GUI Unexpected Impedance Measurement with Cyton

collig14collig14 IE
edited February 2017 in OpenBCI_GUI
Hi

I'm trying to measure skin-electrode impedance with the OpenBCI GUI with my OpenBCI V3-32 board.
Except I'm getting some strange results.

Firstly, I have shorted negative pin 1 and SRB1 to see if there's a safety resistor or something in there. 
When the pins are shorted directly, there is an impedance of 3.6 kOhm shown. That's fine.

So I tried inserting a 5.5 kOhm resistor directly between N1 and SRB1 expecting to see 9.1 kOhm.
Instead I get 7.48 kOhm.

Can someone explain?

Here's an image to show what I'm seeing.

image

Thanks

Comments

  • wjcroftwjcroft Mount Shasta, CA
    Did you mean SRB2 (closest to board)? That is the default reference.
  • collig14collig14 IE
    Apologies,
    Yes I do.

    N1 shorted with SRB2 gives an impedance of 3.6 kOhm.
    Very consistently 3.6 kOhm exactly too. So I see in Chip's blog that known lead-off current (~ 6 nA) of the ADS1299 is only accurate to +/- 20 %, but that's not the issue as repeating the measurement leads to the exact same value.

    But adding a 5.5 kOhm resistor gives me 7.48 kOhm instead of 9.1 kOhm then.

  • wjcroftwjcroft Mount Shasta, CA
    edited June 2019
    It sounds like a glitch got into the GUI program. Because this used to checkout as you would expect. Can you open an issue here,


    Are you using the V1 GUI or V2?

    Thanks again for reporting this. And if you want some reassurance, impedances of up to 40K work fine with high impedance input amps such as the ADS1299 (1 gigaohm),


    Regards, 

    William

  • wjcroftwjcroft Mount Shasta, CA
    edited February 2017
    7.48 - 5.5 = 1.98 Kohms error.

    1.98 / 5.5 = 36% error from expected readout of 5.5 ohms.

    The complete short of channel to reference never happens with skin resistance. The residual 3.6K that you see can be considered a sort of side effect of the RC network at the front end of the ADS1299. So the actual percent error would be calculated as above, not adding the resistances as you mention.

    I do think this agreed closer with the 20% error figure in the past. It's possible the impedance calculation code got slightly tweeked or morphed over time; and was not rechecked.

Sign In or Register to comment.

© Vanilla Keystone Theme 2026