Hi Maithil,
We did some temperature testing, and the AD7699 went along for the ride. In other words, we have the temp of the chamber, and the temp of another component on the board, but not the actual temp of the AD7699. But the results are still very interesting.
The product has two mezzanine boards, with three AD7699 on each, Ux49 is near the edge of the boards.
[ U149 U148 U147 ..... ] [ U249 U248 U247 .... ] with air flow <--->
Here is chamber temp vs temp sensor voltage, using internal reference, all 8 inputs and sensor sequenced.
Here are linear fits of the lines, using chamber temperature.
| Device | mV/C | mV@0C | mV@25C |
| U147 | 1.39 | 366 | 401 |
| U148 | 1.75 | 373 | 417 |
| U149 | 0.70 | 323 | 340 |
| Device | mV/C | mV@0C | mV@25C |
| U247 | 1.67 | 349 | 390 |
| U248 | 1.72 | 358 | 401 |
| U249 | 0.72 | 317 | 335 |
Here are linear fits of the lines, using actual temp of a nearby component.
| Device | mV/C | mV@0C | mV@25C |
| U147 | 1.29 | 348 | 381 |
| U148 | 1.63 | 350 | 390 |
| U149 | 0.66 | 313 | 329 |
| Device | mV/C | mV@0C | mV@25C |
| U247 | 1.56 | 335 | 374 |
| U248 | 1.61 | 345 | 385 |
| U249 | 0.68 | 311 | 328 |
I can't begin to guess why Ux49 has a different slope from the others, or why the 25C voltages are so far from the data sheet value of 283 mV. The parts are not warm to the touch (when operated at room temp), so temperature can't explain that large difference. I would be curious if you can explain this better.
In any case, I now see that the device is not useful as a temperature monitor unless individually calibrated in the target system. I would respectfully suggest that the data sheet should delete the "typical" values, and have a footnote explaining this.
Regards,
Frank