11sORbust

I have ran for several months without a maf. It has been though weather changes of 20-90*f. The car runs really good. Not perfect. But good. The wb readings are centered around 14.7. Closed loop is working...

I have a few questions for senior members . How does ecm know the desired PW based on "calculated" GR/sec(no maf)? I'll just start with that one.....

Low C1500

I don't know MAF very good, but it is suprising to me that the motor runs fine without one. It would be running pretty "blind".

Grumpy

It uses default values.
I guess I need a cane now for answering....

11sORbust

don't forget the wood pipe

All jokes aside, any help would be great. I think Rbob doen't have time to answer all the questions via PM. He has helped me a lot so far. It seems that the rpm, tps and iac is calculated into a gr/sec value. That gr/sec value is then used with injector constant and something else to spit out the pulse width. Thats the part that I'm stuck at.


I was suprised how well it ran without a maf on the stock code. It goes into closed loop.uses blms. Even still displays a gr/sec through aldl(that's the calculated gr/sec). The car has been working fine for a long time. When I bought the car the maf was bad. Once I unplugged it the car ran good. I just need to figure out how to control fueling....

11sORbust

blms are close to maxed but the ecm still controls the a/f ratio

11sORbust

closed loop flag is checked...

11sORbust

for some reason the mat is not working. I'm going to fix that and log some more data. Here is the sensor data.....

11sORbust

I disable the high maf code. If the mat was woking right then I guess the ses lite would not turn on. Wonder what happens in the code when "turned off" the maf error codes. Does that just mean the light will not come on. Or does that mean.....

11sORbust

Another odd thing is that I can't run 3.9k mode only 10k. In 10k my idle rpm is 1000.

88TPI406GTA

It would seem like a bad idea to me to run without a MAF...just like running without a MAP. At this point your computer is a modified Alpha-N, simply running off of default values and the tuning that you have done.

I would say that if it indeed is going into closed loop...something is wrong. You should have the Load sensor (MAF) in the loop. Maybe your ECM is bad...it could account for the wierd behavior.

Unless you like it this way

11sORbust

I don't think so. I have 3 ecms laying around.....

11sORbust

I guess that I'll keep looking at the hack and try to find the correct way to change bpw. It would seem to be easy if the bpw vs load table was used. Then have the table look at tps % instead of load value. from what I understand there is no default fuel table right now.Wonder how that works..

I'm a little sad that nobody has really helped this thread out . It seems like a cool idea. I know the car will run ok without the maf because I have logged maybe 10,000 miles without it. That's including countless races...

back to the hack..........

RBob

iTrader: (1)

Here it is, an un-reduced equation that converts airflow, AFR, RPM & Injector flow rate to a PW in msec.

{edit} A correction: the constant 0.1526 in the equation and examples needs to be 0.01526 (move decimal point one place). This value is from the inverse of 64KHz which is the clock frequency of the ECM counters.

The .01526 is the msec time span between clock ticks.
{end edit}

RBob.

11sORbust

what about without the maf? Then tps,rpm and iac calculation replace the actual gr/sec ?

RBob

iTrader: (1)

Yes, the no-MAF default creates an airflow term that is used in place of the one measured and reported by the MAF.

Does this equation answer one of your two questions?

RBob.

88TPI406GTA

That equation is pretty cool...

So basically then, without the MAF, you are tuning the exact parameters for the engine without a load calculation...is there a difference between this and actual Alpha-N?

Also, this really isn't like MAP because there is no load-sensing device in play...but don't the sensor inputs create the "load" that the ECM needs to see?

I am just surprised that the car runs so well without the MAF...

11sORbust

That answers one question. I have a few more.....

So the calculated gr/sec replaces maf input. Then how is load sensed? There is LV8 and gr/sec displayed through aldl. Guess I need to know what does LV8 really mean and how is it calculated?


I found out that my mat was unplugged and not bad at all. Guess that doesn't matter on a maf car. I mean the blm (log) posted will stay the same after the mat is working?(the reason I say that is Glenn might have hinted the MAf system uses mat for something else besides egr.)

