Voltage spike protection

You'll need to define your requirements for "protect." No over-voltage protection can react in zero time to clamp the voltage to (V-limit plus exactly 0 volts).

Knowing how to do it cheaply requires more knowledge of what device and in what environment. Guessing the operating and maximum allowable parameters of an arbitrary device isn't productive.

"Efthimios" skrev i meddelelsen news: snipped-for-privacy@3g2000yqs.googlegroups.com...

limit the current into the device to a safe value

-The spike has a duration of 0.2 to 0.3 seconds.

-The device to be protected is a sensor with operation voltage of 1 Volt and maximum allowed 1.5 Volts.

-The sensor draws a less than 0.1 milliamp.

-The device works under normal room temperature and humidity conditions

-Cost preferably about a dollar.

Yes, but we frown on doing people's homework for them. Maybe check the other side of google for "transient protection" or so.

Good Luck! Rich

Protection of power rails and signal lines pose differing problems. You'd have to be more specific about the actual node on the specific device that was threatened.

RL

"Efthimios"

-The spike has a duration of 0.2 to 0.3 seconds.

-The device to be protected is a sensor with operation voltage of 1 Volt and maximum allowed 1.5 Volts.

-The sensor draws a less than 0.1 milliamp.

-The device works under normal room temperature and humidity conditions

-Cost preferably about a dollar.

Just short the + and - terminals together and connect the AC ones across your sensor - this will clamp the DC rail to 1.5 vols. A 1 or 2 amp rated bridge costs about $1 ( DIL ones are cheapest ) and will survive 5 amps or more for 0.3 seconds.

Your problem if there is no series impedance limiting the current to a few amps.

..... Phil

---+---[DEVICE]-+---[fuse]--- | | +-->|--->|---+ | | +--|

So, you need to put a network between the source and the sensor that passes 1V and 0.1 milliamp accurately (so less than 10 mV drop and less than 1 uA current error), but protects a sensor that undergoes some kind of failure at 1.5V, against inputs in the 100V range.

Is that +100V? or could it be -100V?

Usually, one would put a triggerable shunt across the sensor, and a small triac (or pair of SCRs) might do it. Most packaged surge protectors operate at much higher voltages, though, and sense-with-an-amplifier schemes have a few microseconds of delay.

If 1.5 volt is the max allowed, why not use the simple components that are out there already? 2 sets of diodes to form 1.5 average break down joined back to back.

this gives you a nice bi-directional clamp via some components that should be able to handle that through some kind of current limiting source, like a resistor.

They also make TVS diodes...