Unexpected behavior with ACE 128 encoder

LetItLearn

Hello again. A new puzzle came up when connecting an ACE 128 encoder.

I’m following the wiring recommended in this document: https://www.bourns.com/docs/Product-Datasheets/ace.pdf#page3

(As I’m unsure how to turn on the board’s internal pull ups, it uses external pull ups from 3.3v. The pull up resistors I’m using are 10k)

What’s odd is that in about half of the rotational positions, a clean reading can be taken and converted into a decimal value (using C++). The other half… the board disconnects, reconnects, disconnects, reconnects, ad infinitum. The lights never turn off, and as soon as I disconnect the encoder from ground OR find a “good” rotational position again… the IDE reconnects within 5 seconds and we’re happy again.

I’m unsure how to interpret the behavior, but I’ll be trying out a few things tomorrow:
- switching to 2k2 resistors (I believe this was recommended in an earlier post?)
- removing pins 1-8 and reconnecting them until a failure happens

If anyone recognizes this behavior or has a clue on where to investigate, I’m all ears. 🙂

Cheers!

p.s. Encoder pins 1-8 go to D2-D9

giuliomoro

LetItLearn he other half… the board disconnects, reconnects, disconnects, reconnects, ad infinitum.

uh that's not nice ... I suspect a miswiring could be the culprit here. 10k should be fine and not the cause of the issue you are having. Can you show a picture of your wiring?

LetItLearn The other half… the board disconnects, reconnects, disconnects, reconnects, ad infinitum. The lights never turn off, and as soon as I disconnect the encoder from ground OR find a “good” rotational position again… the IDE reconnects within 5 seconds and we’re happy again.

are you perhaps printing every time a new value is read? It could be that miswiring or a software error are causing for one or more "bad" positions a steram of prints,. which in turn choke the board.

LetItLearn - removing pins 1-8 and reconnecting them until a failure happens

this will definitely help pinpointing the issue.

giuliomoro

Weeell ... what I am saying is that the 10k resistors won't be the cause of "stress" on the power supply. It affects rise time and when compounded with noise this may cause reading a few alternating 0 and 1s upon switching, but unless there is something really time consuming happening in your code every time a different value is read (is there? You didn't tell us!), this shouldn't cause problems.

lokki maybe one of the 10k resistors was bad... never happened to me but who knows.

that's a possibility ... I heard years ago that it is possible that a resistor has a different value than what is marked on it, surely an extremely rare occurrence and maybe not happening anymore in modern days ... maybe you could try measuring the resistance of all those "10k" resistors? Or swapping them out one at a time until things become problematic again?

However

LetItLearn the board disconnects, reconnects, disconnects, reconnects, ad infinitum. The lights never turn off, and as soon as I disconnect the encoder from ground OR find a “good” rotational position again… the IDE reconnects within 5 seconds and we’re happy again.

This suggests that we are not actually somehow shorting 3v3 to gnd because of a very low value resistor in there: in that case the board would shut down and reboot and so it would take up to a minute for it to reconnect, not just 5 seconds as reported here.

LetItLearn

Switching to 2k resistors did the trick though I’m not sure why. I did have eight 10k resistors in parallel, and I don’t have an intuition for Ohm’s law but perhaps that was creating stress on the power? I may switch for something closer to the 4.7k resistor array mentioned in the documentation. Thanks for your feedback, and I am curious for anyone’s thoughts on what may have gone wrong.

(I’d like to post a picture of the circuit w/ 10k resistors, but the website says “Error” — debugging never ends does it)

giuliomoro

LetItLearn but the website says “Error” — debugging never ends does it)

I think it doesn't like images larger than about 2.5MB

LetItLearn y. I did have eight 10k resistors in parallel, and I don’t have an intuition for Ohm’s law but perhaps that was creating stress on the power?

No, it is smaller values that may cause "stress" when connected between a power rail and gorund, but larger values don't. It may be that you had some wiring mistake that you fixed while swapping resistors? Who knows!

LetItLearn

I’m inclined to agree that it was a wiring issue — the only thing that leaves me wondering is that the set up was rewired several times including on a new breadboard and with a second encoder (and checking for shorts between the resistors), and the only apparent change which precipitated the breakthrough was the switch to 2k. Odd, but I’m relieved it’s working.

Thank you for your for the feedback along the way! Here are some resized pics of the 10k circuit.

alt text

lokki

maybe one of the 10k resistors was bad... never happened to me but who knows.

also 10k resistors will switch a bit slower then 2.2k, but if @giuliomoro says it's fine it usually is 🙂

LetItLearn

giuliomoro read (is there? You didn't tell us!),

Ah sorry! I ran a stripped down version of the code and the same behavior happened — no rt_printf and nothing cpu intensive.

For posterity I will see if I have time to test out the 10k’s… but no promises! 🙂

Thank you again for your communicativeness here! I’m quite enamored with the community already.