11sORbust

That finally explains why there is no main fuel tables with 165......and maf table changes tend to be caused by "sensor calibration error"

RBob

iTrader: (1)

Mmmm, without the MAF the ECM has tables that model the engine airflow requirements. In using the RPM & TPS as the major inputs it is an Alpha-N system. Although crude as it is only intended to be used to allow the engine to run and the driver to get home.

The load variable LV8, is created from the airflow and RPM terms. This is done whether the airflow term is a default term or a true measured MAF term.

I too was surprised one day when I found out via data logs (and a SES light) that I was running on default MAP values. At higher loads/RPM wasn't as good as a real MAP sensor. But at lower speeds, around-town driving, it was decent.

RBob.

RBob

iTrader: (1)

I've covered LV8 many times already. For load the ECM divides the airflow by the RPM and scales it. The greater the airflow, or the lower the RPM, the greater the load (LV8 term).

Or, same airflow and higher RPM, the lower the load.

RBob.

RBob

iTrader: (1)

Yep, the MAF scalar tables are the calibration of the MAF sensor. They directly convert the MAF input (freq or voltage) into an airflow term (gms/sec).

Change the MAF calibration (screens, gutting, hacksaws) and the MAF scalar tables will need to be changed to match.

RBob.


P.S. Correction on the equation, the constant 0.1526 in the equation and examples needs to be 0.01526 (move decimal point one place). This value is from the inverse of 64KHz which is the clock frequency of the ECM counters.

The .01526 is the msec time span between clock ticks.

11sORbust

That's what I'm saying. alpha-n doesn't run too bad. That's why I would like to try and tune it. Can't do that without any fuel table.... just wondering how can I do this? What about my idea on replacing gr/sec with tps% for the bpw vs load? If I could use that table it would be easy.

Hey, maybe running in open loop is "cool". But alpha-n is hardcore,lol..

Grumpy

If you want to play this game, why not take a 730 and replace the MAP input with the TPS. Then cancel out enough other stuff so that'll run.
At least then you have the tables to start with.
Might try doing with the 58 code since that'll give you an TPS/RPM/AE table.

You'll probably be stepping all over with one table to cover another.

If you want to see what an Apla-N setup is like get a ride in a Holley dial adjustable TBI setup.

RBob

iTrader: (1)

Default mode Alpha-N doesn't run nearly as well as real sensors. I stated that in a previous reply to this thread. There is another current thread where I stated what I needed to do to get an Alpha-N system running almost equal to a sensor based system.

I don't think you give open loop enough credit either. GM has released hundreds of thousands (if not a million) of EFI vehicles running the same code (among other masks) as we are discussing in open loop mode. The cal has close loop coolant enable threshold set to max and the O2 sensor wire is cut.

If you want to run Alpha-N mode then you need to model the engine airflow much better then the stock default mode does. That is the key.

In that same thread where I mentioned what I had done for an Alpha-N mode there is a link to a paper that describes a model based manifold pressure observer. Upon reading it I found that it closely resembled what I had done.

RBob.

11sORbust

Guess the main thing is I don't care much for the engine in my red car. It's a 305/tpi with 190,000 miles on it. So it would be fun to try anything. .

I'll dig for that thread....

11sORbust

the bit about hardcore was a joke......

TPIgirl

Could someone please say where this default table is exactly and how to go about changing it?

There is nothing in tunerpro $6E that shows this default table.

11sORbust

There is no default table. Just a calculation, from what I undrstand. But that's what I'm waiting for a response on(how to change). I don't have the ability to modify/add tables.....

11sORbust

it's just like the maf tables. It's alot simpler calculation than SD. The stuff Rbob posted explains the ecm uses the direct measurement of air to figure out pulse width. No fueling tables needed. That's why I said the maf tables are sensor calibration tables but can be indirectly use like a fuel table. The stuff Rbob posted explains exactly why the ecm doesn't need fuel tables(aside from bpw vs load).

RBob

iTrader: (1)

The table and parameters to create a default gms/sec value are in the diagnostic calibration area. They follow the malf 33 and malf 34 test parameters.

As for changing them I would use empirical data from data logs while using the MAF. Observe the RPM, IAC, TPS% and gms/sec then back calculate that into the calibration terms.

