High Voltage protection for NI-DAQ

[mattoeding]

I am currently designing a device that will measure the leakage voltage through a patient while undergoing defibrillation. The defibrillation pulse is biphasic(bipolar) and can reach up to 4kV. The DAQ module we have (NI-9215) has an input of +-10V with overvoltage protection of +-30V. I am resistively attenuating the signal down to this range but I will require additional safeguarding incase of a fault in the attenuation.

My questions are as follows:

• Is it possible for the NI-DAQ to safely operate without the use of an isolation amplifier? (See Figure 1)
• Is it possible to use a series of bi-directional varistor’s or transorb’s be used to shunt the voltage?
• Will there be a problem with isolation between channels if isolation amplifiers aren’t used?

See attached for my proposed project.

[johnsold]

Having been a paramedic as well as an engineer, your project is of interest to me.

I think you could operate without isolation amplifiers, but then you have a quite different measurement problem. Without the isolation amplifiers all measurements will be referred to some common point. In your image the rescuer and the environment do not have defined potentials referred to the patient.  So where is the common point? If the rescuer is standing on a wool carpet, static charges of 10s of kV could easily be generated.

For the safety of all personnel and equipment and the integrity of the measurements, I think isolation amplifiers are appropriate.

If you use isolation amplifiers, then the amount of DAQ protection required, if any, depends on how good the isolation amplifiers are.

I prefer a layered approach to protection.  For example I would probably use a string of resistors for the attenautor so that no one resistor ever sees more than about 300 V.  I would probably use at least two stages of attenuation with overvoltage protection after each stage. The first stage output might be about 100 V. Varistors work nicely but do not have sharp knees in their resistance-voltage curves so they must be selected carefully so that their leakage currents do not degrade the measurements, especially with very high series resistances.  Varistors also tend to fail open. So after a few activations they may stop working with no way to detect the loss of protection.  I would not use them in measurement circuits. Transorbs are better.  Zener diodes may be fine for the last stage of the attenuator.

Lynn