Memory dump of ECU - Page 37

b3lha · #**541** 12-12-2007, 03:10 AM

Quote:

Originally Posted by Calum

Your patents in pdf form should be attached. Pre-emptive your welcome too in case I don't wander across this forum for another week or two.

If you use it, please make a donation, its really really cool. I used to pay money for a service like this-

http://www.pat2pdf.org/

That's great. Thanks Calum.

Dessertrunner · #**542** 12-13-2007, 03:10 AM

I can't get Upnorth and I can't be sure that he tested the unit before he shipped so I not sure what next other then to try a different USB cable which I will do. The cable wiring is differnt on the output side to the car plug then in the circuts in the write up. Is there any way I can test that the correct pins are in the right plugs?
Tony

b3lha · #**543** 12-13-2007, 04:23 AM

Quote:

Originally Posted by Dessertrunner

I can't get Upnorth and I can't be sure that he tested the unit before he shipped so I not sure what next other then to try a different USB cable which I will do. The cable wiring is differnt on the output side to the car plug then in the circuts in the write up. Is there any way I can test that the correct pins are in the right plugs?
Tony

That's the first time I've seen one of Upnorth's boards. On the left is a ready made RS232 to TTL board, on the right he has added Calum's upgrade board and a voltage regulator to power it all. The wiring looks sensible to me, but I can't really see from the photo.

I have a suggestion for you Tony: Maybe the little white potentiometer needs adjusting.
Do the hex comm tool thing as before.
Send the hex command: 78123400
turn the potentiometer a little
Send 78123400 again
turn the potentiometer a little
Send 78123400 again
turn the potentiometer a little
Keep doing this until you get a response from the ECU. The correct position should be about 2/5 of the way round.

EDIT: An easier way to do the above is to run the SelectMonitorDump program from the VWRX guys. Click start and then turn the potentiometer until you get data that isn't zeroes.

There seems to be some LEDS on the top left of the green board. Do they light up at all?

Dessertrunner · #**544** 12-13-2007, 12:06 PM

No don't get any lights on the borad at all, do get lights on the USB cable when I send a code to the memory. Would it be correct to say I need to get some lights on the RS232 before on adjust the pot? Upnorth was suppose to supply with a cable but it never turned up.
I will have a go on the weekend again. Thanks for your help.
Tony

b3lha · #**545** 12-13-2007, 04:32 PM

Quote:

Originally Posted by Dessertrunner

No don't get any lights on the borad at all, do get lights on the USB cable when I send a code to the memory. Would it be correct to say I need to get some lights on the RS232 before on adjust the pot? Upnorth was suppose to supply with a cable but it never turned up.
I will have a go on the weekend again. Thanks for your help.
Tony

Yes. I would expect to see the lights on the RS232 board if everything is working.

It sounds like the board may not have power. Can you put a voltmeter across the red and black wires to check that the car is supplying 12v and ground.

Dessertrunner · #**546** 12-13-2007, 06:43 PM

Phil when you say red & black wire do you mean on the unit? Is it correct that the 12 v is coming from the car?

b3lha · #**547** 12-14-2007, 02:35 AM

Quote:

Originally Posted by Dessertrunner

Phil when you say red & black wire do you mean on the unit? Is it correct that the 12 v is coming from the car?

Yes. The power and ground come from the car. The ignition key must be in the on position.

On the right side of the picture is a three pin voltage regulator. The middle pin is ground, one side is 12v input from the car and the other side is 5v output to the circuit. You can test this with your voltmeter.

Dessertrunner · #**548** 12-14-2007, 02:09 PM

Thanks Phil I will give it a go in the next or so and let you know.
Tony

Dessertrunner · #**549** 12-16-2007, 11:39 PM

Okay Phil checked out the unit as you suggested and all is good so I ordered a better quality cable and I will see how that goes and let you know. If that fails I will move a desk computer that has a serial port out to the car and try that.
Tony

b3lha · #**550** 12-17-2007, 11:19 AM

Quote:

Originally Posted by Calum

I worked out what combination of peripherals the 7791 has. Its got two UARTs, but 3 less timers and 1 less PWM than the 7790. Everything else is the same. I've uploaded the changed disSVX.cpp code and new disassembly (SVX.txt). The source code lists the correct arrangement of registers and interrupt vector table.

