|
I bought a case with fan speed control. (12V, 7V, 5V)
Define R4
http://www.amazon.com/Fractal-Design-Define-Arctic-FD-CA-DEF-R4-WH-W/dp/B008HD3EFA/ref=sr_1_3?ie=UTF8&qid=1364641179&sr=8-3&keywords=define r4
The case has connectors for up to 3 fans plus a LP4.
I tried plugging in only the fan connectors only and of course they wouldn't run.
I did this because I thought the LP4 connector might have been for an LP4 fan, while power to the fan might have been provided through some other connection that I couldn't see in the case's hardware.
I figured the LP4 connection was actually for power, and plugged it into the power cable directly from the power supply (the cable was of course a 6 pin to LP4 cable).
One of the two black wires started smoking, melting, and overheating.
I immediately turned off the computer. The wires are exposed, and has melted insulation now.
There is a simple switch board/chip in front of the case, where the fan control switch is.
In order to test if the chip was got burnt, I disconnected the LP4 connection and turned the computer on. None of the fans connected to this variable speed switchboard was functioning.
Did I make a mistake in concluding the LP4 connection was not for power? Was the problem due to some faulty manufacturing/assembly? Or did I completely miss something and make a wrong connection to overload/overvolt the wire?
|
That LP4 Molex connector is indeed for power and should run directly from your power supply, just as you had it.. Does your psu have any native molex connectors, or is it all 6 pin to molex? It shouldn't make a difference, but it is possible. I'd try with another cable first if that's possible (preferably on a different rail), and if that doesn't work it's probably the controller, and you should contact Newegg first, then Fractal.
Edit: also make sure you didn't derp and plug the molex in backward. Unlike most connections, it is sort of possible to do it backward, which would explain a lot.
|
The LP4 connector is keyed so that it's not possible to plug it in the wrong way.
The PSU is all modular so it's all 6 pin to LP4 molex.
I'm not sure how to do a different rail. did you mean different connection slot? The PSU is single rail. Not sure which output belongs to which voltage level, although it should be taking care of that by itself, since there is no way to distinguish between 6 pin outputs from the PSU (nor should you have to).
LP4 connectors don't have their contact items very positionally stable. When I first plugged it in without problems, the 3rd pin had gotten pushed out. So I then pushed it back in so that it wasn't sticking out, and that all 4 pins made contact. I then connected it, turned it on, and ran it. Was the 3rd pin supposed to be forced out? It didn't come that way. I can't imagine that being part of the design.
|
Molex LP4 connector pins:
Pin # Color Function
1 Yellow +12 V
2 Black Ground
3 Black Ground
4 Red +5 V
It was #3 that burned. I'm not quite sure how only that one would burn. It's ground. If somehow there was a short between pin2 and pin3, it shouldn't affect anything since they're both ground.
If pin2 is the ground for the 12V pin1, and pin3 is the ground for the 5V pin 4, then maybe pin1 acted as the source and pin4 acted as the sink along with pin2 and pin3? Even if so, I don't see why only pin3 would start overheating.
Somehow there must not have been enough cross sectional area in the wire to carry even the mild current, perhaps due to some damage in the wire that resulted in a partial reduction of cross sectional area. Still, how much amps would this loop draw? It is a 400W PSU, but it's just a fan connection.
Maybe the soldering in the little PCB switch behind the front panel was bad?
I don't see why it should have start overheating, or that I could have done anything more safely. I'm not Nikola Tesla, but I have some understanding of circuitry and general engineering safety design. If some computer was putting put this together and it started smoking.... the point is if what I did was incorrect it was a very easy mistake to make.
|
Power draws here are low; you really shouldn't get anything burning without a short, unintentional connection. Could happen from some kind of miswiring or manufacturing defect.
I wonder how they're getting +7V. The easiest way is to just use the +12V as hot terminal and +5V as cold terminal, giving you +7V difference. Maybe through some faulty wiring somewhere in some fan, or the front panel, or maybe the molex connector, +12V or +5V got connected through to pin 3.
Or maybe pin 3 just got loose and if there was current flowing through (especially if there was a short), the higher contact resistance from the loose connection meant more power losses and heating up at the connector.
|
i think what you're saying is on the right track.
question: how does a loose pin translate to higher contact resistance? is this different from saying the cross sectional area at the contact was effectively too small for the current draw?
Point to note: the entire 3rd wire (black) got melted the entire length of the wire (from molex connection end to the switch PCB end), while all the other wires are fine.
|
Resistance of say a wire depends on material, cross-sectional area, length, maybe some effect depending on material of other things like temperature. If you have two conductors rubbing against each other, there is some contact resistance seen by electrons across the materials. The usual models may not adequately describe the behavior there; or the parameters are difficult to know without measuring anyway. Not every material and junction out there is Ohmic. for a basic overview: http://en.wikipedia.org/wiki/Contact_resistance
If the connector didn't melt and especially seeing that the whole third wire was overheating, it's probably not a matter of a loose connection or contact resistance (which, if it caused heating, would be at the place of poor contact) or anything like that.
The weird thing is how the other wires are all fine. If there were a lot of current going through the third wire, where did it come from and why did the wire(s) involved with the rest of the loop not also overheat? This might imply that there was some unusual manufacturing defect about the third wire only that would make it have a much higher resistance than the others.
|
so there's a good chance i'll be able to get this covered under warranty?
It should not be that easy to have the circuitry destroyed like that. I think it would be more than reasonable that this be covered.
But i have to wait until monday, to call the manufacturer, blah.
Newegg won't cover both ways of shipping
|
|
|
|
|
|
EPT
