Fred's firmware development diary

[Fred]

viewtopic.php?f=41&t=1103&p=34459#p34459 be good contributors thread updated for those with more than one thing on their plate.

[Fred]

Just gone 5am. Just pushed out some new code and features/updates:

• Alpha-N mode now uses its own correctly scaled table, NOT the VE table, which never made any sense...
• Speed Density mode now calculates VE only if SD is active and not otherwise.
• Configured load is only used for lambda and ignition timing, SD uses MAP, AN uses TPS, always.
• New blended mode and 2d RPM vs blend table that uses both SD AND AN at the same time (more blending options in future...)
• Appropriate updates to EMStudio to take advantage of the above, to be pulled in by Mike in the morning (he went to bed)
• A number of bugs found/updated in EMStudio while doing fundamental/basic testing of my own changes (one not by me) see below.

To use only one of SD and AN, but have both calculated, set the 2D blend table to all 100% (AN) or all 0% (SD) and enjoy.

One note about AN mode: The values required can be pretty low because the comprise the raw airflow for the engine setup. So the maximum airflow is a VERY large amount of fuel, so use small numbers at first.

Bugs:

http://issues.freeems.org/view.php?id=838 < updated a few times
http://issues.freeems.org/view.php?id=848
http://issues.freeems.org/view.php?id=849
http://issues.freeems.org/view.php?id=850
http://issues.freeems.org/view.php?id=851
http://issues.freeems.org/view.php?id=852

This should:

A) Make life easier for the owners of engine 20
B) Give Sean an excuse to not use a MAF sensor
C) Take away Sean's excuse for using a MAF sensor
D) All of the above :-p

Fred/Bed.

[Fred]

Spent a bit of time tonight doing something I should have done years ago: Learning to use Kicad. Stumbling my way around the PCB designer part at the moment. Thanks to Andy for the real-time n00b QA session :-)

Spent a bit of time earlier today helping someone get their repo in order and their changes rebased on top of the latest dev branch content. That went OK in the end.

Spent a bit of time in the last few days trying to coach some basic common sense and quality development practice into the group of people who I once looked up to and have, over the years, learned a lot from. Staggered. Baffled. Lost for words. What a complete and utter waste of my time. It never ceases to amaze me how badly run some of the fundamental projects that hold the dev world together are. And it always breaks my heart.

Fred.

[Fred]

viewtopic.php?f=41&t=2161 < Someone I was advising over skype earlier tonight suggested that I do podcasts. I've been thinking about it anyway, so that thread is a feeler for interest. Let me know if it's a waste of time, or not.

Fred.

[Fred]

Make no mistake, I take versioning very seriously indeed! viewtopic.php?f=45&t=2114&p=34504#p34504

[Fred]

There is a new thread stickied somewhere on this forum, if you're a regular, you may want to hunt it down and check it out. Just FYI. No rush, not really important.

[Fred]

viewtopic.php?f=41&t=2165 Support FreeEMS, build an awesome car/do an awesome thing, claim it's the EMS, not your skill/money/build. :-p I hate stupid people.

[Fred]

Last night I was doing over-skype pair-programming with a friend who's working on some paddle-shift code for their use of FreeEMS. After about 2 or 3 hours he had it looking pretty good. Nice work! You know who you are. I've also written a gear-detection function that uses a single ADC reading to determine which gear, and a corresponding speed calculation function that uses RPM and a ratio. I'll add a tyre-stretch correction to it later, too. The two blocks of code are cleanly separated and easy to integrate and move around, so expect them to pop up at some point, though not soon.

Once done with the code review, pair programming and education session, it was 2am, so I did a bit of work on ULV to improve its usefulness. It now has selectable parameters for the simple histogram, which is nice.

I won't be pushing any major updates to the firmware until some other things have been taken care of. The worst of these have been outstanding for more than four months. Only one has successfully ended up here in proper form, so far, though that should change soon. Once those things are sorted and finalised, I'll finish off the new RPM code and push it up for general testing and use. I also won't be pushing anymore firmware code at all until Sean tests the dev branch on his LT1 Camaro and gets me an EMStudio log of it. Make it happen, Sean, it's just you now, all other variants are tested OK. Then I'll consider that code adequately tested for general consumption and migrate it to the master branch. There's currently far too big of a gap between them... that's got to change! Ball is your court, my friend!

In somewhat sadder news, I'm being forced to take a chunk of time away from the firmware and other community stuff to remedy a situation caused by one person's short-sightedness and selfishness. I guess that I can achieve what I need to in the order of 2 to 4 weeks, but it's time I'd much rather be spending on *anything* else. None of the alternatives to spending this time are acceptable to me, and many of them aren't acceptable to you, either. This just generally sucks, but like always, we'll pull through the shitty time and come out stronger and better off.

Today is my day "off", so I'm going to lay in the sun for a while and then do some personal stuff later. Tomorrow will be flat out assisting a friend with their tuning remotely and other skype calls that are planned.

Fred.

[Fred]

FreeEMS on the dyno! Fuck yeah! :-) viewtopic.php?f=55&t=2123&p=34549#p34549

[Fred]

OLV 33 seconds to crash with insufficient memory. ULV 7 seconds to finish with enough memory left over to do it 10 times more. Same memory settings for both apps. OLV needs a LOT of love.