Fred's firmware development diary

[Fred]

Today I knocked out a memory management document that should help everyone understand it as well as I do now. I also spent more time researching the pin out stuff by scouring the docs and the reg dumps. Finally, I spent the rest working on the JSON files.

I'm uber excited about tomorrow as word on the grape vine is that Ben and Aaron both have some treats in store for me. I can't wait!!

Bed now,

Fred.

[Fred]

Worked on the JSON files again, and when I got home merged in and tested Seans flash stuff. Would like to have tested bens work, but he didn't push it to the repo, shame on you ben :-)

Tomorrow night will be on hardware again, and tomorrow day/train on the json stuff.

I need to nail a solid serial document set and code set so that Aaron and Ben can get on with building a fully functional tuning app for it :-)

Fred.

[Fred]

Wrote more serial code today, and may write some more soon too. I still need to keep working on the hardware stuff. I notice Dave doesn't log in much. This concerns me. Perhaps I was too talkative (read big mouthed) when we met in London? LOL, anyway, needs to be done, as does the serial stuff and the interface docs for both.

All progress is good, but the progress that allows others to work is probably most important. Right now it's :

1. Hardware docs
2. Serial/data docs and code
3. Datalogging
4. Scheduler for wheel decoding
5. Math stuff
6. Toyota decoder (real version)
7. PIT ignition
8. Everything else

That list may vary, but those top three are definitely top.

Fred.

[Fred]

Implemented the three cell adjusting routines tonight. Can't test yet easily, but may have a try manually in the morning. Tried out the tuner, still buggy, but better :-)

Wrote up some json lists for location ids and error ids to be parsed and configuration based.

Fixed a few silly issues in the serial code along the way, none functional, just stylistic, but saved some flash and speed doing it :-)

Logging new task for ben now :-)

Fred.

[Fred]

I'm in the kitchen cleaning up after dinner with guests last night, but between loads of dishes I'm implementing basic datalogging. The guests are asleep on the lounge floor so I can't use the big screen or test on the hardware just yet :-)

Another task for Ben in the works! ;-)

Fred.

[Fred]

Datalogging is partly done. Basic datalogs can be got async and sync and can be turned on and off for async. The amount of log to send can be adjusted on the fly too.

Even better, I've proved that the duplex stuff works, ie, it is sending constantly, and you can interrupt it with a packet to tell it to do something else.

Even better still :

http://www.google.com/search?hl=en&safe ... tnG=Search

We are getting 80+ % bandwidth from it :-) Sweet.

The packets are about 150 bytes long at the moment. This equates to 67 logs per second or so, which I think young Cam will be happy with.

http://www.google.com/search?hl=en&safe ... tnG=Search

Much documentation and some tidy up to do. I'm not happy with some of the code as there is gross duplication, but not sure how/if it can be solved yet. Certainly it more or less works right now, I just don't like the style of it at the moment...

Bed time, via kitchen (ohhhh nooooo)

Fred.

[Fred]

(10:26:28 PM) Fred: hey man
(10:26:38 PM) Fred: how many logs per second does MS give you?
(10:27:33 PM) mops (Matthew): mmmm
(10:27:45 PM) mops (Matthew): you mean log entries
(10:27:48 PM) Fred: yep
(10:27:49 PM) mops (Matthew): i havent measured
(10:28:01 PM) mops (Matthew): but probably more than 10 ?
(10:28:05 PM) mops (Matthew): and also
(10:28:12 PM) mops (Matthew): there's a 'burst mode'
(10:28:17 PM) Fred: how fast is taht
(10:28:17 PM) mops (Matthew): where it loggs faster
(10:28:24 PM) mops (Matthew): i dunno never used it
(10:28:31 PM) mops (Matthew): normal logging is fast enough
(10:28:41 PM) mops (Matthew): you see the delay bewteen PW and afr
(10:28:55 PM) Fred: so, if i told you i could deliver 67 logs /second?
(10:28:57 PM) Fred: what would you say

Not only are our comms more reliable (utterly reliable) but they are also faster LOL. Brilliant.

Fred.

[Fred]

Last night I stayed up exceptionally late...

Trying to nail some bugs with Aaron. Some of them turned out to be his... and some didn't...

I'm somehow loosing the start byte off the front of the messages... not good for decoding obviously...

I'll have to have a read of the chip docs and find out why this is happening.

A mantis needs to be raised against me for this too.

Fred.

[Fred]

Serial bug fixed! (In a slightly dirty way)

Datalogging occurring :

http://i260.photobucket.com/albums/ii15 ... gFixed.png

Not exactly fast, more like glacial, but its happening.

It definitely receives packets (can hear the hard resets occur) but I can't see the data packets for the big stuff come back at the same time as the logs. Not sure if i just miss it, of if something is wrong. Code looks ok, but... you never really know.

Fred.

[Fred]

Full table dumps work too :

http://i260.photobucket.com/albums/ii15 ... Packet.png

But only if they have no escaped bytes in them...

That could be my fault, or Aaron's fault, checking...



Aaron's fault :-) Yay, I can go to bed soon :-)

Fred.