How to detect electrical breakdown?

Discussion in 'The Projects Forum' started by Klimto, Dec 21, 2010.

1. Klimto Thread Starter New Member

Dec 21, 2010
4
0
Hello people,

I am working on a project that has to detect an electrical breakdown. The system works as follows:

A high voltage is beign applied to the top of a metal, this metal has a protective layer on it. If the protective layer has pores, breakdown will occur between the top and the bottom, because the air inbetween will start to conduct. The bottom is grounded.

The problem is, the protective layer combined with the metal forms a capacitor. This means there will be a current when the voltage is being applied, because the capacitor has to charge up first.

When a breakdown occurs, the high voltage will only fall about halfway down and then it will charge up again. Is it possible to detect a breakdown, and not detect when a breakdown does not happen? Because of the charge up current it will always detect something initially.

Is there a way At this moment i use a monostable multivibrator to detect a change in voltage, and i have coupled this capacativly (don't really know the english word) to the high voltage. Problem is, the coupler will try to draw current from the ground when a breakdown occurs so the current will flow away from the detector. I'm really stuck, i have tried a lot of things, maybe someone here knows something that could help me?

2. Klimto Thread Starter New Member

Dec 21, 2010
4
0
I have tried this, only i have used the 74121. The problem is that i might use a voltage up to 500 volt. This is way to much for a single IC, so i had to use a voltage divider combined with a zener. Problem with this is that when the breakdown occurs there will be a stable voltage because of the zener. This is because the high voltage does not totally dissapear with a breakdown.

Jul 7, 2009
1,585
142
If you provide a description of the system and what exactly you're trying to accomplish (in addition to the engineering details of the solution you're currently using), there may be other solutions you haven't considered. There are a number of people with a wealth of experience who might provide insights that could lead to a better solution. In other words, don't assume the method you're using is the best choice -- there might be better solutions and a fresh eye might see them. Also provide information on what constraints you have (e.g., technical ability, costs, time frame, etc.).

4. Klimto Thread Starter New Member

Dec 21, 2010
4
0
I am very limited in supplies, i can only use certain IC's (I may not use programmable IC's for instance). I would actually love to see another solution to this problem. Fact of the matter is, i can not share the exact details of what i'm doing, because it is a company secret.

I can define the main problem as follows, i must detect a fall of voltage from 500 volt to about 300 volt. Because the product itself acts as a capacitor the initial charging current must not be detected. I can only use CMOS and TTL IC chips. I also tried to couple the bottom of the metal optically to a detector, but this also detects the charge up current.

5. marshallf3 Well-Known Member

Jul 26, 2010
2,358
203
Company secret eh? I hate to to tell you but this forum isn't in the business of helping out companies design their own products but I may suggest something that may or may not be of use.

Put a large string of high voltage zeners together and figure it out from there. It is one solution.

6. blueroomelectronics AAC Fanatic!

Jul 22, 2007
1,758
98
Why's that?
Sounds more like homework than secret company stuff.
Yep, gotta be homework. Why would a company care what you used as long as it works.

7. Klimto Thread Starter New Member

Dec 21, 2010
4
0
Well it's pretty much something i have to do for a company, but school ordered me to do it. So it's somewhat like homework yes

8. kubeek Expert

Sep 20, 2005
5,350
1,010
You could use a comparator to find the drop in voltage, and use some RC time constant to inhibit the output until the inductance is charged.

Because you don´t tell us how exactly is the procedure of measuring done, or how much is it automated vs. manual work, the rest is for yourself.

Last edited: Dec 22, 2010