1 2
Dusterbd13
Dusterbd13 UberDork
11/19/15 8:32 a.m.

So with the changes in employment and income that I am facing in the next few months, I've realized that I need to scale back my plans for the duster. I can't afford to pay the man to tune it.

I'll be running a gm ecm from a 1991 corvette. On a solid roller mopar small block. Full mpfi build.

I had planned to pay the man after my attempts to teach myself how to tune years ago on something a lot less radical. I still have all my datalogging/chip burning/prom editing stuff. But I never really learned how to use it.

I guess what im looking for is a good step by step guide of first this, then this, and this does this. Something that takes me from a faily broad but basic understanding of the parts of the program, and teaches me what they do and what to do with them.

Any ideas?

81cpcamaro
81cpcamaro Dork
11/19/15 8:50 a.m.

http://www.gearhead-efi.com/ is a good place for info on tuning, if you haven't been there already. They helped me get the TBI swap working in my 79 Dually.

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/19/15 8:53 a.m.

Just remember, from the highest level, it's just fuel, air, and spark. And for almost all aftermarket modules, and that '91 module, that's all there pretty much is.

The additional stuff are just details.

captdownshift
captdownshift GRM+ Memberand UltraDork
11/19/15 9:05 a.m.

In reply to alfadriver:

And you start by reading the amount of air that you have coming in, then trimming your fuel to as close as you are comfortable with on the safe side, then controlling your spark and timing for optimal combustion, after which you may need to make slight adjustments to your fuel trim.

Dusterbd13
Dusterbd13 UberDork
11/19/15 9:34 a.m.
81cpcamaro wrote: http://www.gearhead-efi.com/ is a good place for info on tuning, if you haven't been there already. They helped me get the TBI swap working in my 79 Dually.

this exact same topic is cross posted over there. between these two boards, I should be able to learn enough to be dangerous.

Dusterbd13
Dusterbd13 UberDork
11/19/15 9:38 a.m.
alfadriver wrote: Just remember, from the highest level, it's just fuel, air, and spark. And for almost all aftermarket modules, and that '91 module, that's all there pretty much is. The additional stuff are just details.

could we break down for my semi-stupid self just which tables do which? like fuel enrichment table is effectively the accelerator pump, map table is air, etc? I know this is spoon feeding me, but I want to make sure that my basic assumptions are correct. all the articles I have been able to find assume a basic knowledge of some things, and I think I have that basic knowledge, but I want to confirm it. I understand tuning a carb and ignition (to a point. its so crude that I can never seem tp get happy with it), and I understand each sensor function and purpose, but the actual tables aren't labeled in something that makes sense to me.

Strike_Zero
Strike_Zero UltraDork
11/19/15 9:55 a.m.

Hmmmm this looks promising . . .

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/19/15 11:52 a.m.

In reply to Dusterbd13:

I have a few meetings, but I'll try to put some high level thoughts together for this afternoon.

curtis73
curtis73 GRM+ Memberand PowerDork
11/19/15 11:59 a.m.

This is relevant to my interests....

Dusterbd13
Dusterbd13 UberDork
11/19/15 12:11 p.m.
alfadriver wrote: In reply to Dusterbd13: I have a few meetings, but I'll try to put some high level thoughts together for this afternoon.

I'm going to owe you a case of beer.

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/19/15 2:17 p.m.

So lets start with air. or some kind of measure of air...

I'm going to start with the easy one to understand- a MAF sensor- which measures the Mass Air Flow. Since we mostly know what air fuel to run, if you measure the mass of air, you then know what the fuel mass you need to deliver. It is as simple as that. There are a lot of small calculations that change that down to an event basis, so that an individual injector will inject the right amount of fuel, as the same air mass at 1500 rpm vs. 3000rpm, the amount of fuel injected is twice as much as the other.

When doing the individual event calculations, it's basic thermodynamics to estimate the relative amount of air in the cylinder vs. what an ideal charge would be. And the ratio of actual in cylinder/ideal in cylinder is pretty universally called load. That load information can be also used for spark advance- but more on that later.

The thermo is pV = mRT, Pressure * Volume = mass if gas * gas constant * gas temperature (in absolute amounts).

With that, you can guesstimate the ideal mass amount if pressure is atmospheric, volume is the total cylinder swept volume, and T is normally 100C/373K. And if you know what m should be based on the MAF and corrected down to an event basis, you can estimate P....

