Multiple device setup and bias

msnejbjergmsnejbjerg
edited April 2017 in Other Platforms
Hi
This page is awesome! Fantastic work.

I'm designing af 32 channel (also ref and rld) electrode impedance measurement tester for some electrode caps used for research. It is going to be used for checking impedance before actually moving the patient into a MEG scanner where an commercial scanner takes over (no impedance check).
What mode should I use? Daisy or Standard
How do I setup the bias hardware in these modes?

Or should I just add 8 (4:1) multiplexers on the 8 inputs of a single ADS1299?

best regards
Martin

  

Comments

  • chipaudettechipaudette
    I have not yet played with impedance checking on the new V3 boards, but on the old V1 and V2  16-channel boards, the impedance checking had some behavior that might complicate your measurements.

    The V1 and V2 boards (like the V3 boards?) implemented 16 channels by daisy chaining two ADS1299.  Fine.  The interesting behavior that results (at least from that old V1/V2 implementation) is that each daisy chained unit mirrors the settings of the first unit.  So, if you activate the impedance check on the first channel of the first board, you're also activating the impedance check of the first channel of all the daisy chained boards.  For the regular electrodes, this is probably fine, but for the ref and DRL, some funny results might occur depending upon how you're implementing your ref and your DRL.

    So, based on this discussion, one might be inclined to avoid the complications and use the 4:1 multiplexers.  For low-level analog signals, multiplexers make me nervous.  If you've had positive experience using multiplexers for this kind of signal, go for it.

    Otherwise, I'd move forward with using the ADS1299 in daisy chain mode, while recognizing that you might need to spend some time trying out different methods of implementing the ref and DRL to achieve the behavior that you desire.

    Chip
  • msnejbjergmsnejbjerg
    Hi Chip
    Thanks for your comments. I think I will try the multiplexer solution. I'm also a little nervous about low-level signals and multiplexers but there is a first time for everything. Something like analog devices ADG759 with 100 p leakage current and 3 ohm on resistance might do the trick.
    Martin
  • ferdinandferdinand
    Hey Martin,

    you can try to connect all your ADS1299s to one SPI bus but give each it's own CS pin on your micro. That way you can access them separately.

    Ferdinand
  • msnejbjergmsnejbjerg
    Hi Ferdinand,
    I'm aware of this, but how do I handle the bias in this setting? Can I use the bias on the first chip while measureing impedance on the third chip?
    Martin
  • ferdinandferdinand
    Hi Martin,

    so you want to use the lead-off detection mode to measure the impedances of your electrodes, right? In that case you should be fine with just one BIAS connected.

    Regarding the bias I suggest you take a look at page 53, figure 51. Basically you can just connect the BIASINV pins of all devices together, but you have to power down all but one amplifier. For this to work you have to either
    • use the daisy-chain configuration, but supply each device with a separate DIN signal,
    • or give each device its own CS - which I think is the better solution.

    Also, if you're just doing impedance measurements I would suggest you take a look at the ADS1298. Its specs are a little worse than the ADS1299's, but it's less expensive.

    Ferdinand

  • wjcroftwjcroft Mount Shasta, CA
    edited September 2014
    I'm pretty sure Joel changed the Chip Select on the V3 boards, so each ADS1299 can be separately addressed.  Email below was part of our impedance discussions last year.

    William

    ---------- Forwarded message ----------
    From: Joel Murphy
    Date: Fri, Dec 20, 2013 at 7:42 AM
    Subject: Re: your ADS1299 impedance measurement solutions
    To: K
    Cc: William Croft, Chip Audette, Conor Russomanno, James Cotton

    Hi K,

    Yes, you're right. The Daisy-Chain implementation that allows for control of all cascaded devices via individual CS pins is much more versatile configuration. In our first cascaded design, we had to work with the Daisy_In to D_Out setup because we could not 'see' the individual ADS CS pins through the isolation barrier. Our design for the 16 channel system will use the cascade design instead of the daisy design.

    The data sheet says that using START pin is recommended with this, however we will also test sending the START SPI command to synchronize them. Have you tried this?

    Joel

  • biomurphbiomurph Brooklyn, NY
    Hi,

    The V3 boards with the Daisy Module are set up to use different CS pins for each. That makes it possible to adjust settings for each individual pin (gain, input type, SRB connection, Bias generation inclusion, etc). We call it the Daisy Module, but it is daisy in name only :)

    As said above, we are connecting the BIAS_INV pins of each, which is how the Bias generation is able to include any and all of the input channels.

