Trigger related page for MembersPage/EricN
truck is a 6 cyl. I am firing cop coils and using a distributer mounted trigger wheel. Tooth count is dependent on what sensor I get working
Board is a v3.2 serial number=110
MembersPage/EricN/Trigger/Optical
Going to try this for now.
- so much for that idea.
This page hopefully all garbage. I cant get the input trigger to stop cutting out (happens at any rpm all the way to redline) no matter what I try.
OPTICAL TRIGGER DROPS OUT LIKE VR DOES
Built an optical trigger. 6 teeth, set to coil type ignition. Use rising or falling edge makes no difference. Use filtering or not, makes no difference. Trigger drops out at any rpm (tested up to 4000) is worse at low speed. Loose fuel pump, injectors and ign.
Used signal generater, run at 250 to 500 rpm according to lcd. Fuel relay cuts out in the same manner. Tested signal at atmega128 pin 29, looks clean.
Is there anything config related that could cause vems to cut out? Keeping in mind the test bench has no sensors hooked up at all, just a input trigger and fuel pump relay.
- fuelcut_min_kpa-0
- fuelcut_max_kpa-ff
- TPS > 60% (at cranking, that is RPM / 100 <= cranking_thres )
- firmware version ? an upgrade might help
- 1.0.23, I have used every single version since 1.0.14 and it is identical.
- MegaTune log could help
- strange format (not xls). However, it can be opened as tab separated csv (just some numbers show as dates). Perfer real xls format.
- Odd, it works fine for me. Will do another and repost.
- it can be verified that RPM drops to 0 immediately in 2 cases (first it's just one line! stays that way on the second occasion) Trigger dropped out at least 10 times, only 1 showed up, second occasion was me letting go of the drill before disabling the log, signal drop is extremely quick. Time does not fall back or advance. It's unlikely to be comm error. Engine state changes from 3 to 0 at the same time. Wheel error =0?
- mcd,mct (always needed to reproduce):
- are the GND and (at least 3) GND5 signals all connected ?
- 2 GND and 4 GND5 wires, only load is the fuel pump relay running on an injfet channel.
- InputTrigger/TriggerLog could provide more insight into the pattern the ECU sees (problem at missing tooth, or random).
- http://www.vems.hu/filethingie.php?subdir=EricN I cant understand why I am not able to make this readable and I am at my wits end. I cant even attach a link other than this one. It is terminal.log. The megatune datalog from before is there too, see if I screwed up somehow when I attached the link before.
Upgraded to 1.0.36, it didnt seem to work right, lots of megatune errors but I tried it anyways, trigger problem the same. Found the oldest possible megatune/firmware package (r14 I think) tried it, problem is way worse but I think that was something else. Finally settled on the megatune package with 1.0.30, problem the same. Only thing I was hoping to accomplish was to trash the config so it would make everything gone and I could start from scratch. I wrote down all my config variables on a peice of paper, erased every possible place that I have it saved on my computer and the stupid thing still gets loaded into the controller somehow. I upload the firmware using Avr studio 4, since the batch programs that come with the release just say "no hello from controller, click any key to continue"
Trying to log and read my input trigger using the mde40 command.
1. In the terminal program click "start log" selected the file name as the default "terminal".
2. Type Man then mde40.
3. Start spinning trigger wheel, odd looking characters and symbols start running across the Terminal program screen.
4. Stop spinning trigger wheel, click "end log" in terminal program.
I assume that is correct so far, now how do i make it look like something? If I just double click on the terminal.log file, it opens in notepad and shows some dates then shows the exact same characters I saw in the terminal program.
- need to run perl bin/binary.pl < toothtimes.bin > toothtimes.txt
- but I just get a message saying cant find file: which file ? If you saved to terminal.log, use terminal.log not toothtimes.bin !
- I substitute all the "toothtimes" text for "terminal" but get the same message.
- watch out for the extension. Turn on "view extension" otherwise it's easy to make files like whatever.bin.log or similar. Verify with dir command before you run the perl bin/....
- why don't you upload terminal.log (file manager at DocsPage bottom)? We can easily format it for you and check how it goes
I navigate to that folder through the command prompt, and type it again, it still says it cant find the file.
I have saved the log file in the file thingy under EricN>terminal.log
I now get an error that says "cannot open perl script "bin/binary.pl": No such file or directory. I do end up with a file called toothtimes.txt but it is empty.
- do it from the right directory (same dir as the firmware). Search for binary.pl. (and go up one dir so bin/binary.pl will work)
- See the terminal.log below: perl /svn/firmware/firmware/trunk/bin/binary.pl <terminal.log >terminal.txt
the terminal.log file is in the file thingy, under EricN. I try and click it to attach the link but it just opens as a file with all the garbled characters
Tried something else to attach the link, see if the next one works.
- http://www.vems.hu/files/EricN/datalog200604072201.xls
- http://www.vems.hu/files/EricN/terminal.log
- the same formatted: http://www.vems.hu/files/MembersPage/EricN/terminal.txt From position 134, it seems that times are 4000 .. 16000 (*4 usec) but there are no missing tooth in this (looks like a coil-type trigger with 3 pulses per crank rotation, 4000 * 4 usec = 16000 usec = 16 msec => 60000 / 48 = 1250 RPM at max, does this look right for when the mde40 log was captured ?)
- Trigger is currently only 6 slots attached to distributer as stated above, no missing tooth, set as coil type ignition. 1250rpm sounds like the max rpm I would have been up to.
- the same formatted: http://www.vems.hu/files/MembersPage/EricN/terminal.txt From position 134, it seems that times are 4000 .. 16000 (*4 usec) but there are no missing tooth in this (looks like a coil-type trigger with 3 pulses per crank rotation, 4000 * 4 usec = 16000 usec = 16 msec => 60000 / 48 = 1250 RPM at max, does this look right for when the mde40 log was captured ?)
Note that it's basically impossible to tell from the above (=> chance_to_help=0) how you configured your trigger (you assembled the board yourself, right?). It's probably something small like R181 missing (HALL-pullup too weak, etc) or similar, but it's important to go back to step 0 and check the trigger HW configuration if it's cutting out.
- Board is assembled myself. First assembled with a vr primary trigger, problem was present. Did all the mods suggested to make it work (this is the important thing, it needs to be checked again. Make an exact list of all the parts at the trigger section, near LM1815 chip, preferrably take pics too!) but nothing changed the way it dropped out. Changed to optical trigger to get rid of vr circuit, problem the same. I will go through the board again and try to find a bad solder joint somewhere.
- As for bad pullup, I get almost 5V at the trigger wire without anything attached if that means anything. Also get 5.02V at the jumper with no sensor connected.
- it's hard to tell if you're using VR or HALL hardware setup now (this would be the most basic thing to start with). Really, start an empty page, and write everything, component by component you have there at the time of the measurement. Preferrably also measure at the shorted jumper (because that's what the processor sees. The input has at least some protecting resistors where there can be voltage drop of interest).
- also meausure with a known (eg. 510 Ohm or 1k) resistor between input and GND
- So I assume that there is no indication of the trigger signal disappearing in that mde40 produced file?
- only at the beginning and the end, which can be normal