EEVblog #607 – Agilent B2912A Source Measure Unit SMU Teardown

What’s inside a $13K Agilent Source Measure Unit capable of 15fA and 100nV resolution?
Plus triaxial cables, and low current measurement connection.
Data Sheet
Connection Guide
Component Datasheets:
OPA1611 Audio Opamp
ISO7240A Isolators
ADS8202 Diff Amp
ADS1675 24bit ADC

Forum HERE

High Resolution Teardown Photos:

About EEVblog

Check Also

EEVblog #1334 – Mystery Dumpster Teardown

Mystery dumpster teardown time! With the most amazing mechanical mains power switch you’ll ever see! ...


  1. Very Nice!
    I have one on my bench waiting for some time to start measuring stuff, and was really nice to see the guts of it! All I had time for was to measure the reverse current of a diode under 200V.
    One thing I learned from that is there is absolutely no correlation whatsoever with the devices that I measured and the datasheet!!!
    A video of it functioning would be really nice, as most people will not see one of these beasts functioning.


  2. Lovely SMU.
    Get it working and shoot usage video about shmuuu 🙂 Since you have dual-channel unit, prolly best would be to trace some curves for diodes, LEDs, zener and transistor 🙂 Or make a curve of your uCurrent, voltage to current dependency or something like that.

  3. Hi,
    two comments on this nice insight video.

    I noticed you mentioned the guard to connect to the internal common. Usually in SMU test systems one would have the guard being driven via a buffer directly by the signal voltage. This way there is absolutely no leakage out of the center conductor towards ground, an absolute must for very high precision low current measurements at elevated voltages. That is also the reason why triaxial cables can be quite dangerous to work with if you don’t know your business… there are a lot of different BNC/Triax adapters and some of them expose guard to the outer BNC shield which can be deadly.

    Being a physics scientist myself, I can attest that SMUs are the backbone of research. We rely on them every day, delivering exactly what is displayed and being very versatile when it comes to device characterization. BUT a lot of scientists are really not good electrical engineers and lack an understanding of such devices even though they use them on a daily basis. Just as an example, there was a recent case in the prestigious journal Nature (see where experimental errors where exposed and led to the admission of false data interpretation. It basically boiled down to scientists starting to use an SMU just like two independent devices, in this case a current source and a voltage meter. The internal voltage meter is designed to measure a potential that is close to the one put out by the current source… basically some internal diodes prevent large potential differences. One wouldn’t expect it, but a trap for young players and senior scientists as well 😉

  4. Agilent is really sneaky. They got you to do an hour of free advertising for them. If they produced an hour long video, how many people would watch it? Would someone they hired have seemed trustworthy? How much would the production have cost? A free device to you is a small price to pay. Keep it up.

    • I don’t get to keep these units, they go back to Agilent.
      Yes, they know the game, and I’m happy to play along. I get high tech fodder for my videos and my own curiosity, and they get advertising and views they can’t buy.

  5. All nice and good.

    I was an EE for 25 years.
    Made good money, had fun making hundreds of circuits and black boxes. But, in the end
    I was a hired hand, and nothing I made will be remembered. I realized this and became an artist. I paint now, oil and acrylics.
    My paintings are all over Minneapolis MN, USA.
    These paintings will out live me, while all the work I did as an engineer will end up in a landfill.
    A cautionary tail.

  6. Hi Dave, very nice! I did not even know that triaxial connectors exists ;-). It would be nice to see how some parts correlate to their datasheets.

  7. Ooh, it’s the Agilent P600 CPU platform development board!

    I actually worked on the software part for that board.

    If it sounds familiar, it’s probably related to the CPU in those Agilent 2000 and 3000 series oscilloscopes, except those weren’t separate boards.

    Heck, even the boot screen is familiar, except the LCD is upside-down (the bar going across the top is the progress bar for Windows CE bootloader startup – it is supposed to be on the bottom running across left-to-right). Obviously the screen is upside down and they’ve just rotated the Agilent logo for this. Hence why the bar is on the top and runs right-to-left.

    Some things never change.

  8. Just wish to say your article is as astounding. The clearness in your submit is simply excellent and that i can assume you’re a professional in this subject. Well along with your permission allow me to seize your RSS feed to keep updated with approaching post. Thank you one million and please keep up the rewarding work.| а

Leave a Reply

Your email address will not be published. Required fields are marked *