DIYEFI.org Forum

First start : http://www.youtube.com/watch?v=5KkexS29Seg
Driving : http://www.youtube.com/watch?v=gg15SRHHhbQ

Congratulations man, very nice work :-)

Fred.

Thanks Fred, you are the best leader ever :-)
Keep going like this.
You rock!

I'll upload some videos when the ECU is running properly.

Thanks again for the support

Marcos

Wow! That's pretty damned nice. Way to go, watching the driving vid now.

Here is your reward:
http://www.youtube.com/watch?v=D9Qoff9e ... re=related

How old is the 4k rpm bug, still working on it? I'd guess its the input circuits - I remember one of my input filters was so strict it wouldn't let a motor rev over 5 or 6 k. What sort of sensors does it use?

nitrousnrg wrote:Thanks Fred, you are the best leader ever :-)
Keep going like this.
You rock!

I'll upload some videos when the ECU is running properly.

Thanks again for the support

You are most welcome! My pleasure :-)

Fred.

8InchesFlacid wrote:Wow! That's pretty damned nice. Way to go, watching the driving vid now.

Here is your reward:
http://www.youtube.com/watch?v=D9Qoff9e ... re=related

How old is the 4k rpm bug, still working on it? I'd guess its the input circuits - I remember one of my input filters was so strict it wouldn't let a motor rev over 5 or 6 k. What sort of sensors does it use?

Hi!
its one week old, probably i'll have to rewrite part of the code. The problem is that i have only one timer, and is quite difficult to schedule all the four injections while keeping count of the time it takes to complete a crankshaft revolution.

The input circuit for the rpm is something similar to the lm1815, i tested it and works fine.
I'm using the stock reluctor, MAP, and air temp. Water correction and TPS are implemented, but i can't find the TPS in the car (drive by wire), and i didn't connect the thermistor.

There are still many things to fix, but once the rpm bug is fixed, i'll take a try at the dragstrip :-)

It might be a lot of work, but perhaps you could find an external timer chip of some sort, to interupt you when it's ready?

8InchesFlacid wrote:It might be a lot of work, but perhaps you could find an external timer chip of some sort, to interupt you when it's ready?

Hi!
Don't worry. It's fixed now. I only have a small RPM oscilation near 5500rpm, but the VE table is very flat there, so the injection pulses are quite the same.

Here it is working just fine:
The crazy bar means nothing, i was debugging a timing-related variable.
http://www.youtube.com/watch?v=3yXYihwEvos
Maybe I'll go racing tomorrow. There is a dragstrip a few miles away, I only need a helmet

I'm working on this:

I've datalogged that yesterday. The ugly spike is when i missed 3rd

TODO: scale, zoom, etc
It would be nice if i can show the data as a function of other variable instead of time, like A/F vs RPM.

I'm thinking how can I start writing the communication protocol for freeEMS if I don't have an ECU to see if something is wrong.
I can translate jems's communications to ECUmanager, but i can't test it.
Can i get a binary file, containing every byte received from the ecu?
That way i'll read and parse from a file instead of a serial port, and it would be the same thing.
Send data to the ECU should be the next step.

Well... byebye

nitrousnrg wrote:I'm working on this:

http://img90.imageshack.us/img90/7506/screen3xf7.png

I've datalogged that yesterday. The ugly spike is when i missed 3rd :oops:

TODO: scale, zoom, etc
It would be nice if i can show the data as a function of other variable instead of time, like A/F vs RPM.

Wow, cool :-) Which license will you use for it?

I'm thinking how can I start writing the communication protocol for freeEMS if I don't have an ECU to see if something is wrong. :?

I'm more than happy to test and provide bug reports for you :-) Plus, you can do what Aaron is doing which is to embed responses in a test function so to the rest of the app it looks like you are hooked up.

I can translate jems's communications to ECUmanager, but i can't test it.

The JEMS code was never finished. Aarons python stuff is the only working implementation, but you might be better off using the code from the firmware to base yours on as you are in C too.

Can i get a binary file, containing every byte received from the ecu?
That way i'll read and parse from a file instead of a serial port, and it would be the same thing.

Yes you can. However the ECU is quiet unless you talk to it, with the exception of datalogs which come asynchronously.

Send data to the ECU should be the next step.

Just like receiving it, but different :-)

If you structure your code well it should be quite reusable :-)

I need to get the comms stuff nailed down for you guys. I'm close now, but still not there yet.

Fred.

Wow, cool :-) Which license will you use for it?

It's embedded in ECUmanager, so it's GPL

Plus, you can do what Aaron is doing which is to embed responses in a test function so to the rest of the app it looks like you are hooked up.

.
Cool, i'll take a look.

The JEMS code was never finished. Aarons python stuff is the only working implementation, but you might be better off using the code from the firmware to base yours on as you are in C too.

Hehe, i've to catch up with the project's status.
I'm reading the firmware, i never used a protocol like that, so i'll need some time.

However the ECU is quiet unless you talk to it, with the exception of datalogs which come asynchronously.

Okey, thats seems fine.

I'll keep reading, and when you get the communication stuff finished, or close enough, i'll start writing some code, wish me luck on that hehe.

Marcos

nitrousnrg wrote:I'm reading the firmware, i never used a protocol like that, so i'll need some time.

I'll keep reading, and when you get the communication stuff finished, or close enough, i'll start writing some code, wish me luck on that hehe.

I'm sure you'll be fine :-)

Just make sure you are reading the latest code.

The best way is to setup Git on your box and do a 'git clone' on my repo. Then each time before you sit down you can just type 'git pull' and it will be updated.

Feel free to talk with Aaron about understanding it and implementation details etc. I'll send you his gmail address so you can chat with him. I think he has it pretty under control these days :-)

Fred.