ADS1198 Discussion

Hi Jamie,

The ADS129x family of devices have been used in a variety of medical applications, both clinical and consumer.

Using a 4V reference, you should have no problem reproducing the ECG signal whose max amplitude is anywhere from 1-5mV. Most customers use the internal 2.4V reference and have no problem reproducing the ECG signal. The ECG skin electrode "offset" voltage can be subtracted and you should be able to see the ECG signal riding on top. We have dealt with some applications which AC couple the inputs to remove the DC offset from the electrodes and bias the inputs using other methods.

You should not need a secondary external amplifier to add more gain. That post you were referring to was talking about the ADS1298 for an EEG application in which the signal levels may be 1mV max amplitude.

You may be having communication issues with the ADC if you are getting back all zeroes. Are you using our EVM board or your own board?

Regards,

Tony Calabria

Thanks for the response

Tony Calabria said:

The ADS129x family of devices have been used in a variety of medical applications, both clinical and consumer.

You mention the ADS1298. How about the ADS1198?

Section 4 of AFE for ECG shows a block diagram with an input gain stage. Is this the preferred block diagram for the 16-bit ADS1198?

And section 5 in the same document does not have the input gain stage for the 24-bit ADS1298.

Can I read this to mean that the ADS1198 requires an extra gain stage but the ADS1298 does not.

Even if the gain is set to 12, a typical ECG voltage will only be 5mV * 12 = 60mV max.

Using 2.4V reference would still only use 2.5% of the range. This doesn't make much sense to me? I would think that you would want your signal to use as much of the range as possible and if the expected signal for an ECG is 5mV, then why would there not be more gain stages?

Thanks

Hi Jamie,

The link to AFE for ECG document you are referring to seems to be broken. Could you send me the URL and I will see what you are talking about. Having an additional gain for the 16 bit ADS1198 is dependent on what you want to do and analyze. If you are doing simple R detection, you will not need the additional gain. If you want to do more clinical measurements and analysis, such as ST detection/QRS period detection/etc, you may want an additional gain stage or use the ADS1298. Adding an additional gain stage in going to add noise to the system and as you mentioned, we are dealing with signals in the mV or less. Your design is going to be dependent on what you want to analyze.

You are right, that the ECG measurement over time is only using a fraction of the total range. At a 24 bit level, this is sufficient at the clinical level. Also, remember that you will experience DC baseline drive and electrode offset which need to be accounted for in the total range (assuming you are not AC coupling your inputs). This may be on the level of hundreds of mV and can be digitally filtered in the post processing.

Regards,

Tony Calabria

The link I was referring to is here:

http://www.ti.com/lit/an/sbaa160a/sbaa160a.pdf

 Thanks

Hi,

I have the difficulty of getting an ECG signal from 1198.

We have a product with an ECG built with discrete components and has higher gain.

The ECG skin level is +/- 0.5-2.5 mV.

Here is a non-sense,;  a gain of 12 brings the signal into  30mV range and the  reference voltage  of 2400 mV means:

30/2400=1.12%.

We have a ADC of 16 bit , or even 24 bit resolution on ADS1298 just to get into 30 mV range?

Your first stage opamp  it’s good and having a gain up to 12 makes sense for  removing  the DC component and the noise floor level.

We need a second stage opamp to bring the signal into a decent 16 bit resolution.

Usually for  ECG signal, the first stage opAmp is a differential instrumentation opamp which removes the DC component.

Now,  having the dynamic component of ECG we need the second stage opamp to bring into a range at least 160-300mV

If I inject a high level signal into ADS1198 , I get a nice and clean output.

I compared the sensitivity of ADS1198 against our current ECG product and it’s not even close range.

Do you have any test results regarding a human body used for the component verification, please?

