Fred's firmware development diary

[Fred]

Last night I read through all of the doxygen docs and it was *really* useful even for me, so I guess it's a good thing I put the time into that for the last release! I'll continue to look over stuff today, and over the weekend I'll get mantis back up and then fill it with goodness - all the stuff that was in my head that I need to fix and change and have nearly forgotten.

Priority next week = rearrange memory for xgate experimentation. After that, scheduling stuff and comms again, need to nail comms down better and get a stable gui compatible with it :-)

In other exciting news : I'm going to the nurburgring again!! next weekend :-) Can't wait, there is quite the atmosphere there, esp when the helicopter takes away the dead and mangled!

In still more exciting news : I resigned yesterday, so in 6 weeks I'll be unemployed and working full time-ish on this again. Well, that and my CV and websites and packing my things and repairing the extractor and and and. 6 weeks or so of that, then a trip into euro land, then home!!!!!!! wow, I thought this day would never come, seems to have come quite quick despite all the pain and sorrow england has caused me.

When I get home I'll be working on and driving my MS powered truck - sure to be good motivation to get this running and ditch the MS. I know that other members on here are motivated in this way :-)

Till next time,

Fred.

[Fred]

OK, so, last night we cracked out Abe's board and loaded the firmware up into it. I read my own documentation and threads on the forum to figure out what to do and get started and it was mostly pretty good. Abe's card now has the latest code on it and is ready for serial comms functionality testing.

There is a lot of work from Aaron's side to do on the tuning app to allow us to adjust the tables etc but the basics more or less work.

My plan is to get my deb usb install up to a usable spec because i can't run python 2.6 on the eee with xandros/asus hacks. I will also be writing a simple single pulse "decoder" for running one injection per received pulse.

Perhaps we can get some serious testing in if Abe and Aaron have working hardware.

Time to write some code soon perhaps!

Fred.

[Fred]

I'm posting this from a USB installation of Debian Unstable running on my EEE PC. I intend to run only this distro from now on until I get it near perfect and then install to the SSD inside and ditch Xandros completely.

I just did a full firmware build on this box and it took sub 9 seconds, not bad for 5 different firmwares built!

Next up is getting python and the tuner up to speed and then test it against the dev board i have here.

More on this tomorrow as I make good progress in Phoenix Arizona :-)

Fred - Developing on the road!

[Fred]

The tuner now runs, though due to debian being a bit on the slow side with python 2.6 it was a slight hassle - thanks Aaron.

Fred.

[Fred]

The tuner spoke to the latest firmware this afternoon after a few patches from Aaron. I've been soldering and installing all afternoon and should have some more progress to report soon. I have eclipse up and running on my eee pc usb debian, and lots of other goodies too. Tick tock.

Fred.

[Fred]

I've been constructing a sort of "on the road" test rig here at Sean's place. It now has a map sensor and battery voltage reference and powersupply all setup and working nicely. I'm using a serial adapter that Abe gave me in San Diego and I should have some other circuits built tomorrow morning some time with a bit of luck.

Aaron has built a real time data display and added binary logging to the tuner. Sean is working on a parser to produce MLV logs from the binary stream (easy with a solid protocol!! ;-)

Unfortunately with Aaron's code up and running and some sensors attached it looks like my code is all wrong! The BRV figure is way out and permanently offset by 32000 or so and similar for MAP - this is fundamental and simple stuff though, so it shouldn't take long to sort out tomorrow now that I'm sure of where the issue lies.

Nearly 3am and it si time to hit the sack - via the rather yummy Breyers Strawberry Icecream - England has a shockingly shit selection of icecream and althought America doesn't have my favourite flavours it does have a good variety and good quality of produce in that area :-) Yay for yankee icecream :-)

Fred.

[Fred]

Fixed!! With some debug help from the tuner (thanks for making it easy to hack!!) I managed to find the issue and fix it. Just a simple one line change - out of order struct init/def lines... woops. All things working as expected again now.

Also fixed, a tuner bug - need to cast to floating point and truncate or round values. I made the decimals long for you though mate.

Back to the soldering iron and email changes out to Aaron and fix up my git repo stuff, it's a mess :-/

Fred.

[Fred]

I've just about completed the first physical FreeEMS ECU and it's sitting here on Sean's desk beside me. I need to do some more study on the car tomorrow before completing the circuitry for it, but it's one input and one resistor value away from ready to bolt in and study/learn from.

We've located the fuel computer and wiring for it in the Volvo and are ready to splice into that piggy back style tomorrow morning. By tomorrow afternoon I plan to be logging real data from the engine into binary files that we can parse and analyse later. I should also be able to view that data in real time with the tuner! Sweet :-)

Thanks very much to Abe, Sean, Aaron in particular for helping get to this point today. A special mention for Gross Poluter and Josh from Yoshifab too for the inspiration to hook up to the volvo and try some stuff out.

More news as it comes to hand!

Fred.

[Fred]

Just a quick note to say that a simple but complete FreeEMS is now installed in my Volvo (piggy back/datalogger only) and I have recorded about 1/3 of a gig of datalog from it driving up from phoenix to slc.

In other news, the tuner has been getting some real world useage. Hopefully i can use it in earnest soon! :-)

Fred.

[Fred]

*No* comments? Really? Short post, but pretty big deal IMO, pretty close to T- = 0