Bench Digital Multi Meter

[Conrad Hoffman]

The GR1650B is plus or minus 1%, confirmed in a GR catalog. Good enough? Depends on what you're doing. I'm usually matching parts for two channels, so the absolute accuracy isn't that big a deal. The most fussy thing might be an RIAA network- 1% is pretty good, but if you're a fanatic, a better bridge might be called for. OTOH, you might never be able to select components much closer, unless you buy a lot or make up combinations to dial them in. Until you get to the transformer bridges like the 1615, .05% is near the top of the heap without spending a fortune. A good rule is "test it the way you'll use it", so for RF work you might want an RF bridge. There's really no such thing as extreme accuracy with RF, because of strays, and if no core is present, an audio measurement might be good enough. If you're dealing with, say, RF toroids on various cores, it absolutely won't be good enough. Resonate the thing with a suitable capacitor at the frequency of interest and make a manual measurement, or maybe find an old Q-meter. As for the pitfalls of digital bridges (which aren't really bridges at all), it comes down to cost. A good one, like the latest $16k Agilent, or many of the better GR Digibridges, will be stable, accurate, and give you D and Q over a huge frequency range. I have the basic 1 kHz GR1657 Digibridge, which is handy, but the more versatile models are still way beyond my price range. Unfortunately, what most people end up with are the inexpensive hand held digitals that lack D and Q, only test at one frequency and one amplitude and might not even give a choice of the parallel or series model. A cheap used 1650 will let you apply your own signal at the frequency and amplitude of your choice, and addresses the other issues as well. If it breaks, it's easy to fix. It is, however only useful over the audio range. Very few traditional bridges allow RF measurements unless they were specifically designed for RF.

(I can easily adjust a 1650 to spec, but remember that if it hasn't seen a calibration in some decades, or if somebody's messed with it, it might not be to catalog spec.)