I spent the day under the dock hooking up the permanent cabling for the various post caps, and finalizing what is there.

I implemented the low pass filter last night, of course, until I fix the ESP stuff (connecting to wifi is temperamental) I can't test it so I will need to do that. I have decided that while I have the main controller connected to the programmer, I will make sure that it can do the broadcast thing, using the SDR to confirm the output. Then I will do receiving the broadcast on the light controller, I decided that I don't actually need to a great deal on the light controller, just "simply" turn on the second rx pipe and configure it's address.

ADC is working.

I can read the duration on the main controller fairly accurately, I will take the number and feed it into a mathematical low pass filter to remove some of the jitter, and then figure out how to convert it to volts. I will probably end up plotting the data in excel and figuring it out that way. The graph should be linear so the math won't be that hard :)

Tonight I got the uC that is measuring either the battery or panel to work. I had to use separate controller that output a pulse train with a frequency that is input voltage dependent which is fed through an opto isolator so the grounds are completely independent.

For my next trick, I am going to have a broadcast message from the main controller to the lighting... It "should" be relatively easy, I will simply make every light controller listen to the same address on a separate channel ('dock0' on channel 2) then the main controller can simply turn off acks and send to everyone.

This means everything will be the same, but super useful for the "hey the battery is dead shut it down" event.

One of the problems I have been facing is the lacklustre wifi signal quality on the dock. Keeping in mind that 802.11 is the sole house to dock communication method it's fairly important that the main controller can get and maintain a good connection to the wifi.

As I mentioned in the previous post, I got a new-to-me router on freecycle to mess with and fortunately it has a non-integrated antenna. I also have this d-link antenna from a previous project.

The gazebo is mounted (the led strips where still there from the rehearsal dinner years ago) but one of the strips had a failure about halfway down. Fortunately I had a replacement strip that I could mount. I also used the hot glue, and gooped them down into final position. Hot glue is a pretty good adhesive, as well as a waterproof potting like compound.

So today, I got it working, I am not sure what the causes of my deaf NRF were, but everything has started working and I was able to make this video. I will have to keep an eye on it as an old email signature of mine once read "If you didn't fix it, it's not fixed" essentially meaning if you don't understand why it's no longer broken, it's not fixed. I suspect that it was a combination of output power being incorrectly set, as well as them being too close and causing front end overload.

I tried some more things on the light controller to see if I could get something to happen, no joy.

However, and I'm gonna be upset if this is the problem, I did realize that I may have got the power output bits muddled, and instead of 0dBm, I might be using -18dBm. I am not 100% sure this means for actual output, but I do know it works out to be around 0.016 of the power, I wouldn't have expected it to be an issue considering the antenna's are about 100mm apart, but you never know. I will turn it up and see what happens. I might look at the differences on the SDR first.

The number one thing that had me hosed...

PWR_UP not set! I think what was going on was the first bytes of SPI were going out too soon. The fix was to do..while with the end condition being a correct read back. Once that first one works, it seems to be good thereafter


Subscribe to RSS - blogs