J-series Acura/Honda

[Fred]

EDIT: Skip to end to save time. Will run correct test soon.

I just spent half an hour poking around my log of your pattern and the code and I'm confused.

The error I get out is 75, MatchedPairNarrowWidePlusMask meaning 3 evenly spaced events followed by the gap and an event, when we were expecting four evenly spaced events. The mask thing differentiates it from other possibilities of the same code. The trouble is that the sync lost on event number is 4... you don't have an event number 4, only 0,1,2,3. The code that does the incrementing reads like this:

Code: Select all

lastEvent = KeyUserDebugs.currentEvent;
KeyUserDebugs.currentEvent++;
if(KeyUserDebugs.currentEvent == numberOfRealEvents){
    KeyUserDebugs.currentEvent = 0;
}

http://docs.freeems.org/doxygen-html/d7 ... tml#l00157

So you'd expect that the only way for it to end up 4 would be to have numberOfRealEvents wrong, however I checked the ppc file and there it is = (6 - 2), so I'll be stuffed if I know what's going on. I'll keep working on this today until I understand the issue and can bench test it effectively.

DOH!!! I found the issue... <blushes> I was running the wrong firmware...

Code: Select all

################################################################################
###################### Initiating Interrogation Sequence!! #####################
################################################################################
Interface Version: IFreeEMS Vanilla 0.0.0
Firmware Version:  0.2.0-SNAPSHOT-83-g836c162
Decoder Name:      MissingTeeth-Crank-60minus2
Build Date:        Sat Jan 28 19:33:12 CET 2012
Compiler Version:  3.3.6-m68hc1x-20060122
Build OS Type:     Linux
Maximum Size:      2064
Echo length:       10

Fred.

[Fred]

MUCH better:



OK, the four effects that you can see are these:

• Curved grey profile at far ends of screen shot: linear change in RPM = non linear percentage change in RPM.
• Linear grey profile increasing towards the centre is the error from my FredStim setup which gets worse with RPM.
• Stepped nature of the blue RPM trace towards 32768 RPM (yes, you read right) is due to the rough RPM calculations in current FreeEMS compounded with the inaccurate input signal.
• The step in the middle is the RPM calculation rolling over and starting again at zero. That behaviour is a little ugly as the scheduling will be totally wrong at that point too, but you're unlikely to reach it on planet earth, so no biggy for now.

This proves my theory about your wheel and the metal structure of it. If you want to manually do what MissingTeeth.h does and then tweak the values until you get a good compromise between low RPM and high RPM that will smooth out your log a bit. Otherwise wire it up and do some injection.

Hmmm, can you please do a starting log or two:

Start logging, start engine, give it a minor rev, stop engine, stop logging.

Repeat a couple of times, binary logs, please. I'd like to verify that you will be able to start it, as it may not sync well at low RPM while cranking due to how coarse it is. Might need some tweaks to get that right.

First decoder tested to the limit of the rest of the system! I recommend you don't exceed twenty thousand RPM for the time being

Fred.

[Peter]

I recommend you don't exceed twenty thousand RPM for the time being

I don't think that'll be too much of a problem. For the cranking log I took out the ignition coil fuse out to crank on it. I also found out that the circuit I got the power out of loses power when it's cranking.

[Fred]

I didn't open the other three yet, but take a look at this from cranking.log :



EDIT: Blue = stable 154 RPM, Green = time tolerance 962 to 1021 range, orange is sync state and red is pulse count.

It clearly shows the perfectly consistent uneven pattern from your wheel. The only question is whether the settings that make it "right" for cranking will still be right at 5k. I can take a stab at correct angles, if you like. Hard coding the array generation for a temporary JSeries cam only decoder is fine by me and looks like it will work perfectly.

Let me know if you want me to have a crack at it, or if you're happy to fiddle yourself. I'll still do you a special purpose decoder soon, but it'd be nice if your RPM wasn't wildly fluctuating with higher RPMs and you could safely run it on fuel now.

http://issues.freeems.org/view.php?id=517

Please name the files .bin for me in future. The cranking log is a success, btw. It syncs up and stays synced fine. I wasn't sure if it would.

Fred.

[Fred]

Might be a bit dozy to start, though, by the looks, however we can loosen the tolerance up a bit to help with that. It's very tight and fussy right now for a -2 pattern.

[Peter]

Let me know if you want me to have a crack at it, or if you're happy to fiddle yourself.

I'll play with it, that way you can spend your time on the more important issues. I should have gotten it wired up yesterday, but my soldering iron died, and there's no way I'm going to use smash connecters on it. Plus I was feeling pretty lazy. It should be all wired in tomorrow. I assume this viewtopic.php?f=54&t=1085 is probably the best thread to look at to setup the initial tune?

[Fred]

Yeah, although it may not necessarily be 100% up to date. Also check out all of the ifdef and elif sections in the initialisers directory and the headers which they use. There are quite a few examples now, and will be more soon, though I'm going to have to cap that and provide some sort of tool or different method as it's currently a maintenance bottle neck.

[Peter]

So I found something I don't like , and I want you opinion on. The wire from the ECT sensor splices to the ecu and the air conditioning. Should the air conditioning circuit be removed from this so that it doesn't ruin my bias?

[Fred]

There will only be one bias source, probably, the other device will just watch the signal, which was an option for you too. Try feeding a 100k into the air con and moving the voltage at your end up and down and see if it drops a voltage. ECT = CHT = Engine Temperature, right?

[Peter]

ECT = CHT = Engine Temperature, right?

Ya.

All wired up. Just need to run a tube to the MAP sensor, and setup some base maps.