Newbie losing the plot!

[Cobra Dave]

I have searched until my head hurts and clicked on every tab I can find in the Editor and after 12 hours I'm losing the will to live.

What I have, and what I'd like to do:-

Engine: LS3 crate motor.
VCM and harness and sensors all GM for this crate motor.

In the Uk we have to meet the following emission limits:-

Idle;
CO........0.30 max

Fast idle (2500-3000rpm)
CO........0.20 max
HC........200 max
Lambda....... 0.97 to 1.03

currently the idle is fine.

The fast idle is where the problem is.
between 2500 and 3000 rpm the lambda struggles to get higher than 0.97 which pushes the CO to 1.14+

If I rev the engine to 3200rpm the lambda suddenly jumps to 1.00 and the CO clears up nicely. But 3200rpm is too high for our authorities.

I was hoping to be able to pull up some tables (AF or VE etc) and drag around some cells and sort it out but I can't find anything like this to work with.
And being a novice I don't know what else i dare play around with to get the figures I need.

I'm sure this is really simple to somebody but sadly at this stage it isn't me.
Ireally want to get my head around how to sort this GM VCM as I have made it my personal mission, even though as a business it would be cheaper to go aftermarket at this rate.

Thanks for reading,
Simpleton terms in reply please.

[IDRIVEAG8GT]

OK, well I see your dilemma to meet emissions here.... Um!

I need a few technical specs,

Are we running a stock cam?

What KIND of Cats?

Do you have proper injector data?

What Compression Ratio?

VVT yes/no?

And can you post up the current tune?

Somethings that may help alot with your.... "Fast Idle" problem would be to work over your EQ ratio tables and put alot more timing to it.

[Cobra Dave]

Hi.
Stock 430hp cam
aftermarket none race cats
"Injector data"? They are the ones that come with the crate notor.
Stock CR (10.7 I believe)
no VVT.

"EQ Tables"?

Cheers.
Dave.

[pontisteve]

The number one job of the PCM is to measure the airflow coming in. In GM PCMs, this is done in two way. Its directly measured by the Mass Airflow Sensor, and it's also estimated by the Manifold Absolute Pressure sensor (with help from the Intake Air Temp sensor which corrects for temperature). GM uses a blended model, where it looks at both sensors at idle and low speeds, and at higher RPM (4000+ usually), it only looks at the MAF sensor.

Below 4000, it still looks mostly at the MAF sensor, but it does blend the two systems. The systems are known as speed density (MAP sensor) and Mass Air (MAF sensor).

The #1 best thing you can do is get the MAF xfer function correct in the PCM. That, and characterize the injectors correctly. If the PCM knows how much air is coming in, and if it knows what injectors are in the car, it will calculate the fuel correctly.

Youll also find that GM refers to the MAF sensor system as "steady state", due to its nice, steady condition readings. The MAP sensor is referred to as Dynamic Airflow, since it best estimates what happens when the motor is in a dynamic airflow situation. By that, I mean "moving", or changing, or transient condition. In other words, when you change the position of the gas pedal, the motor is temporarily in the transient state. When you hold the pedal steady, it's in a steady state.

You generally measure the air/fuel ratio with a wideband, command the motor to put out a specific air/fuel ratio (in open loop, with all fuel adders zero'd out), and measure the % of error in the commanded air/fuel ratio to the actual air/fuel ratio measured. If you command 14.7:1 and you actually measure a 14.0:1, then you would take the actual and divide it by the commanded to get the correction factor required.

Apply this correction factor at the point where you are in the MAF xfer function. This will be in Hertz output of the MAF sensor. So at idle, you're at about 3000 Hz or so. If you are running too rich at the current MAF output of 3000 hz, you would reduce the airflow number in the 3000hz cell by applying the correction factor to this cell.

