attendant large inrush current and big spark (many Amps).

• Try unplugging the converter and see if the big spark goes away.

If so, you can use a 48V relay to run the converter. I have a similar setup on my bike and use a 48V relay from a golf cart lighting system for this same purpose. See this post for the relay and socket info.

 

[mistercrash]

There's not an easy way for me to see if the spark is there or not. The reason I found out that a spark was happening was to destroy the old ignition switch to see what went on inside and find evidence that it happened. I saw your thread (awesome build) and then tried to find a 48V golf cart relay. Many popped up and I'm not sure which one I should use. I'm not sure either on how to install it in the electrical system.

EDIT: I think I found the relay you used here http://www.ebay.com/itm/48-VOLT-ELE...296?pt=LH_DefaultDomain_0&hash=item256d886148 But I'm still not sure how to wire it to the system above.

 

[teklektik]

Yep - good find on the relay. That should solve your issue.
I'll hack up a little wiring diagram tomorrow.

 

[teklektik]

The relay has this pin out:


There are a couple of ways to do this but this may be the most straightforward - sort of depends on where the switch and converter are located to make the wiring mods. Here's a simplified subset of your original wiring diagram and a mod to hook up the relay:


In this mod the switch drives the relay coil and the relay contacts take over the original switch connection to run the controller plus the converter. Not much to it.

An alternative is to have the switch run the controller plus the relay with the relay driving the converter. In this mod the bike will still run if the relay dies for some reason:


So - either one will work fine - your choice. You can either mount the relay with the provided tab with a screw or just cable tie it to a wiring bundle. Things may be easier if you can locate a mating socket as I mentioned in my build thread, but you can use 1/4" Faston connectors to make the relay connection if you wish. If using the socket, I recommend cutting off the hook-shaped retainer thingie with an Xacto/utility knife, otherwise it's extremely difficult to deeply depress the release while pulling the socket and relay apart (3 hand operation).

Hope this helps...

• BTW - It turns out the relay you located is from the same Golf Cart guys that I got mine. I originally made a special request purchase back in 2011 for the relay that was part of their 48V golf cart lighting kit then posted them as a source in my thread. Looks like they had enough requests that they started selling them separately with 'bikes EV' in the eBay title...

 

[mistercrash]

Your post is helping tremendously. You explained everything very clearly, I know exactly what I have to do to fix this problem. Thank you so much. I researched the golf cart company from your build thread and found they had an ebay store. From that it was easy for me to find that relay. I'll probably end up getting a 12V relay that has the socket I need. I will also share your solution and drawings to the company I got the scooter from.

Thanks again for all your trouble.
Ray

 

[teklektik]

Glad it looks doable at your end.

I think the only thing that I might add is a snubber diode across the relay. This wasn't really much of a consideration for the first version above where the relay is just buffering the switch, but it might be a good plan in the second version where the controller ignition wire remains hooked to the relay to get a spike when the relay switches off. I would add them for safety in any case - only takes a few minutes.

You can use a 1N4004, 1N4005, 1N4006 as available - the higher numbers have higher voltage ratings.

View attachment 3
The idea is to hook it up as close to the relay socket as possible in reverse polarity so the induced spike will be squelched before it propagates downstream to electronics. I used a small pointy tool to push back the tang holding the relay socket contact in place and withdrew the coil contacts from the plastic shroud. The snubber leads were just bent in a U and soldered directly to the crimped contact neck, then the contacts shoved back into the shroud. Depending on your skills you can do that or some other variation to get the diode hooked across the leads as close as reasonably convenient to the relay. Use heat shrink or similar to avoid exposed diode leads. Also - be careful about the relay wiring polarity once you have these in place since they'll be zapped if you hook the relay up so the diodes conduct when power is applied. The diodes are commonly available at Radio Shack, eBay, etc.

I've redrawn the diagrams here:




Apologies for any confusion - this is a simple addition to make your installation more bulletproof...

 

[chucho]

Great diagrams, talked for not loosing them :wink:

 

[mistercrash]

teklektik I appreciate your detailed instructions. I'm waiting for tomorrow to call a couple companies to see if I can get the relay locally, the ebay store does not ship to Canada. If I can't find it locally then I'll have to make a call to Florida and ask them if some arrangements can be made for them to ship to Canada.

One thing I thought of and I want to ask if it is possible. The way I understand by what you explained is that the DC/DC is sending a rush of current and voltage back to the ignition switch creating the spark, am I right? Is it just to simple to put a diode on the wire marked "red +48V (switched)"? I don't know much about electronics but a diode sends electricity only one way right, so I thought maybe a diode would keep any power from going back to the ignition switch. It's so simple I feel silly asking but I have to ask.

Thanks
Ray

 

[teklektik]

Not exactly. The issue is that the converter has large capacitors on the front end. When power is switched on via the switch in their initial discharged state, they briefly appear as a short circuit or very high current load. As they charge up (happens very quickly) they effectively disappear as a load until the converter calls for power (you switch on the lights). So - this means that even if the switch is rated to handle the normal running load from the converter, the initial power needed to charge the caps exceeds that level many fold.

This is the exact same idea as your controller. You can see that it is always connected and the caps are always across the battery - but in a quiescent state they draw no power. The 'ignition wire' that you see being switched just turns on the microprocessor and other small bits of electronics that allow the controller to begin operating. However, if you were to disconnect the battery and come back tomorrow and hook it up again after the caps discharged, you would get a big honker spark as the caps charged up initially - just like your converter.

Here are three gimmicks to fix this capacitor-on-the-front-end problem:

1. separate the logic in the device using the ignition wire strategy and just leave the capacitor part of the unit powered all the time like your controller, or
2. use a pre-charge button that you press before switching on the device - the pre-charge button charges the caps slowly through a resistor to limit the in-rush current (no spark) so they are all charged up before you throw the real power switch (already charged = no in-rush current), or
3. get a switch that can handle the inrush current.

For convenience, your scooter controller used the ignition wire approach in lieu of the annoying pre-charge button. We're just using the higher rated relay contacts as a means to do (3) for your converter although we might have gone with (2) as a distant second choice...

Your diode solution would not work because it does not address the high in-rush issue. In fact, it would need to be sized to handle the giant in-rush current since it would appear to be connected to a short. So - you made a good try considering the perhaps overly 'thrifty' explanation I gave in the earlier post, but it doesn't really address the core issue.

 

[mistercrash]

Ok then, I will do as you said before my master Thank you.