What about to use some 74hct14 between the encoders and the '245? It
could reshape the pulses nicely and the '245 would not receive "analog"
intermediate signals...  '14 comes in 6 inverters per package, what
could serve up to 3 encoders.

Some optical encoders use a simple diode+photodiode as the movement
sensing element, and the photodiode doesn't need to be necessarily a
sharp and fast switching unit... in real it switches slowly as the
movement wheel starts to allow the light from the emitter to reach the
sensor, it goes increasing in lumnes, what causes the photo-diode to
make a curved slope and not a sharp one...  the '14 could solve it.

Probably if you turn the encoders in high speed, with your actual
configuration, the '245 would work since the photo diode transitions
would be faster.

Wagner

Richard Ramsey wrote:
>
> Howdy,
>
>     I got a problem. Let me give you some backgorund. I circuit with 4
> encoders tied to a pic84 and and parallel to a 74F245 which is tied to a
> pic 877. The '84 monitors the encoders and interrupts the 877 on change.
> The '877 enables the '245 and reads the encoders.
>
>                      encoders
>                       | | | | | | | |
>                     pic84 + 74f245 <--\
>                         |         | | | | | | | |   |
>                          \--->  pic877 --/
>
> I noticed that the '245 goes into a oscillation on the inputs and
> outputs when the encoders are in transition. The '245 ocisalltion causes
> the '84 to see phantom changes. This problem is more pronounced the
> slower the encoders move. If I remove the '245 the '84 can read the
> encoders with no problem. I also added a 74F373 in between the encoders
> and the '84 and '245 thinking that the problem was timing differences
> between the 84 and 245.
>
> My conclusion is that the rise time on the optical encoders is to slow.
> The specs on the encoders reads up to a max of <30ms transition time
> varying based on the rpm of the encoder.
>
> Does anybody have any ideas on how to fix the problem?
>
> Thanks,
> Richard Ramsey