ADS1298: How to properly use the RLD functionality

Mr. Smith,

Thank you for the response!

Upon looking over the documentation, I still have some questions.  

1) As far as common-mode rejection (CMR), does that take place through an electrode connected to RLD_OUT?  Or where do I send that CMR signal?  Is this the signal I can also send through a different electrode in case the RLD electrode becomes disconnected? (I care less about the RLD becoming disconnected and more about how to send RLD_OUT through a different channel).

2) What is the purpose of CONFIG3 bit 4? (RLD_MEAS)

3) If I have RLD_SENSP and RLD_SENSN at 0x00, do I have any RLD signal?  Or is it this that explicitly generates the signal?

Clarifications:

• I am trying to get EMG data from a forearm, and therefore all the electrodes will be on the forearm.
• I will, in theory, be using this to control prosthetics.
• I was able to watch the entire video series and read some of the extended documentation in the BIOFAQ.

Thank you again for all your help.  

Best,

Spencer

Mr. Smith,

Thanks for responding via my number system.  It helps keep track of whether questions are answered or not.

When I look at figure 34 in the datasheet, I see RLD_OUT going into RLD_IN.  After RLD_IN, I see it essentially replacing IN8P and IN8N electrodes so you can read the signal.

I'm hoping to use the RLD signal to help get rid of 60 Hz noise, but my current board doesn't offer access to RLD_OUT.  I don't fully understand the RLD signal and how to use the electrode.

1) Is the RLD electrode an input or an output for my device?

2) If I understand the process correctly, I take the RLD signal, apply it to the patient via the RLD electrode.  This helps eliminate 60 Hz noise.

Is 2) correct?  Or am I missing how the RLD_electrode works?

3) Is there a way to send the signal through a different pin besides RLD_OUT?   I thought in an earlier post (of yours) that it was possible to send this signal out a different channel.  Is this shown in figure 33?  MUX bits either 110 or 111?

Thanks for your help in this project and your timely responses.

Best,

Spencer

Mr. Smith,

The following is a sampling of data using all the information we've talked about.

I believe that the red data is correct EMG data.  However, I cannot figure out why the other 3 channels shown seem to not follow the same pattern.  The channels shown are channels 1-4.

My registers read the following in hexidecimal...

CONFIG 1 85
CONFIG 2 10
CONFIG 3 E4
Channel 1 00
Channel 2 00
Channel 3 00
Channel 4 00
Channel 5 06
Channel 6 00
Channel 7 00
Channel 8 07
RLD_SENSP 07
RLD_SENSN 07

I'm trying to generate an RLD signal based off an average of the first three channels.  I think I have CONFIG 3 correct.

Any thoughts on the rest of the signals?

Thanks,

Spencer

Hi Spencer,

Where exactly are the electrodes connected on the body? 

If you change their position, Ie swap channel 1 and 4 do you see the same signal pattern? 

I would like to know if this is a channel configuration or an electrode location issue. 

How does it look if you change CONFIG3 to EC to generate the RLD signal internally?

Mr. Smith

I use a sleeve of 34 electrodes which are spaced most places on the forearm. I can essentially use any for a set of differential pairs. My current setup involves using one electrode near the elbow/ulna as my RLDOUT signal (Currently channel 8 N). Another electrode near the elbow/ulna is the Negative IN for channels 1-7. The other electrodes are placed sporadically around the forearm in no particular order (and I can move them to more or less any location around the forearm).
The following series of figures shows data I collected in one sitting. Figure 1 shows the noise when connected to an AC power source. I then disconnected my laptop from the wall for the remaining figures.

Figure 2 shows that when using CONFIG 3 = 0xE4.
Figure 3 shows that at some points, I can get two channels working.
Figure 4 shows that without any further commands to anything, the two channels can change.
Figure 5 shows that when a channel drifts into another channel (or area), it changes from good EMG data to noise.
Figure 6 shows that one of the two channels has been lost to noise.
Figure 7 shows that all the channels have been lost to noise (because of the drift mentioned in figure 5).
Figure 8 shows a different two channels with good EMG data.

I feel as if I have tried several different channel settings for both the configs and the channels themselves.
If you have any suggestions, I would happily take them. In the meantime, while I wait for your response, I will more heavily document any patterns I can find.

1. Is there a good way to reduce drift?

Thank you for your extended help in understanding RLD with the ADS1298.

Best,

Spencer

Figure 1

Hi Spencer, 

Very interesting application. No rush, but when you get the chance if you could re-post the pictures that would be great. 

What kind of drift are you referring to? Is the DC level of the signal slowly changing over time? 

Try placing the RLD electrode in a more central location and see if that helps. 

That's odd the pictures only semi-uploaded.  I'll give it another shot.

The drift I'm referring to I believe is DC offset, but it can move significantly over the course of 5 minutes as you'll see from the figures.

I'll move the RLD and see if it changes anything.

Thanks,

Spencer

Hi Spencer,

It happens. 

Alternatively, you may want to try blue-wiring a mid supply common-mode configuration like what is depicted in this post: