Hi Jacci,

Dealing with noise is not something that can be done well from a distance. But I feel pretty sorry for your situation, so I'll offer what help I can.

You have intermittent noise, which is the worst to deal with, and part of me thinks that you're pretty much stuck with the situation until you can figure out what makes the noise come and go.

That said, there are some general principles you need to stick with.

All of the electrical power needs to come from a single power plug. You don't know whether two plugs directly beside each other actually connect to the same physical ground. If one plug goes to ground in one place, and one plug goes to ground in another place, you will get a ground loop, which causes noise.

Similarly, you need to make sure that everything you ground (with the exception of the bath ground) all goes to the same place, normally a grounding pin on the back of your amplifier. Similarly, all metal objects in your rig should be connected to that ground with less than 1 ohm resistance. If I ever trouble shoot a rig, I walk around with a multimeter, measuring the resistance from the ground pin on the amplifier to everything metal, the microscope, the antivibration table legs, the faraday cage, the manipulators. Everything needs to be grounded. And you want to avoid situations where Item A is grounded by a connection to item B, which is connected to Item C.. and a long wire connects it to item D and it's item D who is actually connected to the amplifier ground.

Stimulators are a common source of noise. Make sure the wires to your stimulating electrode are wrapped in a grounded insulator.

The aCSF in and out lines are a common source of intermittent noise. You need these clean, and the aCSF in line should have a nice air gap in it, if the aCSF is coming in from outside the faraday cage. Likewise, air gaps in the aCSF out line (just caused by air bubbles) are good too. Otherwise, a continuous tube of saline makes for a great antenna.

But if you're serious about it, what you need to do is rebuild the rig. Start with nothing but your amplifier, you headstage, an oscilloscope and a model cell. Confirm that you get negligable noise. Then replace the model cell with a real electrode in a 35 mm dish of saline. Confirm no noise. Then bring in the DAQ and a PC. Confirm no noise. Then bring in the microscope, then the aCSF, then heating elements etc etc etc until you have a full functioning rig. The problem is, with intermittant noise, it can be hard to see be certain whether you don't have any noise.

What kind of recording electrode are you using? It sounds like you're using glass. Can you measure it's resistance with your setup? Your noise will be directly proportional to your electrode resistance. And they can get blocked with time.

Oh, and P.S. you shouldn't need a humbug. If you need a humbug something has gone wrong.

P.P.S This " once I have them placed, I do a test pulse to see if I get the correct waveform. I do this a couple times, and then the amplifier becomes overloaded, prohibiting me from getting a waveform at all. " this sounds pretty... non-ideal. You sure your amplifier isn't just broken?

Some additional things to consider to the wonderful answer above:

How is your perfusion system set up? Are you using a vacuum trap for the outflow? If so, is the flow steady? Intermittent noise could arise if the fluid level in the bath is varying. Additionally, if your bath is overflowing salt crystals could be forming on your bath ground wire above the area where you've chlorided the wire.

Are you using a heating system in your bath? Is it set on manual or automatic? If the heating unit is going on and off at intermittent periods that could be another potential source for intermittent noise as the heater comes on and off to maintain a constant temperature.

Has anyone told you about the wonders of aluminum foil and banana clips? You can make a small noise "shield" with aluminum foil and double-ended banana clip wires. It's a quick and dirty method to try and isolate where the noise is coming from. You cover the item you suspect of causing noise with some foil and clip the foil to a trusted ground. If the noise level drops it suggests the item you've "shielded" is contributing to the noise. For items too large to wrap w/ foil, just clipping a metal portion of the item to a trusted ground also can show you if an item is contributing noise. And of course, turn off everything you can safely turn off to see if any of these reduce the noise.

Good luck. Every electrophysiologist can sympathize with you about noise the frustrations of noise.

You may try to clean out any tubing (e.g., water) inside your rig. Water inside perfusion system needs to be cleaned completely in order to remove AC noise.

Thank you all so much for your highly detailed and informative answers! I really appreciate your time. I wish that I could spend some time in any of your labs to really learn this stuff.

William M Connelly , I am not positive that my amplifier isn't broken, and I'm not quite sure how to find out. I would love to rebuild the rig, but I don't have an oscilloscope...I can only see the noise when the Humbug blinks or when I turn on my PC with AxoGraph (I know, I know, this situation is not ideal in so many ways!!). The noise is intermittent, anyway, like you mentioned. It came on immediately when I turned the rig on yesterday, throwing my previous theory regarding its onset after the first run away in the trash! When you say that all my power needs to come from one power plug, would it be OK to plug my DAC/Humbug/amplifier into a powerstrip and then into the wall? Would that help? I know that powerstrips are notorious for adding noise, too. I also don't have a multimeter, are they expensive? Perhaps I can get the PI to buy one. Money and space are issues in this lab. Yes, I'm using glass electrodes made from a pipette puller, and the resistance is about 5 MOhms. Regarding the lines, there are air gaps in both in in and out lines, and no water. Regarding my stimulating electrode, do you mean the wires to the actual electrode? what kind of grounded insulator can I use to shield the wires? And I agree with you about the Humbug, it doesn't actually help. When there is no noise, there is no noise. When I do get noise, it just reduces the amplitude by a little bit.

