Fred's firmware development diary

[Fred]

OK! Preston is running my old car on FreeEMS again! Woo Hoo. And in the last 48 hours I've given him 7 historical and 1 current version of the firmware to try. Due to the decoder he is using and the nature of his beast, it has some trouble starting, but, once started, runs fine on the latest stuff. The issue is that he has picked up about 10 degrees advance from the old code with no change in tune. So:

• I need to determine what caused the timing change (historical testing narrowed that down a lot for me)
• I need to fix the code such that starting is not difficult for him
• I need to calibrate his map sensor
• I need to add a temporary hack for his burnt ADC pin problem (Presto, you need to fix the CPU at some point, as when config goes to file based through a tuner or other app, this support WILL be removed)
• I need to fix the timing code or recalibrate his timing (Sean, beware, you run without checking with a timing light. You could be getting screwed here...)

Once we have the starting and timing sorted out, we can start making some more general changes and getting them tested. Hopefully by that point Sim's project will be running too and I'll have a sufficiently diverse test group to check core code changes with.

In other news, Preston has been doing a lot of work on the Spin 1 Puma BOM situation and the results of that are here:

viewtopic.php?f=53&t=1294

This is key to getting more testers on board. So far 2/37 printed boards have been used and it's been 9 months since they were printed... This needs to change!

Fred.

[Fred]

And just like that, number 7! Sim is running his 1979 Volvo 245 on COP with LS1 coils and DSM CAS and 24/1 disk in Canada right now! Congratulations to him for that achievement! :-)

Fred.

[Fred]

viewtopic.php?f=11&t=1301

Practical improvements left right and centre! Fixed a temporary 3% timing inaccuracy hack from a while back and improved the logging situation a lot while also improving MTX display.

I have more changes for MTX coming along well, but I have a train to catch!

Fred.

[Fred]

Check out this awesome log specimen that clearly shows the first three cylinders firing and lifting the RPM of the engine! The engine took 360 degrees to find sync, then 120 degrees to fire the first cylinder, by the end of 2 revolutions from hitting the key, it was running. Worst case for this engine assuming all else is equal is 720 to sync and 1080 to running. (2 revs to sync and 3 to running).

EDIT: This is a 24+1 disk in a CAS housing in the side of FreeEMS engine number 7 (Sim's Volvo). The blue dots are each of the 24, where 24 = 720 degrees. You can count them to verify what I said above.

Also, the first step, and plateau, is the cranking region at 280 RPM (below that is zero). The next step is 580, the next 740, then 920, then it becomes unclear and there are less logs than outer electrical pulses.



The power of awesome logging quality is starting to be revealed :-)

Thanks to Sim for the log and to Gufi for the viewer :-)

Fred.

[Fred]

More logs from the massive user base (3) today:

Electrical noise on the RPM input but ONLY during cranking:



The hotel decoder has no timestamp checking, which all others do have, and it should have, but it was only for my car, so... it results in bad/wrong and EXTRA scheduling of ignition pulses and WILD RPM readings:

Scale is 0 - 32768 RPM



Scaled to 0 - 3000 RPM so you can see the smooth running post crank



I'll fix this soon such that it won't run and Preston will fix the wiring such that it doesn't happen and will run.

Then it's time for some more general sync/loss/rpm/sched/etc changes to allow everything to behave better under all reasonable AND unreasonable conditions.

Fred.

[Fred]

MegaTunix for FreeEMS now has a full set of 42 gauges in AutoMeter SportComp style.



All that is needed is a couple of dash designs, easy stuff. I might do one or two tomorrow :-)

Fred.

[Fred]

Dash design done, I finished last night at 4am and got to sleep by 5am. Hard work gets things done! :-)



It's VERY easy to load this design into the editor and remove all of the excess ones and add a few more and maintain the layout more or less. This is somewhat maximal.

A full set of GP UINT and SINT and ADC10 gauges are available too, for debugging purposes.

Fred.

[Fred]

May I proudly present FreeEMS engine number seven, Sim's 1979 Volvo 245 wagon! Take it away, Sim! :-)

http://www.youtube.com/watch?v=I5K-6jdSUeo

[Fred]

Today someone tried to use FreeEMS to receive a text message:

viewtopic.php?f=62&t=1314

Which gave me a good idea:

viewtopic.php?p=17721#p17721

So all is right in the world! :-)

Fred.

[Fred]

viewtopic.php?p=17785#p17785

JimStim updated, time to do some updates to the 4/1 decoder, the generic hall/opto decoder, and then maybe check the DSM/Miata pattern and add a decoder for that. Any of the above 'cept for Preston's updates to the generic hall/opto decoder can and likely will get interrupted by Sim and his temperature sensors and associated timing adjustment features.

EDIT: This is POST NUMBER 500 in this thread!!! I think I need to celebrate the 500 post milestone somehow :-)



Fred.