ADS1298ECGFE-PDK: Lead off detection circuitry adds Offset to the measured Value !

Hi,

May I ask are you using an evaluation kit/board(EVM)?

Are you using DC resistor or DC current or AC Lead-off detection method?

Thanks 

We observed the behaviour in both EVM board and also in a custom board which we developed.

We used DC resistor based Lead off detection method.

We also have a doubt regarding should we allow this DC offset to be present as any offset in LA,LL and RA will affect the WCT calculation. This offset of 10mV is measured / observed differentially. So we don't know how much is the individual offset present in LA,LL and RA. Since WCT is calculated based on LA, LL and RA and this is then used as a reference for measuring V1  to V6 leads.

Could you please explain regarding how should we go about it.    

Hi,

Yes, there are DC offset when using DC resistor Lead detection method.

So, some work around for Resistor Lead Off detection might be

1. do some calibration and then subtract the DC offset.

and/or

2. use resistor lead off detection to detect Lead On/Off before acquiring the ECG signal. Once Lead On is detected and confirmed for x seconds, then disable the Lead On/Off detection to acquire the ECG signal without the Lead detection offset.

and/or 

3. 

Enable lead on/off detection and acquire for just short period of time(e.g. x seconds) and then disable the resistor lead on/off detection, but continue acquiring ECG signal. And, it leaves developer's choice what they want to do with the x seconds of ECG signal with offsets.

or

4. Use DC current Lead detection could result much smaller offset than the Resistor method.

or

5. Use AC current Lead detection method, and then developers will need to design/develop some LPF, HPF or BPF DSP filter afterwards.

Thanks

We planned to use the method 4 already but still our doubt is how this offset in LA LL & RA affect WCT and in turn how this offset will affect V1 to V6 lead measurement ?

Option 2 seems okay but in our case we are using DMA for acquisition and setting up of this logic will require DMA to handover the SPI lines back to core.

If the effect of Offset is Linear and same in all the V1 to V6 leads then most preferable for us will be to subtract this offset from measured values. For that we don't know the relation ship of offset in LA, LL and RA  to the WCT and then to the measured V1 to V6 w.r.t WCT.

Hi,

According to datasheet 9.3.1.7.3 Wilson Central Terminal (WCT) and Chest Lead and Figure 35. WCT Voltage, the offset seen in the 3 primary leads may not be the same as the V1 to V6 as the the WCT node voltage is generated/calculated differently from the 3 primary leads(differential voltage of any two electrodes) vs WCT(average of 3 electrode voltages) and the V1~V6 are calculated/generated by the differential voltage between any one of the chest electrode with respect to the WCT node voltage.

With above being said, there might be different offset that customer/developer need to subtract/compensate/calibrate digitally in firmware or software.

May I ask have you tested or will you use DC couple or AC couple between electrodes to the ADS1298 input pins? 

Thanks,

Chien

Currently we plan to use DC couple only. Do you have any general strategy for how V1 to V6  and  LA, LL & RA compensation be handled ?

Hi,

are you referring how to compensate or correct/calibrate the DC offset?

If yes, something you could try are injecting some known precise and steady DC voltage e.g. from -2 mV to 2 mV range should be sufficient for ECG applications. and then try different gain settings to note down the ADC code/count/value and/or the converted mV unit.

From there, you shall be able to come up a simple linear equation, either just an offset or a coefficient with an offset per channel.  It leaves developers/designer's choices whether developers/designers want to use only offset or coefficient with offset and whether to just use one equation of all channels or each channel could have its own calibration/compensation equation.

Note, the equations may depend on customers PCBA designs, system designs, product designs, wires, cables, electrodes, etc, so customers will need to do a series and reiterate of V&V and qualification tests.

Thanks

Hi,

We have already followed this approach for LA,LL and RA i.e Lead I, Lead II measured using ADS1298 Channel 2 and 3.We obtained some linear relationship. We shall proceed with same approach for V1 to V6 lead.

Could you please clarify that weather using unipolar or bipolar supply configuration is good for a typical ECG measurement system ?.

We observed that when in unipolar supply configuration (AVDD +3.0V and AVSS 0V) this system is unable to measure a differential signal of  +/- 300mV + 1mV sine @ 16Hz at channel 2P  and  0v at Channel 2N.

We are able to measure this signal clearly when we switch to bipolar supply configuration in ADS1298 PDK. Bipolar supply puts AVDD at +2.5v and AVSS at -2.5V. 

My analysis is that in Unipolar supply configuration, the test signal provided i.e (+/- 300mV + 1mVsine @16Hz) at channel 2P and 0mV at RA forms a single ended input with common mode 0V which is out of range of acceptable Common mode voltage 1.25V< Common mode voltage <1.75V.

But in Bipolar configuration the test setups common mode voltage falls within the range of the bipolar configuration -1.25V<CM<1.25V.

Hi,

Your observation might be valid as in datasheet

section 7.1, Analog Input voltage is restricted/bound by AVSS – 0.3 to AVDD + 0.3

section 7.3 Negative input (VREFN) AVSS V ; Positive input (VREFP) AVSS + 2.5 V

9.3.1.4 Analog Input . Assuming PGA = 1, the differential input (INP – INN) can span between –VREF to VREF. The absolute range for INP and INN must be between AVSS – 0.3 V and AVDD + 0.3 V.

9.3.1.5.1 Input Common-Mode Range

The usable input common-mode range of the front end depends on various parameters, including the maximum differential input signal, supply voltage, PGA gain, and more. This range is described in Equation 2:

and there are more.

-----------------------------------------------

With above being stated, any thing(e.g. bounds, limits, ranges) that depend on AVSS and/or AVDD-AVSS and/or AVDD will change respectively.

Thanks.