ADS1298R: How to make the defibrillator protection circuits for the respiration inputs of ADS1298R?

Hello Frank

To answer your question about defib protection on the respiration inputs,

Remember that the ECG and Resp inputs both share the connection to the patient electrode inputs.

So if you add a high voltage clamp right at the patient electrode inputs, you will protect the ECG and the Resp circuits.

The best way to show you is using the schematic for the TI development board as a reference.

You can download the ADS1298R development board users manual from the TI website,

On page 1 of the development board schematic, J1 is connected to the Patient Electrode wires.

At each input of J1, add a 1Kohm 1Watt resistor in series with the patient electrode wire.

Now on the board add a SMBJ14 bidirectional diode from each input pin of J1 to ground.

(Using Bidirectional diodes to protect in both discharge polarities.)

The SMBJ14 diodes will limit the input voltage at the board to about 25V, and the 1Kohm resistors withstand the defib energy.

(I have found that large carbon comp resistors work the best.)

These 2 parts added to each input will protect the entire board from any input voltage going above or below 25V.

The existing input circuits on page 1 of the development board will then limit the input voltage more to protect the ADS1298R.

I hope you find this information helpful and can apply the methods to your project.

Bill Burris

Hello Frank!
I can answer your first question; Yes, I think you get it, you are asking good questions.
1. Yes, the schematic changes you made are almost correct. You need to add a BAT99T to the junction of C18 and C19 to protect the Resp circuits.
2. When the defib 5KV discharge is applied there is a very short period of time before the SMBJ14 diode conducts. This transient high voltage spike can arch over a 401 resistor body. I suggest using at least a 0603 resistor and make the pads as far apart as possible to maximize the gap distance. In addition, C18 should be a 0603 100V part at minimum for the same reasons.
3. The location of the 1K1W resistors must be before the connector, in the wire path. The 5KV defib discharge voltage will arch across the connector pins and shunt energy from the patient. If the 1K1W resistors are at the device end, the wire must withstand 5KV. If the 1K1w resistors are at the patient electrode connection the wires only have to withstand the 25V.
4. The reason for using 1K1W resistors is to assure the device can never shunt more than 10% of the defib energy from the patient, every lead wire should have the 1K1W resistor added. The reason it is only 1Kohm and not higher is for respiration sensing. The resistors are in the patient impedance path, sensing a 1 ohm change in body impedance is challenging with the 1Kohm and would be impossible with higher resistance.
5. For the RL connection I would suggest removing C92, C31, and R38. The SMBJ takes the place of C31, the 1K1W takes the place of R38. Resistor R37 should be the same as the other patient input resistors.
6. C92 is not needed.

And one more note;
The defib energy is at 5KV and about 400 joules. The 1K1W resistors must be able to withstand 10% of that energy which is 5KV at about 40 joules. I would recommend Leaded Carbon Comp resistors, they are the only parts I found that work for this application. Check out Ohmite OX series.
The SMBJ14 diodes can withstand the defib energy, but if you have room on the board, you may want the larger SMCJ14 instead to be safe.

Exciting Engineering!
Bill Burris

Hello Frank

Yes the resistors can be on the board as long as the ECG wires can withstand the 5KV.

Where the wires attach to the board, make sure there is sufficient gap between the connectors and resistors for 5KV. (About 8mm between conductors)

I would not suggest using different input resistor values, CMRR is VERY important in ECG monitoring.

The body resistance does not change much on the patient between any of the leads. Body impedance is usually less than 100ohm.

The issue with mismatched components on an ECG front end is phase delay differences at the inputs.

So keep the components matched as much as possible.

And I suggest testing the CMRR performance to make sure it is good.

A good ECG design goal is -90db CMRR, and the absolute minimum performance must be -60db CMRR.

(Keep in mind, a device that barely passes -60db usually does not work very well in real life.)

To measure the CMRR performance of an ECG monitor, I apply a 1Vp-p sinewave directly to all input snap leads in reference with the RL snap lead.

Then I slowly sweep the input frequency from 0.67Hz to 150Hz and assure there is never more than 1mV of signal on the ECG output display. (1V in and 1mV out is -60db)

I know this test is different than the test described in IEC but if you can pass this test you shouldn't have any problem passing the IEC test.

(This is how an EE should test ECG CMRR performance, just like it is a Diff-Amp.)

Good luck with your design!

Bill Burris