Differential input filtering deviance from datasheet

NateNate United States
in Cyton
Hi all,

I did a quick search of the forum but wasn't able to find this topic, if this has been brought up before please link me to the correct thread. I'm going to be using a Cyton board as a test platform for the ADS1299 front end soon (already have the board). When going through the schematic, our dedicated hardware designer noticed a discrepancy in the input filtering. Specifically, the last bullet point in the ADS1299 datasheet section 12.1 (Layout, page 72) states that analog inputs with differential connections must have a capacitor placed differential across the inputs. It is essentially saying that the input filtering for a differential channel must also be differential, and it shows that on the layout example on page 73. However in the Cyton schematic it shows that all inputs, both positive and negative on each channel, are filtered to ground instead. This would work fine for single ended channels, but violates the layout restriction for differential channels.

Unfortunately I need to use differential inputs to obtain the data I want. So given that this board is advertised to support differential inputs, how is the board getting around that layout restriction? Does the board inherently accept a reduced sensitivity by going the ground filtering route, and if so by how much? Or is this a case where the datasheet claims a must but in practice it has been proven to not be required?

Kind regards,
Nate

Comments

  • techsupporttechsupport United States
    Do you have an evaluation board? Would love to see the difference.
  • NateNate United States
    Interesting, I took a look at the eval board schematic and as far as the datasheet is concerned it would also be in violation. So our hardware engineer took a look at the TI forum and found a thread where a TI apps engineer said that they originally thought that single ended capacitors were best at the time they created their eval board, but "learned otherwise" and updated their datasheet to reflect the real best configuration. Unfortunately he doesn't quantify how much difference there is between the two configurations. 

    So I guess its something to be aware of for future designs. Here's a link to the post, part 1 of first answer. 

  • wjcroftwjcroft Mount Shasta, CA
    Nate, hi.

    Here is my unofficial observation. The per-channel front end RC filter (regardless of exact schematic placement) is a low pass filter, intended to dissipate / block environmental-noise high frequencies. So in the 'optimal' placement shown in the data sheet, it might be very slightly more effective at this blocking than in the configuration used by both OpenBCI and the TI Eval board.

    Your question was: "Does the board inherently accept a reduced sensitivity by going the ground filtering route, and if so by how much?" I think the answer to this is clear. The RC filter placement has NO effect on the sensitivity to EEG signals for the channel, regardless of whether the channel is used differentially or referentially. The RC filter placement will only do a slightly better job of high frequency environmental noise rejection, with the data sheet configuration. 

    One advantage of the configuration used by the TI Eval board and OpenBCI is that it can be used easily in both differential and referential configurations without any circuit changes.

    As far as sensitivity comparisons of the Cyton with other medical grade amplifiers, see this paper by Jeremy Frey,


    Regards,

    William


  • NateNate United States
    William,

    That's fair enough, thanks for responding. That's a really unfortunate wording of the datasheet with the use of the word "must", when really it's just an optimal placement for slightly better noise rejection. My question's pretty much answered at this point, have a great day. 

    Nate
Sign In or Register to comment.

© Vanilla Keystone Theme 2026