Strange electrical noise - help?

[JamieH]

Hello everyone,

I'm a neuroscientist and I'm currently trying to start a project using EEG.

Anyway to cut a long story short, we have been doing EEG for years with no problems, then about 5 weeks ago we encountered a strange source of noise in the data. The noise is at about 31 Hz and it looks like this (see image attached.) It's a weird waveform.

Our EEG technicians have not seen this kind of noise before. It is clearly distinct from the 50 Hz noise that you get from mains power (this is in the UK).

We think that the noise is somehow coming from the ground. This is because it does not appear in "free" pairs of electrodes recording from the air. These pick up lots of 50 Hz noise, but not the 31 Hz.

Anyway we are stumped. We're assuming it is some kind of equipment or electrical device somewhere nearby which is causing this. But it's a busy hospital area so there's equipment all over the place. It would help to know what kind of thing we're looking for... so if anyone could suggest what might be making this waveform?

Thanks!

Jamie

 

[poor mystic]

I've seen streetlights that oscillate a bit like that.

 

[poor mystic]

I see that the 31Hz noise suddenly gets worse at the end of the graph. Is that worsening usually related to the slower waveform on which the 31Hz is superposed?

It's a DC fan

 

[shrtrnd]

1)Find-out what new equipment was INSTALLED in your area when this started. It's
possible the new instrument is not well-isolated electrically, and is inducing the signal
into your EEG. When you've narrowed the list of probables, DISCONNECT power to the
suspect instrument, and run the EEG.
Sometimes these things only happen when the problem instrument is running, sometime
it happens as long as the instrument is plugged-in, and on stand-by.
2) Twice in my experience, I found the grounding for the building electrical had either
been weakened from wiring corrosion, or disconnected, allowing the ground potential to 'float'.
3) Look around your area for any new electrical items, that could be inducing this signal.
New compact flourescent bulbs, new computer or computer monitor, somebody's new
communications device that's always on, anything new.
4)Possible failure in EEG itself, that needs to be looked at.
Good luck

 

[daddles]

As I'm sure you've surmised, it looks like a switching noise. Poor mystic's suggestions could be good starting points. When doing such finicky measurements, One possibility would be to take down the experimental setup and having a fresh person rewire things, looking explicitly for any ground loops. Grounding is terribly important, so don't assume anything -- hence the need for a "fresh" person unfamiliar with the way things are usually done.

Another thing to do would be to contact hospital maintenance and ask them to search their records for what changed in the building 5 weeks ago.