[TowerTalk] synchro to RS232

[jimlux]

On 11/19/18 1:41 PM, Al Kozakiewicz wrote:
> The defeat of the Nazi's assisted by this analog technology has almost passed completely from living memory.  Rotary encoders are a fraction of the price and already signal compatible with digital systems.
> 
> Or maybe I am missing something?
> 
> Al


Synchros (and these days, resolvers) are still widely used as sensors - 
they're really rugged, inherently a balanced ratiometric device, so wire 
length isn't a big deal, and its all twisted pairs (or triples) so noise 
pickup is less of an issue.  They laugh at ESD or transients - it's a 
transformer - unless the discharge is big enough to destroy the winding, 
it still works - you can punch holes in the insulation with HV 
transients all day.  They are also inherently dust/moisture/liquid 
insensitive - the accuracy is more about the quality of the shaft 
bearings and how precisely they can locate the rotor within the stator 
windings.

You can easily transformer isolate them for galvanic isolation (a real 
issue with long sensor wire runs - galvanic isolation is good)

Unlike a quadrature encoder, they're an absolute position sensor - no 
need to "find home" and count pulses.   Yes, there are absolute rotary 
encoders, but they don't have 16 bit accuracy, without a geared scheme 
and two encoders.   16bit accuracy is achievable off the shelf with a 
resolver at moderate cost (a few kilobucks, brand new, for resolver+chip 
to turn it into a digital number).

You'll also see the linear equivalent called a LVDT (Linear Variable 
Differential Transformer)- same basic idea, a transformer made with a 
stator with multiple windings, and  a slider that has an excitation 
winding.  Used in the same sort of hostile industrial environments.  You 
don't need three phases (or quadrature) for an LVDT, because the motion 
is constrained - no need for "unwrapping"