"Voltage Fuse" .... ?

[Zben3129]

I am trying to think of a way to make a "fuse" of sorts to protectmy computer when I am using it as an o-scope.

I have heard that any voltage over 1.7v can damage the soundcard, so I want to make something that will not conduct if the voltage is over 1.7v.

For DC, would a diode with a voltage drop of 1.7v from + to gnd work?

Im stumped on the AC part completely


Thanks

Zach

[George Giblet]

You can go along way with something like this:

probe tip -> Series Cap -> 2k2 resistor -> Red Led Clipper to ground (2xled in parallel) -> sound card in

Note there is no resistor to ground a the probe tip, this prevents inadvertent circuit loading.
If you want to put one in use say a 10MEG resistor.

You could also add say a 1MEG resistor across the LEDs.

The next option is to place a preamp protection at the input - better but more mucking around, and you will need a power supply.

[Zben3129]

You can go along way with something like this:

probe tip -> Series Cap -> 2k2 resistor -> Red Led Clipper to ground (2xled in parallel) -> sound card in

Note there is no resistor to ground a the probe tip, this prevents inadvertent circuit loading.
If you want to put one in use say a 10MEG resistor.

You could also add say a 1MEG resistor across the LEDs.

The next option is to place a preamp protection at the input - better but more mucking around, and you will need a power supply.

What does that setup do? Wouldn't the series cap block any DC voltage I was trying to probe? Does the 2k2 resistor limit the current/voltage? What do clipping LEDs do?


Thanks for the help

Zach

[George Giblet]

Most sound cards will not convert DC, the PC input is internally AC coupled to the A to D converters.

The circuit I gave is for AC,  the LED provides clipping so that you cannot blast the PC input circuits.  The capacitor is so there is no DC path when you probe circuits with a DC component - you need the cap in there.

[gez]

George's second suggestion of (diode) protecting the input of the sound card might be the better option.  If you're using the sound card to run a virtual scope (or whatever the phrase is), then sticking capacitance in series with the input could result in distortion of the signals you're measuring, leading to false results.  It's also possible that your card has some protection.

I vaguely remember either a post or article by RG about protecting a soundcard.  Try geofex, or a search in the archives?

[R.G.]

Yeah. What you do is use four diodes in series/parallel so that it conducts at +/-1.2 to 1.4V depending on the diode. Then you use some resistance in series with the input AND ground. About 120 ohms is right.

The diodes clamp the voltage across the input to be less than 1.7V, and the resistors keep the current down to a level that does not burn out the diodes. If you use 1N400x series diodes, you can withstand an ampere of current, and a 120 ohm resistor keeps the current down to 1A even if you decide to (and please don't do this intentionally) to measure an AC wall socket. It would be 60 ohms per side, but the ground side needs at least 120 ohms because depending on what you touch, there may be 120Vac just on the ground side, and that needs to be limited. 120 both sides for balance.

The 120Vac thing is one of those low-likelihood things that are hard to protect against, but also hard to say will never happen. Unfortunately, you're betting your computer's motherboard that you will never do it.

Unfortunately, it would take a 120W resistor to withstand this current indefinitely, so use a 1/8W resistor. This will burn open and act as a current-limiting fuse.

In most cases...

An external high impedance divider followed up with a buffer/amplifier to get the signal level back is even better. This can be designed to use high resistances at the input for the divider; the followup buffer regains the signal and can be designed to not be able to put out more than +/- 1.6V and have a sharp, undistorted output up to that level.

[George Giblet]

I tend to use LEDs because the turn-on is sharper so there is less distortion of the signal than silicons.  They light-up when they conduct which is a good warning system.

(You shouldn't try to use the PC to measure mains!)

[R.G.]

I use silicons because the fault current capability is higher. If I want them to be really sharp, I'll go active to sharpen them up or sidestep the issue with a divider in front.