Use arduinomorse instead of Jled #38

2023-09-13T20:13:41-05:00

W1CDN commented

2023-09-13 20:13:41 -05:00

Branching off a branch to see if a custom version of arduinomorse (https://github.com/markfickett/arduinomorse) is easier to work with than Jled.

Note that I have a working copy of arduinomorse locally in Sites/Arduino library testing/arduinomorse.

Todo:

start sending on schedule
repeat n times on cycle
wait right amount before starting again
add continuous

Branching off a branch to see if a custom version of arduinomorse (https://github.com/markfickett/arduinomorse) is easier to work with than Jled. Note that I have a working copy of arduinomorse locally in `Sites/Arduino library testing/arduinomorse`. Todo: - [x] start sending on schedule - [x] repeat n times on cycle - [x] wait right amount before starting again - [x] add continuous

W1CDN added the enhancement label 2023-09-13 20:13:41 -05:00

W1CDN changed title from ~~Use ardiunomorse instead of Jled~~ to Use arduinomorse instead of Jled

2023-09-13 20:13:54 -05:00

mattbk added 1 commit 2023-09-13 20:25:01 -05:00

Add arduinomorse and dependencies. 087ad27c8c

mattbk added 1 commit 2023-09-13 21:06:20 -05:00

Shelve. dd085635a2

W1CDN commented

2023-09-13 21:06:43 -05:00

Ugh, I can't get it to send on cycle.

Maybe try continuous first?

Ugh, I can't get it to send on cycle. Maybe try continuous first?

W1CDN commented

2023-09-13 21:35:52 -05:00

I think Jled and arduinomorse were fighting for the gpio. I commented all the Jled stuff out and at least got one message out on startup.

mattbk added 1 commit 2023-09-13 21:54:48 -05:00

Start on schedule and able to stop. 06e69bd8c0

W1CDN commented

2023-09-13 21:57:33 -05:00

I can start on schedule and run forever, then stop by switching the program to off. Seems like a good place to stop.

Repeating n times seems like a for or while loop.

Waiting the right amount of time might be a while loop with millis() around the outside??

Might need duration of the message to figure out repeats. Looked around a bit and then just asked: https://github.com/markfickett/arduinomorse/issues/13

I can start on schedule and run forever, then stop by switching the program to off. Seems like a good place to stop. Repeating n times seems like a `for` or `while` loop. Waiting the right amount of time might be a `while` loop with `millis()` around the outside?? Might need duration of the message to figure out repeats. Looked around a bit and then just asked: https://github.com/markfickett/arduinomorse/issues/13

mattbk added 1 commit 2023-09-14 12:44:22 -05:00

Cycle appropriately. 73a6b694a9

mattbk added 1 commit 2023-09-14 12:46:39 -05:00

Clean up. a47cd1465a

W1CDN commented

2023-09-14 12:48:13 -05:00

I did it without getting the duration, using millis(). It seems to start/end 1 second slow, though. Will figure that out next.

This has the bonus effect of sending right up until the end of each cycle (sending fractional messages) rather than leaving blank time if the message won't fit (as in the Jled implementation).

I did it without getting the duration, using `millis()`. It seems to start/end 1 second slow, though. Will figure that out next. This has the bonus effect of sending right up until the end of each cycle (sending fractional messages) rather than leaving blank time if the message won't fit (as in the Jled implementation).

W1CDN commented

2023-09-14 16:04:14 -05:00

RTC isn't keeping time, again.

mattbk added 1 commit 2023-09-14 16:36:26 -05:00

Add back ability to send continuously. c8f9d823da

W1CDN commented

2023-09-14 16:37:29 -05:00

So cool. I wouldn't understand this if not for all the time with Jled, but it's a lot more nimble.

The last bit is to wire up these so they actually work:

So cool. I wouldn't understand this if not for all the time with Jled, but it's a lot more nimble. The last bit is to wire up these so they actually work: ![image](/attachments/7d48c2e3-507f-49b1-8d07-3a0f2694c098)

image.png

24 KiB

mattbk added 1 commit 2023-09-14 17:10:00 -05:00

Clean out old Jled and arduino-timer code. f5aa43ef76

W1CDN commented

2023-09-14 17:32:48 -05:00

Woops, need to use a non-constant cycle ID and n transmitters.

Started on this, haven't committed yet. Need to shut down the sender when not using it. LED (gpio) stays on sometimes.

Woops, need to use a non-constant cycle ID and n transmitters. Started on this, haven't committed yet. Need to shut down the sender when not using it. LED (gpio) stays on sometimes.

W1CDN commented

2023-09-14 20:46:16 -05:00

Sending extra space at beginning and end because my math/logic isn't right.

W1CDN commented

2023-09-14 21:02:11 -05:00

Need to take a break, but this doesn't work if yourCycleID == 1, because there is no extra space at the beginning, so it shows up yourInputStepLength early. So I need to draw a picture of the timing (or use a spreadsheet).

start_millis = millis() + ((yourInputCycleID - 1) * yourInputStepLength);
top_millis = start_millis + yourInputStepLength;
// Subtract 2 rather than 1 here to account for start_millis duration at beginning of repeat.
pause_until_millis = stop_millis + (yourInputStepLength * (yourInputNtransmitters - 2));

Need to take a break, but this doesn't work if `yourCycleID == 1`, because there *is* no extra space at the beginning, so it shows up `yourInputStepLength` early. So I need to draw a picture of the timing (or use a spreadsheet). ``` start_millis = millis() + ((yourInputCycleID - 1) * yourInputStepLength); top_millis = start_millis + yourInputStepLength; // Subtract 2 rather than 1 here to account for start_millis duration at beginning of repeat. pause_until_millis = stop_millis + (yourInputStepLength * (yourInputNtransmitters - 2)); ```

mattbk added 1 commit 2023-09-14 21:02:34 -05:00

Shelve. 4380a56055

mattbk added 1 commit 2023-09-15 12:22:36 -05:00

Do the math right on cycle timing. 6c1a718a98

W1CDN commented

2023-09-15 12:23:35 -05:00

Got it. Pumped for this. Will merge this when I have time today.

Add these to issue queue:

make sure user can't input larger cycle ID than number of transmitters - #42
why are cycles starting ~1s too late? - #41

Got it. Pumped for this. Will merge this when I have time today. Add these to issue queue: - make sure user can't input larger cycle ID than number of transmitters - #42 - why are cycles starting ~1s too late? - #41

W1CDN referenced this pull request

2023-09-15 12:47:22 -05:00

Set up scheduled program cycles #24

W1CDN merged commit a93498ac84 into program-cycle

2023-09-15 16:35:54 -05:00

W1CDN referenced this issue from a commit

2023-09-15 16:35:55 -05:00

Merge pull request 'Use arduinomorse instead of Jled' (#38) from arduinomorse into program-cycle

W1CDN deleted branch arduinomorse

2023-09-15 16:36:02 -05:00

Sign in to join this conversation.