Curious about rotary encoder pulses and detents

Study the _relative_phases_ for CW and CCW rotation. A proper decoder has two outputs... one pulse per "detent" and a "0" or a "1" depending on CW or CCW. ...Jim Thompson

Possibly to save power. The pullup/pulldown resistors could be low values but not use current if both switches are open at detents.

Den mandag den 27. november 2017 kl. 21.46.07 UTC+1 skrev Jim Thompson:

coder

per

ough

and

with

and

uld

cess

al

no

a proper encoder has two outputs in quadrature and a simplistic decoder tha t uses one signal as clock and the other as direction rather that decoding quadrature is why sometimes when you turn a knob it jumps in counts and som etimes i the wrong direction if you don't turn it just right

A good decode algorithm always gets it right. It can dither one count at the transitions, as the switch bounces... that's all.

My Tek DPO2024 scope misses about 10% of the knob detents, just ignores them. Three clicks clockwise then three CCW should get you back where you started, but often doesn't.

Piece-o-junk. I'll get a Rigol next.

Rigol has a special on now. Buy the 4-ch 350 MHz scope, and get the upgrade to 500 MHz and some serial decode stuff for free.

Seems like a sort of always-on special maybe.

It's a great application for a simple lookup table.

The approach I've used takes as input the two quadrature bits from the current sampling, the two from the previous sampling, and one bit of stored state ("I was going up" or "I was going down"). The output is one new bit of state to store, and a three-way "increment, decrement, no change" delta for the accumulator.

With this approach it can dither back and forth across a quadrature boundary (as frequently as "every sampling time"), without the accumulator value changing. Only when there are two successive quadrature transitions in the same direction will the accumulator change.

Easy to do in even a very simple low-speed microcontroller, or in a tiny little PLD if anybody's still making those.

Yes, but I was asking whether there is ever a reason to select an encoder which has different pulses and detents per revolution. So far, nobody has mentioned one.

These look cool:

Well if it's any consolation to you, I have the same problem with my DSO150 Shell scope kit. I figure - if you can get bad encoder servicing routines for $21, why pay hundreds of dollars more for the same thing. :-)

I would hope so; for the scopes without digital channels the price jump from the 2 channel 350MHz model to the 4 channel is wacky.

The DS4032 seems to be the price/performance sweet spot

I kinda like decoding it by ... feeding a stepper motor. An old-school mouse or trackball makes a good remote control for a variety of gotta-make-adjustment gizmos. Just amplify the outputs and it drives the motor.

You can usually get the 4-channel 500 MHz scope for very close to $4K. I just ordered one.