Based on information posted by Doctor J... specifically, this:

http://temp.corvetteforum.net/c4/doctorj/carwb.htm

and this:

http://www.dataq.com/products/startkit/di194rs.htm

I have ordered one of the data loggers to help tune an AE stumble I have when cold. The ALDL data stream just isn't fast enough (only about 1.5 frames a second) to see what's really happening over a period of about two seconds...

I was wondering if Doctor J (or anyone else!) can comment on any experience they've had tuning AE this way? I'll be hooking this up to a DIYWB, the TPS, and MAP. Seems like it should provide a great way to see what the AFR is actually doing over a brief period, not to mention provide a lot better sampling rate for WOT tuning too and seeing real trends while crossing the VE cells as the RPMs wind up.

 

I just use one of Bruce Roe's LED displays.
Depending on where the sensor is, there is some delay, 4' can make a difference. So if you want to do that, and be *real* time, you'll have to mount the sensor right at the exhaust port, and that can generate a bunch of headaches in itself.

The actual AE AFR is going to vary under about all circumstances, since the fuel is a fixed value in most masks and not corelated to a TPS vs RPM, or MAP vs RPM.

 

This is one thing I don't like. I would also like compensation from CTS input. I've had to tune the choke AFR down some and use the open loop AFR to help "hide" slightly lean accel tip in until engine is warm. Since this only is an issue with below freezing temps it really isn't much of an issue 98% of the cars life. But for that 2% it really is a shame. Being limited to a darn onboard ROM ecm stinks .
Now if only I could convince my friend that works for Honda to give me the documented source code for those new 64mhz computers.....
BTW, happy new year to all (and their cars).

 

The 8746 has no CTS correction for AE?.
Well, just another reason to fab up an adapter and try a 749.

Having actually done a ROMless from 0 calibration, you'll want it to be plug and play with another ecm, cause the hours to get the intial tune up and running can take MANY WEEKS.

Too bad not more people want to get involved with a Source Code project. It can make all the difference.

 

The data will help a lot. Once you have an area logged import it into Excel and graph it. Makes it a lot easier.

On another note your '730 (the L98?) should be giving you 10-11 frames per second.

RBob.

 

Check the table at location $D332.

Source code is where it's at, no doubt about it. Seems an AE vs RPM table is useful too. I just add functionality as required.

RBob.

 

91L98Z28-

Glad you liked the page. Here are a couple of construction notes, if you are interested, to be used with
your own good judgement:

My Dataq is a 2+year old model (haven't bought one of the newer ones yet). Old 8-bit design allowed
me to choose the input voltage ranges, so I have 1 channel @ 0-5v and 1 channel @ 0-10v for the WB &
Tach inputs. That takes maximum advantage of the ADC's response. With 2 channels active, I get ~120
samples/sec. Not sure if it's limited by the s/w or the serial device Dataq put on the board, but the more
channels I use, the slower the sample rate.

I use shielded signal cable, trying to minimize noise pick-up. I also LOVE fuses - so I put them EVERY-
WHERE when I add circuits to the car.

The rig I built turned out to be surprisingly fast - for example, below is a misfire situation I happened to
capture one afternoon. Chart shows the start of an uphill pull, when one plug wire leaned over against a
header pipe & began to burn through.

You can see what happened to the WB output voltage when one cyl lost spark - the jagged line was recorded
between ~2 & 2.5 K rpm. The misfire was also picked up on the RPM trace to a smaller extent. Samples are
~0.008 sec apart with 2 channels running, so you lose the ability to see individual spark events somewhere
above 2 K rpm (V8)...

Thus, looking at AE is no problem at all. It sticks out like a sore thumb from the normal (flat) stoich voltage
trace in part-throttle driving. Using a 0-5v scale, the 2.5v stoich WB output is always on the center (horizontal)
line of the chart, making it easy to pick out the pump shot, DFCO, etc.

I haven't tried to record the MAP or TPS signals with Dataq. MAP is probably not a problem, since it is an
active electronic output. TPS I'm not sure of - I don't know the input impedance of the Dataq off the top of
my head. As long as it's high enough not to interfere with the TPS (voltage divider) it should be OK - but
you will want to check BOTH before you head out into traffic. Either signal might(?) need to be buffered to
avoid horsing up the ECM inputs.

SAFETY of what you are testing is your responsibility. Note that I chose to record two signals that are
isolated from the ECM. Both units can fail, and not affect operation of the engine. Just something to
keep in mind, word to the wise, etc etc.

If you have other questions drop me a note off-line.

DrJ

 

Log of spark misfire:

 

Yikes. Really? I am using Diacom. I must have something mucked up or misconfigured, then... But I get about 100 records per elapsed 1 minute period in the output log. 10 frames per second would be awesome. I will have to see what I'm doing wrong, or if this is just a limitation of diacom. maybe it's time i switch to the moates software once and for all.

 

I use Datamaster and it runs 10 frames per second all the time.

91L98Z28

Are you in so-cal ? if so you can try my setup


Jerry

 

Nope, sacramento.

I think it must be either a configuration option i have in diacom is set wrong, or perhaps diacom intentionally limits the output so that it will work on low end hardware. i'm booting up that notebook now to see..

 

well, i looked in the diacom software. it does have a "sample frequency divisor" but I had it set to 1, so it shoudln't be reducing my output at all. the notebook is a piii 650mhz so it should have plenty of power to log the full stream of data.
guess i need to look into other options than diacom.

 

I know Diacom will run faster, But I don't have it to tell you how to set it up. On Datamaster you set it to 10 if you set it at 1 you get 1 frame per second.

Jerry

 

Not sure if I could contribute, but if it would be for a project that would be compatible with a 2 injector tbi setup, I would be interested. I've got a background in software engineering, but mostly with higher level languages. I've done some assembly, and would be happy to learn more.