The horse has bolted

Posted in Uncategorized on January 18, 2020 by asteriondaedalus

Microsoft promises to be carbon negative by 2030. In an around the extinction of wild rhinos and elephants. Just short of the last glaciers disappearing. Not to mention the raft of predicted extinctions because of plastics in the ocean.

Qantas dog killers

Posted in Uncategorized on January 18, 2020 by asteriondaedalus

Go figure. The Qantas response to its killing a dog by leaving the dog for an hour on the tarmac in 40 degree Celsius heat was to ban the breed. If the owner had killed the dog that way, the owner would be charges by the police The ground staff involved need to be named. The staff who did not act need to be publically named. Sackngs are in order. Qantas personnel are amoral and unethical. Snub nosed breed owners should shame Qantas. Unite. Post “Shame Qantas shame”

Oh the pain

Posted in Docker, Home Assistant, HypriotOS on January 12, 2020 by asteriondaedalus

Trying to get Clicksend integrated with Home Assistant appeared to be easier, once you find all the goss, but its finding the goss.

The clicksend config is at Home Assistant site. Straight forward enough.

To set up an action the trick turned out to be use a “Call Service” with:

alias: ''
data:
message: Zone 1 turned on
service: notify.clicksend

Now the hint was at the Clicksend integration page for Home Assistant. That it, it was a service call.

That pieced together something for me that didn’t stand out from the mess of help at Home Assistant. That is:

notify:
   - platform: clicksend
     name: ClickSend
     username: CLICKSEND_USERNAME
     api_key: CLICKSEND_API_KEY
     recipient: PHONE_NO 

Leads to a service call, in the actions and likely scripts, of:

notify.clicksend

Makes eminent sense.

Google smoogle

Posted in Rant on January 7, 2020 by asteriondaedalus

Google “warns” us that having your hands smacked for breaching IP and copyright laws will somehow “upend” software development!

Bullshit.

The premise is that, given opensource evolved out of the criminality of software piracy, succeeding simple due to the weight of infringement overwhelming the legal system’s capacity to respond, then somehow Google can use the “software is art” mantra to steal licenced IP. Oh, masquerading as loss of “openness”?

Bullshit.

If openness was so precious, why steal IP in the first place, simply to jeopardise the outcome?

OF note also, the turn around in so-called opensource community, where people are advocating “monetization” of their efforts, failing to remember that they are likely the same people that engaged in software piracy and therefore raping the software developer’s of that era of their “monetization” efforts of their work.

Bullshit.

So, since licences are still pertinent. Since “monetization” is back in. Pay up Google you wankers. Grow some ethical standards you prats. Give up on your “monetization” efforts if you are so philanthropic.

Google, you are transparent.

Look inward at the treatment of your employees. That would require “openness”.

Hint

Posted in OpenSprinklette on January 4, 2020 by asteriondaedalus

If you are using a module such as the one pictured below, to drive a single sprinkler solenoid, then you will note it will accept 7-30VDC in. So you will need a AC-DC conversion. A bridge and a 1000uF cap will do. If you are wanting something neater then aliexpress comes to the rescue with a kit option.

Final Opensprinklette designs (for now)

Posted in OpenSprinklette on January 3, 2020 by asteriondaedalus

Presenting the quad board. Added additional terminal blocks to help manage the wiring. Added prototyping area to allow for additional hardware hacking, including four terminals for additional I/O. Notice GND, +5V and +3.3V power rails.

Top side of Opensprinklette quad board
Bottom side of Opensprinklette quad board

Presenting the single board. Added additional terminal blocks to help manage the wiring. Brought as many pins out as I dared to a pair of headers (J1 an J2). J2 also has associated solder jumper (on bottom of board) to help with option of V1 or V2 of Wemos mini relay board. The solder jumper on the Opensprinklette V2.1 sets the mode of J2 pin 2.

If you are using a version V1 of the Wemos mini relay board you don’t have options, as D1 drives the relay and you don’t get then to use I2C. Jumper defaults J2 to Pin 1=D2, Pin 2= D3 and Pin 3= D4. You lose D0 in any event since its set aside for V2 of the Wemos D1 mini relay board.

If you are using a version V2 of the Wemos mini relay board you change the jumper on Opensprinklette single to set J2 to Pin 1=D2, Pin 2= D1 and Pin 3= D4 . Effectively you lose D3. You need also to change the relay driver pin on the Wemos D1 mini relay board to D0 (from D1), using the solder jumpers on the Wemos D1 mini relay board.

Top side of Opensprinklette single board
Bottom side of Opensprinklette single board

J1 pinouts

PinUse
1+5V
2GND
3+3.3V
4A0
5D5
6D6
7D7
8D8

D0 and D3 are not that exciting to lose in either event. D5..D8 gives you access to SPI so you either have SPI if using V1 Wemos mini relay or SPI and I2C if using V2 Wemos mini relay. Not a bad trade off. Not to mention UART is also brought out.

RxTx pinouts

PinUse
1GND
2Tx out
3Rx in

Et Voila!

Posted in OpenSprinklette on December 29, 2019 by asteriondaedalus

So, I rebuilt a brand new single relay opensprinklette board. I did notice, as I stripped the original, that the end of the electrolytic was a tad “swollen”, oh dear, oh well.

I had a new batch of 47V TVS, physically smaller than another batch I had previously bought. So, that means I don’t need hack at the boards to fit them. I thought the original problem was that I had was using a lower voltage TVS than I should have been. Only marginally I thought. It didn’t turn out that the slightly lower rated TVS was the problem, though hacking the board to fit the larger and thicker leaded TVS probably helped me fry that board. I likely jumpered lead to the wrong place or inverted something.

No matter. Turns out a clean build, with all the properly rated components, still sees the brownout after a few cycles.

Hmmmmmm.

So, I shorted the shutdown pin to the VIN as recommended.

I then set up a node-red test that turned the relay on/off with a 50% duty cycle with 5 seconds in each state. It has been running now for about an hour without failure. So the quirk is that the Pololu 5V down converter seems more sensitive than the 9V down converter.

I will still need to short the shutdown on the quad board to be sure. I am now ready for final versions of the two boards, with minor tweaks including additional terminal blocks to keep the wiring tidy.

I am in the process of playing with a Kotlin front end for Android to manage the home setup. That is really just a prototype as the home setup will start with a single quad unit. I will build HMI system up as I get a hang of Kotlin development.