IoT#30 RPTC-Q LTC2984 Initial Power Up and Repair

In today’s video, we are going to fire-up our LTC2984 Multi-Sensor High Accuracy Digital Temperature Measurement System with EEPROM, setup the SPI, and see what kind of temperature readings we can get out of a PT-100 RTD, a PT-1000 RTD and an off-chip “diode”; and I use diode in quote marks. Because this video got to be over an hour long, I went ahead and have split the SPI code deep dive into the next video, #31. It took me over 2 days to get this chip up and running; ugh! I’ll show you why!
To show your support for the Muxall Open IoT Video Blog donate with credit card or Paypal: www.paypal.com...
Buy Muxall Products: shop.muxall.com
Muxall LLC is an Engineering company that offers: Network Consulting, Automation Development, and Training Videos. Muxall produces: The Movement Clock, Automated Turntable, Network Engineering Video Blog, and Muxall Open IoT Video Blog. Call or write us today!
Michael Crane
Muxall LLC
Dallas, Texas, USA
Phone: [phone]
Email: i n f o (at) m u x a l l . c o m
Web: www.muxall.com
Muxall Stuff Website: mx3.muxall.com
Network Engineering Video Blog: / networkengineeringvide...
Muxall Open IoT Video Blog: / @muxallopeniot9194
Auto-Subscribe
Muxall Open IoT Video Blog: / @muxallopeniot9194

[@qwerty1234158]

Bit of a concern that there is a 1 degree difference between the 2 probes. At 25 degrees that is a 4% difference. A future video showing how accurate the 2 probes are compared to your fluke temperature probe as the temperature ramps up from room temperature to BBQ temperatures.

[@NetworkEngineeringVideoBlog]

Yes, 1 degree C is an order of magnitude out of spec. Here are some possible causes; some I can fix, some I cannot: If you recall from the video, one probe is going through input protection (RC filter) and the other probe I removed the input protection and soldered in some crude jumpers. There may be a board layout issue since some of the traces run almost across the length of the PCB. According to the LTC2984 spec, the chip may need strain relief for the voltage reference if positioned in the middle of the PCB (Pg 64 Reference Considerations). And, there are most certainly errors inherent in the RTD probes, wires, and connectors. My Fluke 87V has a 1% + 10 count accuracy (not including error in the thermocouple) with a 0.1 C resolution. I can certainly use it for verification but the specs on the LTC2984 are better than the 87V. So, when we start measuring tenths of a degree C, which one do we trust? A wise man once told me: “A man with two watches never knows what time it is.”

[@qwerty1234158]

@@NetworkEngineeringVideoBlog The 3 wire RTD is designed to compensate for differences in the wires connected to the probe. The circuit uses the fact that 2 of the wires are terminated at the same side of the probe to adjust for issues with the wires. Agree with the Fluke's accuracy of 1% at 25 degrees that means you are +- .25 degrees and if the LTC2984 is more accurate I would expect them to read within the Flukes accuracy. Also since the LTC2984 supports so many different probe types you can test it using the Flukes thermocouple. But for a BBQ grill, how accurate does the temperature measurement really need to be?

[@muxallopeniot9194]

@@qwerty1234158 Yep, you are dead-on about the 3-wire RTD. For a true Kelvin resistance measurement, we need a 4-wire RTD. For a BBQ, the accuracy is already good enough. But, I will make it better. I will also make the reading faster for the meat probe so it can be used as an instant read thermometer.