  • adeeladeel
    Hi, I have designed a 32 channel EEG system using Open BCI 32 bit v3 mother board..
     I developed a 24 channel daughter board with three ADS1299 in cascaded form ... making CLK Sclk Din Dout Reset pins shared with all ADS of daughter to mother BCI board.. and allocating separate CS pins from microcontroller. ... and BIASINV is also commen for all ADS chips... the designed project is working but from daughter board it doesn't gives me the desired signal ... can any one help me that where is the problem...???
  • wjcroftwjcroft Mount Shasta, CA
    It sounds like you've made some firmware modifications as well. One question I have is how do you get all 32 channels through the RFduino serial port, since that was the bottleneck that reduced the sample rate from 250 hz (8 channels) to 125 hz (16 channels).

    http://openbci.com/forum/index.php?p=/discussion/504/expand-from-16-to-32-channel

  • biomurphbiomurph Brooklyn, NY
    @adeel
    Can you share your work with us? Code repo? Images of the work you did? That will help us help you figure out the problem.
  • adeeladeel
    yes sure , I will,the hardware is complete and right now I'm working on the coding for 32 channels, As looking to the hardware i am facing some problems because the ADS datasheet and BCI mother board circuits configuration are not matching .. like common SRB and Bias .. and BiasINV pin is not having feedback loop ... because the voltage on Motherboard Bias DRV is -2.25   and for my daughter board its +2.25 .. from which i conclude that the reason is only with the BIAS DRV. and which I also think that in BCI board its noth that BIAS which are expecting to be there.. From this point I may say that the results of BCI board will be more efficient for taking the BIAS INV as mention in ADS1299 Datasheet, Figure 46 page 48. The BiasINV having feedback back with BIAS IN and out ... Here is the problem in my problem and iam stuck here so can you please explain me why there is no feedback loop in BCI mother and daisy board
  • adeeladeel
    @wjcroft yes dear I will think about it but this is not the serious problem . the main problem with me rightnow is I'm not getting the right analog signal using my daughter board and BCI motherboard, Even for 16 channels with my board (8 for my board and 8 for BCI v3 Board)
  • adeeladeel
    Actually I'm doing my masters thesis and my objective is to make 32 channel EEG system. For this I bought BCI R&D 16 channel kit ., MY core is to make new 24 ch daugther board which you named as daisy board( actually not daisy :) ). and use this 24 channel board with BCI v3 motherboard to have total of 32 channels. but i didn't understand that how 1 Megaohm resitor is providing a BIAS DRV
  • biomurphbiomurph Brooklyn, NY
    @adeel
    the figure you mention refers to re-routing the bias drive to different channels. We have not tried that. Our system is set up to use the Bias as a DRL connection with the subject. In this case, the BIAS_IN is connected to provide a feedback to the system for monitoring the BIAS output (you can connect it to a channel and then monitor it internally) 

    The BIAS_INV is connected to the Daisy (yes, I know, it's not 'really' daisy) and therefore the channels selected on the Daisy to be a part of the Bias generation are all input into the mother board Bias drive. In this case, only use the mother board Bias pin for DRL, as it will collect all of the electrode inputs thru the BIAS INV.

    I don't understand that the voltages are -2.25 and +2.25 on the various bias pins. 
  • adeeladeel
    edited January 2016
    Yes! you are right Sir, but I'm saying that if you are using mother board's BIAS as DRL and also you connected the daughter boards BIAS to the bias In and out and by opening this pin will also disturb the bias signal ( in case of 16 channel ). And how 1 megaohm resistor gives you the correct bias Drive ?? As i mention the page 48 figur 46 in the ADS datasheet ..

    Because The BIAS IN and Out and BIAS_INV are all connected to the amplifier internally so daughter board's BIAS_INV(which is connected with mother board) will also have comes from the BIAS IN and OUT. This is little bit confusing me .

    Thank you alot for your answers Its really appreciating and I'm learning more from you.
    I want to share the picture of my designed board but don't know how to attach image here, ther is no such button :( :(
  • wjcroftwjcroft Mount Shasta, CA
    To link to an image, store your jpeg on an external file server. Then use the "TV Screen" icon in the menu icon bar and enter the URL of the image.

  • adeeladeel


    Thanks @wjcroft

    https://fhscript.com/Dm4/10603436_1550106501981676_6228247410065775798_n.jpg
  • wjcroftwjcroft Mount Shasta, CA
    Beautiful job @adeel! Keep us posted.
  • adeeladeel
    Thankyou @wjcroft . I'm designing now a 24 channel board of same size as BCI mother board. 
  • adeeladeel
    https://fhscript.com/Dm9/WP_20160106_22_24_50_Pro_(2).jpg

    This is what actually confusing me :p which i was i think unable to explain .:)

  • wjcroftwjcroft Mount Shasta, CA
    There is a thread explaining the Bias refinement on the TI E2E forum for the ADS1299. I'll see if I can get a direct link for that. Also look at,

    http://openbci.com/forum/index.php?p=/discussion/172/understanding-bias-signal-generation

Sign In or Register to comment.

© Vanilla Keystone Theme 2026