Disclaimer: I'm not a EE professional (I have a background in physics though), so please bear with me if the question seems banal, obvious or just plain stupid.

We have a Keysight 34465A DMM, which, according to the data-sheet, has DC-I measurement accuracy of 50 pA (typ) at it's lowest range of 1 uA when the input is close to zero. When the device is set to this setting (auto-zero enabled, 10 PLC) and nothing connected, the display fluctuates within a few pA around zero. As soon as a probe cable is connected to the common port (labelled "Input LO"), the display shows a measurement value of between +2.5 and +3.0 nA DC. If it is connected to the current sense port (labelled "Input 3A"), it shows values of around -50 to -150 nA DC! The probe is a simple passive test lead (Keysight 34138A). If two such leads are connected to these two ports, the displayed value remains in a similar range, unless they are shorted, where the measured current is again within a few pA around zero.

To me, i's absolutely puzzling to see any DC current beyond the specified accuracy of the device when "it is not connected to anything". The best I could come up with is an AC current due to EMI that somehow reduces the device's accuracy via the induced noise. However, the values seem very high to me.

So here are my questions: What is the origin of these currents? Is this to be expected or is it a hint to a problem with the device or the workplace? In view of those issues, how do I measure currents with nA accuracy coming from a high-impedance source?

  • \$\begingroup\$ How high is your high impedance? \$\endgroup\$
    – Andy aka
    Commented Jan 14, 2019 at 13:05
  • 2
    \$\begingroup\$ How close are you to the nearest broadcasting tower? \$\endgroup\$ Commented Jan 14, 2019 at 13:13
  • 1
    \$\begingroup\$ how do I measure currents with nA accuracy coming from a high-impedance source? Very, very carefully, asking the sorts of questions you're asking. You may need a battery-powered instrument with shorter leads, or shielded leads. Rectified EMI sounds plausible. A small battery powered I to F oscillator with the output coupled through an opto isolator might give you a 'front end' you can build easily and calibrate to give a reading on a frequency meter, to check other approaches. \$\endgroup\$
    – Neil_UK
    Commented Jan 14, 2019 at 13:45
  • \$\begingroup\$ See this question/answer of just prior days electronics.stackexchange.com/questions/416812/… \$\endgroup\$ Commented Jan 14, 2019 at 14:13
  • 1
    \$\begingroup\$ From using analog scopes in the lab, under fluorescent lights, the 10Meg Ohm probes easily pickup 10 or 20 volts FROM THE AIR. Read the prior comment, where I work thru math on Efield coupling, and predict 50 nanoAmperes of 60Hz into a 1meter wire. Keysight/Agilent/HP operates these precision meters at integer-multiples of the 60Hz power, to best ignore the impinging interference. For clean measurements, you'll need shielded (perhaps DOUBLE shielded) wiring. AND your device-under-test may need full shielding. \$\endgroup\$ Commented Jan 14, 2019 at 14:18

1 Answer 1


What is the origin of these currents?: Every cable is an antenna. I think you measure noise (FM radio, cell phone, WIFI, Bluetooth, DECT, ...)

How do I measure currents with nA accuracy? Use shielded cables; shield your DUT; Use a battery as supply for your DUT; Eventually use a shielding enclosure for the whole setup (DUT + DMM); increase the number of samples the DMM uses to build the RMS value;

Maybe this free book helps: "Precision DC Current, Voltage,and Resistance Measurements" https://download.tek.com/document/LowLevelHandbook_7Ed.pdf

  • \$\begingroup\$ Thanks for the tips, and particularly for the book! \$\endgroup\$
    – burnpanck
    Commented Jan 14, 2019 at 15:29
  • \$\begingroup\$ Hi, Can you please edit the answer to clarify exactly what you mean by "Ev. use a shielding enclosure [..]" Does the "Ev." mean evaluate, or even, or eventually, or ... ? Thanks :-) \$\endgroup\$
    – SamGibson
    Commented Jan 14, 2019 at 20:49
  • \$\begingroup\$ When you where able to eliminate all the noise sources and still measure random values (noise) you could eventually try a shielding chamber \$\endgroup\$
    – schwdk
    Commented Jan 15, 2019 at 19:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.