1
\$\begingroup\$

I'm designing a resistance measurement device using an LT3092 current source. I've used a DIP switch to select different current ranges, which switches in different sense resistors.

The LT3092 uses a 10uA reference to set the output current, and the datasheet recommends a guard ring (tied to the output pin) around the set pin, on all board layers. This gives me two issues:

  1. The DIP switch means that there's more traces than usual connected to the set pin, meaning the guard ring has to be large.

  2. The set pin (5) and power in (+24V, pins 7,8) are next to each other, separated by an NC pin (6). This leaves the power pins inside the guard ring, which seems less than ideal.

I've attached screenshots of the PCB layout and render (for context). What's the best way to proceed here? Thanks!

enter image description here

enter image description here

\$\endgroup\$
6
  • \$\begingroup\$ @Bimpelrekkie - the reason for the guard ring is local leakage currents. 10nA of leakage creates a 0.1% reference error. I would try and keep all the set resistors as close as possible to the set pin so it can be guarded more easily. That would require moving things about quite a bit. \$\endgroup\$ Jun 13, 2019 at 9:58
  • \$\begingroup\$ @Bimpelrekkie it's on its own PCB, but it's a piece of measurement equipment so will be sitting on the end of a long cable and possibly in noisy environments. \$\endgroup\$ Jun 13, 2019 at 10:01
  • \$\begingroup\$ @PeterSmith would that all be for nothing with the proximity of the 24V pins? \$\endgroup\$ Jun 13, 2019 at 10:02
  • \$\begingroup\$ @PeterSmith OK, I didn't read that part of the datasheet. OP: Instead of the DIP switches maybe you could use jumpers (these have a smaller footprint). 10 uA in a noisy environment might be asking for issues. Can the PCB be placed in a metal (shielded) enclosure? If possible I would use a chip with a much larger reference current. \$\endgroup\$ Jun 13, 2019 at 10:09
  • \$\begingroup\$ @JoeofLoath This part is used in the DC2132A demonstration unit - look at the design files to see how (then) Linear Tech did the guard ring. analog.com/en/design-center/evaluation-hardware-and-software/… \$\endgroup\$ Jun 13, 2019 at 12:36

1 Answer 1

1
\$\begingroup\$

The idea with guard rings is to prevent leakage current. FR4 has a volume resistivity (varies from manufacturer) as low as 10^14Ω cm. What you have is most likely sufficient for the leakage current you may have, but since it's easier to run traces than to calculate the leakage, I'd add another one around the pins of the dip switch.

Make sure you pull the ground plane underneath the fence and add a guard trace underneath also in the layer below the top layer. (while maintaining a proper ground to the LT3092)

enter image description here

Minimize board leakage by encircling the SET pin and circuitry with a guard ring operated at a potential close to itself; tie the guard ring to the OUT pin.

This is what I would do (because the pin 6 is an NC pin and not a DNC pin):

enter image description here

\$\endgroup\$

Your Answer

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

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