legacycentral bbs

mike, tell your brother to ask on the mp3car.org forums for software. There were people who got ride of their intrument cluster in favor for a carputer, an lcd, and something to interface with the ODB-II port.

Cool. I'll point him there. Thanks!

I hacked together a wiring harness and DB25 connector and got it working successfully on a Thinkpad 380ED (P166MMX).

I need to show this to Trey Cobb on Friday

Nice, Dave.

Cool Dave!

Hey Vikash,

I was wondering if it would be difficult to have a peak hold feature? Maybe a second and third figure that shows the highest / lowest values for the various readings. Nothing fancy like remembering them when you switched sensors.

But say for instance, I wanted to make sure I was getting the proper amount of boost (yes, I'm still wondering). Having it show the highest boost reading during a run would be pretty useful. Obviously I've got a 2nd person along, but if the values change quickly enough, they might not notice it.

I dunno, just wondering through my fingers....

Oh! I managed to get my laptop running off the cigarette lighter. I finally have a use for this 486 doorstop.

Something to work on... it's a little weird because every so often it gets a spurious weird reading because the timings don't always sync up right, so you might get artificially high peaks.

That is weird.

Oh! While I'm here, what does the fuel trim represent? I was at 0% the entire 10 minute drive I did. Of course, my O2 sensor isn't plugged in right now.

It represents how much the ECU is deviating from the static map at the time. I believe positive numbers mean the ECU is adding fuel and negative mean it's removing fuel.

Without an oxygen sensor, fuel trim will always be 0, since it's the oxygen sensor that provides the ECU the feedback with which it determines how far to deviate from the static map.

I really don't know too much about assembly language, but might i suggest that in order to make this some sort of datalogging program, having it output to Excel would be great. That way, we could make graphs based upon the data collected easily. Thanks for this little program, i can't wait until finals are over so i can build the connector.

You're welcome.

I think it would make sense for any logs to be CSV (comma-separated values, ASCII text with commas between the fields). Any self-respecting number crunching software should be able to read it.

Oh, and I've discovered that this tool won't fully work on at least some early 1991 models; the injector pulse width mode doesn't work. I think I've figured out how to make it kind of work, but the granularity will kind of suck.

Mike, did the 486 work right? I don't have a good idea of what the minimum processing requirements are. What speed is the 486? SX or DX? Clock multiplied?

Hey Vikash,

I had to run outside to my car to check.

Its a 486DX4 100. I don't think its possible to change the clockspeed. Looks like a run of the mill, non ECP/EPP parallel port. If we can figure out how to underclock it, I'll see how low I can get the laptop to run, and let you know what happens.

I haven't noticed any problems with it pulling and displaying values, although I haven't had the engine past 4000 rpm with it connected.

Its strange that it doesn't work properly on early 91's. The ECU code must be different then.

Cool. I think that computer is the slowest yet tested, although its clock speed is the same as my Libretto's. If it works at all then it should work all the time though.

This setup doesn't require a special (bidirectional, ECP, EPP) parallel port.

The ECU code is indeed different. I believe with 90-91 models the Select Monitor actually directly measures the duration of the injector pulse instead of being told by the ECU. I think I've found another place in the ECU I can read pulse width from but it just isn't going to be quite as precise.

is that for manuals only or autos only or both? Aren't the injectors for autos and manuals different within those years?

It's true for all 90-91 models. The injectors are different but that isn't really relevant to this issue.

Okay, new version up. I think this one'll work on non-turbos as well as turbos, and on 91 turbos too. Test it out and let me know how it goes!

Just tested the new version Vikash.

Some observations: ROM ID isn't shown until you switch to another value and then go back. Previously it showed up right away. This isn't critical, just something I noticed.

When I turn my A/C on, the values for A/C relay and Rad Fan stay at OFF. Now, I know my A/C clutch isn't engaging, so maybe that has something to do with it. But I thought I would at least get a change in the Rad Fan value. Or is that when the ECU triggers the fans because of the coolant temperature?

Otherwise, it looks good! I like the addition of the new values (IAC, boost solenoid, etc).

are the fans actually on when you hit the ac button? Is your ac actually on....or does your compressor cycle?

Hmm, I'll try to fix that ROM ID issue. I changed how it works; in the first version it just showed you the ROM ID at the beginning; in this version it's considered just another one of the parameters so you can cycle back to it and stuff.

