Molex Switch

  • Thread starter Thread starter Tux Fan
  • Start date Start date
T

Tux Fan

Were I can purchase a molex DPST (Double Pole, Single Throw)
switch that will power off both the 12V and 5V power?
 
Tux said:
Were I can purchase a molex DPST (Double Pole, Single Throw)
switch that will power off both the 12V and 5V power?

Do you mean, you want a length of four wire cable, with
Molex 1x4 male on one end, 1x4 female on the other end,
and a double pole single throw switch in the center of
the cable assembly ?

They don't make something like that. You'd buy a "Y" cable,
cut a section of wires in two, then get a Radio Shack
switch and solder it to the two ends. (In a pinch, you
could go to the hardware store, and use a 120V light
switch or something similar, if there is no Radio
Shack nearby.)

/
------- +12 --------x x------------->
------- GND ------------------------->
------- GND ------------------------->
------- +5 --------X X------------->
\

Example of a switch here - not perfect, but the
right general idea. The double throw means you won't be
using one row of pins on it. This one is rated 20 amps,
and the Molex connectors are good for about 8 amps or so.
It all depends on what you're going to power, as to what
the best switch might be.

http://www.radioshack.com/product/index.jsp?productId=2062531

There is an example of a Y cable here. They call the Molex
an "LP4" in this advert. These will get harder to find,
now that IDE drives are about to be discontinued. The
suppliers will assume there is no further business with
the old connector.

http://www.startech.com/item/PYO2L-Power-Y-splitter-Cable-3-Large-Ends.aspx

Paul
 
Do you mean, you want a length of four wire cable, with
Molex 1x4 male on one end, 1x4 female on the other end,
and a double pole single throw switch in the center of
the cable assembly ?

They don't make something like that.

Might be smart to mention the 'why' here: Such a beast could harm
both the PC and the device attached to the switch. Although the
actual risk of hardware damage is rather remote, this would
create a situation where the PC sees an unpowered device and the
device receives data on it's data connector while not being
powered.

A more serious risk:
Given the fact that Windows is the predominant OS and that
Windows is unable to cope with devices that suddenly seem to
disapear, you're in for a disaster :-)
Even devices that are designed to be unplugged (USB) require
prior warning (safely remove hardware) in Windows.
 
Gerard said:
Might be smart to mention the 'why' here: Such a beast could harm
both the PC and the device attached to the switch. Although the
actual risk of hardware damage is rather remote, this would
create a situation where the PC sees an unpowered device and the
device receives data on it's data connector while not being
powered.

A more serious risk:
Given the fact that Windows is the predominant OS and that
Windows is unable to cope with devices that suddenly seem to
disapear, you're in for a disaster :-)
Even devices that are designed to be unplugged (USB) require
prior warning (safely remove hardware) in Windows.

I can think of yet another reason it might not be a good
idea.

I've had my system crash, just by plugging in a 120mm fan
into an auxiliary power connector. Depending on the load
and its characteristics, it can cause the power to glitch
enough, to upset system operation.

A device like a disk drive, when initially powered, doesn't
draw current on +12V immediately. The motor is started
after the controller has done some initialization. So the
current flow is "shaped". USB is the same way, in that
intelligent USB devices (not your battery charger), are
only allowed to draw 100mA initially, and 500mA after
configuration is done.

An example of something that might upset the computer, would
be if the switched subsystem, had large bypass capacitors of
its own. The act of charging that capacitance, can drop the
rails significantly.

So there is some thinking to be done, about the type of
load, and whether it has inrush limiting, or inherently
limits transient startup current.

Paul
 
Might be smart to mention the 'why' here: Such a beast could harm both
the PC and the device attached to the switch. Although the actual risk
of hardware damage is rather remote, this would create a situation where
the PC sees an unpowered device and the device receives data on it's
data connector while not being powered.

A more serious risk:
Given the fact that Windows is the predominant OS and that Windows is
unable to cope with devices that suddenly seem to disapear, you're in
for a disaster :-) Even devices that are designed to be unplugged (USB)
require prior warning (safely remove hardware) in Windows.

I have an internal USB hard drive that I need to power off when not in use.
I have been pulling the plug when I am through using it, but a switch would be
easier.
 
Tux said:
I have an internal USB hard drive that I need to power off when not in use.
I have been pulling the plug when I am through using it, but a switch would be
easier.

http://pinouts.ru/Slots/USB_pinout.shtml

USB uses differential data lines to ensure noisefree signal transmission
and reception. That means there is a D+ and D- signal line. There is
also the 5-volt line. You would need a 3-pole ganged switch to
open/close all 3 lines at once (which they won't be all at once because
switches always will make-break contacts a bit different by micro- to
milliseconds; i.e., they do NOT all instantly connect or disconnect at
the same time).

By inserting a switch inline with the cable, you open the shielding to
introduce noise. I haven't investigated into the internals of switches
to look for those that incorporate additional internal shielding but
soldering to the pins outside of the switch undoes the shielding
provided by the cable so you may have to design a grounded shield around
the switch. For USB cabling, the ground line is supposed to be
connected only to the host's ground, not to the device's ground. This
is to eliminate ground loops that can introduce noise. You will have to
run the switch shield back to the host's ground (and that is NOT the
same as chassis ground). Also, some USB cables use twisting to further
nullify induced noise, and you will be untwisting them at the point you
introduce the switch. Some USB cables don't use twisting.

Rather than add a switch, why not go into the BIOS and disable the
controller for the USB ports. USB ports are paired at the controller
(i.e., it manages 2 channels). If you disable a USB controller, you
disable the pair of ports. If you only have the one internal USB drive
attached to the 8-pin mobo header then it's the only device on that USB
controller. After all, it's not really the power you need to remove.
It's the accessibility to the drive. My guess is that you do not want
malware, malicious or ignorant users, or accidental backups to gain
access to the drive except at the times that you choose to allow access
to the drive. A switch could be flipped by anyone. Add a password to
the BIOS so only you can get into it to enable/disable the USB
controller for the USB drive.

Because of the fast data transfer rate for USB, just slicing apart a USB
cable to slap in a mechanical switch could introduce noise, lag, and
reflection that generates poor quality transmission. A non-professional
hack job could result in flaky operation. Rather than open the case to
unplug the USB cable when you want the drive isolated, why not just push
out a punch-out in the back of the case (provided for serial or parallel
connectors) or leave out a card blank and run the USB cable to the
outside to plug into a USB port. You'd still be having to plug and
unplug a cable but you wouldn't have to be getting inside the case.
Instead you just have a cable dangling out the back.
 
Back
Top