Archive for July, 2019

Moldy but goody but ..

Posted in Buyer beware! on July 29, 2019 by asteriondaedalus

Dusted off my Acer Aspire One. Couldn’t find power adapter. I had thrown a bunch of adapters out because I had so many. So I found a place to order one from. It turned up with the wrong country pins. Then I bothered to look and found the original charger. DOH!

So, it plugged it in to charge it. It had Debian 8 installed so it wanted to update 500 plus packages. So I left it running over night.

Rebooting over my morning coffee, I was excited at having that little hack back online.

Unplugged it and logged onto wifi. Battled through a zillion expired certicate warnings on the dopey old browser Conquerer. DOH!

5 minutes into browsing, 6 minutes into unplugging. It turned itself off. Now I remember. Battery’d died. DOH!

Still. I need a terminal for my cluster.

Oh well.

Seems good …

Posted in OpenSprinklette on July 27, 2019 by asteriondaedalus

I did a retake on the components required for the 24VAC to xVDC boards, used in opensprinklette, and realized that I had bought 100uF instead of 1000uF electrolytic. Little wonder it wasn’t helping in filtering the spikes caused in hot end because of solenoid activation/deactivation.

So, swapping out the 100uF with a 1000uF fixes the crashing wemos!

Almost. It runs for a while then … wemos fires up, I have a looping test in node read that turns the relay on/off every 5 seconds. The LWT will pop up after the widget goes quiet after a few cycles. I found it would reboot and start again, though would only cycle a few times before crashing. More often than not, I had to pull the powerpack from power board and restart.

Runs a treat otherwise, as long as you don’t turn the relay on. All 24VAC elements appear on the correct side (the other side) of the protection, so its a little perplexing.

3 options. 1) the cap voltage rating needs boosting, 2) the TVS is too tight versus the RMS of the 24VAC input, 3) the circuit I found is a dud.

I am actually waiting on delivery of electrolytics of a higher voltage rating. Jaycar only had 60V caps. The TVS rating was on a batch I got based on an example circuit for OpenSprinkler. It it works for that then surely it should here? In any event, I have a slightly higher rated TVS batch coming.

The design I borrowed appeared to opt for a 40V design target so the cap really should have room. It will likely be the TVS (I am hoping) and then we can move on.

Bigger IS better 😉

Not idea-L

Posted in Android, Jetbrains IDEA, kotln on July 21, 2019 by asteriondaedalus


I have been doing some kotlin courses, to get over the hump.

I have been running Jetbrain’s IDEA on my PC and that has been smooth.

I thought I would set it up on my Debian 10 laptop, as I need feign being social with the wife as I can tinker while we sit in the vicinity of one another.

Tribulations then.

  • IDEA installed.
  • No SDK reported when creating a kotlin project!
  • Option to add SDK via JB SDK bin addin.
  • JB SDK bin adding installed fine.
  • When running JB SDK bin addin, it kept failing with not enough memory.
  • Bumped up memory.
  • Kept failing.
  • Went searching for JDK via SDK widget.
  • Found opensdk V8.
  • Restart saw IDEA fail to start with error that class path could not be found.
  • Found info on NET that config and system directories need to be deleted to recover and got back up and running.
  • There was a link provided to Oracle’s JDK under the pesky SDK setup box.
  • Downloaded and installed Oracle JDK (V12).
  • Oracle JDK V12 came up in SDK box next restart – so did error that class path was kaput.
  • Ran SDK installer again. No errors this time. Without having fidgeted with memory. Great mystery – wooooooooooooooooo (using ghostly warbling voice).
  • Pulled down V11. SDK widget prompts me to choose between V12 and V11. Chose V11.
  • Now setup!

However, no clear cut reason SDK installer did not work first pass.

No real point pointing to Oracle JDK if, when installed, it breaks. Certainly reported same message that openjdk did raise.

No real clarity why, if openjdk V8 was already installed, it didn’t turn up in SDK selection. It would have sat outside the JB SDK install mechanism in the same way the Oracle JDK V12 did.

No curiosity in me to “solve” any left over mysteries. Don’t feel I have learnt any lessons then.

