Shielding + Measuring noise on oscilloscopes

[Mike Parker]

Does anyone know if it is necessary to earth a shield or is simply connecting the shield to the common of the source sufficient? Is there some fundamental difference, does one give more protection than the other...

Ive tried measuring with an oscilloscope but you cant see the difference as the oscilloscope is earthed anyway.. annoyingly. On both occasions you need to connect the shield to the ground terminal on the front of the oscilloscope otherwise you see a lot more noise. I presume that terminal is there to give a better reference for the oscilloscope to your local ground.. it just happens to be earthed as well.

Mike

[sahpe]

Hello Mike,

I was looking for a quick answer to your question and came upon this article which I believe you will find very useful and informative:

http://zone.ni.com/devzone/cda/tut/p/id/3344

Please have a look and let me know if it is helpful.

Kind Regards,

Michael S.

Applications Engineer

NI UK & Ireland

[Mike Parker]

hi mike,

i recognise the top bit of that article but not seen the lower stuff before. is very good actually, is prompting lots of questions. im currently in the process of reducing the noise in my measurement system.. hours and hours ive spent fiddling gradually reducing the noise but was wondering what to do about the magnetic signals its picking up.

some questions arising from that doc

1. refering to the bit on capacitive coupling.. if you notice the vnoise source is grounded, i wonder if the source of the noise wasnt referenced to the same ground whether you would still get interference.. if not then would that be one of the benefits of having the measurements system floating?
2. for the bit on measuring floating sources, i read somewhere about using the shield as a return path for bias currents being better than sending them back down the signal cables themselves. ie take figure 17 and disconnect the ground at the source. also i wonder if this would work for grounded signals although would need to go through a resistor to limit any potential ground loop current.
3. not come accross the concept of magnetic shielding before.. is this essentially building a faraday cage around either the source thats made from iron. if so does this i guess trap the flux within the cage..?
4. ive often considered using balanced signals for my measurement system but have no idea just how much difference this would actually make. i imagine you would have to have a lot of noise to justify the added circutary. also if you're going to add circuitary to invert your signal you may aswell boost it at the same time and add another amp to boost the +ve version too. also for fig 23 i think you could also use the shield for bias current return path again.

seriously though thanks so much for that link.. is really informative. measurement noise is such an under-studied subject.. most people know about twisted pairs and that you need to ground sheilds. but grounding them to the common at the source is the essential thing.. before i connected them to earth which added a horrific amount of noise to my signals.

many thanks

mike

[mickeyw]

Hi Mike,

To answer one of your questions, instrument mechanisms cannot be shielded against magnetic flux, therefore it is necessary to employ some means of directing flux around the instrument. This is done by placing a soft iron case about the instrument. Because the flux is established more readily through the iron (even though the path is longer) than through the air inside the case, the instrument is effectively shielded. I will continue working on your other queries.

Best Regards,

[mickeyw]

There is also some really useful information on this link on noise reduction and shielding. I hope this helps.