Fred's firmware development diary

[Fred]

http://www.youtube.com/watch?v=67mR38gXjPg

Sweet, the Deuce Coupe is running nicely now :-) Once his wideband arrives and he gets some tune onto it it'll run even more nicely. Yippee.

Fred.

[Fred]

Tonight I added priming pulse code that with light testing in the US seems to work. I also rescaled the warm up code to have double the range for cold places. I had a minor fail on the oven mods son didn't complete those. And I educated myself on some web tech to achieve a personal goal. Not a bad day all in all. Not as good as I'd have liked, though, either.

Point being that all current builds will now get a shot of fuel on their injection channels and should be tuned to not over or under do it. I chose something that I felt was reasonable, so it shouldn't be too bad, but your mileage may vary. You've been warned.

Fred.

[Fred]

The priming pulse code is now fixed and working correctly, tested by Peter and available on master FYI. Anyone running injection is advised to configure and tune it ASAP as it could be useless or too aggressive for use as is depending on injector size and cylinder size and engine design. Get to it. Get your configuration changes to me via temp commits to a fork on github properly rebased to latest so that I can easily use them.

[Fred]

Tonight I can't sleep. I was browsing the roadmap and came across issues 499 and 500 which are both benign, however they should be nailed and soon.

I was also reviewing a header and came across some ugly inclusion of data in a code file. This reminded me that I want to remove all the literals and use macro calcs for all config vars in future.

So, my ordered list:

Finish rev limiter commit
Fix 499 and 500
Tidy up the config style hugely in terms of literals/macros 289
Tidy up the literals in code and move it all to config 180 and friends

In that order.

http://issues.freeems.org/view.php?id=499 < i saw the code that causes this and then lost it again
http://issues.freeems.org/view.php?id=500 < this may take some time to figure out, but needs to be done
http://issues.freeems.org/view.php?id=289 < macro/def/conf issue
http://issues.freeems.org/view.php?id=180 < config out of code
http://docs.freeems.org/doxygen-html/d7 ... gs_8h.html < shouldn't have defines in it
http://docs.freeems.org/doxygen-html/d2 ... ource.html < shouldn't have TRUE/FALSE in it
http://docs.freeems.org/doxygen-html/d1 ... ource.html < shouldn't have TRUE/FALSE in it

Fred.

[Fred]

The limiters commit is now public on dev branch and unlikely to be dangerous to use. This closes 163 and 144 assuming it tests out well. Feel free to upgrade to it if you're not Andy who is busy testing other stuff for the time being.

[Fred]

Issue 499 is solved: viewtopic.php?f=9&t=1675

[Fred]

I resolved issue 289 today after putting in a huge 2 day marathon to complete the required work. 26 files and 700 lines of change, 3 new files, not a single byte of binary difference. That's care and attention for you! Configuration is now direct and human readable, thus easier to follow and configure and more reliable and robust in the process.

More work to come on that tomorrow and other related work, too.

Fred.

[Fred]

Nailed issue 546 today, remove all google search links and replace with http://duckduckgo.com equivalents! Why?

http://dontbubble.us
http://donttrack.us

It's fast, clean and gives decent results that aren't skewed :-/ When it can't provide what you need, it's straight forward to try google instead. Adios google!

Also, I just put up this thread on future concurrency issues with XGATE, and even some existing code. Thoughts are welcome: viewtopic.php?f=8&t=1676

Planning to nail two more issues by the end of the day, though we'll see how I get on.

Fred.

[Fred]

Decided to do some soldering instead:

• Fixed my EEE power supply wiring by soldering a new cable directly to the back of the old socket just like the last one that eventually broke inside and failed.
• Soldered a knock chip to a board for testing on the bench
• Fixed my CAS logic adapter cable and verified that it works. This provides a real world way of developing a decoder for the 4G63 and MX-5/Miata as well as many many many other mitsi and mazda engines.

I also dealt with some drama from one of the team members. I don't like such drama at all, as anyone who has been around for a while well knows. Sometimes we might be better off with them only engaged in a social way. This might be one of those times, or it might not, only time will tell. In the last few months this individual has mostly cost time and not done a whole lot anyway. This is in contrast to the past when they both cost time and did a lot, possibly second only to me in work done. I personally hope that this person calms down and reconsiders their behaviour. I have no control, and never did have any control, over that, though. Best of luck to you, if you come across this.

Hopefully, with the drama over with for the short term, or so it seems, I can get on and get some work done tomorrow. Fingers crossed no more entirely unnecessary interruptions occur.

Fred.

[Fred]

As an experiment this morning, and because I'll need to do the same in a week or two, I decided to hand solder a 0.635mm QFN package with 0.3mm pads. Here is the result, pretty good IMO. :-) No liquid flux was used in the making of these photographs! Just raw skill :-p

http://stuff.fredcooke.com/soldering/qfn/



Apologies for the large image! It was necessary to show scale :-)

Fred.