Kotlin compiles – through very slowly on this clunky old laptop.

What a wanK!

Posted in hass(le)io, The downside of Opensource on July 15, 2019 by asteriondaedalus

So, to date. Things that don’t seem to work out of box on Raspberry Pi 3B+ hass(le)io (either 32 or 64 bit) are:

  • cloud9 ide
  • node-red
  • portainer
  • a few others that I can’t be bothered listing

Now the problem might be mine. It seems though that the setting up of the config files is trickier than it should be. This is aided and abetted by “help” files that don’t.

Idiomatically, a templated warning on many “help” files is “Don’t use this config file, write your own”. The example config in the “help” file acting, as it were, as taunt.

The problem turns out, if you cut and paste the example config files, from the “help”, many of them cause errors! They don’t, therefore, act as a example of a straight forward, no bells’n’whistles, get you started type of example.

What seems to be missing, in peoples puny attempts at “help”, is:

  1. Help should not assume users have full architect status in the project, and understand not everyone has the priori knowledge of an architect OR even the plug-in developer.
  2. See point 1.

Certainly, when some of the error logs are reporting potentially missing files, nested deep in the Docker image, then the fact that the Docker image is not ready for production, and shouldn’t be on the streets, is laid bare.

Some of this is the open source, the user is the tester. Or worse, the user is a developer. I know that there is a derogatory name, in the open source community, for users who don’t contribute. I admit I am one. What I marvel at is the pressure to contribute is a counter argument to the utility of open source, since if I can’t use the software due to useability, installability, understandability, portability or other software quality problems then I am not likely to be much help because of the inherent problems, generally, with open source in the areas of useability, installability, understandability, portability.

That is, if using the software is my goal, getting involved in all the open source code bases I might opt to use for my projects is not an aspiration, BECAUSE I am officially bored with the quality of open source software in terms of useability, installability, understandability, portability.

You should be too. Especially now that open source projects are whining over “pay me for my time”. YOU FREAKS! You were the types that set up the free software rort. You would have been passing floppies of stolen software around, putting companies out of business. SUCK IT UP YOU PRINCESSES!

Pardon my grumpy pants. I will actually buy a beer, if the option is there, if the software meets my expectations on useability, installability, understandability, portability. Not, however, if in the first hour I am having to debug what should be straight forward, or clearly explained.

I did managed to get the Aircast running, no changes to config needed other than, it seems, turning SSL flag to false. That was a punt I took, as the default (SSL=true) raise a error when you tried the web UI link. That “fix”, applied to the other apps above, did not fix them.

Aircast is great, since I have a C.H.I.P. running shairpoint-sync under the television. So, as chrome-cast is attached to television, I can free up a airplay speaker.

The other app that seems to run fine is motioneye, which is a CCTV monitoring thing. Same thing, came up after setting SSL=false. CCTV around the place was my next project for the home, so sweet.

The configurator works, so at least I can play with the scripting.

Otherwise, I am running the hass(le)io based RPI3B+ side by side my OPiZ.

The OPiZ is running Mosquito and node-red just fine, except for the system going down about once a month.


Posted in hass(le)io on July 14, 2019 by asteriondaedalus

Herein, with this category, will blog my trials and tribulations with hassio.

So far, not so good

Posted in OpenSprinklette on July 14, 2019 by asteriondaedalus

All testing good until I wire in a solenoid.

When you go to turn on solenoid, the wemos crashes.

I think I designed in a whoops, since the TVS protects from spikes on power from other solenoids, but not the one the wemos is controlling (I managed to have it on the wrong side of the TVS.

So far, so good.

Posted in OpenSprinklette on July 13, 2019 by asteriondaedalus

So. I will need to redesign, as I used too big a cap template. Otherwise, it seems to power the board fine. I will need bomb opensprinklette onto the wemos for some testing.

Out of shot is the 1 AMP 24VAC power pack.

Missing is still the single relay shield, to be plugged into the Wemos D1 mini.

The redesign will shrink the 24VAC – 5VDC section.

And now, it’s on!

Posted in OpenSprinklette on July 13, 2019 by asteriondaedalus

The prototype opensprinklette boards are here! So, the final sprint to pull everything together.