Is it possible to run more than one pin to an interrupt for sleep/wake purposes?
-
@karl261 It is a pretty hard to find where a possible cause is.. :confused:
What should happen to generate an interrupt is a change in one of the inputs. In a standard application the outputs (i.e. rows or colums) will be set to low and the inputs (colums or rows) pulled-up.
I can't figure out what the state of the row's/ column's is in the idle state from the library. YOu can probably measure if the rows en colums have a different level Sorry for now...@AWI I don't recall that columns and rows have a different level. Of what I measured yesterday was that all 8 pins of the keyboard were at 3.3 V. So the only thing that happens is that there will be a connection made between column and row when a key is pressed. It seems that nothing is pulled up or down. :-/
So you think the ic can do it? Then we have to re-program the library... :-(
http://www.ti.com/lit/ds/symlink/pcf8574.pdf
On page 15 is a wiring example. Maybe we can learn something from this. Some pull up missing?EDIT: No, still can"t get it to work. Buh.
-
No progress....
Maybe this library is better to understand the pcf...
https://github.com/RobTillaart/Arduino/tree/master/libraries/PCF8574
Also here are some explanations (in German) and some code examples (in English)...
http://www.mikrocontroller.net/articles/Port-Expander_PCF8574
-
No progress....
Maybe this library is better to understand the pcf...
https://github.com/RobTillaart/Arduino/tree/master/libraries/PCF8574
Also here are some explanations (in German) and some code examples (in English)...
http://www.mikrocontroller.net/articles/Port-Expander_PCF8574
-
@AWI I think the problem is with the keypad. Because it is not really putting a signal to its pins. The only thing it does is short circuiting them when a key is pressed... No change in high or low.
I tried for example to pull row 1 to high and column 1 to low. Then the interrupt fires ( I think) but there is no key press detected any more. But I am not sure if the interrupt fires correctly.
I tested some variations last night, but nothing works. It's all still set up on the breadboard, so if you have some suggestions I am happy to try.
-
@AWI I think the problem is with the keypad. Because it is not really putting a signal to its pins. The only thing it does is short circuiting them when a key is pressed... No change in high or low.
I tried for example to pull row 1 to high and column 1 to low. Then the interrupt fires ( I think) but there is no key press detected any more. But I am not sure if the interrupt fires correctly.
I tested some variations last night, but nothing works. It's all still set up on the breadboard, so if you have some suggestions I am happy to try.
@karl261 That's how it is supposed to be. When the library "scans" the keyboard it succesively pulls one of the rows (or colums) and check in the colums (or rows) if there is a connection. To have interrupts detected either the rows of colums should be set to low to detect a change. That would mean adapting the library (if possible) -or- (more fun) build your own routine by using Rob Tillaarts basic library from your previous post.
I don't have any PCF8574 to play with... -
@karl261 That's how it is supposed to be. When the library "scans" the keyboard it succesively pulls one of the rows (or colums) and check in the colums (or rows) if there is a connection. To have interrupts detected either the rows of colums should be set to low to detect a change. That would mean adapting the library (if possible) -or- (more fun) build your own routine by using Rob Tillaarts basic library from your previous post.
I don't have any PCF8574 to play with...@AWI Yes, basically, 4 inputs should be set to high in the library, and 4 inputs to low. Rows and columns. I had a look at the library examples yesterday, but so far it is beyond my programming skills. As far as I understand from all the docs this is theoretically possible.
Then, if done so, the pcf should detect a change in the low/high of the input pins and also detect what key was pressed. Or better the library interprets the received data by i2c correctly.
-
-
@karl261 That's how it is supposed to be. When the library "scans" the keyboard it succesively pulls one of the rows (or colums) and check in the colums (or rows) if there is a connection. To have interrupts detected either the rows of colums should be set to low to detect a change. That would mean adapting the library (if possible) -or- (more fun) build your own routine by using Rob Tillaarts basic library from your previous post.
I don't have any PCF8574 to play with... -
@AWI I think this Library sounds very promising. What do you think? At least it talks about high and low.
