Headstage with noise only on home-built open ephys acquisition board

[tom....@gmail.com]

I completed my home-built open ephys acquisition board v2.2. The lights are all on the board. I checked all the voltages at the test points and they are all correct. I have no difficulty connecting to the board in Open-Ephys Gui. However, the signal appears to just be noise (see screenshot).

I would like to record ECOG and LFP (0-500 Hz), and was expecting to see a signal that appears to be like an EEG in this range. However, with either the headstage at rest (not connected to an animal), or connected to the animal,  I am just seeing noise in the GUI in all the channels. I get the same result with a home-built 16-channel intan omnetics headstage (https://github.com/open-ephys/headstage/tree/master/1x16_Omnetics_Intan), as well as with a brand new intan headstage I ordered from intan (I bought one since I thought it may have just been my headstage).

I'm not sure what the issue is. Everything else seems to be perfect; I seem to be able to record with the ADC, and I can control the LEDs. I have reviewed the board under a microscope, and don't see any solder bridges. I am happy to test out the logic, but I'm not sure what to test for.

Any help would be appreciated!

--Tom

[Jon Newman]

Good job making your own board!

I guess the first thing I would start looking at is power. I would do this with an oscilloscope if you have one -- a rail might look fine with a voltage meter and be totally unstable at the same time.

The noise looks quantised as well indicating a potential issue with the digital communication with the headstage. This could be caused by power as well, though. 

Do you see noise on the onboard ADCs as well? -- they have a completely different A/D path comparted headstage.

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/514fdfa9-b418-4428-8e12-f943352b6c12n%40googlegroups.com.

--

Jonathan Newman

Postdoctoral Fellow, MIT

http://www.mit.edu/~jpnewman/

[Jon Newman]

Also, those neopixels are notorious for having the same footprint and different pin functions across parts with terrible documentation. So make sure those are correct. They are not required for board function.

[tom....@gmail.com]

Thanks; I'll check the rails with a scope, and check the ADC.

The Neopixels seem to work, changing colors appropriately when connecting headstages, etc.. I checked the datasheet (http://cdn.sparkfun.com/datasheets/Components/LED/WS2812.pdf) and it appears to be consistent with the schematic.

Also, does the headstage have to be connected to A? For some reason, the headstage is not detected in the GUI if I leave A empty and connect the headstage to B-D. 

Thanks!

[Josh Siegle]

The headstages should be detectable on any port, so that’s probably a clue. I would recommend checking the soldering of the Omnetics connectors under a microscope, to make sure there aren’t any loose connections or bridges.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/4d6c75e2-3a98-42a3-b4a9-81613fcda753n%40googlegroups.com.

[tom....@gmail.com]

Thank you so much for the replies. The ADC is working perfectly; I checked with a voltage pulse train and I get accurate recordings in all 8 channels.

I don't see any loose connections or solder bridges at the connectors. I have also confirmed there are no shorts between the pins. It's weird that all 3 (B-D) all are not working. I'll try removing the omnetic connectors for B-D in case there might be a short somewhere in one of those connections that I can't see.

Now when I connect, I'm just getting diagonal red/white lines. It definitely seems to be an issue with serial communication to the headstage.

Thanks!

[Josh Siegle]

When the software checks for connected headstages, it starts with A, then goes to B, C, and D. So I think it’s less likely that B-D are messed up, and instead there’s a problem with A that’s preventing it from detecting the other ports.

I’d try removing the A connector first.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/950e3dee-9350-4f47-9192-4459c88929f2n%40googlegroups.com.