Fred's firmware development diary

[Fred]

And its 5am and we've just discussed the crap out of freeems comms! More work to do tomorrow, lots more.

viewtopic.php?f=8&t=997&p=12154#p12154

Fred.

[Fred]

Progress from Dave:

http://stuff.fredcooke.com/FreeEMS.Inte ... by.MTX.png

Fred.

[Fred]

Day 4 consisted of cleaning the house and tools up (it had to be done, and it wasn't my choice.) as well as lots of talk on comms with Dave and others. IE, I got nothing tangible done, but a lot of thinking went on around the world.

Day 5, today, consisted of implimenting some of the ideas discussed yesterday and more discussions and various tidy up and documentation. All docs are up to date, out of date drafts/ideas are moved to a directory of that name and all test packets function as described. I will shortly take a shower and complete the tool/lounge cleanup afterwards.

Tomorrow I will work with Dave on the comms stuff and work on the truck and if there is time, the scheduler. I'll also be testing SeanK's loader and EssEss' testing framework soon, I can't wait for either! :-)

More news as it comes to hand :-)

Fred.

[Fred]

Day 6 I finished cleaning up my tools, fixed a couple of bugs from yesterday, implemented a lookup service with details, got it partly working, emailed and chatted about various things, talked to the mrs, had some more crazy news, prepared an open source wideband for publication.

Tomorrow I'll try to make the AND service work, and update all the metadata till it looks to be correct. Fix any bugs Dave finds, talk to the wideband man, work on the ute, and start the FreeMS2 project in a simple way (toggle outputs at variable duty/speed for bench testing ms2 units.

Note, it's 5am, and I didn't give up until I had something for dave to work with. Tomorrow vs today are a bit blurred.

Fred.

[Fred]

Day 7, realised that the and thing WAS working, created test cases and some other minor work. I also went to the lake for a swim and gave a friend a driving lesson.

Day 8, completed the metadata service stuff, made the code more robust and efficient and clean, now all that is required to be done for dave is provide data structure defs. he may want this in xml/json, so it may be time to write a utility or find something that can create arbitrary output based on some simple rules from xml or json. with that we can generate C as part of the build process and automatically keep the xml/json and C in sync. I also visited a friend who is here from WA and another from Welly for a couple of hours.

Tomorrow, ute! ute, bug fixes, data structure def discussions, sched work.

It's now 2am and I'm well ready for bed after spending half of lats night awake worrying about someone.

Fred.

[Fred]

Day 9, chatted to other devs, talked to the good woman, Wrote a thread on the final stage of communication framework required for a truly reliable system. Hacked on a new project for a few hours and made a big commit, still more to do, but getting close to trying it out soon.

Tomorrow, clean lifters, head, start to reassemble the head, clean up the garage, tools, etc. Sched work, bug fixes, work on the data def thread, maybe try some ideas out. More hacking on the new project too very much a side effort, but I need to use it a little bit soon for some testing.

There are 13 days till the rescheduled event I had hoped to attend in the ute. I need to find a sump and spend 2 or 3 days putting it all back together etc. Then its tweak req_fuel (vomits...) and start hacking on code and hw to make it run on FreeEMS.

Just shy of midnight, but i'm so so tired... shower, bed.

Fred.

[Fred]

Day 10, another day, another thread!

Yesterday's thread: viewtopic.php?f=8&t=1004
Today's thread: viewtopic.php?f=8&t=1007
Another with less thought/effort put into it: viewtopic.php?f=41&t=1009

Plus, I got the other 12 lifters cleaned out, they were full of metal flakes.. Plus the usual chatter on IRC with the other devs. And it's not quite over yet, either, so I might get some more misc work done tonight before bed.

Tomorrow is clean the head day. I could start assembling the engine again once that's clean. Plus the usual planning and thinking and any fixes, etc. Hopefully dave implements the universal hex editor function soon. Sched and data defs are on the menu in general, too. Ute first, though, as it must be done before I leave again one of these months... AND because I need it to test Puma! :-)

Fred.

[Fred]

Niceties:

TODO: Add all interesting data blocks to locationID DB so they can be pulled back generically. Error counts, runtime stats, clocks, core vars, ADCs, etc etc. Most should be marked readonly and only be in RAM.

TODO: Add ability to manipulate ADC values and core values and output values directly bypassing the usual math and logic for code testing purposes and bench testing purposes. Possibly related to the previous TODO.

TODO: Add fake decoder that reads an RPM figure somewhere and behaves accordingly. Possibly with some preprogrammed pattern, or possibly default to boring 36-1 or something. Combined with the previous TODO this will allow full testing of all core comms visible functionality.

Sched and data def and all things in the 0.1.2 goal thread are more important, but I may squeeze these in if it makes sense or for fun. The last one could come in as part of the sched change anyway. The first one is just a matter of documenting stuff thats already there. The middle one requires more thought and engineering.

EDIT TODO: Make configurable datalogs general and possible and feasible using locationID details. Possibly include power to provide lists, either in call, or in config, such that combinations of different fragments and specific variables can be retrieved rapidly in a stream or once in a while by request. This would require some fairly significant changes tuner side, but would be perfectly flexible and reliable and generic and lovely. This change depends upon the first one listed above.

Fred.

[Fred]

I'm running MegaTunix/Tunix/MTX on a 13" MacBook Air attached to FreeEMS and viewing the ADC values etc in the "runtime display"! It needs some mapping file to make the fancy dash work, and some data map files to make tuning possible, so we're really close to full FreeEMS wide support by MTX, which is FCUKING FANTASTIC!!!

THANKS DAVE!!!! :-)

Note built on MacPorts, not fink, though I may try that soon. I'll try it against an ms2 soon too to test for stability issues due to the large changes involved to support wildly different (and better) comms such as FreeEMS and SECU-3 have.

Fred.

[Fred]

Day 11: I cleaned the head, including the oil galleries, and the bungs for them, and found two valves are either bent or have shit sitting under the seats. I'll be pulling those, checking for straightness, cleaning, installing them, or some spares, cutting the seats with paste, etc, soon. I also went nuts on MTX and tested anything I could. Along with lots of discussion and a bug fix and a "it's not my fault" verification. To do that I had to test it on the mac. It built clean on macports as noted above.

Today: I'm going to attack my old sump with an angle grinder and remove the flange, straighten the side up, cut the flange off a fwd sump and get them close enough to weld. Later I might fix those valve issues quietly in the evening as I don't have the tool right now and will then. I've already tried to build/test the loader on the mac, no dice, SeanK is working on it. Marcos' boards arrived and I did a little publicity work on them. Images are here:

http://www.facebook.com/FreeEMS

Or just here:

http://img508.imageshack.us/img508/5382 ... dfront.png
http://img211.imageshack.us/img211/8642 ... edback.png

Pretty things, aren't they! :-)

Data discussions and thoughts continue too. Sched and that are still being thought about and worked on all the time.

Right, off to the ute! :-)

Will update before bed later.

Fred.