EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: TimNJ on September 29, 2017, 10:33:01 pm
-
Hi all,
I scooped a Keithley 179A on EBay for about $40. The seller characterized it with a Fluke 382A and it was pretty good but not spot on.
Since the calibration procedure is completely analog, you don't need fancy calibration gear to produce the reference levels and make adjustments. The manual recommends 190mV, 1.9V and 19V (as well as 1kV but who cares) for the DCV ranges. Looks like the DCV adjustment also applies to DCA, but not 100% certain. Additionally, it recommends a 1.9k and 190k resistor for resistance calibration.
Here's my thoughts on how to get reasonably good reference voltages/resistances:
+10V 0.003% Reference (such as http://www.voltagestandard.com/-.html (http://www.voltagestandard.com/-.html))
+0.01% resistors (<5ppm):
100 ohm, 500 ohm,1K, 5K, 100K
- Use 10V (instead of 19V) to test the 20V range.Might be good enough?
- Produce 1.6666V (instead of 1.9V) using 100ohm/500ohm resistor divider
- Produce 0.19607 (instead of 0.19V) using 100ohm/5K resistor divider
- Use 1K (instead of 1.9K) to test 2Kohm range
- Use 100K (instead of 190K) to test 200Kohm range
Now, none of the reference levels are exactly what the 179A manual recommends. Is it a bad idea to not perform the calibration at the top end of a range?
Also, we have to be careful of the loading effects of the multimeter's input (10Meg) impedance in parallel with the resistive divider. For the 100R/5K combination, I calculated a 0.00980% uncertainty. For the 100R/500R, I calculated 0.008333% uncertainty. By the way, should I calibrate the meter to the calculated (unloaded) voltage or to the adjusted (loaded) voltage? The 10Meg resistor is 0.1%.
Additionally, the 10V reference has a 10ppm/mA (0.001%/mA) load regulation spec. The 100R/5K divider draws 0.196mA (0.000196%). The 100R/500R divider draws 1.66mA (0.0166%).
I finally calculated the total uncertainty as %loading + %load-regulation +%10V accuracy +%resistors. I got 0.0329% uncertainty for the 100R/500R (1.666V) divider and (interestingly) also 0.0329% uncertainty for the 100R/5K (0.19607V) divider. (I counted each 0.01% resistor for a total of 0.02% uncertainty; Is that correct?)
The basic DC accuracy of the meter is 0.04%. Ideally, we'd like to be better than 0.01% total accuracy (4:1 accuracy)...so maybe this wouldn't be the soundest idea. Thoughts? Does my math look sound?
Anyway, I'm sorry for such a long post; I just wanted to be thorough.
Thanks so much!