What to Tune First

[jfpilla]

You stay in SD when dialing in the maf?? I thought youd dial it in with it reenabled, because in SD its disabled.

It's not really disabled. The functions that rely on the MAF are disabled. If you look at a SD log you will see the MAF still reports, but without any trim interference.

You can calibrate the MAF by way of LTFTs, but it's tedious and old school.
Joecar can probably give a more technical explanation.
I should have mentioned this part is CL.

[SSpdDmon]

But how exactly are you logging the IFR? I read:



..I'm just not understanding what you are actually logging? When I open a new map in Scantool for example, I use the MANVAC pid as my row, use the BEN LC1 factor as the data but what am I using for the column, and how is the commanded AFR not influencing the final amount of fuel injected? Or should I be setting the same commanded AFR for an entire section of B3605 (say setting all rows 60-80*C ECT at 12.5 AFR) and then driving the car and hitting as many vacuum points as I can and then using the difference in commanded AFR and actual AFR to set the IFR amounts? (Did I just answer my own question? )

You're logging your BEN as the data point, MANVAC as indicated in the IFR, and a single RPM column (use label ,8000 - yes use the comma). The difference is, the change you need to make to the IFR table is the inverse of your BEN. In other words, if your BEN for the 45kPa cell in your log is 1.02 (indicating a 2% lean situation), you would apply a 2% reduction to the 45kPa cell in the IFR table. This is because the computer is assuming that the injector needs to stay open 'x' miliseconds based on the flow rate provided when really it needs to stay open a little longer. The way you get the injector to stay open longer is to tell the PCM it flows less. Really, that's what you're trying to fix. If the MAF is saying 'y' grams of airflow per second, the PCM calculates 'z' grams per cylinder of air. Then, it delivers the appropriate amount of fuel. Well, if you're 2% lean, then one of the variables is wrong. If the MAF is flowbench tested and R&D'd out the wazoo like most GM guys will claim, I say change a different variable. In this case, that would be injector flow rate.

After your IFR has been constructed off your MAF, then you go back and dial in the VE for that IFR table. This is what sets this type of IFR tuning apart from the 2hour 'pro' dyno tune.

[joecar]

The IFR table column is PID-less and unitless, so you can use an already logged PID like RPM.

A calc pid can be made for calculating 1/BEN and this is what is paste-multiplied to the IFR table, as SSpdDmon said (these is no paste-divide-by).

[redhardsupra]

so you're setting IFR from MAF? and how do you determine MAF? fueling calculations using fuel flow and wideband? oh wait, we're back to IFR...

jeff, what was first, the chicken or the egg?

[jfpilla]

This is an arguement that goes on and on.

I've always been told and read that the MAF is "bench calibrated" but for the injectors it will mate up with. There is nothing sacred about the stock calibration.
Tuning via IFR always ends up with eneven values(curve). I've never seen anything but a straight factory curve.
Suggest that Injectors be flow tested, it's cheap insurance towards a good tune, set the IFR and match the MAF to the injectors.

Just my opinion.

[jfpilla]

so you're setting IFR from MAF? and how do you determine MAF? fueling calculations using fuel flow and wideband? oh wait, we're back to IFR...

jeff, what was first, the chicken or the egg?

Very cute

[SSpdDmon]

so you're setting IFR from MAF? and how do you determine MAF? fueling calculations using fuel flow and wideband? oh wait, we're back to IFR...

jeff, what was first, the chicken or the egg?

Answer: Neither...theropods came first. If you want to look at it less technically, the egg came first since the chicken evolved to what it is today.

http://www.pbs.org/lifeofbirds/evolution/index.html



As for the MAF method I've been pushing, I still claim it's no more right or wrong than any other method. It's just different...and remember, just because something's different doesn't mean it's deficient. Also, I never claimed you would tune yourself to a linear result in the IFR. You can get close though - then find the best line that satisifies most of the cells within a few percent and then tweak with the MAF. End result, a fairly stock looking MAF curve and adhearance to the timing table. Blow the MAF out of proportion and your granularity changes in the timing tables. Plus, with this way you don't max out the MAF by 10,000hz.

[jfpilla]

Also, I never claimed you would tune yourself to a linear result in the IFR.

My point was the opposite. GM's IFR curve is always linear. I mean to use that point as evidence that the IFR curve should not be anything but linear.
I'm missing your meaning about granularity?

[SSpdDmon]

Also, I never claimed you would tune yourself to a linear result in the IFR.

My point was the opposite. GM's IFR curve is always linear. I mean to use that point as evidence that the IFR curve should not be anything but linear.
I'm missing your meaning about granularity?

Oh, I must have misunderstood your IFR point.

Stock MAF curve tuned, my '01 H/C SS saw ~.76 grams/cyl IIRC. Using the method seen in the tutorial (set IFR, tune VE, tune MAF), I bumped up the frequency so much to get the AFR right that the PCM was now showing .86~.88 grams/cyl. Look at the timing table and see what it does....it shifts everything to the right (i.e. lower timing). Maybe a better analogy would be - it's like stretching an image on your desktop from 600x600 to 800x600. It also means you jump across the cells a little more. In my experience, the more you jump across the timing cells at low RPM, the probability of surging/bucking occuring increases.

[redhardsupra]

both IFR and MAF are smooth functions out of factory. the fact that we must screw at least one of them up to get a decent tune means (to me) that there's a hidden variable (or a dozen).

i'm not saying that one method is better than another either, however, i did say they're all equivalent, and we need a solid starting point. this is my way of arguing chicken vs egg: http://redhardsupra.blogspot.com/200...ss-models.html