My Toolbox — Oscilloscope
Why do I have an oscilloscope?Â The real answer is that scopes are cool, and many years ago, I couldn’t resist when Fry’s had Fluke Scopemeters on sale — at 50% off!Â So I own a 100MHz dual trace Fluke 196 Scopemeter.Â Later I bought the yellow the Pelican case for it; the two are a great combination.
Since software ties everything together, I’m also the “systems” guy, and thus I should be able to troubleshoot the entire machine.Â I do use my multi-meter much more than my scope, but the oscilloscope is essential when troubleshooting encoders, hall sensors, and the occasional electrical glitch which won’t show up on a DMM.
I’ve used other scopes, such as the Tek THS720 handheld, the Tek TDS 210 desktop, Tek analog scopes, and a Heathkit analog scope.Â Recently, I participated in a hands-on demo of the new Rhode and Schwartz RTO1014 1GHZ oscilloscope.Â It’s quite impressive, even intimidating at first, with all the buttons and things it can do.Â Fortunately, such high powered scopes aren’t needed for typical factory troubleshooting; a dual channel, 20-100MHz real time digital does fine most of the time.
There is uncommon feature that’s really handy: isolated channels (each channel’s ground is isolated from the others).Â With a typical dual channel scope such as the TDS 210, I can only view one differential encoder signal (channel 1 probe on A+ signal, channel 2 on A- signal, display mode set to combine channels).Â But with a dual isolated channels, I can view two differential signals (channel 1 probe on A+, channel 1 ground on A-, channel 2 probe on B+, channel 2 ground on B-).
Hand-held scopes, such as the Fluke 196 and Tek THS 720, typically have isolated channels, along with ruggedness, battery power, small size, and a hefty price tag.Â However, a 4 channel desktop scope is often cheaper than a 2 channel isolated channel scope.
I once spent a lot of extra time because I was too enamored with isolated channels.Â The system had a servo motor with the differential encoder output split, with one set of signals going to a motion controller and the other set going to a custom board.Â The motion controller was counting correctly, but the custom board wasn’t.
I set up my Fluke 196 to look at the differential signals.Â The B+/B- signal did look smaller (less voltage swing) than the A+/A- signal, but it had to be OK because the motion controller was picking it up.Â So I started swapping out components — and kept getting the same problem.
I figured it out when I looked at each signal (A+, A-, B+, B-) individually; the B- signal was dead (because of a broken wire in the cable from the motor).Â The motion controller worked because it was less sensitive, managing to count the smaller transitions.Â My lesson learned: check everything, and don’t make unsupported assumptions.