19 - 25 Apr '21
- What has been confusing me about what should be straight forward is that the IOPi HAT is faster than the Wabbit. The Wabbit does not change the Clear and Thrown optic-isolators for a turnout at exactly the same time. When the Wabbit receives a command, and say changes the Clear isolator, the IOPi sees the change, captures the port and generates the interrupt. At some time after the port, at interrupt time, was read the Wabbit changes the Thrown isolator. So following the example and getting the "Interrupt value" identifies one change but not the other. If I have the interrupt procedure read the port a few milliseconds later, both changes are seen.
I do not know it the difference between the two changes is a Wabbit or DCC clock cycle or just the difference in response time for the Clear and Thrown circuits on the Wabbit I am testing, but the results are the same. To complicate the issue, each interrupt/port/eight bits monitor 2 Wabbits or 4 turnouts. Using a 'smart rout' 3 of the 4 turnouts could be commanded to change. With 6 bits changing only one will be reflected in the interrupt value the other 5 occurring sometime later.
- When I get the Raspberry Pi installed and the Wabbit FB connected, I will need to do some testing to identify what software delays are need to adjust for the hardware.
- Much filing on the status LED bracket. The unanswered question is "Will I be able to bind the metal to the shape I want. Or will I need to buy a MIG welder to hold things in place. There are some for less than $1,000.
- Minor CSS HTML software update.
- Updated/added Pydoc files
- Located the Raspberry Pi under the layout. After connecting the external 5vdc, All Wabbits report status as expected.
- Think I will fix the wiring error on the Wabbit interface board. In the end this can be resolved in software when sensors are assigned to input pins. But during testing it is going to drive me nuts.
- After looking at the wiring on the feedback board, I made the testing patch in software.
- Software development, either locally or remotely is going to be a pain.
- The state of the turnouts reported to the RPi screen is a mess. Much work to define if it is:
- Wabbit settings
- Wiring between Wabbit and RPI
- Software
- New mount, wiring, for the two Unicorn HD HATs. By rotating each Unicorn 90ยบ they fit closer together. Better. map of track and signals.
- Yet another small board to hold resistors and headers to bring together wires for where-ever that go to the status-LED/switch -bracket.
- Need to learn more about exception handling so RPi software won't lockup during startup or shutdown.