You can use a wideband, or for a lesser quality but still decent fix at part throttle, you can just use the factory correction factors that are determined in the Long Term Fuel Trims output by the PCM. These are logged in the HP Scanner. If the LTFT's are adding 5% to get the motor rich enough to run lambda 1.00, then you need to add 5% to the tune in that spot, so it doesn't have to. Generally, with O2 sensors active, if you get within 20% of accurate, the O2 sensors will fix the rest for you good enough to return the car to lambda 1. While not perfect, that is a easy way to get where you're trying to go.

Tune the MAF, read Greg Banish's first book, and read this forum.

[Cobra Dave]

Hi steve.
Thanks for the reply.

Am I to do this with the laptop connected to the engine?

Where do I find the "MAF xfer function"?

I have a wideband LM-1, so I can see what is coming out of the tailpipe ok.

How do I characterize the injectors?

Cheers.
Dave.

P.S.
I am reading the forum as best I can, but usually within about 3-4 posts in a thread I am on information overload and can't find the relevent tabs to click on anyway.

I am thinking of buying the Learn at Home tutorials but if they go so in depth as the forum threads do straight away then I'm fooffed.

At the moment this is the only issue I have an interest in sorting as I cannot get the cars registered unless it is done. If my customers want to take their cars to rolling road dyno sessions later then that will be their choice.

[IDRIVEAG8GT]

See attachment Number 1.

See where the red circle is? This is bad. Smooth that curve out. That will help a bunch. (Don't make it look like the table on the right, that is just you high MAF table that I was looking through.)

See attachment Number 2.

This is the proper way to put a vehicle in Pure MAF mode.

See attachment Number 3.

This is purely experimental but is what the Australian Holden's run on their O2 Sensors to pass emissions. Try it, if it doesn't work just go back to what you had.

[Cobra Dave]

Thank you very much
And in a language I can understand.....Pictures.
I'll give it a go tomorrow. It's 9:30pm here at the moment, (Don't want to annoy the neighbours) and let you know.

With this tweak is the engine safe to keep at these settings, or is this just a "test Pass" tune?


Thanks again.
Dave.

[IDRIVEAG8GT]

Thank you very much
And in a language I can understand.....Pictures.
I'll give it a go tomorrow. It's 9:30pm here at the moment, (Don't want to annoy the neighbours) and let you know.

With this tweak is the engine safe to keep at these settings, or is this just a "test Pass" tune?


Thanks again.
Dave.

You should have no problem running those changes from here on out IF they work.

You have lots of timing so we can't improve a whole lot there anyways.



-Robbie.

[pontisteve]

Here's an easy way to fix this up close. Go to the MAF Low table pictured, and highlight the cells between 2950 and 3700. Then click on the Interpolate Between Horizontal Bounds button in the toolbar.

Somebody jacked with this tune to do that. It obviously was smooth when it was new. Anyway, check your Short Term and Long Term fuel trims and report back what they are after fixing this and running the car.

Warm the car up, and let it idle for a minute, then increase rpm by 250 and hold for 15 seconds. Repeat that all the way up to 3000 rpms. That will give the O2 sensors plenty of time to learn. At each step, after the 15 seconds, record the Long Term fuel trims that the scanner displays.

I don't get it though. Why mess with these cars with HP tuners if you've got a stock crate motor, stock PCM that matches, stock everything?

[Cobra Dave]

Hi Steve.

I will have a go at the 250rpm test.

I don't know how the "trim" was out.
The whole engine, sensors, loom and PCM were all brand new in the GM packaging??????
As soon as my HP Tuners stuff arrived I wrestled around installing it and the drivers and then did the scan.
I saved this and made a copy for safety and that was it.

My engine doesn't have the stock cast headers, it has tubular ones and the O2 sensor is in the collector on the side pipes.

I know two others who have failed the emissions, one with the same problem as me and another who's figures were all over the place. The second guy's figures never came good at any revs though. He's now gone aftermarket (Omex). These two people have used the stock cast headers with the O2 sensor bung in them??????? which only makes things even more confusing.

