Page 12 of 17

Re: Firmware developent status

Posted: Sat Apr 01, 2017 11:09 pm
by stefanst
Anything <=0 should be an error.
And log(x) == x should never be true for any number x.

Re: Firmware developent status

Posted: Sun Apr 02, 2017 5:11 am
by AndreyB
2017-04-02 01_10: EngineState: FATAL error: Natural logarithm logf() is broken: 18000.00
2017-04-02 01_10: EngineState: FATAL error: Natural logarithm logf() is broken: 18000.00
2017-04-02 01_10: EngineState: FATAL error: Natural logarithm logf() is broken: 18000.00
2017-04-02 01_10: EngineState: FATAL error: Natural logarithm logf() is broken: 18000.00
2017-04-02 01_10: EngineState: FATAL error: Natural logarithm logf() is broken: 18000.00
log(18000.0)==18000.0

:(

Re: Firmware developent status

Posted: Sun Apr 02, 2017 5:41 am
by AndreyB
mobyfab wrote:You're using softfp in the Makefile, so it's not fully used.
It has to copy register data from regular to FP registers everytime you need FP.
Looks like this natural logarithm issue is caused by switching from softfp to hard. We are not doing something right I guess, rolling that change back for now.

See https://github.com/rusefi/rusefi/issues/375
See original ticket from a while back https://sourceforge.net/p/rusefi/tickets/149/

Re: Firmware developent status

Posted: Tue Apr 04, 2017 12:58 am
by AndreyB
ChibiOS3 migration is pretty much done, looking for help with last step: anyone to start a real vehicle using http://rusefi.com/temp/test_chibios_3.zip ? Download/uncompress/flash and publish some log file if car runs fine?

My test mule is waiting for new alternator. Looks like alternators do not like to be excited while output wire is not hooked up.

Re: Firmware developent status

Posted: Tue Apr 04, 2017 1:20 pm
by AndreyB
Firmware help needed: https://github.com/rusefi/rusefi/issues/382

that's an effort to migrate from three separate log entry namings in three different places to using the same constants everywhere by means of wider code generation. All the tools are ready, the remaining part is to actually go locate all the names and extract into the common file.

Re: Firmware developent status

Posted: Fri Apr 07, 2017 1:01 am
by AndreyB
Major change: @ has migrated rusEfi to ChibiOS 3.1

Re: Firmware developent status

Posted: Fri Apr 28, 2017 3:46 am
by AndreyB
new issue - "integration with TunerStudio: 3.0.20 and 3.0.21 something is broken" - 3.0.20 and 3.0.21 are pretty unuseable with rusEfi. I am emailing EFI Analytics
https://github.com/rusefi/rusefi/issues/403

Latest known good version http://tunerstudio.com/downloads2/TunerStudioMS_Setup_v3.0.18.exe

Re: Firmware developent status

Posted: Fri Apr 28, 2017 4:32 pm
by AndreyB
https://github.com/rusefi/rusefi/issues/403 closed - TS 3.0.22 fixes the issue

Re: Firmware developent status

Posted: Fri May 05, 2017 7:57 pm
by AndreyB
Yet another defect reported by @ fixed - https://github.com/rusefi/rusefi/issues/404

Re: Firmware developent status

Posted: Sun May 07, 2017 12:24 pm
by AndreyB
TunerStudio 3.0.22 knows rusEFI signature now - in a couple of weeks I will switch from older test signature into proper signature, this would mean everyone would need to migrate to 3.0.22 or newer to use rusEfi with TunerStudio.

Usability progress - version is now displayed in TunerStudio title.

Re: Firmware developent status

Posted: Mon May 08, 2017 1:00 am
by Abricos
Small jump congratulations ...

Re: Firmware developent status

Posted: Sat May 27, 2017 1:01 pm
by AndreyB
https://github.com/rusefi/rusefi/issues/429 is an incompatible protocol change - you would have to use latest console or TS project with latest firmware.

Re: Firmware developent status

Posted: Sat Jun 03, 2017 1:52 am
by AndreyB
@ has discovered a bug in PID implementation - https://github.com/rusefi/rusefi/issues/435

It's about minValue/maxValue handing which is currently implemented by artifically setting iTerm to a funny value, even while iFactor is zero, which value then stays forever?!

https://github.com/rusefi/rusefi/blob/master/firmware/controllers/math/pid.cpp
/**
* If we have exceeded the ability of the controlled device to hit target, the I factor will keep accumulating and approach infinity.
* Here we limit the I-term #353
*/
if (iTerm > pid->maxValue - (pTerm + dTerm + pid->offset))
iTerm = pid->maxValue - (pTerm + dTerm + pid->offset);

if (iTerm < pid->minValue - (pTerm + dTerm + pid->offset))
iTerm = pid->minValue - (pTerm + dTerm + pid->offset);

and here is a unit test which shows wrong behaviour:

Code: Select all

	pidS.pFactor = 1;
	pidS.iFactor = 0;
	pidS.dFactor = 0;
	pidS.offset = 0;
	pidS.minValue = 0;
	pidS.maxValue = 100;
	pidS.period = 1;

	pid.reset();

	assertEqualsM("target=50, input=0", 50, pid.getValue(/*target*/50, /*input*/0));
	assertEqualsM("target=50, input=0 iTerm", 0, pid.iTerm);

	assertEqualsM("target=50, input=70", 0, pid.getValue(/*target*/50, /*input*/70));
	assertEqualsM("target=50, input=70 iTerm", 20, pid.iTerm);

	assertEqualsM("target=50, input=70 #2", 0, pid.getValue(/*target*/50, /*input*/70));
	// WOW, we are getting non-zero iTerm while iFactor is zero?!
	assertEqualsM("target=50, input=70 iTerm #2", 20, pid.iTerm);

	// and now we inherit this iTerm even for cases where targer==input?! NOT RIGHT
	assertEqualsM("target=50, input=50", 20, pid.getValue(/*target*/50, /*input*/50));
	assertEqualsM("target=50, input=50 iTerm", 20, pid.iTerm);

Re: Firmware developent status

Posted: Fri Jun 09, 2017 12:36 am
by AndreyB

Re: Firmware developent status

Posted: Fri Jun 09, 2017 1:00 am
by puff
It's not April's fool, is it? :D

Re: Firmware developent status

Posted: Sat Jun 17, 2017 2:29 pm
by AndreyB
I wanted to add mass storage drivers into the firmware for a while but I though that I've lost the working sample. I've found it!

Would anyone with a MAC try the sample firmware from ARMCM4-STM32F407-DISCOVERY-USB-MASS-STORAGE-SPI.zip from https://github.com/rusefi/rusefi/issues/140? That's without rusEfi but using rusEfi default SD card pinout.

Re: Firmware developent status

Posted: Sat Jun 17, 2017 6:58 pm
by AndreyB
For first step my idea is two use the user button to switch between USB port modes: normal mode serial USB, is reset while holding button it would be mass storage instead.

Theoretically a composite device could be possible but that's above my current USB skills.

Re: Firmware developent status

Posted: Mon Jun 26, 2017 6:43 am
by AndreyB
I did not realize that TS supports text fields!

Re: Firmware developent status

Posted: Tue Jun 27, 2017 3:10 pm
by Dron_Gus
russian wrote:For first step my idea is two use the user button to switch between USB port modes: normal mode serial USB, is reset while holding button it would be mass storage instead.

Theoretically a composite device could be possible but that's above my current USB skills.
I can try implementing composite device. CDC + MSD.

Re: Firmware developent status

Posted: Wed Jun 28, 2017 1:10 am
by mobyfab
Dron_Gus wrote:
russian wrote:For first step my idea is two use the user button to switch between USB port modes: normal mode serial USB, is reset while holding button it would be mass storage instead.

Theoretically a composite device could be possible but that's above my current USB skills.
I can try implementing composite device. CDC + MSD.
This might help: https://github.com/fpoussin/MotoLink/blob/master/code/common/usb_config.c

Re: Firmware developent status

Posted: Wed Jun 28, 2017 7:49 am
by AndreyB
Another week of vacation :)

Re: Firmware developent status

Posted: Sat Nov 04, 2017 12:31 am
by AndreyB
A lot of life in the way of firmware progress these days.

Re: Firmware developent status

Posted: Sat Nov 04, 2017 12:51 am
by puff
!!! congrats!!!
your first kid?

Re: Firmware developent status

Posted: Sat Nov 04, 2017 12:52 am
by AndreyB
Thank you :)
It's complicated :)

Re: Firmware developent status

Posted: Sat Nov 04, 2017 11:06 am
by kb1gtt
Such a cute little one. Assuming that's yours and not a rental, your going to have allot less time in the near future.

Re: Firmware developent status

Posted: Sat Nov 04, 2017 12:19 pm
by matt
Congratulations!!

Re: Firmware developent status

Posted: Sat Nov 04, 2017 3:22 pm
by mobyfab
Congrats!!

Re: Firmware developent status

Posted: Sun Nov 19, 2017 6:40 pm
by AndreyB
Making an incompatible configuration change today - https://github.com/rusefi/rusefi/issues/495

Re: Firmware developent status

Posted: Thu Jan 04, 2018 11:46 pm
by AndreyB
While I am not making too much progress these days, @ and @ are really pushing cold start and idle control forward!

Re: Firmware developent status

Posted: Wed Jan 24, 2018 3:44 am
by AndreyB
Just migrated to latest stable ChibiOS - see https://github.com/rusefi/rusefi/issues/463

minor issue with it - need to fix https://github.com/rusefi/rusefi/issues/551 but that should be easy