Michael Matott I'm using a peristaltic pump with vacuum outflow, and the aCSF level seems pretty steady. I've tried cleaning the well and the grounding wire and pellet already, so I'm not sure it could be due to salt buildup. I haven't added the heating element yet, as I wanted to see if I could fix this noise situation before I added another confounder. I've tried the foil shielding, but I have not tried grounding it! What can I say is a "trusted ground?" I'm not sure I trust anything in the lab right now! ;-)

Jacci, if you follow William's detailed suggestions you should be able to identify a reliable ground. If no one around you doing e-phys is having the same issue the odds are pretty good that your building's ground (via the electrical outlet) is sufficient to eliminate ambient electrical noise.

Then again, it might not be. My PhD mentor had to get our physical plant folks to drive grounding rods into the ground outside our lab and run a grounding wire into the building for us to eliminate 60 Hz noise due to the shoddy wiring. (This problem was thankfully eliminated when they renovated the portion of the building we were in and installed GFI outlets.)

If the noise is as big as you are describing, just grounding to a stable object in the room - like your rack or air table - should be sufficient of a ground to at least determine likely noise sources, even if the grounding doesn't completely eliminate it.

Thanks Michael Matott . The other two people in my lab with rigs are patch-clampers, and they have their own isolated rooms. The plugs and copper grounding plates on the walls seem to be enough to keep their noise levels stable. Do you think that the fact that I am in the main lab room with everyone else has anything to do with the noise? Today, for example, I'm getting constant 10 -15 mV 50 hz noise. There were 8 people including myself in the lab at the time(it isn't very big), and they were doing lots of electric-based things, with much other foot traffic in and out of the room. I am thinking about coming in really early in the morning when no one else is here and seeing if that makes a difference.

You may also use "Notch Filter" set at 50 Hz which would appear in software such PClamp or LabChart for removal of such noise.

Re all power into one plug: Yes, get a big power strip, plug that into the wall, and connect everything in your rig to that power strip.

Re multimeter. I completely understand regarding money, but a good multimeter is worth it's weight in gold. I have no money for my lab, but I still bought a proper Fluke brand meter. I'm not exactly sure how much they'd cost you in the US, but i'd guess around $200. You could get a Fluke 114, which is the cheapest Fluke meter worth buying, but for a bit more you could get a Fluke 117, and that is a fully fledged meter that will last longer than your lab will. If you're super out of money, then you can work with cheaper, but anything worth less than $50 is a toy.

"Regarding my stimulating electrode, do you mean the wires to the actual electrode? what kind of grounded insulator can I use to shield the wires? "

So in a perfect world, you'd buy cabling that had two wires running down the middle, and a shield running around the outside. That's actually quite hard to find, but "triax" or "triaxial" wire is relatively common (google it to see what I'm talking about). But if you're stimulating electrode is the source of the noise, it will be easy to check, just pull the stimulating electrode out of the bath. If the noise amplitude drops, that is your problem. Which brings us back to my original advice...

So you don't have an oscilloscope. Well that's a crime, but let's do the next best thing. Take everything out of your rig apart from your amplifier, your DAQ, and your PC. Connect a model cell to the head stage and confirm you have minimal noise. Then work in an electrode and a static bath, confirm you have no noise... etc etc etc.

Thanks so much, everyone! My PI moved my rig to a small room that I share with a patch clamper, but he works later hours than I do so we don't interfere with one another. No more stampeding undergrads and random electrical interference! I was able to set up my rig using all your suggestions, and (cross my fingers) no noise all week.

Mike: Phil here. Just like EMG's, connect all your electronic eqipment to a solid brass grounding rod that is driven into the earth. Dont rely on grounding through the wall socket. We had to bore a hole through a lab wall to run the heavy copper wire to such a rod. Also, turn off all Fluorescent lights.

Phil Motta

Dear Jacci,

welcome,

Sure i came late to your forum but my answer may help other colleagues who have troubles with interference on their electrodes. The problem with such electrodes is that they have high impedance loops and therefore they are subjected to the pick up of high frequency nearby fields. 

So, you have to isolate and shield them very well. You can filter out the pick up of the mains by a notch filter.

To complement the answer please see the correct procedure for stimulation and recording in the paper at the link:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5291724/

The authors gave the precautions to avoid interference from the environment.

Best wishes

Dear Jacci: One of the simply ways is to use “Notch filter” at the center frequency of 50 Hz to remove annoying noise.

Dear Sheng-Nan,

I'm trying to remove 50 Hz noise using Notch filter in Clampit.

However, could you perhaps give me your wise comment on how to decide '-3dB width (Hz)'? Thank you.

The value of "-3dB with Hz" in Clampfit could be variable. You might try to use the range between 10 and 50 Hz to minimize the background noise.

This thread is very helpful!

We have just got a brand new setup in our lab. We also attempted hippocampal extracellular recordings for the first time yesterday, but when we were ready to stiumlate and record, discovered a 50 hz noise on the baseline of approximately same amplitude of the expected signal, so cannot be ignored. We discovered that pulling out the stimulating electrode of the water bath reduced the noise to an acceptable level, revealing that the stimulating electrode is the main source of the noise. We will try to improve grounding and insulation to resolve the issue.