MS2/Extra pre-3.3.2 beta 6 released

[jsmcortina]

during compile I get this error:
collect2: ld returned 1 exit status

Suggestions very welcome.

You need to look for the actual error message that comes before this. The part you posted means that linking failed, but doesn't show why. (Most likely is that you overflowed a section.)

James

[jsmcortina]

Something is going wrong here; whether it's TS, the ini, or a PEBKAC malfunction has yet to be determined. I downloaded the latest TS beta last night as well as this MS beta, and while trying to create a new project got the same errors again and again, in all three ini versions. Screenshots attached; let me know if other info is needed.

Please post this in the TunerStudio section if you haven't already.

James

[jsmcortina]

James:
I have checked the processor pin connections continuity with IAC1 and IAC2 and confirmed that the schematics are pretty much very consistent...

Look harder at the schematics in the current manual http://www.msextra.com/doc/general/ms2v3schems.html
See where PT6 and PT7 connect from the CPU to the stepper chip, how they route out of the stepper chip and what the mainboard jumpers are.

James

[Wergilius]

You need to look for the actual error message that comes before this. The part you posted means that linking failed, but doesn't show why. (Most likely is that you overflowed a section.)

At first i did not find it but after a compare of the compile result I think this is the error:

At this long row (shortened a bit here)
/usr/bin/m68hc11-elf-gcc -g -Wall -Werror -Wsign-compare -Wshadow -O -fomit-frame-pointer -m68hcs12 -mshort -msoft-reg-count=5 -mauto-incdec -fsigned-char -Wl,-defsym,vectors_addr=0xff80,-m,m68hc12elfb,-T,m68hc12elfb.x,--no-trampoline -o megasquirt2.elf ms2_extra_main.o ....... interp.o

/usr/bin/../lib/gcc-lib/m68hc11-elf/3.3.6-9s12x-20120502/../../../../m68hc11-elf/bin/ld: megasquirt2.elf section `.text3b' will not fit in region `text3b'
/usr/bin/../lib/gcc-lib/m68hc11-elf/3.3.6-9s12x-20120502/../../../../m68hc11-elf/bin/ld: section .text3c loaded at [00100000,0010215f] overlaps section .text3b loaded at [000fc000,00100007]
/usr/bin/../lib/gcc-lib/m68hc11-elf/3.3.6-9s12x-20120502/../../../../m68hc11-elf/bin/ld: region `text3b' overflowed by 8 bytes

What does that mean, that it's running out of memory space?

[jsmcortina]

As it says - region text3b overflowed. That means that the code you added is more than the space available in that section.
Section sizes cannot be changed (defined in memory.x)
Functions are assigned to sections in the prototypes section of ms2_extra.h e.g. FAR_TEXT3b_ATTR

James

[slyrye]

James:
I have checked the processor pin connections continuity with IAC1 and IAC2 and confirmed that the schematics are pretty much very consistent...

Look harder at the schematics in the current manual http://www.msextra.com/doc/general/ms2v3schems.html
See where PT6 and PT7 connect from the CPU to the stepper chip, how they route out of the stepper chip and what the mainboard jumpers are.

James

James I believe there's nothing faulty with the schematics, what i'm having trouble with is, wrong status port reporting on Tunerstudio with PT6 and PT7

[jsmcortina]

James I believe there's nothing faulty with the schematics

and I believe you are wrong. On MS2, IAC1=PT7 and IAC2=PT6

ms2 iac schems slyrye.png

I've spent enough time on this report already and won't be looking at it again unless someone else believes there's a problem.

James

[slyrye]

James I believe there's nothing faulty with the schematics

and I believe you are wrong. On MS2, IAC1=PT7 and IAC2=PT6

ms2 iac schems slyrye.png

I've spent enough time on this report already and won't be looking at it again unless someone else believes there's a problem.

James

I see it more clearly now....!!! PHASE1 and PHASE2 connections haha

[Spilly]

After upgrading to the current stable version and the beta 3.3.2 I have a bunch of gibberish for my EAE tables.

I am upgrading from 2.5... something. I'm not sure how to find this but I can upload my current tunerstudio folder if necessary.

I opened the example tune included with the beta, changed the control algorithm to ITB, and enabled AE EAE. I can see logical values for the EAE tables. However, they appear to be exactly the same as the speed density values. I'm assuming sense the % levels for ITB are user configurable, the EAE values are not going to be anywhere close.

Is it possible to use speed density for EAE? The previous firmware's EAE seemed to be pretty close.

Also, I believe priming pulse and few other startup and related tables contained gibberish as well. Sorry I do not remember which ones. Settings flooded engine. May want to include looking over every table in the readme notes. I did not see any mention of it.

[jsmcortina]

After upgrading to the current stable version and the beta 3.3.2 I have a bunch of gibberish for my EAE tables.

I am upgrading from 2.5... something. I'm not sure how to find this but I can upload my current tunerstudio folder if necessary.

I opened the example tune included with the beta,

Don't do that.

Load the firmware.
Working online (i.e. Megasquirt powered up and "talking to" TunerStudio)
Load in the MSQ your were previously using.
Review settings mentioned in the README files.

Ka,es

[Spilly]

I'm about 98% sure that's exactly what I did.

I did receive lots of errors after updating my tunerstudio project to the updated firmware and it auto opened the current tune msq.

[jsmcortina]

I just made a fresh project with TS 2.5.17 and MS2 3.3.2 b5, and had no issues. That suggests it's not TS or PEBKAC.

Is this resolved with TS 2.5.20 ?

James

[jsmcortina]

After upgrading to the current stable version and the beta 3.3.2 I have a bunch of gibberish for my EAE tables.

Yes, the EAE curves in the MSQ you posted are garbage.
Please post your previous working MSQ.

James

[jsmcortina]

Still exists on this Beta release... Under Output IO testmode "LED15/SprkC and LED16/SprkB" test output is swapped...!

Should be corrected in pre-3.3.2 beta 7 which I just released.

James

[Fabius72]

Given up waiting any feedback on the significant change in pre-3.3.2 beta1 which was a change to the "Toothed wheel" decoder for oddfire and odd cylinder count engines.

James, I'm sorry for the late, but I have just closed the engine and made some changes on the MS2, because of the lost sync trouble while cranking.
I've finally pulled out a tooth to have a 36-2 wheel and now the flywheel is also balanced.
I've changed the VR conditioner with a MAX9926 circuit.
I also changed ignition and injection drivers with bip373 and vnp5n07, so I need to find the new dead time for injectors.

I have tried on the bench this new firmware at various RPMs with CrankWheelPulser and all seems ok.
Tomorrow I will do some logs to check crank&cam, tooth, and composite, but only during cranking (without fuel injectors now).
I would like to confirm that all is fine about "Toothed wheel" decoder, so how can I find something wrong? What I exactly must do?

Thank you.
Fabio

James, you probably didn't see this post.
20 of this month I will come back home, so I'll try something to help and give some answer about odd fire code... but please I need that you tell me what I must do to check.
W/O injectors, I have tryed crancking about 50 times discharging three times the little battery (20Ah) and I have not had no one lost sync. With the old firmware, old 36-1 wheel and standard V3 VR circuit, 90% of crancking made had one (only one but really annoying) lost sync. In this way I don't know which the problem was, but I'm happy is gone.
Are there some particular tests I could make?