[pontisteve]

This tune is in MAF only mode, which is caused by the Dynamic airflow high RPM disable being set to 200 rpm. So the PCM is relying solely on the readings from the MAF. I suppose it is possible then, that GM needed to perhaps play with the MAF xfer function some to get it to deal with reversion at idle on this particular motor. Its very unusual to see a MAF curve that's stock that isn't pretty much smooth. Then again, most cars are not run in MAF only mode until over 3800 RPM.

You can try what I suggested before and see if moving the MAF straight helps the trims. Make sure you have a copy or two of the stock tune laying around, preferably on separate computers, in case you have to go back to that!!!

Make sure the O2 sensor is reading an average of all 4 cylinders, by putting it in the collector a few inches back from the primaries. If the O2 is reading any one cylinder too much, it can be problematic and possibly erratic.

Did this engine come with an airflow meter and intake tubing? Are you running it like that, stock? Placement of the MAF and shape of the intake tubing can change the MAF curve requirements. The MAF is the major player in the a/f department.

If that stuff doesn't help, try turning the Engine>Airflow>Dynamic> High RPM Disable to 2500 and the High RPM Re-Enable to 2400. This will blend in speed density at lower RPMs.

[Cobra Dave]

Hi Steve.
the MAF is 10" from the throttle body in a straight piece of 4" tube, with a big K+N filter stuck on the end of it at 3" forward of the MAF.
The electrics kit came with instructions saying to mount it like this.

Another thought (If plan "A" doesn't work). The air filter is just behind the radiator, so the IAT could be getting a bit too warm , eslecially when the fan kicks in. Your thoughts?

[pontisteve]

The IAT getting heat soaked is a known issue on GMs. Most guys relocate it further up the intake tubing, usually to an airbox or something that's not in line with the radiator throwing heat towards it. With your short intake tubing, that might be difficult. You could extend the tubing forward of the MAF by about 10" or more, then bend it 90 degrees and run it towards a fender, mount the K&N on the end of it somewhere near the fender, and install the IAT right behind the filter.

One other slight concern (although less of one with a K&N filter) is the fans blowing directly into the MAF sensor. That could skew the data, particularly at idle since there isn't much airflow at that speed. I'm talking about blowing air into it, or into the air filter. The PCM could register this as extra airflow into the motor, and therefore richen it up. I think setting the Dynamic Airflow settings as above might help. Try that first.


The trick there is to leave enough straight tubing in front of the MAF that any air making the 90 degree bend has time to straighten up before going thru the MAF.

The IAT sensor mostly affects spark timing when the motor is hot or cold. It is also used to calculate speed density (MAP) airflow. I doubt any hot air on the IAT would cause an emissions idle problem.

Another possible fix is to run the straight intake tubing right over the top of the radiator and point it downward to fit behind the bumper cover. I've seen this done on a 98-02 Camaro, again IAT relocated as well, and it worked well. Intake temps dropped considerably, and power improved.

[Cobra Dave]

Hi Steve.
here's a picture of my radiator/air filter situation.
the space over the top of the rad is not good for fitting a pipe. Hood hinge in the way.
I have been thinking of making a shroud to deflect rad heat down from the air filter so that at least has a better chance of drawing colder air from over the top of the rad.

Time will tell.


Edited to add:-
I piece of cardboard slid under the air filter and angled up to the top of the rad core and bingo! IAT falls well into its comfort zone.
Must make a pretty one out of aluminum now.

I did the screen captures of the 250rpm test and they don't really show how much the indicators are moving up and down, so I thought I'd run it again and record a video clip (This also shows my wideband display)........Well, kind of. The clip is rather blurred.

