novembersierra28 wrote:Runaways are my greatest fear thanks to my BIG error with brushed tech there's just no telling when that capacitor will blow...
I read on 4qd about how high currents will inevitably damage the caps causing them to pop...so ever since then I've kept it at low speeds..(is there any way of knowing if a cap is on it's way out?)
Only way is to test them periodically, which in general will require removal from the device they're in. You can visually check for swelling, but they're typically already toast by the time you can see any sort of failure. The test would be with a capacitance meter, or better yet a, LCR meter that can test them for capacitance and internal resistance (ESR). Off-the-shelf units that do this for large capacitors like a controller will have aren't cheap; the DIY versions could be done a lot cheaper but may require you to be able to program an MCU (not to write the program, but to flash a blank one).
I wouldn't really worry about it--the thing is really not so much the currents in the caps, in the enclosed controller case like the Curtis, but rather the heat from the FETs and having no airflow inside the controller, that adds wear to the caps even faster than the high currents. Keeping the controller case itself cool, with a good heatsink and fans, is the only practical way to deal with that in anything but totally dry weather. In dry weather, you could open both ends of the controller and stick a fan on one end, and that would help way more than just keeping the outside cool.
As for speeds, well, realistically whenever the controller is actively doing PWM (at lower throttle settings) instead of being at full throttle, it's more likely to be having high current flow in the caps. At full throttle the PWM shuts down and it just passes power thru all the time, so the caps aren't having to filter anything, really, and have just about no current flow. This is true of brushed controllers and probably of most brushless, too.
Also, at full throttle, you have less FET heating from switching, although more current in total passes thru the controller, less heat is usually developed within it.
What capacitors did you use when you mended your curtis brushed motor controller?
I didn't have to replace anything in there; it was just a bad solder joint (well, two or three, actually, between the control PCB and the power PCB).
When I fixed the ScootNGo controller, I tried out several caps; the best results were from photoflash caps out of old cameras.
With the 2QD controller, I forget what caps I tried in it, but I think one out of a high-wattage power supply from a laserprinter/copier was what I ended up with. Not sure now.
how long did your brushed controllers generally last irrelevant of driving style?
The only one I had major problems with was the ScootNGo controller, which started out quite fried, and I didnt' help it with my further experiments with various drives on this bike and others, leading up to the large FETs and drivers on separate external heatsink, with the v2.0.2 friction drive that worked great till I broke a motor shaft.
Most of the problems I had were me using the wrong kinds of FETs, with bad layouts and insufficient heatsinking.
I even moved the controller over to CrazyBike2 at first, and it worked well enough there until something damaged the op-amp chip at the heart of it, and I figured that replacing that op-amp was going to destroy the PCB since it had already been repaired so many times.
I never did go back and fix it again.
Thats when I moved to the 2QD, and the only thing that failed on that was after the chain came off, jammed, tacoed the chainrings, pulled the wheel out of the dropouts, twisted and ripped apart some chain links, and then FINALLY the controller couldn't take it anymore and blew two FETs (of four). Given that I had a tiny tiny heatsink on there, I don't really blame them--at 153A, maybe 3500W? it was already taking way more than it was designed for. http://electricle.blogspot.com/2009/12/ ... sible.html
Is it true that, if you place an ammeter, this creates inductance down to the controller causing back EMF which kills the controller?
Not that I am aware of, but it would depend on the meter and the wiring. If the meter installation requires adding a long loop of wire (a few feet, perhaps) up to the handlebars, that isn't normally there, then yes, it adds resistance and inductance. How much? Not a lot, but some. Enough to kill a controller? I guess that depends on the conditions. At high currents, a little inductance causes very large voltage spikes. At lower currents it matters less. The more switching that's going on (at lower throttles but higher loads), the worse these spikes will be.
If you have to loop the wire up there, I would recommend a twisted pair of wires (both pos and neg, even if you aren't using both of them for the meter) of equal length, and that will help cancel out some of the problem.