1
\$\begingroup\$
  1. What are your personal policies for selecting which nodes should have testpoints?

  2. Can a testpoint be used as a node for powering a circuit?

  3. Does a testpoint on a line create a stub, subject to interference or crosstalk, making it better to not have a testpoint in some situations?

  4. Do you put testpoint on I2C and UART RX/TX lines for debugging?

\$\endgroup\$
  • \$\begingroup\$ 1. That's a personal policy for a reason. 2. Yes, this happens regularly in bed of nails and other in-circuit testing. 3. It can, it may mean you need a smaller test point on top of the signal trace. 4. In designs that may need it, yes. \$\endgroup\$ – Asmyldof Jun 6 '16 at 7:04
  • 1
    \$\begingroup\$ There are no general rules, you just have to do what seems the most appropriate in your case. However, in addition to test points, one thing often overlooked, but useful for test boards (not production), is to place straps (or 0ohm resistors) on the supply traces of the various functional blocks on your PCB, so you can much more easily check for the consumption of each part, if necessary. \$\endgroup\$ – dim lost faith in SE Jun 6 '16 at 8:45
  • \$\begingroup\$ @dim good points \$\endgroup\$ – chris Jun 6 '16 at 10:48
2
\$\begingroup\$
  1. ground should always be a test point, power supplies are good test points too, because if the power is bad that can cause unusual behaviours. other signals that pass between circuit blocks make good test points too.

  2. some circuits can be powered by the test points

  3. stubs, yes at several gigahertz, not usually a problem except on fast circuits like PCIe

  4. yes, especially if those lines do not emerge on external connectors.

|improve this answer|||||
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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