RBob.

TPIgirl

That makes sense but any hint of where these values are in the bin? Do you know what locations and how many locations they go across? Are they all single bite values?

11sORbust

This is the table but I have a few simple questions....

11sORbust

Why the gr/sec column doesn't display gr/sec. It look like TPS voltage?

What is so bad about alpha-n? You still have load values. The only thing is the ecm doesn't know how much air is coming in the engine. That doesn't matter if you program the right amount. I can see grumpy's point. The tables(AE and such) would need some work. If everything is done right then why not run without a maf?

11sORbust

if that is the table then where is the IAC taken into the calc? After that table,right?

TPIgirl

Thank you.
I think what they mean is that technically TPS vs RPM is not a load value at all. I think you'd have to have a MAF or MAP input for it to come close to actually determining load on the engine.

11sORbust

If I could do that then I would.....

Since I can't do anything to the code I guess this thread is over. I'm going to do the 730 swap, again.

TPIgirl

I was still hoping someone would answer this.

Grumpy

The whole point to EFI is giving the engine, most accurately what it wants. You could have a VE table with 2 cells for MAP, and 2 for RPM. But, what would you think runs better a table with 4 cells or one with 256?. Same with coolant temp corrections you can get by with a 2 axis table, but do you think a 3 axis table would be more precise?.

The trick, IMO, is finding out the min amount of corrections to get the best job done. While the oems have all kinds of stuff that's excessive for what we're doing, the aftermarket ecms in an effort to be so friendly haven't enough. If you look at the newer aftermarket stuff it's alot better, ie sophisticated then the earlier stuff. But, it seems they're also missing the boat on what an engine actually needs, and are working on systems that are so universally used, that they can't be configured as well as the oem stuff.

Which gets us right back to needing to have actual source code where the misc things like some of the IAC stuff, can be removed, and ideas from one ecm used with another. And more importantly, IMO, the start-up routines redone. Plus again, IMO, the AE stuff has a serious flaw in it. I can tell you as fact, once you have an easy code to optimise, the car takes on a new life for what it drives like. Not to mention getting the best it can in the way of mileage, and I suspect emissions. The only reason I use the word suspect is that with out a multi gas analyser, I can't state it as being fact.

So while you may think there is a easier way to do EFI, there is, it's just that your looking at it from the wrong way. The easy stuff's been done, what's needed now is some real grunt work. Which few (so far no one) really wants to get their hands dirty doing. I've had exactly ONE person volunteer to get serious with the basic source code I've let out.

This ain't a rant, just a statement of what the direction is folks should be taking. My car runs fine.

TPIgirl

Grumpy, who's question were you answering there?
I hope it wasn't mine

RBob

iTrader: (1)

IMHO Grumpy was making a general statement. There is more to getting a good responsive engine then just a few tweaks and a minimum system. It is worth reading his reply.

TPIgirl, as far as your question, the $6E mask has the default parameters/table as such:
RBob.

Grumpy

It wasn't in reply to any one person.
It was again about where we as a list need to head, if we're to progress.

11sORbust

I understand and you are 100% RIGHT.....

TPIgirl

Thank you Rbob.
Do you know why the table is limited to 230 gps and 4800 RPM?
Where is the TPS values? and how does it factor into this table?

RBob

iTrader: (1)

To create the default gms/sec value the ECM uses this calculation:

airflow = LC217[RPM] * min(TPS%, LC213) + (IAC * LC214) + LC215

Note that these locations are from the BUA bin, the $6E code uses the same exact code, just the actual locations are different:

airflow = LC222[RPM] * min(TPS%, LC21E) + (IAC * LC21F) + LC220

The min() statement means that the ECM will use the lessor of the two values: actual tps%, or 65%.


Plugging in the cal terms we have this:

airflow = Lookup[RPM] * min(TPS%, 167) + (IAC * 21) + 1024

Then for these particular settings:

2400 RPM, 20% TPS (51), 56 steps

airflow = 175 * 51 + (56 * 21) + 1024

= 11125 (as used by the ECM, gms/sec has a fractional term).

