I recently acquired an EEVblog uCurrent Gold and have been using it to test= nA currents. This low-cost device is more accurate than my US$1k+ Fluke 88= 45A benchtop meter when it comes to measuring currents smaller than 1uA.=20 While reflecting on the "weak pull-up" dialog we exchanged back in March/Ap= ril this year, I decided today to check the leakage current of input pins o= n a PIC16F1508. Measured leakage current was within the datasheet specific= ations, but it varied more than I thought across my 4 pull-up values. Even= though the datasheet specifies 125nA as the max, it also specifies 5nA is = being "typical." I only got close to "typical" when I used a 1M-ohm pull-u= p! Here are my measurements using the uCurrent Gold and Fluke 8845A: Pull-up Resistance to 3.00V: 0.994M-ohm Current: 2.28nA (Equivalent Input Pin Internal Resistance =3D 1.3G-ohm) Pull-up Resistance to 3.00V: 99.5k-ohm Current: 16.6nA (Equivalent Input Pin Internal Resistance =3D 181M-ohm) Pull-up Resistance to 3.00V: 9.66k-ohm Current: 43.7nA (Equivalent Input Pin Internal Resistance =3D 68.6M-ohm) Pull-up Resistance to 3.00V: 1.00k-ohm Current: 51.8nA (Equivalent Input Pin Internal Resistance =3D 57.9M-ohm) NOTE: Resistor values measured with the 8845A set to maximum 6.5-digits, 10= 0PLC, using the 4-wire probe. Current measurements were taken with the uCu= rrent Gold's output connected to the Fluke 8845A in Volts mode, 6.5-digits,= 100PLC precision. PIC DATASHEET (see p319): http://ww1.microchip.com/downloads/en/DeviceDoc/4= 0001609E.pdf Best, James Wages 2017/03/30 1:00 AM, "Stephen" <>: > 100nA input leakage is within specified range if a bit=20 > on the high side (see param. D060 in the datasheet)=20 > Note that even with (maybe especially with) high=20 > accuracy meters, you are in a measurement range=20 > where every little error becomes significant.=20 > > The datasheet says 125 nA worst case at 85 degC die=20 > temperature). If you design to that you can expect it=20 > work. If this is for DIP switch pull up on I/O's that=20 > don't go off-board and have short traces with well- > defined ground then noise should not be an issue... Tue, 28 Mar 2017, Electron : > Of course it is, it's the gate of FETs, not the base of bipolar transisto= rs. > Theoretically it would be "infinite" impedence, of course it cannot be, b= ut > it's VERY high nonetheless. Of course, there's still capacitance that mus= t > not be forgotten. --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .