Latency is a measurement of delay in response. Deadband is a measurement of positioning accuracy. You need both to be as low as possible so that the servo goes where you want it,when you want it there. Taking an extreme example, suppose you have a 1μs servo with a 10 second latency. You set your trim resolution to 1 μs on a 2048 resolution radio and bump the trim one time, take a sip of coffee, then observe the reaction as you are setting the cup down. If it's a 10 μs deadband rating and zero latency instead, you can go grow the coffee beans, roast them, grind them, make a pot and it still will not have moved.

The JR demonstration with the pointer was a deadband demonstration, showing how 1 microsecond changes in digital trim movements resulted in a "zero deadband" response (not really, but within 1 μs or less at least). TheJR servo picks up each 1 μs change, whereas a servo with a higher deadband value will not. In other words, if you have 2 or3 μs deadband, you can't accurately resolve a 1 μs change. Anyway, I'm not trying to sell JR servos, but they had the demo video available. I looked around a bit for specs for the Airtronics and Futaba, and didn't find them, but it wasn't an exhastive search. Savox does publish their deadbandspec. The Savox looks like a fine servo, but it's just a question of how much accuracy you want, that's all.

Besides JR, I have used some Futaba and Airtronics digitals, andI have observed that theywill also resolve 1 &mu;s changes. That's all I was trying to say.</p>