@karl261 At least it is a more comprehensible library but it does not deal with interrupts (yet).
Are you able to do the following experiment?- Using the library you quoted, execute the following sketch
#include <Wire.h> #include <i2ckeypad.h> #define ROWS 4 #define COLS 3 // With A0, A1 and A2 of PCF8574 to ground I2C address is 0x20 #define PCF8574_ADDR 0x20 i2ckeypad kpd = i2ckeypad(PCF8574_ADDR, ROWS, COLS); void setup() { Serial.begin(9600); Wire.begin(); kpd.init(); Serial.print("Testing keypad/PCF8574 I2C port expander arduino lib\n\n"); pcf8574_write(pcf8574_i2c_addr, 0xf0); } void loop() { }- measure the input/output pins of the pcf8574 (should be 4 pins low/ 4 pins high)
- measure the power consumption of the pcf8574 (hope it is next to nothing)
IF the above is true we can easily rework te sketch to use the interrupt.
-
@karl261 At least it is a more comprehensible library but it does not deal with interrupts (yet).
Are you able to do the following experiment?- Using the library you quoted, execute the following sketch
#include <Wire.h> #include <i2ckeypad.h> #define ROWS 4 #define COLS 3 // With A0, A1 and A2 of PCF8574 to ground I2C address is 0x20 #define PCF8574_ADDR 0x20 i2ckeypad kpd = i2ckeypad(PCF8574_ADDR, ROWS, COLS); void setup() { Serial.begin(9600); Wire.begin(); kpd.init(); Serial.print("Testing keypad/PCF8574 I2C port expander arduino lib\n\n"); pcf8574_write(pcf8574_i2c_addr, 0xf0); } void loop() { }- measure the input/output pins of the pcf8574 (should be 4 pins low/ 4 pins high)
- measure the power consumption of the pcf8574 (hope it is next to nothing)
IF the above is true we can easily rework te sketch to use the interrupt.
@AWI Thanks. I am in the train now and I am spending my time to try to understand how things work. I also made some progress in my brain. My idea is similar, but you already have a sketch ready. :-) I will try it out, I guess on Saturday. Until then I will try to deepen my knowledge of the libraries.
It is also possible with the "original" library, but you are right: the "new" library is much clearer.
-
@karl261 At least it is a more comprehensible library but it does not deal with interrupts (yet).
Are you able to do the following experiment?- Using the library you quoted, execute the following sketch
#include <Wire.h> #include <i2ckeypad.h> #define ROWS 4 #define COLS 3 // With A0, A1 and A2 of PCF8574 to ground I2C address is 0x20 #define PCF8574_ADDR 0x20 i2ckeypad kpd = i2ckeypad(PCF8574_ADDR, ROWS, COLS); void setup() { Serial.begin(9600); Wire.begin(); kpd.init(); Serial.print("Testing keypad/PCF8574 I2C port expander arduino lib\n\n"); pcf8574_write(pcf8574_i2c_addr, 0xf0); } void loop() { }- measure the input/output pins of the pcf8574 (should be 4 pins low/ 4 pins high)
- measure the power consumption of the pcf8574 (hope it is next to nothing)
IF the above is true we can easily rework te sketch to use the interrupt.
@AWI I found this library:
https://github.com/skywodd/pcf8574_arduino_library
There is a lot of talk about interrupt functionality and it seems to be somehow implemented. But I don't fully understand how it works. Maybe you can make something out of it?
Oh, and maybe in even more detail with some examples, I have not read it yet:
https://github.com/RalphBacon/PCF8574-Pin-Extender-I2C
...but it looks like a software interrupt rather than a hardware interrupt??? I am confused.
... at least for the last example: It does not need any library. It's all quite clear in one sketch.
-
@AWI I found this library:
https://github.com/skywodd/pcf8574_arduino_library
There is a lot of talk about interrupt functionality and it seems to be somehow implemented. But I don't fully understand how it works. Maybe you can make something out of it?
