In the MAX31856, if a series resistance is used to reduce noise from the thermocouple (as in Figure 8 of datasheet), can the BIAS output be connected directly to T-?

No, the BIAS should be connected on the thermocouple side of the resistor as it is used to bias the thermocouple.

There is a OVUV Fault flagged in the MAX31856. Could Open Circuit Detection be engaged to pull the circuit out of the Fault state?

No, this would not be an option. When an OVUV Fault is present in the MAX31856 all conversions stop. The presence of the OVUV fault will suspend conversions and the ability of the MAX31856 to detect other faults until the fault is no longer present.

How do I interpret the full scale error of the MAX31856?

The internal ADC has a full scale input voltage of 625mV. The PGA (programmable gain amplifier) in front of the ADC has configurable gains of 8x or 32x, which reduce the input voltage range to 78.125mV (625mV/8) or 19.531mV (625mV/32). These full scale voltage ranges only come in to play if using an unsupported thermocouple type. If using a B, E, J, K, N, R, S, or T type thermocouple, linearization and cold junction compensation should happen automatically based on internal look-up tables. Looking at Configuration Register 1, the TC TYPE[3:0] bit field is used to select either a supported thermocouple type, or a voltage mode conversion with specific gain value. The full scale and INL error is reported based on applying a 78.125mV or 19.531mV signal in voltage mode and observing the change in output while sweeping the temperature. The linearization error specification for each supported thermocouple type should be viewed on page 4 of the datasheet for table data, and pages 8/9 for graph data (for various cold-junction temperatures).

MAX31856 have voltage mode and code formula mentioned on datasheet P20.Could you please explain code formula theory?What is 1.6 and 2^17 mean of this formula?

MAX31856 input voltage range is +/-78.125mV for GAIN=8, the input voltage amplifes 8 and then input to ADC, so the ADC input voltage range is +/-78.125*8, Let's only consider positeve section, the resolution is 78.125*8/2^18(As it 19 bits resolution and with bipolar input), so, the output code at input voltage is Vin should be Vin*8/Resolution(Vin is input at MAX31856, the voltage to ADC is 8*Vin). that is: code=Vin*8*2^18*/(0.078125*8)=Vin*8*2^18*1.6.

Using the MAX31856 and have a OVUV Fault which is still there even when the fault is removed. What is the issue?

Please check the Fault bit in the configuration register. If it is set to 1, the Fault would be locked in once the overvoltage condition was sensed. A 1 needs to be sent to the Clear Fault status register in order to remove it.

What is causing the MAX31856 to experience large voltages/oscillations?

In some applications, the MAX31856 can experience large voltages on the T+ and T- pins due to interfering sources. Ceramic heating elements or coils are known to cause large amounts of radiated EMI at high temperatures, causing the OverVoltage or UnderVoltage protections to kick in. A more common source of noise is the AC Mains (50/60Hz) being picked up by the thermocouple cable. The oscillations can be seen on either T+, T- or both. This can be suppressed most of the time using a shielded thermocouple cable. Adding a ferrite bead can also suppress much of the noise experienced as the ferrite bead can act as a low pass filter with a capacitor connected to ground. If the issue persists, a TVS diode can clamp the higher voltages.

What could be causing erroneous faults conditions and random temperature shifts in the MAX31856?

If you have multiple devices setup on the same SPI bus then it is possible that there are collisions occurring on the communication path between consecutive readings from the micro-controller. Make sure that you include a sufficient delay between readings to avoid corrupting the data. Other sources of this kind of error can be interference with the thermocouple or weak connections at the cold junction.

To lower system operation current, the MAX31856 is designated as single conversion during the design phase (ON Duty 1/60). However, for most components the standby currents are between 10uA -- 500uA. The data sheet mentions that the standby current is 5.25

The component can be configured to Standby mode by designating CMODE=0 and 1SHOT=1. If the sampling period is 1/60 for a 50Hz filter setup, the sampling duration is 185ms. Therefore the required time gap between sampling is 185ms*59=10.9s, which means a single conversion is initiated per 11 second. The measurement of standby current should be conducted after the component completes conversion. The other method for measuring standby current is to directly measure the power supply's current when the component is connected to power. Do not setup any register at the moment. The component has to be in standby mode at this time.

What is the purpose of the MAX31856EVKIT 150 ohm series resistors on the digital lines?

These are series termination resistors which are recommended if the SPI communication lines are long. This means putting 100R resistors as close to the driving source as possible. Thus we need to put resistors in CLK, MOSI and CS\ lines close to the host. At the slave we would use a resistor in MISO. These resistors series terminate the lines to reduce or eliminate undershoot and overshoot at the receiving end. Care must be taken with multiple SPI slaves though, series terminating is best with point to point connections.

MAX31856 Open Circuit Fault Detection not detected upon startup with no thermocouple attached

Example Issue: When the MAX31856 is powered on without a thermocouple connected, the MAX31856 does not consistently identify an open circuit fault. Temperature measurement conversions result in a temperature of exaclty 0C. If we then connect a thermocouple to the connector and disconnect it, the MAX31856 correctly identifies the open circuit. Solution: It is likely this is actually due to an Over-Voltage/Under-Voltage (OVUV) error. A conversion result of exactly 0C indicates conversions are not occuring. In auto-conversion mode, conversions happen every 100ms and open-circuit detections in this mode occur every 16 conversions. If there are no conversions happening (due to this OVUV error), then there would be no automatic open-circuit detection. With the TC pins floating, it is possible that the potentials detected on T+ and T- may have floated above VCC or below GND. Placing large (Mohms) resistors from the T terminals to GND provides a very weak reference path to GND to prevent those nodes from floating out of range.

Can I interface InfraRed Thermocouple to MAX31856?

MAX31856 is not compatible with Infra Red thermocouple.

Does the /CS pin on the MAX31856 have an internal pull up? Can it be left floating?

SPI devices use a push-pull configuration and no pull up resistors are required. You should not leave the /CS pins floating if using a device like a multiplexor to choose different devices. For this case, the unused devices /CS pin should be high and the used devices /CS pin should be low.

Does the MAX31856 use a sample and hold at the input?

No, MAX31856 does not have sample and hold at the input as the ADC in MAX31856 is Sigma Delta ADC.

How to deal with 50Hz noise for Thermocouple?

In thermocouple measurement design, 50/60Hz noise impact the system measurement heavily. Even tried internal 50/60Hz filter in MAX31856 and other company's devices, the performance is not good enough. According to this situation, we recommend a glitch filte circuit which based on MAX4230. It has around 44dB attenuation for 50Hz. After adjust parameters, it can be used for 60Hz also. This filter has been verified and got good system performance.

Can I connect BIAS to TC+ instead of TC- on MAX31856, will this affect accuracy?

Connecting BIAS to TC+ won't make a big difference in accuracy as long as it is within the part's voltage ratings. If TC- is connected to BIAS, then TC+ will be at V BIAS + V difference . If TC+ is connected to BIAS, then TC- will be at V BIAS - V difference . The voltage of TC+ and TC- will be shifted around depending on which input you bias, but the voltage difference will remain the same.