Just so you know, any key other than Space, BackSpace, or Escape will re-initialize the parameter, which is kind of equivalent to going to another parameter and back.

The Rad Fan bit ought to turn on whenever your radiator fans are on, for whatever reason. Are they on when you hit the A/C button?

I'm not sure how the A/C relay bit works... my A/C is inoperative so I can't test it. I believe, though, that it's on whenever the A/C is actually operating, which should be any time the A/C or Defrost is operating and you're not near WOT. I don't know how a malfunctioning A/C clutch would affect that.

I tested the non-turbo functionality by putting non-turbo ECUs in my car... it's kind of cool how adaptive the ECUs are. For example, despite having the wrong kind of IAC valve and much bigger injectors, a 1990 Hitachi ECU was able to learn to idle my engine. It did end up bumping IAC duty cycle up to like 90% though. And a 1993 JECS ECU learned to idle my engine despite having the wrong MAF sensor and bigger injectors. It pulled ignition advance back by like 30 degrees though.

Legacy777 wrote:are the fans actually on when you hit the ac button? Is your ac actually on....or does your compressor cycle?

Yup, both fans come on when I hit the A/C switch. I haven't done any real troubleshooting with it yet, just something I noticed while testing Vikash's program. I'll try to get the car out on the highway this afternoon to see what happens with the Rad Fan bit after the ECU triggers the fans itself.

As for the ROM ID thing, it was also something I noticed. Its not critical, as I know the connections to the Select Monitor port are correct when I see the ROM ID screen come up.

That is cool that they're adaptive. It makes me wonder now why some cars have such trouble idling and such, if they're able to learn so well.

'kay, give the new new version a try, Mike.

I don't know what the deal is with the fans and A/C. I'd be interested in seeing what a real Select Monitor reads.

ROM ID comes up right away again Vikash.

Did you make any other changes? I didn't notice anything else different.

Still no change in the A/C & Rad stuff, but I want to do some more investigating with my car first. Hopefully someone with A/C that actually works properly can give us some insight into whether the display changes or not.

All I did was add 2 lines to make the ROM ID come up.

What's your ROM ID, Mike?

ROM ID is 72.36.22

Just got back from a drive. The Rad Fan bit does change when the ECU needs to turn the fans on. They must be wired to turn on with the A/C switch. I'm still not getting a change fromt the A/C relay, so I'm going to do some checks on the electrical system. I'm hoping its a blown fuse.

My timing correction reads 0, but I just reset the ECU this morning. Does anyone know how long it takes to start advancing?

I noticed my injector pulse width was 35-40 ms, but Josh's scan of the FSM says 2-3.5 ms (at idle). Did you do some formatting of the values?

Finally (wew!), my boost control duty cycle read 1.56% the entire drive. The FSM says 70% at idle. So... Is my duty cycle actually sitting at 1.56% the entire time, or did you transpose some digits Vikash? Probably the best thing would be to have someone else check theirs, and report back.

I'm just going to quickly do some conversions on the torr values for boost, and I'll finally be able to see whether my boost/vacuum gauge is borked or not. Edit: Damn... 280 torr really does equal just under 6 psi. Stupid car

Thanks again Vikash!

Oops. I messed up the injector pulse width on 92-94 turbos when I added pulse width for 90-91 models. 35-40 milliseconds would be insane! It's actually giving you a number 16 times too large.

Okay, I fixed it. Yay, the third version today!

I don't know how long it takes for timing corrections to start happening... I'm not even sure what exactly the timing correction parameter represents. I mean, it's possible that it only shows knock-induced correction and not long-term learning. I have no idea. Hopefully after I get enough bugs ironed out of this scan tool we'll get input from a lot more people.

I'm recalling the info in this paragraph from memory: 1.56% is the lowest number you should ever see for the boost control duty cycle; when the ECU uses that low a duty cycle the solenoid never actually clicks "on" at all. An the FSM is wrong; at idle it should be 1.56%. Just off idle, though (if you blip the throttle), it should rise up to 70-80%. It's possible that the ECU has decided to turn off its boost control for some reason... I've seen that happen on my car before, and haven't figured out why.

Do you still have stock boost control setup? What boost level do you reach?

Thanks for the help testing this.