Impossible to hold the revs below 1750rpm as if you back off slightly then it drops to tick over.

here's the youtube clip of what i have so far:-
http://www.youtube.com/watch?v=K0huhiUWBqc

Wehey! finally worked out how to "Log" a scan:-


And the screen caps:-

[pontisteve]

It wouldn't hurt to shield the air filter from the radiator fan. You could make custom tubing to bring the air inlet tubing into an inner fender well area maybe, or perhaps a carbon fiber cover that covers the entire open area under the air filter, from fender to fender, front to back. Either way, you get the idea about that now. Personally, I think the extended air filter and tubing looks kind of funny as-is. There are all sorts of creative air filter housings that might go with the car theme too, like double K&N's or something like that. You could also lower the radiator an inch or two, build an aluminum box that goes right over the radiator and inward. Think 02 Ram Air Trans Am.

Now that you've figured out how to log, I want you to go to the Table Display in the scanner, right click and Insert the following Engine parameter: Mass Air Flow Sensor Hertz.

You may also want to consider getting yourself familiar with the Histograms, because they'll really take the work out of what you will be doing. Ultimately, you'll be setting up your own histrogram based on a table of Mass Airflow vs Frequency - Low, and you'll be datalogging a sensor to overlay on this table. That sensor will either be STFT's, LTFT's, or WBO2. Do you have a wideband?

If you just need to fix this and forget it, we can probably help you thru that. But if you want to start really understanding tuning, I would suggest an Innovate LM-1 or NGK AFK wideband sensor. (Once you see how those things work, you'll be able to fix any air/fuel problems with any cars, carbureted or injected!).

Now, looking at your log real quick, look at the table display. On the right side, you'll see that there are short and long term fuel trims. STFT's are what the O2 sensor is telling the computer to modify the fuel curve to right now. Long Term trims are basically short term trims learned slowly over time. So if the computer sees a -5% stft for any significant amount of time, it will make it a ltft instead, and the stft starts over at zero. The STFTs are ALWAYS moving around some, but the LTFTs are more slow and steady, and better reflect what's really needing to happen.

Looking at your logs, you can see that your LTFTs are pulling about 5 - 10% fuel at the various RPMs you held it at. Ultimately, we'll be going into the tune and pulling out that amount by modifying the MAF Frequency vs output low table. For now, I would flash in a tune that smoothes that curve out exactly as how I told you to above, and relog the car. (Don't forget to add MAF Frequency Hertz to the table display!). See if your LTFT's go down.

After each change to the tune, you'll want to reset the fuel trims, since we're most interested in whatever new trims are being produced. (Over time, these LTFTs learn more and more by averaging lots of data over time, and this is how cars learn to run perfectly, and this is what is reset when you disconnect a battery). You can reset the LTFT's by going in the scanner to VCM Controls>Fuel and Spark>Fuel Trim Learn>Reset Fuel Trims.

You'll probably find lots of other cool things to play with in the VCM controls as well. These are basically bi-directional scantool controls, meaning the scan tool (in our case the HP Scanner) can tell the PCM to temporarily do certain things, like turn on fans, change air/fuel ratio, run in a certain trans gear, change idle speed, go to a fixed timing advance, etc.

Make the MAF curve change I told you about, and do another log like this one.

[Cobra Dave]

Hi Steve.

"Mass Air Flow sensor (Hertz)" done. I need a lie down now!

Yup, got a wideband. (LM-1) unfortunately I have long since lost the software disc and the cables to connect it to my Pc or the HPT gear.
On the youtube clip it is sat alongside the laptop screen.

For the time being just being able to know how to correct this emissions problem will be enough. Full scale tuning might be something I can get into at a later stage.

All the cars I build (Cobra replica Kit cars) have to pass the same emission test (it's a politics thing in the UK). I have become a bit of an expert at getting old school iron block Chevys through this test with carbs fitted.
Because they are spanners and screwdrivers I'm fine. This Laptop based stuff is all new to me.

The above log recordings are with your tune mods, The Wideband gives the same results as before though unfortunately. I see what you mean about that curve though. Should I change everything else back and leave that in?


