Complete Lack of Weatherproofing on Crystalyte Controllers

[smithinparis]

I will first start off by stating that I am completely to blame for my procrastination of providing more protection of my controller. I've been planning on sealing it off for sometime, but given that I always watch the weather, and don't bike in the rain, (60 km round trip in the rain as a commute to work is something I've tried to avoid).
Anyway, on to the problem ...
Yesterday it was calling for sun, but I ended up getting some patches of wet flurries on my way to work. We aren't talking about a storm by any stretch of the imagination ... my jacket and pants got wet, but I was in no way soaked. I got on my bike at the end of the day to head home, and managed to get about 100 meters, when the low battery light came on. Thinking that was odd, I switch battery packs, and managed to get a little further when the low battery light started to flicker and I felt a rumbling in the wheel and then it went dead (power light is off too). Unfortunately, I was in the left hand turning lane (front of the pack none the less) of a busy intersection when it went dead, and trying to peddle that bike across the intersection was like hauling a ton of bricks. So I got across the intersection and unplugged the motor from the controller (which loading the motor), and rode home.
So at home, I took the controller apart and the insides were completely soaked (dripping wet), and to make things worse, it was really dirty because of the trail that I ride on (which kicks up lot of dust). I haven't yet seen any obvious areas where it has discharged and let the magical smoke out that makes things work, but I did see some questionable connectors. I cleaned the assembly with rubbing alcohol, and dried it up, but I haven't tested it out yet to see if it's ok - I'll do that tonight.
So now for my rant ... this controller is for a *bike* and bikes are ridden outdoors where they are at the mercy of the environment. Nobody would accept it if the bearings on a bike would seize because of a little rain, or your speedometer would fail because it got damp, or if brakes stopped working completely. A little give and take is expected though, eg clean your chain occasionally, don't store your bike wet for long periods of time, don't expect anything electronic to survive a huricane, and expect the brakes to not work at 100% in the rain - but at least the manufacturers make an attempt at making sure their products are protected to some degree.
Why are controllers *any* different? Why is it ok to leave gaping holes and entirely unsealed gaps so that dust and debrie can get in completely unrestricted, along with a little bit of water, in areas that contain sensitive electronics which are mixed with extremely high currents? The manufacturers of the Crystalyte controller made absolutely no attempt at offering any real protection other than a porous case.
Anyway, let this note be a warning to anyone out there like me ... don't expect any protection from the elements. Make sure you shelter the electronics somehow. Seal it with silicon, put it in battery bag or case - just keep it safe.
Does anyone have any advice to offer? What do you do to protect your electronics? Does anybody keep in in a bike bag? If so, what about heat sinking, does it become a problem?

 

[fechter]

Yes, that's a problem with many controllers. You don't want to seal it up inside an enclosure or heat will be a problem. The best approaches I've seen are:

Arrange the controller so all the wires come out of the bottom side. Water tends to follow wires. Water goes downhill. Seal the top end, but leave an open drain/vent hole on the bottom so water that does get in can't accumulate. Some kind of shield or "umbrella" is good to keep rain off when parked, but it needs to allow air flow when riding. Fenders or spray guards are good to keep tire spray off when riding.

 

[smithinparis]

I did hook the controller back up last night, and it was a no go. Dead as a doornail.
I didn't have a chance to start probing around yet, I'll give it a go tonight. I'm keeping my fingers crossed that it's just the regulator that ended up getting toasted, since the power indicator won't even light up, the fault has to be pretty far up the chain.
Are there any repositories anywhere that have a bunch of schematics for Crystalyte controllers? Mine is a 36 volt, 20 amp immediate start that's got connections for brake switches, cruise, and the other usual suspects. I'll grab the board number when I get a chance and post it.
So like I said I'll be probing for voltages tonight. Off the top of anyones head, given that nothing is getting power, and nothing is obviously blown, and also nothing is getting the least bit warm plugged in currently - what would you first look for?

 

[smithinparis]

Fechter,
BTW, I'm reading over your controller repair and mod thread. I'd read it over quite some time ago and forgot all the awesome detail that was in there. I was overlooking it before because it's not the same controller as mine, but you did go into quite a lot of general detail as well.
Anyway, I'll be keeping that thread handy tonight.
So a big thank you to you for putting that up for people like me!

 

[smithinparis]

So in my digging through the thread "Crystalyte Controllers - Repair and Modification information", I managed to find a picture of my controller, and it is infact the Crystalyte V1 20A start immediated controller

 

[johnb]

Totally agree with the rant. I'm in the UK and it rains here a lot. My controller got wet and went crazy before I 'got around to' sealing it. It's not the same controller as yours but probably the same case.

Anyway, I fixed the controller and took it out of its case and put it into a cast alloy case with a lid that seals. It sits at an angle and the wires enter at the lower end. I also have a small drain hole at this end so I agree with Fechter on his ideas. Not had a problem with water since, even through several downpours.

 

[smithinparis]

johnb thanks for the feedback!
I played around with it for a bit, and plugged it back in and managed to get the power light to come on, but with it was a very big flash, and some smoke from the underside of the board.
The power light still goes strong, but there is no power to the motor.
I turned the board over to take a look at where the smoke came from and it was from R46 which completely blew out. Unfortunately the board isn't symmetric so it's not immediately obvious what the value was, but I think it was a 101 smd resistor.
And upon further inspection 2 of the IRFB4710 have also completely blown out!
Doh!
So 3 obviously dead parts that lost their magical smoke! But the big question is ... is the fault still up the chain from R46?
A schematic would be helpful ... I'll see about tracing out a schematic for my V1 board.

