touchscreen triggers istouched after backlight turned off

Question

I'm trying to setup a timeout to turn the backlight display off after a predetermined amount of time. The timeout works but the routine I use to re-enable the backlight (isTouched) triggers immediately after turning off the backlight so then re-enables the backlight. Any idea why this is happening. I have commented out the ESP sleep so the only thing that should happen after the sleep timeout is that the backlight LED should turn off

Thanks in advance,

Fish

Answer 1

It is tricky, and basically the idea of deep sleep / wake doesn't work. At least on all my tries. You can see in my fork a workaround by keeping all alive but only OLED off after timeout. Then turning on on touching lower part of the screen. Also see cxandy's sketches.

Comments

What I don't understand is why isTouched, after I shut off the display and LED, always returns true?? I ended up doing pretty much the same as you and had to specify a region to ignore because the x and y returned on a false wake was pretty consistent. Shouldn't isTouched return false if the screen isn't touched?

---

Just a thought: would it help if you drain the touch buffer before turning the backlight display off?

---

I cannot get exactly your thoughts about "drain the touch buffer" , but any better solution is welcome. The touch chip+sensor layer seem to have an internal wakeup processing on first touch after inactivity timeout. See the datasheet:
http://www.buydisplay.com/download/ic/XPT2046.pdf

---

My thought was to empty the touch buffer with dummy readings before shutting down the backlight:  while (!ts.bufferEmpty()) p = ts.getPoint();
hoping that (ts.isTouched && !ts.bufferEmpty()) as a trigger for re-enabling the backlight returns false, unless a real touch event occurs.

(I have to do something similar when using a STMPE160 touch controller + Adafruit library, otherwise it behaves unpredictably.)

Answer 2

Hi catfish, I found this thread that may explain anomalies on OLED wake up.

Touch chip seems to be sensitive on voltage dropping of 3V3 supply line.

Comments

Hi lorol,

I was starting to suspect something like that, especially after reading the XPT datasheet and noticing various options to read the data out for stability. At the moment I have it working ok, not great. I used the lower screen detection (was for 12h/24 h selection) and that appears more stable. I had added a print to output the p.x and p.y after a trigger and noticed that it was always the same screen location being detected. The  p.y < 80 avoided that so it kind of working.

Now, I'm finding I have a different issue. The weatherstation appears to work properly but then I find that the current temperature stops updating after a few hours. If I reset the ESP8266 then it updates to the current conditions but a few hours later it stops updating. The time always updates correctly so I'm not sure whats going on. I'll have to dig in further.

Thanks for the info.