|
|
|
EEPROM storage and snow
Tuesday, December 2 2025
The day started on a good note when a woman from the New York Department of Labor called in response to the message I'd sent via the website. Had she not called, I was going to have to maybe go to the brick & mortar unemployment office in Kingston and try to get them to fix my problem. But now I no longer had to think about it.
I turned my attention back to the menorah I am supposed to be making. I cut out all the pieces needed for a simple branching (but not reticulate) menorah. I then put the pieces all together with flux and took it out in the garage to solder the bulk of it together. In an effort to make it as orthogonal as possible (particlarly to get eight of the nine candle holders to form a line) I stood it upside down before soldering it. But then when I went to solder it, there was some problem with the flux I was using. It was behaving more like vasoline than flux, and the molten solder was refusing to flow into the joints. I'd never seen anything like it. I tried other flux, but by now it was too late. I was going to have to tear most of the joints apart and reflux them with flux that doesn't suck.
So then I resumed work on the thing I wanted to be working on anyway, further improving the firmware for my ESP8266 Remote Control system. Now that I had the refactored configuration working correctly, I wanted to be able to store this configuration information in the EEPROM of a slave Arduino attached to the ESP8266 via I2C. To do this succesfully, a lot of things had to work correctly, as the ESP8266 can only hand off the data to the slave, but the slave must do the storing and retrieving. I'd had ChatGPT build out some functions for both the slave and the master, but these were all untested. When I went to test them today, I found that they didn't seem to be working. But why they weren't working was hard to tell. At the time I was flashing my slave firmware using a USBTinyISP on a separate computer, and this was clunky procedure that didn't make debugging software on it easy. So I spent a little time setting things up so that I could flash the slave software from a nearby laptop while it was in-circuit with the master (which I was flashing with Woodchuck, my main computer). This allowed me to monitor the serial outputs of both devices simultaneously and to make corrections to both the slave and master interactively as needed. With the help of ChatGPT, I gradually mopped up little issues. For example, the slave was expecting little-endian data, whereas the master was sending big-endian data. There was also an issue where a read from EEPROM could only retrieve one byte per request. But even after both of these issues were fixed, I still had a weird problem where my communication with the slave always failed during the Arduino setup() function but worked fine from the loop() function. When I finally figured out why this had been happening, I felt stupid: in setup(), I was trying to use functions for accessing the slave before I'd set the configuration for what the slave's I2C address was. So the master didn't know how to reach the slave at that point. Once I'd fixed that and ironed out an issue related to the saving of empty strings, the saving and retrieving of configuration data became extremely reliable, and I could now not only change configuration data during run time, but I could also alter the default configuration for an ESP8266, meaning I could make it boot up believing itself to be a different device with different capabilities.
I don't think I went outside even once today, which is very unusual. The main reason for this was that we received six to eight inches of snow throughout the day, so I couldn't do most of the things I would normally go outside to do.
For linking purposes this article's URL is:
http://asecular.com/blog.php?251202 feedback
previous | next |
