Ground plane for test lab

[TMD]

Can anyone direct me to reference material for designing a grounded floor and earth ground for a high voltage (100 kV) test lab? Something on the web is best of course.

 

[dpc]

http://www.erico.com/products/SRG.asp

Might try a Google search on "signal reference grids".

 

[peebee]

Are you more interested in quiet signals or safety grounds? What kind of power levels are we talking about here?

If your 100kV source has any ampacity behind it, you might do well to check the IEEE Green Book.

If you're talking high frequency and low power, that's a whole different can of beans. Usually, your SRG or Faraday cage or whatever you're building wants to have a maximum grid spacing of about 1/3 the wavelength of your highest frequency in free air to ensure standing waves can't ride on your grid.

 

[TMD]

Thanks peebee. To answer your questions:
signal or safety - both.
power - transformer is 100 kV, 5 kVA, so not much.

We'll be measuring 1.2/50 us impulses, so noise is the primary concern.

 

[peebee]

What's the actual bandwidth that you're interested in? Not exactly sure what you mean by 1.2/50us, please clarify.

I'll assume you're talking about 1.2us pulses, so that would be a 2.4us cycle, or 400kHz -- although I suspect that you might be interested in frequencies orders of magnitude higher than that depending on harmonics, my assumptions, etc.

Anyway, if I recall correctly and my math is right, wavelength x frequency = speed of light. So wavelength = (3x10^8m/s)/400kHz = 750m. To damp noise out at 400kHz, you'd want a maximum grid size of about 1/3 of that, or 250m. So, you don't need much of a grid. A nice grounded loop of wire running around the perimeter of your lab, say perhaps #8 bare stranded copper or THHN insulated copper, bonded to your building grounding system, should do nicely.

Note that if instead you are interested in frequencies up to 4MHz or 40MHz or 400MHz, you'd instead wand a grid sized at 25m or or 2.5m or 0.25m. Even at 25m or 2.5m, you're probably still talking constructing something out of building wire. At 0.25m, or about 10 inches, you might do better to buy prefabricated copper mesh from a company like erico (see dpc's link above), or even just using 1/2" chicken wire from home depot.

Hope this helps.

 

[peebee]

One more note -- if you're looking for a quiet ground to reference your equipment too, then construct this grid as a plane at the floor (our around the perimeter walls) and bond everything to it as a signal reference point.

If your are looking to shield your lab, then line walls, floor & ceiling with the grid.

If you're looking for both shielding and signal reference, than line walls, floors & ceiling & bond all equipment to it.

 

[TMD]

The 1.2/50 us is a standard impulse test for electrical equipment. For instance, for equipment rated 11 kV, a sample must withstand a 75 kV pulse, having a rise time of 1.2 us and a fall to 50% of peak of 50 us. Standards vary, so the amplitude and number of positive and negative impulses can vary. Refer to IEEE-4 and IEC 60060 for lots more.

This is what we intend to do. We need a good ground to prevent noise when measuring this impulse.

We do not intend to do partial discharge testing, at least not at low levels of PD, so a shielded room is not required.