Calum, if you're still here.
Could you shed any light on how the various timers work? I have a theory that Timers A1 through A6 are used to generate the duty cycle for the 6 injectors. (Does that sound reasonable to you?) I am having a difficult time reconciling the instructions in the code with the information on the 7733 datasheet.
Phil.

Calum · #**551** 12-17-2007, 09:22 PM

Thats a good question. The problem is that the 77xx series had a pretty long run as microcontrollers are sometimes apt to do (its actually still in production btw, nevermind the EOL on it), but Mitsu unlike some manufactures changed the peripherals around many times during its run. So comparing how the peripherals work on the later dudes like the 7733 isn't that useful. This is about as close as I've gotten to a datasheet on one of the models in the same era as the 779x guys-

http://www.calumsult.com/calumsu/dis...ocs/M37700.pdf

The brief info sheet on the 7791 actually does tell you a bit too if you study it, although not specific down to how the different registers actually work. I also know I've got the registers 100% correct for the 7790 and 7791.

The 7700 datasheet does describe timers that are of the same type as the 779x uses, and the 7791 datasheet names its different timers. Thats about as good as I can do.

On figuring out which timer goes with which actual function for the car, the easiest way is to just back track the circuit. While your at it, make a spreadsheet of what every pin on the mcu is connected to. The way I do this is use an engine simulator I've developed (aka 'the simple engine simulator') so the ecu operates normally, clip one oscilloscope lead on the signal I want to hunt down (like injector #1), then probe the mcu until I find a pin thats clocking out the same signal. With the sensors just wiggle the input signal and look for a like wiggle at the ecu to find out what ADC channel is hooked to each sensor. You can use the same idea to work out a concept of what the large asic next to the mcu on the subbie ecus is doing.

Building a bench engine simultator is soooooo worth the development. It makes everything from hunting down how the ecu works to developing software that works with the factory diagnostic port much much easier. Everything except the timing signals (crank angle sensor, distributor, or whatever your car uses) is easy. If you can figure out what the pattern needs to look like I can adapt 'the simple engine simulator' to support your ecu (well, its missing two injectors, but thats easy to ham up).

Of course, you've got to find a donor wiring harness to hack up, and a spare ecu (and tcu etc if your working with those too). Easy enough when your working with a Nissan Sentra (Sunny in the UK), but might be a bit of work for the SVX.

Calum · #**552** 12-18-2007, 09:30 AM

You know what would be nice is to have all the known bit fields in the disassembler and let it autocomment when the fields are being modified.

b3lha · #**553** 12-18-2007, 10:17 AM

Quote:

Originally Posted by Calum

You know what would be nice is to have all the known bit fields in the disassembler and let it autocomment when the fields are being modified.

Thanks for the datasheet. I haven't had time to study it properly, but at first glance it seems to fit better with the code I'm looking at.

On my linux box I have a directory of "clean" code, split into separate files for each individual subroutine. Then I have a "sed" script that runs through each file, comments it and writes it to a "processed" directory. That way, when I mislabel a variable, I don't screw up the clean copy of the code. I can just edit the script and rerun it.

I may look into building an engine simulator sometime next year.

genzzel · #**554** 12-19-2007, 06:30 AM

Well--ordered cable from UpNorth on Oct 27th and paid with PayPal. Still have not got cable although he said all were shipped in November. Have tried to Pm, call cell and file complaint with PayPal--No response from anything!!!! Had to file claim with PayPal on 14th--still no response from this guy. All I want is my cable--What a chump!!!!!! Never thought I would get screwed here

Hocrest · #**555** 12-19-2007, 10:58 AM

VRG over on the Legacy board made up something new...
http://bbs.legacycentral.org/viewtopic.php?p=236151