11125 / 256 = 43.5 gms/sec

RBob.

ghost_man

I am not a senior member, but been following and reading alot about the ECMs in these cars. Grumpy's injector link to the GN site got me super interested in learning more and understanding these systems.

Some of the guys in the GN post I have been tracking down and watching.....take a look at this stuff.

24# injectors, MAF, and 1 5/8" headers going 10.68@128MPH!!

How are these guys able to do this stuff?http://forums.corvetteforum.com/zerothread?id=806351:hail:

onebinky

iTrader: (1)

"Who would have guessed....
MAF, 24# injectors, 1 5/8" headers => 10.68@128MPH!!!

A few other on some other boards need to put that in their pipes and smoke it. If it weren't for the graduation, we could have most likely backed them up with some duplicate runs "

That was ski_dwn_it, and the owner of the car is corky! Both former members who were banned

Any one know yet why they were booted? They both seemed pretty knowledgable.

JPrevost

I don't know the details but they both had a hard time understanding static injectors. And worse, they posted data showing there injectors were static but didn't believe it.
It was all about argueing and not having definative reasons and facts. So what resulted was name calling on multiple instances. I think just looking at there "engine break in" procedure shows us all that they don't care about money, only times. Although I was suprised they said the combo's were streetable... maybe a slightly lean condition resulting is smooth acceleration yet hard to modulate wheel spin at the track?!?!?! Very odd that ski admits he might gain some mph and time going to s/d... now if only they took that step and then some by going to injectors that aren't static. Anybody know what fuel pressure they're running .

88TPI406GTA

They do have impressive results...but...

When you combine the static injectors with the toulene mix, it did kinda make me wonder what kind of reliability (over a lot of miles) that they would have.

Besides, I never really remember much discussion about actual streetability. Apparently, Ski drove his car on the street, but I think that my definition of "streetable" is a bit different. Like drive it 30K in a year and it feels like a stocker at part throttle...

Oh well...kinda makes me want to take a stock L98 and throw a 400 HP shot of nitrous on it too...it might actually survive the trip down the strip

ghost_man

guys. Not real sure about all the details either. I did some searching on the forum over there and also sent ski an email asking him about the setups.

Going back through the years, though. These guys go to the track ALOT. Nearly every weekend and hammer their cars with zero problems....so I highly doubt they are ruining much on their cars. In fact, I see where ski and corky both sold their motors 406s to their buddies that race just as frequently as they had.

After reading the post that Grumpy linked to the GN site that brought be to these guys, I am not entirely satisfied that we fully understand what these guys are doing....Some here were saying going back through the archieves here that when ski was at 11.1sec on 24# injectors, he was static and people couldn't explain their times....now Corky is at 10.68@128MPH! Correct me if I am wrong, but if they were out of fuel, they certainly wouldn't make more power right, even if they threw a bigger motor. Maybe the Ets would get better, but not the MPH....7-8 more MPH mean that its making ~80 more hp....

Help me understand all this.....seems to me we are all going in circles, while these guys are going quicker and quicker with every post. Maybe they don't even go to the track... Based on the verbage here, that might be the case, since I can't image that many people here being wrong. Right? But it seems to me that many people know these guys personally and someone eventually would call BS on them.

11sORbust

Exactly!! You can run fast on any injectors. That's not impressive though. Like I say, why shoot yourself in the foot. Those guys was clueless. Then the tried to bait people in an injector debate,lol.

I would like to clear a few things up with that static injector bit. Injectors can only stay open 100% of the time. Once the duty cycle exceeds that there is no ecm control. Who would want that? We are prom burning to do things right. It's not '89 when fuel pressure changes was the only way to add fuel. That is why large injectors have a bad rep. Because of old hat tuning proceedures. Why it worked is simple. When the injectors are open all the time, they "turn into" nossles. Just like a garden hose. Flow will be directly related to pressure. So your only choice to add fuel is to raise fuel pressure. It all works but with the f/p raised, part throttle fueling is now off. Or what if you need to lower pressure. then you run a risk of going lean on your blms and gettin wot fuel correction. Not the best for consistancy.

There is a million reasons to overshoot the injector size. Only one reason to run smaller injectors.....money. Do you want to just get by or do things right?