SOLVED Cathode switch wiring fried motherboard

Hi all,

So I'm in the process of building a new computer and I'm putting a lot of extra effort into a complete custom wiring job. One of the things I'm trying to do is have a pair of cold cathodes which are powered by the motherboard using a 3 pin fan connector instead of a molex. It's something I've done before, but now I want to have the cathodes controlled by a switch which is mounted on the top of my case. This way I can just flip the switch at night and have a completely dark PC when I go to bed.

So I have the switch and was hooking it all up, experimenting with the wiring using a 9V battery and a spare LED fan to check for connectivity. At one point I decided to test the wiring on my friends computer (he gave me permission), since my computer isn't running yet. Well, I plugged the header into the motherboard with the LED fan connected to the switch, and when I flipped the switch to on, the motherboard exploded... with a loud pop and that telltale burnt transistor smell. Luckily the mobo didn't take any other components with it and we were able to get it replaced pretty quick, but I still can't figure out what caused it.

Here is a diagram of what I'm trying to do. Picture this with the motherboard fan header instead of the battery and no fuse. The switch in the diagram is exactly what I'm using. http://i46.photobucket.com/albums/f135/liljoejoe54/led_rocker_switch_diagram.jpg

I discovered after the motherboard incident that I had gotten careless and accidentally connected the ACC prong of the switch directly to the motherboard ground pin. But I don't understand why that would cause the mobo to blow up.

I redid the wiring and I'm now 100% sure that it is exactly like in the diagram, but it still won't work using a 9V battery, andI'm afraid to test it on another motherboard until I can get it working with the battery. I've tested every solder joint and point to point connection of the wiring with a multimeter and get a good solid connection, but when I put it all together, only the switch's LED lights up, but no fans or cathodes from the ACC prong.

I should mention that I'm using a 3 pin power distribution block from FrozenCPU which is connected to the ACC prong of the switch, which then powers the devices. Eventually I just want the switch to control a single cathode inverter and 2 LEDs on the case.

Can anyone help me figure out what's wrong?

your wiring diagram could be 100% wrong, toss it and rethink more simply, and in steps.

100% manuel switching:
first Direct connect with proper polarity WITHOUT any switch at all.
then manually remove and replace ONE wire in this manuel connection, leaving the other wire connected.
now you have a manuel switch, and your sure that is working right.

Replace Manuel with Switch:
THEN remove only ONE wire connection, and make that same manuel connection through the switch itself.
there now you have a switch that replaces your manually disconnecting.

all the switch should do is break and make one connection , one Pole , in this particular situation. just to remove the connection to the power/battery.
that would be only ONE side of the power is disconnected, just like you would do manually, when you pull off ONE wire.
simple right , and nothing can go wrong yet

ok so now you have it working and powered, and switching. you can ignore trying to light the switch, and you got a switch. (half way there, and with complete hands-on comprehention of what occured)

Just power the LED:
now we add in JUST powering the led light in the switch. the led light in the switch will require both poles of the power going to the Assembly , as shown in the diagram, but that other pole of the power you sure as heck do not want to go to the other pole of the power. the power to the led is just to the led.
As we demonstrated above the switching itself is done via only ONE wire, and all we have to do next is power the led itself Which takes both battery connections.
and that is where the misteak was.

Soo, lets insure that it can never happen again.
as above your switch does only ONE thing, disconnect the power on one side

now use some sort of small fuse (like 1/2-1 amp not 20) inline with the other power line and connect it up to the switch ONLY to power up the led in the switch itself, not to the other side of the switch.
you should see the led go on, if it doesnt you have to remember that Leds will only work in one polarity. So reverse it and try again.

then combine the 2 simple ideas.

by using a very small Fuse , it will break if something goes wrong, using a 20amp fuse would never be usefull in this situation.
any fuse should always break at least ONE side of the power, just like step 0 where you can manually remove the power.

the fuse would go between the power source and ANYTHING ELSE, so the fuse goes closest to the power source or battery.
you can put in 2 fuses TOO, anytime you have possibility for case shorting, and other odities, it would be doubbly safe to have fuseS comming off the power source. at least untill you know more about the interconnections. a fuse is just a small wire, that breaks when needed. in a computer case, your + lead should be fused.

A 9V battery would be a good testing item, but it has very small power, and might not even run the device.
a Voltage meter could come in very handy. it is better to manually trace down what is happening, than to use a curcuit diagram when you dont know what exactally is happening, or can see inside the switch.
.
.

Maybe I'm a little confused, but is there a reason you wish to run the power from a motherboard header instead of a dedicated spare line from your power supply?

Additionally, I'm not 100% certain your wiring diagram is correct. Make sure your jumper connections are appropriate with a DMM. 3 prongs on a switch like that could mean ground-normal open-normal close, or even ground-normal open-power for switch LED. Always check every switch with a multimeter before using.

Thanks for the responses. The reason I want to use a motherboard header is pretty much just because I think molex plugs are ugly and I don't want any in my computer. And it's easier and neater to use a motherboard header than to mod a 3 pin header onto a PSU cable.

As for the switch, I've checked every point of the circuit with a multimeter, and I found that the switch LED lights up when either POWER and GROUND prongs or ACC and GROUND are connected. But earlier I tried having the cathode inbetween the battery and switch in just a single circuit, and it didn't work. I will try it again though. I might have had the wires mixed up at the time.

I will try your suggestion later this afternoon when I have time, Psycogeec. Thanks again for the replies.

Ok I got everything working. Following your directions, as soon as I hooked up the switch, I ran into problems, so I knew the issue was there. There was a connection, but I was only measuring around 2.5V going through the switch, and about 8.5V without the switch. So I undid all my wiring and connections and tried it again by just touching the bare wires together and viola. So it turns out I think the problem was just a weak solder joint somewhere that wasn't allowing enough current to pass through to the devices. Anyway, everything is hooked up and working now, running off a motherboard header and wired according to the diagram I posted above. Thanks for the help.