ADS1298R: ADS1298R VCM

Hi Ryan,

I'm with Jonard and we're supporting a common customer. I have an additional question about the internal RLD summing available on RLD_REF. The difference is that signals for summing is taken after the PGA in ADS1298R, but in ADAS1000, there are taken before the PGA amp. If I understand the derivation depends on the PGA amplifier gain of each channel, where default gain is 6 for the ADS1298R?

Currently I am solving the VCM externally using programmable buffers and 10k resistors for creating average from selected input channels. VCM is then put into negative inputs. For this, the VCM could be derived from the same inputs or from different ones for the testing purposes?

Kind Regards,

Jejomar

Hi Ryan,

Thank you for your detailed response. For my statement clarification:

I have created an independent common mode (VCM) externally (see circuit diagram below), where the input channels INx_P are selected by OPamp with SHDN pin. The averaged (VCM) signal is fed to the INx_N inputs. This independent (external) VCM is created for testing purposes only, because we don’t know, if the common mode derivation will be always derived from the same channes like for the RLD. This independent solution is also used in ADAS1000 from analog devices. We didn’t tested if it will be really usable in some cases for the ECG better signal quality. For our purposes, we need to measure ECG signals unipolarly with different lead systems at the same time (12 standard leads, Frank leads etc.) with the electrodes placed on the chest and back. That's why we need to use two connected ADS1298R devices.

My next questions are related to this:

1. Is the circuit for the independent common mode correctly designed?

2. In case of two cascaded ADS1298R modules (with two chip select (CS) connected to the MCU) is it also possible to use DMA for automatic ping-pong buffers filling without incorporating the host MCU (the ESP32-S3 host MCU supports DMA for SPI).

Kind Regards,

Jejomar

Hello Jejomar,

1. The common-mode derivation circuit above looks correct. However, I would like to ask why the customer is not using the WCT amplifiers contained in the ADS1298? Do they need more than 3 electrodes for common-mode derivation? If not, then they can use the 3 WCT amplifiers for generating a negative reference for the positive electrodes. The same WCT voltage can be shared amongst two ADS1298 devices, if needed.

2. From the ADS1298 device perspective, I don't believe it matters if you are using traditional SPI or DMA to read out the conversion data, so long as you completely read all data between nDRDY interrupts. There is no FIFO on the ADS1298 itself, so all data must be read within one data rate period, or else it is overwritten with the next conversion result.

Regards,

Ryan

Hello Ryan,

1. We would like to use up to 16 input channels (first channel for respiration), where the common mode should be derived from any of the input channels combination. For achieving up to 16 channels we are planning to interconnect two ADS1298R in cascade to configure also RLD from any of 16 input channels derivation.

2. We already have a host MCU software for SPI communication with one ADS1298R device, where reading data is achieved only via DMA channels (DRDY is used as DMA trigger, and other DMA channels are used to fill two ping-pong buffers). The MCU core is only woken up if one of the buffers is full. But there is a question, if the DMA could be implemented also for reading data from two ADS1298R devices connected in cascade without incorporating the MCU core, because there are two separated chip select (CS) signals where each device must be enabled and read the data separately. Or in this case is the only option to connect using daisy-chain? However, in the daisy chain connection, I don't think there is a possibility e.g. to select channels for RLD derivation.

Kind Regards,

Jejomar

Hi Ryan,

May I ask for a follow-up? Thank you!


Kind Regards,

Jejomar

Hello Ryan,

Thank you for your recommendation regarding the daisy-chain! Configuring registers using separate CSs and then simply reading the values ​​in daisy-chain configuration could be a solution for a simpler DMA implementation that I will test.

The proposed ECG measurement system should be configurable for RLD channel derivation setting and also for the VCM channel derivation setting independently. We always assume unipolar measurement for calculating individual leads by SW post-processing.

An example could be a standard 12leads ECG (RA, LA, LL, C1-C6) measurement system, where Einthoven leads (I II, III) will be computed from the limb signals and the Wilson central terminal and V1-V6 leads will be computed from chest signals by the post processing. We assume the RLD will be derived from all the measurement signals in this case, and the RLD will be connected to the body via a separated RL electrode. The VCM (referential potential) will be then measured either by a next separated electrode connected to the body or calculated as an average of selected signals (we currently don’t know if the VCM will also be derived from all the measurement signals or whether only C1-C6 will be selected, for example). The aim is to achieve the best possible signal quality by the RLD and VCM signals derivation.

This system will also be used for other ECG electrode configurations such as Frank lead system or other lead systems that will be tested at our organization for development purposes.

Please let me know if this consideration of the signal derivation separately for the VCM and for the RLD is correct, as also used in ADAS1000. Or, if there is some incorrect assumption, e.g. that the VCM and RLD should always be derived from the same signals.

Kind Regards,

Jejomar