 

[smithinparis]

It turns out R46 sits between the output of the 3 input AND gate (HEF4073B) and the IC driver (IR2101S) which controls the MOSFET gates. And of course they are all part of the same phase. So likely what happened is the MOSFETs somehow ended up both being on at the same time due to a short circuit somewhere, which toasted the two, which then fired the full 36V line current back up the control line blowing R46. That likely means the IR2101S is fried, as well as the AND gate most likely.
I'm having trouble following the lines connected to the input of the 3 input AND gate, but the are pretty close to the brake switch. Does anyone know what the 3 inputs would be for in the 3 input AND gate? One would be for phase switching, and another for the PWM. Would the 3rd possibly be the brake switch?

 

[swbluto]

Where do these guys get so knowledgeable?! :wink:

(Yeah, it's probably the crystalyte repair thread that I've had the fortune of not being forced by circumstance to visit.)

 

[johnb]

So likely what happened is the MOSFETs somehow ended up both being on at the same time due to a short circuit somewhere, which toasted the two

I guess the reason why water and MOSFETs don't mix is that because their gate current is so low, it doesn't take much of a conduction path (supplied by water) to turn them on at the wrong time.

 

[knoxie]

:?

I see a lot of folks on here posting piccies of their bikes with no protection for the controller and it has always worried me, I dont think we can blame xlyte for this as its not too hard to conceal the controller on the bike where it will not get any rain and it can easily be covered if it does rain, its no different to carrying a pair of waterproofs if you like, something to cover it with.

I have mine fitted to a seatpost rack, I got one of those dry bags off Ebay, the bags you use on boats to make sure stuff stays dry, I cut it open and still using the clip at the top made a nice cover that just sits over the top of the controller, its been fine for over 3 years or so. To make the controller waterproof would have meant potting it and sealing it, remember these things are really cheap in all honesty and they do a good job if looked after.

When I first got mine it was obvious by just looking at it that it wasnt really waterproof, but then neither are their motors so they do need a ticking off for that I suppose.

Good luck with the rebuild of the controller, I have fixed loads of them, its not too tricky! follow the guides on here and if you get really stuck PM Richard he knows all there is to know on them.

Knoxie

 

[smithinparis]

knoxie,
Thanks for the feedback. While I do agree with all your points about the cost of the units, and not being hard to protect them on the users side, it's just the concept that the units are intended to be used outdoors, so they should be able to survive outdoors - at least to a certain degree. If you are the manufacturer of a product, you need to understand, and make provisions for the intended environment (yes I'm an engineer ...).
When I first got the controller, I saw that they were far from being waterproof and I should have gotten around to protecting it sooner (I knew better than to leave it be), and for that I totally blame myself.
But as for weatherproofing, watch manufacturers, bike computer manufacturers, bike light manufacturers, gps manufacturers, and so on, have been getting it right for a long time, and they are all sensitive to price point. The most glaring example is my $10 bike light which has a sealed switch and o-ring for the battery compartment. The cost impact of adding those to a $10 unit would have certainly cause some project managers to cringe - but they did it anyway because it was the right thing to do.
Anyway, knoxie, I do truly appreciate your feedback. It's just that I get disappointed when people cut corners at my expense. But also the point of the thread was just a word of warning about just how sensitive the unit is, so other people don't get caught out.

 

[swbluto]

I think it's assumed by the manufacturer that the consumer will assume that's it's "electric" and that it doesn't mix well with water so they won't generally ride out in the rain and so they try to cut costs on that assumption of their "average customer". This, however, completely ignores the more adventurous(Well, that sounds like an excellent euphemism for some sub-groups) customers and when the appropriate warnings aren't added at the point of sale, some get screwed. Sorry for your loss. I'm currently in the process of weather-proofing the individual components(Not the motor, though. It doesn't seem like I have much to worry about since I'm sensorless.) so that it doesn't happen.

 

[smithinparis]

Point well taken...
It's like I said, I knew better - and I'm kicking myself for it ...

 

[smithinparis]

How are you going about weatherproofing the individual components? Silicon?

 

[swbluto]

Heh. Heh. Well... truthfully, I haven't actually done much since my ebike is busted and I'm not planning on repairing it and using it until this January. It's just a "plan" at this point:

Controller:
1. Sprayed polyurethane on the PCB(I still need to check if that's really effective - If not, I'll get some genuine conformal coating.).
2. Route the wires through an appropriately sized "hose" and seal that against the controller's out-port hole.
3. Use some of my neoprene matting to seal off the sides.
4. Silicon around the hose's rubbing against the out-port hole's side and various exposed holes.

Throttle:
1. I'm designing an LED based throttle and I haven't built or completely designed a prototype yet, so that's still to come. Link's easier suggestion was to coat the hall sensor's legs and other electronics inside the throttle(He used hotglue which will probably suit him fine in his Oh-so-sunny Californian environment), but it doesn't seem like I have that kind of ease-of-entry to my throttle's innards. :/

Battery:
Baggy! My panniers are waterproof, so it's pretty good there. The connectors are in the pannier and I should route the cables so they go upward.

And for the various exposed connections, I was thinking about using those "trailer" connectors, but that's still pending investigation as I'm wary of their ability to carry current.

And my environment is somewhat cold and usually wet, so I don't have to worry too much about running "too hot" like those in New Mexico. When the rain starts disappearing during the Summer, its high is usually around 70-80 degrees, nothing to panic about.