At 09:21 AM 2/16/00 +1000, Dean wrote: >I have just completed a PIC based temperature data logger using a NTC >thermistor as a sensor. A PC interface program contains a regression >routine to correct for the non-linear response of the NTC thermistor. I am >aiming to calibrate the device for better than +/- 0.02 C. The zero point >should be no problem. Has anyone experience with variations from 0 C due to >differences in purified water quality (e.g. standard deionised, distilled >or ultrapure)? The freezing point of water varies little with reasonable quality. Even tap water should be within 0.01 or 0.02 of 0.00 C at freezing. Your problem is going to be getting a proper ice bath to the exact temperature. Use shaved ice and just enough water to fill the ice crevices. >The device is only required to measure temperatures between 0 and 50 C. At >100 C, there are too few TMR0 derived counts to use this temperature for >calibration and I think it risky to depend on the accuracy of a >mercury-in-glass thermometer. Does anyone know of a high scale temperature >reference around 50 +/- 0.02 C? The biggest error you are going to have is self-heating effects in the thermistor, which will be about 0.1 C. You can calibrate them out at a single temperature, but you can't fix them over the whole range to this accuracy. Measuring temperature to +/- 0.02 C over a reasonable range of temperatures is a research-grade metrology problem. What possible reason could you have for requiring absolute accuracy to this level? ================================================================ Robert A. LaBudde, PhD, PAS, Dpl. ACAFS e-mail: ral@lcfltd.com Least Cost Formulations, Ltd. URL: http://lcfltd.com/ 824 Timberlake Drive Tel: 757-467-0954 Virginia Beach, VA 23464-3239 Fax: 757-467-2947 "Vere scire est per causae scire" ================================================================