Protecting outputs from high voltage

[Matador]

I've killed two chip-level IC's that I use as simple function generators. Both are simple (cheap) microcontroller boards that are programmed to deliver a 0.5V peak, 440Hz sine wave that I use for testing/verifying circuits. They are essentially DAC outputs that are used to either a) behave as inputs to amps/circuits, or b) used to inject test signals into various points on a circuit.

The output of the DAC runs through an (internal) 47ohm resistor, then through a 1uF, 600V cap so that I can inject into inputs that may be biased up to hundreds of volts. There is also a 1M hanging on the end of the cap to 0V to keep the output around 0V unless I'm actively probing something.

A few times, I've touched the probe to low impedance points that are at high voltages (like tube power supply rails), and it can kill the output(s) on the DAC. I'm guessing that the moment I place down the probe, the cap looks like a short and spikes the output of the DAC to a high voltage for a brief period of time (while the cap charges) which punches through or otherwise destroys the output of the DAC. The last one I touched accidentally to a screen supply and it instantly went distorted, and now the output of the DAC only supplies half cycles (like an internal pull-up transistor got blown).

Aside from adding two clamp diodes from the DAC output to the positive and 0V rails of the DAC board, is there something else I should be adding to the output to make it more robust? I'm guessing a 1N4007 should be beefy enough for this task but am open to other suggestions.

 

[JohnRoberts]

If you can get away with higher than 47 ohms in series bumping that up a couple X could help.

The clamp diodes there do not benefit from being rated for high reverse breakdown voltage. Small signal diodes might clamp faster but I wouldn't worry about that unless and until you lose some ICs with 1n400x clamps in place. Even small signal diodes like 1n4148/1n914 can handle 1A briefly and higher transient currents even more briefly.

[edit- if the PS rails are not robust enough to sink the clamp current, and alternate approach is a couple zener diodes to steer the clamp current to ground. /edit]

JR

 

[sahib]

I've killed two chip-level IC's .....................

Aside from adding two clamp diodes from the DAC output to the positive and 0V rails of the DAC board, is there something else I should be adding to the output to make it more robust? I'm guessing a 1N4007 should be beefy enough for this task but am open to other suggestions.

Will do. All of our modules have overvoltage protection using 1N4003 on both inputs and outputs.