TeamSwift

Richard is right, the only thing the ECU uses the VSS for is to maintain a higher idle when the car is moving faster than 20-30km/h. A disconnected VSS does not have any influence on igniton/fuel maps or open/closed loop. I know this because my VSS-cable broke 3 weeks ago and that didn't have any consequences rather than throwing code 24. I even took the car to the track last weekend... no issues.
I also recall running the car without the VSS when I had the emulator connected, to confirm that speed indeed had influence on idle-speed... nothing else changed.

_________________
MK2/3 PARTS FOR SALE - EVERYTHING MUST GO!

1993 Swift Gti

When I get a chance I'll stick that in and see what I get.

On deceleration the ECU applies an ignition timing correction and increases the ISC to simulate a 'dashpot effect'. The stock ECU is pretty robust and filters out any glitches due to contact bounce.

Very interested in the development here, excellent work! I have been a member here for ages, but assumed no tuning would ever come about for the Swift. I am an experienced tuner, and Auto Sci instructor, I have worked with all the major software (Hondata, AEM, Cobb etc) and run the site www.e-tunez.com.

If you can get me access to the ECU I can help find free HP and MPG for sure, especially with mods. Currently own a 97 1.0L Geo Metro and an 89 Swift GTI, I had just planned on using MS for the GTI, but would much rather play with the ROM.

I have a ROM burner and most of the electro toys... but my time has disappeared as my family has gotten older, and I don't like guessing at hex values to find tables etc, and the work you guys have already put into this takes the leg work out of it. Again, excellent job.

RE a knock sensor, wiring in a knock sensor is fairly easy to do, and spotting knock from knock volume is fairly easy to do as well, if you can get to the tables and the code there is definitely some HP on the table from all this. Most of the NA Hondas I tune I can squeeze another 3-5% WHP just from tuning a current I/H/E setup nevermind once cams are in. For cammed users like XFi or G13B race cams, we can get more MPG and more HP respectively with access to fueling and ignition. Then there is launch control coding... oh the possibilities.

Keep up the good work,

Steve

_________________
There is no replacement for displacement... is an irrelevant statement at 7500RPM and 21 PSI.

I just got an F5 ECU to repair, the chip in it has the ID SKBB113, I don't recall seeing that before, anyone recognise it? I finally got round to ordering the parts to repair my I2, it seemed like a good time to get it done while the tester was set up for a Swift.

Rhinoman, just a question, I have been going through the FSM lately doing some troubleshooting of my own, and the FSM lists a few sensors that would trigger a 'fail safe' or 'limp mode' operation. One of the items is the CPU in the ECM..

How is that possible? Does the UPP do that much of the heavy lifting??

_________________
"Don't argue with stupid people. They just bring you down to their level and then beat you with experience."

Now that I've got a moving car (transmission rebuilt and back in it) and I have my knock sensor wired up, I've had some time to play with my I4 Cultus ECU.

The first thing I noticed driving around is that I was running rich at WOT. I calibrated my wideband again just to make sure, and there is definitely a different scalar somewhere in there, because my whole tune that was pretty close to perfect on the I6 is off now. Part of the issue is that I seem to have more space to play with--before, for whatever reason, even at WOT the ECU never got beyond column 13 along the load axis. Now I'm running all the way to 15 (because the first row is 0...) like you would expect.

And I'm having to subtract fuel like crazy at the high load end of the map, while idle and part-throttle cruise hardly changed at all

_________________

Spanish Inquisition Racing chip burning service--build yourself a custom chip!
http://www.teamswift.net/viewtopic.php?f=17&t=57216

There is a safety processor that runs the ECU if the main processor fails, it is brought in by the hardware watchdog, there are electronic 'switches' that select between the two processors. The UPP handles all the main timing functions like the injector and ignition timing as well as the camshaft sensor decoding. The safety processor has very crude injection and timing algorithms, it also pulls the CEL on permanently.

I wonder if it might have something to do with the fact that a Cultus engine has higher compression. Suzuki tuned the non-cultus to run on 87 octane, and since the cultus is higher compression perhaps they tweaked the computer figuring a higher octane fuel would be used because of the higher compression... Just a thought.

_________________
"Don't argue with stupid people. They just bring you down to their level and then beat you with experience."

Well yes, not to mention the intake and extractors.

However, I had thought that by changing the fuel maps, ignition maps, and scaling tables that I had isolated "the tune" from other aspects of ECU operation.

But I must be missing something because something else in the I6 code is definitely limiting how much of the table space is used (always hitting 12-13 at WOT rather than going all the way to 15 with the I4 code, for example). It's possible the enrichment is due to the altered table scaling, but one way or the other swapping just "the tune" part of the code from my I6 into the I4 runs the car very rich at tip in and sustained WOT even though idle and part throttle driving are almost the same.