Thank you,

Dorel

here is EC11 standard:

4.2.9.1 Input dynamic range

The device shall be capable of responding to and displaying differential voltages of ±5 mV varying at a rate up to 320

mV/s from a dc offset voltage in the range of .300 mV to +300 mV, when applied to any lead. The time-varying

output signal amplitude shall not change by more than ±10 percent over the specified range of dc offset. Changes in

offset voltages at a maximum rate of 1 mV/s shall not cause discontinuities in the output greater than 30 µV, referred

to the input.

The ECG monitoring equipment also shall be provided with means to indicate that the equipment is inoperable due

to an overload or saturation of any part of the amplifier. The amplitude of a ±5 mV signal shall not be reduced below

50 percent of the initial value at voltages less than ±300 mV, and the indicating means shall be fully operative before

the amplitude of the signal is reduced to 50 percent of the initial value. Absence of a visible trace is an acceptable

means of indication

 300mV/5mV = 60 times gain.

The component specs mention EC11 and EC13, but you are not complying

The gain shall be 60 , not 12

thanks,

Dorel

ADS 1198 specs:

Supports AAMI EC11, EC13, IEC60601-1,

IEC60601-2-27, and IEC60601-2-51 Standards

here are my comments:

According to EC13 the gain shall be 62.5.

If we choose a gain of 6 for the first stage, the second stage opam shall have a gain of 10.4. I am good with 10.

An ECG need to be certfied against these standards. The standards are already mentioned in the components specs.

Using your ADS1199 or ADS1298  as today we are not EC11 or EC13  compliant.

Thanks,

Dorel

Hi Dorel,

     The section of the standard 4.2.9.1 doesn't put any requirement on minimum gain. There are two distinct component to the signal the ECG component (+/-5mV)

and the electrode offset (+/-300mV). The section of the standard is only requring that the system should be able to accurately capture ECG component (+/-5mV)

even if the electrode offset is changed from 0 to +/-300mV. In traditional designs you will remove the DC offset after first stage so you can gain the ECG component further

without saturating the analog front-end.

For a system based on ADS1198/ADS1298 which is DC coupled this section of standard puts a constraint on how much gain can be placed on analog front-end.

Assuming a reference of 2.4V maximum gain that can be used is 6. This will results in +/-1800mV( offset) +/-30mV (ECG) of signal. This is still within the ADC full-scale

and doesn't saturate the ADC.  If we increase the gain to 8 we will saturate the ADC once offset is increased to +/-300mV and we will not be compliant with standard.

In summary this section of standard tries to ensure that the system will capture +/5mV of ECG in presence of large DC offsets (+/-300mV). It in facts puts a requirement on maximum gain that can be put on ECG front-end before DC is removed.

Another specification that is needed from this section of the standard is the input referred LSB size. The standard requires that discontinuity in the output is less than 30uV (referred to input). For a 2.4V reference the LSB size referred to input for ADS1198 is 12.2uV which is again in compliance with this part of standard.

With regards,

amit

Hi Jamie,

   The ADS1198 / ADS1298 were intended to be DC coupled systems. This means that DC electrode potential

is not removed in analog domain. The ADC sees both the ECG signal (+/-5mV) as well as the electrode offset volatge (+/-300mV).

Unless the DC offset is removed (for example by AC coupling the inputs) the gain in the signal chain has to be limited to prevent saturation.

For your example the minimum signal you have is 0.5mV.  If you drop the reference to 2.4V and use a gain of 6 the quantization error referred to input for ADS1198

is only 12.2uV which is still 40 times smaller than your minimum signal. You should be able to see a decent signal at output if you give a 500uV signal as input.

If electrode offset is not a concern in your system (or if you can AC couple the input) you can use external smaller reference (say 1.2V) with the highest gain of 12. this will reduce the quantization noise referred to input for ADS1198 to just 3.05uV

With regards,

amit

I'm trying to use the ADS1198 in AC coupling approach, because I need best resolution (with gain of 6 I have 12.2uV of LSB and I want 6.1uV).
Someone could show me a schematic showing this setting to take a 12-lead ECG, (RA, LL, LA, and V1 to V6)
Thank you very much