Oh, and maybe in even more detail with some examples, I have not read it yet:
https://github.com/RalphBacon/PCF8574-Pin-Extender-I2C
...but it looks like a software interrupt rather than a hardware interrupt??? I am confused.
... at least for the last example: It does not need any library. It's all quite clear in one sketch.
@karl261 the last sketch does not work with interrupts. It just mentions them.
I would stick with the previous suggestion. Measure the current when you set some outputs to low (low is active 1)
-
@karl261 the last sketch does not work with interrupts. It just mentions them.
I would stick with the previous suggestion. Measure the current when you set some outputs to low (low is active 1)
-
@karl261 the last sketch does not work with interrupts. It just mentions them.
I would stick with the previous suggestion. Measure the current when you set some outputs to low (low is active 1)
@AWI Finally:
Yes, it works as expected. 4 pins are 0 V and 4 pins are 3.3 V. The chip is using 2.5 uA.
I had to change a little bit the library, it did not compile and was slightly outdated, also the function we wanted to use was private, but I just did the changes in the library.
I also needed to update to the latest arduino from 1.6.5 to 1.6.12, because i got an error: collect2.exe: error: ld returned 5 exit status. Now it works.
Where do we go from here?
-
@AWI Finally:
Yes, it works as expected. 4 pins are 0 V and 4 pins are 3.3 V. The chip is using 2.5 uA.
I had to change a little bit the library, it did not compile and was slightly outdated, also the function we wanted to use was private, but I just did the changes in the library.
I also needed to update to the latest arduino from 1.6.5 to 1.6.12, because i got an error: collect2.exe: error: ld returned 5 exit status. Now it works.
Where do we go from here?
@karl261 Thats good news! What you now have is 4 rows high/ 4 colums low (or v.v.). Now short a column with a row and and expect :baby: an interrupt (= keypress). If that is the case, you can create (or copy) a routine to read the key or keys pushed.
-
@karl261 Thats good news! What you now have is 4 rows high/ 4 colums low (or v.v.). Now short a column with a row and and expect :baby: an interrupt (= keypress). If that is the case, you can create (or copy) a routine to read the key or keys pushed.
@AWI yeah, I kind of thought the same. I'll try to have rows high and colls low and see if the interrupt pin finally triggers. I should have some pull up on the interrupt pin. Then I'll need to see if I need any pull up or down on the other lines..,
Hope I'll get something done this weekend.
-
@karl261 Thats good news! What you now have is 4 rows high/ 4 colums low (or v.v.). Now short a column with a row and and expect :baby: an interrupt (= keypress). If that is the case, you can create (or copy) a routine to read the key or keys pushed.
@AWI Ok, so far so good. If I connect 4 high wires to rows and 4 low wires to columns (or vice versa, no idea) then the interrupt pin indeed changes to 0. If I connect the interrupt pin of the IC to D3of the Arduino and then both through a pull up of 10 kOhm to Vcc, the interrupt is at 3.3 V. If I press a key it goes down to 0.040 V. That should be enough to read it as low.
That sounds good, doesn't it? The question is, whether the key library still works...
-
@AWI Ok, so far so good. If I connect 4 high wires to rows and 4 low wires to columns (or vice versa, no idea) then the interrupt pin indeed changes to 0. If I connect the interrupt pin of the IC to D3of the Arduino and then both through a pull up of 10 kOhm to Vcc, the interrupt is at 3.3 V. If I press a key it goes down to 0.040 V. That should be enough to read it as low.
That sounds good, doesn't it? The question is, whether the key library still works...
@karl261 Certainly and it can be pretty straight forward from here. Stick to the library you use (this), it contains all you need. You need to let MySensors catch the interrupt (in the sleep function) and have the callback routine (receive) execute
char key = kpd.get_key();as described in the example. -
@karl261 Thats good news! What you now have is 4 rows high/ 4 colums low (or v.v.). Now short a column with a row and and expect :baby: an interrupt (= keypress). If that is the case, you can create (or copy) a routine to read the key or keys pushed.