The experiment of the day was to flash the I6-based code that I had been running on my I6 that I knew was more or less spot-on for my car. The result? My car ran exactly like it did with the I6

That includes, believe it or not, the issue I had been having with the car taking another 15-20 minutes of driving after it was already up to temp before enriching properly with the application of throttle. The problem was in the code all along

The I4 code enriches properly immediately. I have a ball-park tune now and the car works great, but it will take a little while to dial everything in. Whatever else that changed in the code (you can bet I'll be looking for it) is definitely worthwhile

_________________

Spanish Inquisition Racing chip burning service--build yourself a custom chip!
http://www.teamswift.net/viewtopic.php?f=17&t=57216

Who sells the Garndi chips these days? I have had someone contact me wanting a turbo chip so who do I pass him onto?

Is this guy Garndi?

http://www.teamswift.net/viewtopic.php?f=15&t=59089

When I saw that post I PM'd him about working out a deal to sell his maps "under license" and pay him a royalty for any chips I sold with his maps, but he hasn't gotten back to me

In the meantime I've made some more progress in isolating the code that runs the knock sensor from code that controls ignition and fuel. I am now able to use the tune I spent so much time dialing in on the I6 and bump the timing safely without having to start over again. I did some further testing of the knock sensor with some interesting results: the sensor can't isolate loud valve train ticking from engine knock. That's not necessarily a bad thing, it was just interesting to see that the nasty ticking I get when I start the car after it has sat for a month or so triggers the knock sensor to drop into the other map until it goes away . I also found I was able to induce (inaudible, at least to me) knock by running extreme timing values, so it's good to know the sensor and ecu are doing what they are supposed to do

I've also figured out at least part of the issue of failing to enrich properly with throttle until > 15 min. of driving, and I'll fix it on all future chips I sell too.

_________________

Spanish Inquisition Racing chip burning service--build yourself a custom chip!
http://www.teamswift.net/viewtopic.php?f=17&t=57216

There hasn't been a lot of juicy action on TeamSwift in a while.

While Rhinoman will probably laugh at me if he ever visits here again, the confusion I had with trying to isolate "the tune" from the other code that seemed to be running me unexplainably rich using the Cultus code on the Cultus ECU was due to the way he had put the table in to the .xdf that I started with. He correctly identified the maf table at address 0800, but made a 2x16 table of 8 bit numbers. In reality, it's a one row table with 64 columns of 16 bit number

Thusly, when I was trying to use the MAF table for my non-cultus engine, not just because I don't have high compression pistons, but also because I didn't want to have to retune just to use the knock sensor, I was only changing a small part of the table--the high load part of the table did not get edited.

If anyone is interested, here's what the two maf tables actually look like--the Cultus curve is in red. The MAFs are identical, but different engines have different flow characteristics...



Now it's time to start playing with some Nissan MAFs...

_________________

Spanish Inquisition Racing chip burning service--build yourself a custom chip!
http://www.teamswift.net/viewtopic.php?f=17&t=57216

Thats sweet! If you've isolated that data, then we can start substituting MAFs and remove that giant air restriction..

_________________
"Don't argue with stupid people. They just bring you down to their level and then beat you with experience."

I found it was worth the time to fabricate parts to convert to MPFI and standalone. Hacking the ECU took a lot more time than I spent in the shop making wiring harnesses and fuel rails.

But, with that said, I hope you get it working

_________________
Ofensus Racing
Rally Racing Geo Metro

You did the right thing, but you had a lot of factors working against you, not the least of which was working with a late model 3 cylinder and trying to go turbo. There aren't enough people tuning the three bangers, and though I'd like to break into the market for chipping them myself, there may be a reason nobody else sells one. I've looked at some photos of 3cylinder ecus on line and some of them have eproms and some do not--some are reflashable and others???? ECUs for essentially the same engine look nothing alike inside.

You would have gotten further "hacking" an ecu intended for a more similar 3 cylinder turbo in the first place, like a Canadian model, but by that time you are already monkeying with a wiring harness, so yeah. Not to mention that they are speed density to begin with, so less tolerant of a less than perfect fuel map, there are separate maps for EGR, and so on.

The G13B ecus already have a long history of chip tuning, so there really isn't a question of if it "will work". I've been playing with the fuel and timing maps on my GT successfully for years now, though I admit there are some limitations I wish I didn't have to work around (the most severe being the limited tuning space at high rpm ) How much extra power could I make by going to sequential injection or coil-on-plug ignition? I don't know, but certainly some. On the other hand, I got my Ostrich for $80 used and the only thing I needed to control my ECU was a socket.

Knowing the MAF curve does open up another level of possibilities too

I'm not even going to bother with speculating about what will work, but I'll definitely post back when I get some results.

_________________

Spanish Inquisition Racing chip burning service--build yourself a custom chip!
http://www.teamswift.net/viewtopic.php?f=17&t=57216