USB Port Surge Protector

Anyone know if such a beast exists? We have guys in the factory that test
power generators by hooking them up to serial ports attached to touch
screens. We are using USB to serial adaptors. When they wire the device
up wrong, it sends 480 volts of juice frying the adaptor. Well, that
wouldn't be a big deal, but each adaptor is $35 a piece and we just went
through 10 like water!

Looking for something with replaceable fuses or a "flip switch".

Thanks,
Adam

Adam Leinss wrote:
> Anyone know if such a beast exists? We have guys in the factory that test
> power generators by hooking them up to serial ports attached to touch
> screens. We are using USB to serial adaptors. When they wire the device
> up wrong, it sends 480 volts of juice frying the adaptor. Well, that
> wouldn't be a big deal, but each adaptor is $35 a piece and we just went
> through 10 like water!
>
> Looking for something with replaceable fuses or a "flip switch".
>
> Thanks,
> Adam

I see a couple of protection possibilities, just looking at
what is available:

This one, clips the difference in voltage, between the signal
lines and a ground screw. If the source of voltage is weak,
then the clamps on the signal lines, will prevent the voltage
on the RS-232 signal lines, from going outside the allowed
range. If the source of voltage is "hard", then the little
clamping devices will be blown to hell. So this had better
be a leakage path, and not a direct connection to 480VAC.

http://www.bb-elec.com/bb-elec/liter...32DSP_4502.pdf

And this one is an optoisolated RS-232. Here, there can be a
large difference between the grounds on the two systems. But
the voltage difference between the RS-232 signals and their
own ground still has to be "normal". I picked this doc, because
PDF page 20 has a partial schematic of the optical isolation.
Note that there is a separate "iVCC" voltage, used to power the
second half of the circuit. It is not clear in this document,
where "iVCC" comes from. There is no connector on the card, for
an isolated power supply to be connected. And in the picture
of the circuit card layout, I don't see a place for a module to
create a transformer isolated source of power. It could be
that the necessary power, is developed from RTS, DTR or whatever,
as in phantom powering, but again, there is no evidence of that
here.

http://www.bb-elec.com/bb-elec/liter...sd1x-3903m.pdf

I only suggest those two, to illustrate they solve two different
problems. The first device assumes an overvoltage on the signal
pins. The second assumes a large difference between grounds. Note
that RS-232 has two grounds, and in the case of the isolated PCI
card, the shell of the connector is still touching the PC ground.
It is the pin 1 ground that has no hard connection to the PC side.
The PDF document above, says there is room for the user to solder
a resistor into the "R5" position on the PCI card, if the user
wants to join the Pin 1 ground to the PC.

The first thing I'd want to understand, is what kind of fault
can arise. And then select a solution for it.

Here is a USB version of optoisolated RS232 output. (You may or
may not want to combine this with the 232DSP device - as it
depends on what problem you are solving.)

http://www.bb-elec.com/bb-elec/liter...ML2_4206ds.pdf

Paul

Adam Leinss wrote:
>
> Anyone know if such a beast exists? We have guys in the factory that test

No. A surge protector has to be "ahead" of anything it protects, NOT
connedected to it.

--
http://www.bootdisk.com/

Plato <|@|.|> wrote
> Adam Leinss wrote

>> Anyone know if such a beast exists? We have guys in the factory that test

> No. A surge protector has to be "ahead" of
> anything it protects, NOT connedected to it.

That is just plain wrong, most obviously with an optically
isolated USB port if such a thing actually exists