Thanks again for your time and patience.
P.S. I have ordered (Should be here on Monday) the beginners course from www.thetuningschool.com

[IDRIVEAG8GT]

Hi Steve.

"Mass Air Flow sensor (Hertz)" done. I need a lie down now!

Yup, got a wideband. (LM-1) unfortunately I have long since lost the software disc and the cables to connect it to my Pc or the HPT gear.
On the youtube clip it is sat alongside the laptop screen.

For the time being just being able to know how to correct this emissions problem will be enough. Full scale tuning might be something I can get into at a later stage.

All the cars I build (Cobra replica Kit cars) have to pass the same emission test (it's a politics thing in the UK). I have become a bit of an expert at getting old school iron block Chevys through this test with carbs fitted.
Because they are spanners and screwdrivers I'm fine. This Laptop based stuff is all new to me.

The above log recordings are with your tune mods, The Wideband gives the same results as before though unfortunately. I see what you mean about that curve though. Should I change everything else back and leave that in?


Thanks again for your time and patience.
P.S. I have ordered (Should be here on Monday) the beginners course from www.thetuningschool.com

Simply put, let's lean this MFer out!

[IDRIVEAG8GT]

Those o2's are switching sooooooooooooo slow. Let me do some research on how to speed those puppies up some.

[pontisteve]

Yeah, I see an issue here somehow. The LTFT's are 5-10% rich, but that's well within the 25% (or so) authority that the O2s have to correct the tune. In other words, the O2s should be switching, and rapidly.

Please check the O2 wiring, and make sure it is all ok. Also, check the color of the wiring and make sure that the two O2 harnesses aren't switched from side to side. In other words, make sure the left O2 gets the left O2 harness and the right gets the right. If the harnesses are crossed, all sorts of chaos ensues.

Make sure the O2s themselves are not damaged, and make sure that the O2 heater circuit is intact and getting power to the sensors. Basically, the O2 sensors must be heated up in order to work. Modern O2s have a heater built into them, and the PCM turns that heater on or off, and can even vary the duty cycle to the heaters to control just how hot they get. These heaters draw a good bit of electricity, so they're usually run on a fused circuit. We need to make sure that fuse is good, and that the O2s are getting heater circuit voltage to them. So get yourself a wiring diagram and make sure the heaters are powered up. If not, find out why. O2 heater circuit fuse would be my first quick check.

The LM1 has an analog output cable (about $10) that plugs into the bottom of the LM1's stereo cable outlet, and splits into 2 wires that plug into the HP Tuner Pro. All software for the LM1 can be downloaded on Innovate Motorsports website. LM Programmer lets you program the analog outputs, and change the AFR range, or allows switching between lambda and AFR, and also lets you select the fuel stoich. I use lambda.

[Cobra Dave]

just a thought? My O2 sensors are at least 3 feet from the cylinder heads due to the header primary lengths. Could this cause the slower switching?

Due to the location of the O2 sensors the GM loom has only just long enough wires to reach them, so fortunately the "Wrong sides" mistake is not possible.
Checked the fuse and that's fine, however, when I pulled the sensors out and turned the ignition on they do not warm up. I have checked the wiring to them and get a ground on the white and both blacks show 12 volts. Could I really have two brand new sensors that are goosed?
Might we finally be onto something here?????

I looked at the lead for the output on the Innovate website....$8... Bargain. I was a bit put off by the $75 shipping costs!!!

With also needing the cable to connect it to my laptop (Available from Innovate) (Could also include this in the same shipping cost, so a bit of good news there.
I would also need and adaptor lead from serial to USB, which Innovate don't list, so this would mean another load of shipping.

I will probably head off to a big local computer/techie stor and see what components I can get and make my own, otherwise i can but a new LM-1 complete kit for around $350.

Edited to add:- complete new LM-1 is on its way to me, with wires and software disc.. I should be able to watch both sides at the same time now


Many thanks again.