Dear Donggyu,

ground problems are a challenge and it will be hard to sort the problems out from outside.

But one rule has to be obtained NOT ONE device must be connected to 2 grounds. When you operate with 2 grounds they have to be completely separated. And I can imagine that this is your problem from the beginning. You had always connections to your plasma device (which was connected to your lab ground or the outside ground) and devices connected to it which are connected to your lab ground. Even in the beginning when also your plasma chamber was connected to your lab ground can you exclude that any device (e.g. computer) was connected to another ground, e.g. via the power line?

So first you have to take your plasma device and make one well defined groung connection. Then add your 'plasma chamber hardware' like the pulse generator. I guess they have to run on the same ground like you plasma chamber - so they have to be isolated from the power grid via a transformer. At this stage you can test if your computer are properly running when the plasma is triggered. the same with the diagnostics. For the diagnostics you have to decide connected to some different ground in the lab or attach to the plasma device. I guess the laser and CCD camera can be operated completely separated from the plasma device ground - but ensure a sufficient isolation there might occur rather high potential (which might be an argument to run at the same reference potential). Any electromagnetic stray radiation is always an issue. Our experience was to connect the shield of BNC wires to the plasma device ground is favorable. Ideal are short wires to some opto coupler and having an optical transmission wherever possible.

Albeit just a rough guess, I hope this helps.

Regards,

Werner

Thank you, Hans.

One thing I'm confused is from ground of the power line. Of course, we have so many power code to provide power to each devices and ground of the power code is obviously to from lab ground (common ground for whole building, rather than direct connect to earth.) If I want only one common ground, Should I have our own power line from new ground (directly go to earth) thus we don't need to touch previous lab power socket? or giving to use new ground and should find some good connection among lab ground?

Dear Donggyu, 

I agree with the advice of Muller. In addition, you should consider that when you create a connection to the ground, it seldom is an ideal connection. Usually, you have a residual inductance L which is intrinsic to the connection. As a consequence, for a given time (proportional to the square root of L) voltage raises and em noise spreads around. You must work to minimize the residual L, e.g., reducing the distance from the device to the ground.

A true earth ground works better than a common building ground.

Finally, to avoid spikes to oscilloscope and detectors, you should cover every cable by a metallic sheath carefully welded to BNCs.

Hope this may help

Paolo

Dear Donggyu,

In the plasma laboratory we are straggling with RF noises and grounding issues daily. It is a little hard to give you a full answer from here, it is a very complicated field which I even have a presentation  about which I have gave in a PW laser conference.

Basically, you should remember that if you have a "short" discharge experiment its ground is connected to the "electrical company" ground by a very long inductance - long-length wire. You need this grounding for safety!

To protect your scope from noise you must (!) use a faraday cage. The Faraday cage will pick the ground noises - will charge a little -since it has  some capacitance will discharge to the ground much slowly- later in time.

There are many rules of how to build a good Faraday cage - if you wish I can send you my talk- just send me your e-mail address,

Bests,

Eyal

Dear Donggyu,

 you will not be available to avaoid any ground from you electric company, e.g. the power supply of your computers. Therefore, it is vital to separate the 2 systems. Common ground of electric devices like your computer and devices attached to the ground of your plasma device which can be connected to the lab ground or the earth (your 2nd option). I would not bed that common ground and lab ground are really identical (as long as it is not proven). If it is the same you should consider to use a separate ground for your plasma device anyway - your colleagues will appreciate it.

When operating with two grounds you need proper isolation of the two systems. E.g. if your trigger unit is attached to the plasma device (which seems the most logical solution to me) it has to be separated from the common ground. I assume it will still need a connection to a power line. This connection has then to be isolated from ground = you have to include a transformer between electricity supply and trigger unit.

Everything else connected to common ground has to be isolated against your plasma device. Best way is to use opto copplers for all signal transmissions between devices connected to common ground and your plasma device.

Then "only" stray radiation problems should remain. All devices which are known to react strongly on stray radiation and magnetic fields should be in metal boxes (in worst case even mu-metal). All wires should be coaxial (I guess you already did it).

But still then some specific issues remain. In our experiment we made the experience for probe measurements that it make a difference if we have the shield of the BNCs connected to the plasnma device side over the whole distance or at half way to the diagnostic device side (albeit both were formal on the same ground). This was changing the noise pick-up already by at least one order of magnitude. Such things you can only figure out by experiments or doing modelling.

Best regards,

Werner

Dear Donggyu

1- using good coductors and short lengh in conection amonge the points in your system (i.e plates from Cu) 

2- using a good insolater as miler.

3- usig the rod from Cu and1.5 m lengh ,put it in  dry sole and mixing with NaCl keept  a good conection with your system.

Best regards

you should follow the tips from the colleagues from Weizmann Inst. and from Vienna. As your discharge current represents a short, large current pulse, it will give rise to large induced emf / voltages that will be picked up around the lab.

You will do yourself a favour by sketching out a diagram of all instruments, electronics and wiring.

Instead of shielding everything inside your laboratory from the effects of the high current pulse, it would be better for you to shield the source, rather than everything that might be affected. Before going to look into isolation transformers and or line filters, for affected equipment, use an oscilloscope and verify the induced noise on the power feed to your equipment. With that baseline, you will be able to tell whether your remedial measures are actually successful. 

Another note: with typical BNC cables you do set up a wonderlul array of ground loops, as the outer conductor of the BNC cable typically connects to an instrument ground on either side. So if both instruments are grounded individually, then via the BNC cable, you will be creating a ground loop. Isolated BNC receptacles and isolation amplifiers are the next things for you to look into.

Your primary reference fro grounding loops should be the classic "arts of electronics" by Horowitz and Hill - you will learn a number of useful things. 

Good luck,  J.