test points

[Alex Eisenhut]

I usually put a 0402 100R resistor in series for voltage rail test points. My reason is that the resistor will blow in case of a short and it can be repaired easily.

For me, test points on rails are just DC checks. If you want to measure noise, you have to go directly to the chip's pins.

But some people flip out and want "direct" test points on the planes.

Who's right?

[SeanB]

The sacrificial resistor is a good idea, and will do very little to DC and low frequencies measurement wise. You might have a problem at 1MHz and above, but for DC and audio it will barely matter.

[Precipice]

I can see both points - but some places run rigid test schemes, where every node needs a test point, and your resistors, while sane, don't do the necessary - in fact, they add test points.
Are you sure that the test regime doesn't do a quick squawk between all nodes designated power before starting, to rule out power-ground and power-power shorts? With hundreds of tiny decouplers on even a small board, it's a reasonably common failure mode.

My short answer: do what makes your test house happy (as long as it's not truly insane). They've got a workflow which (one hopes) works for them...

[David Hess]

Tektronix sometimes did this with resistors to isolate sensitive test nodes from excessive lead capacitance and with high voltage capacitors to act as DC blocks when only ripple and noise would be measured.

High power ballast resistors are often designed or intended to act as fusible links so using a low value resistor to protect a low impedance test point from a catastrophic short just seems like common sense to me.

[Richard Head]

I think the urge to measure the actual DC rail (before the resistor) will cause many people to measure the rail directly anyway.
I would get rid of the resistor.

[tautech]

I am repairing a mid 70's FG PSU that has dedicated test pads on the PCB linked to the + & - rails with 330 Ohm resistors. Never seen this format of test point before, although content in this thread makes me think, why? Circuit protection no doubt, from some aspiring EE with his DMM set on Amps. 

[Neilm]

I can see both points - but some places run rigid test schemes, where every node needs a test point

I work in such a place. Trouble with test points is they take up room and (if the space is tight) can be on a reasonably long stub. When trying to meet EMC regs, if there is any real frequency to the signal then you suddenly have a nice aerial leading to EMI problems.