Anyway- the point is that on a MAF system, it's pretty easy to calculate the amount of fuel to inject.

But many aftermarket systems use a MAP sensor- measuring Manifold Air Pressure.

This is where it can get weird.

The most simple injection systems will have a table of MAP vs. RPM that outputs a fuel pulse width. I've played with one of those, and while doing ONE calibration is easy- doing a lot, or modifying that for another engine, or changing hardware is a major PITA. Still, getting a desk top calibration is a spreadsheet of ideal gas law formulations away. That will get you close, and then you need to measure exhaust a/f somehow to correct. That also means you know the fuel flow of the injectors.

The next most simple system just uses the ideal gas law to estimate the mass. Since pV = mRT, m = pV/RT MAP = pressure. And YES- you do need to measure the manifold gas temp. Once that calculation is done, there's a correction to it- which is kind of called volumetric efficiency. It's not used exactly how it's actually defined, but close enough to not matter. This is how the MS system works- you have a table of numbers of 0-1 (or higher, sometimes)- and the ideal gas law is corrected to the actual cylinder mass. I don't know the specific name of that table- but I will say that it's a pretty elegant solution. Once you have that correction, then you know how much fuel to inject. Done.

Above that- you can use the MAP and the air temp to run a series of regressions of mass vs. MAP. That's what most OEM's do- as the actual in cylinder mass being really close to correct is really important. It's not a hard calculation, but not the easiest thing in the world to calibrate- and is normally done using a dyno. The calculation is mostly a linear one, but it's so easy to include a MAP^2 term in it, it's done. The equation coefficients change with engine speed AND cam timing- and even though that seems like a lot of stuff to look up- it's still less than having a bunch of tables correcting voleff.

Again, the point of that is to calculate the mass of air, which gives you a simple calculation for the mass of fuel.

The way the MS works is really nice- and hope that whatever you get is similar. That way, calibrations are quite portable. You can take a V8 calibration that has a similar cyl volume, and similar cam timings- and start with that as the base calibration for a 4 or a 6 cyl engine. Just getting it going does not require perfection- just something reasonably close.

All of the systems are tuned via- are you getting what you are asking for? If you want 14.6:1 in the exhaust- are you getting that? If not, make a correction to the table where it's wrong. That does require that you trust your injection flow rates- but you have to trust part of it just to get going.

So on the MAF system- if you should be getting 14:1, and it's either delivering 13:1 or correcting the same amount- then on the MAF transfer function, lower the air flow at that point where it's wrong.

On the basic PW table- lower the pulse width at that Speed/MAP point to match the error.

On the vol eff table like MS uses, lower the amount by 13/14 - (or let the auto-tune do that for you).

While tuning the air calculation does take a lot of time- it's important to know that all of the tables should be smooth and continuous. There are no step functions (unless there is some flapper valve opening or closing). Even the point that a VCT system moves from A to B- it will be pretty smooth- just sudden.

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/19/15 2:47 p.m.

Next, the fuel.

So the injector is just a valve. It takes time to open, and it takes time to close. But when open, it will flow it's rated amount at the rated delivery pressure. So a 39lb/hr injector (at 40psi) will deliver 39 lb of fuel each hour.

To deliver less, you open it for less time.

For what I have seen, most aftermarket systems just use the 100% on time, and then linearly lower that- so the 39lb injector will inject 19.5lb/hr at 50% time on, and 3.9lb/hr at 10% time on. It's close enough for most work to be correct, but one must note that at the lower on times- the opening and closing times start to have a larger effect on the mass injected.

For a lot of systems- behind the scenes- the time that an injector has to put in fuel is calculated by how long one cycle is. And the longest of the longest time is 720 deg of cycle- which is 100%. (this is for a 4 cycle system- intake-compression-ignition-exhaust- 2 revolutions of the crank).

At 1500 rpm, that's 25 revolutions per second. Or 0.04 seconds per revolution. So a full cycle is 0.08 seconds.

At 3000rpm, thats' 0.02 seconds/rev or 0.04 sec per cycle.

So that's the 100% on time that can be used.

How much fuel can you put in over that time?

Our 39lb/hr injector is 0.65lb/min or 0.0108333 lb/sec. If I only have 0.08 sec to inject fuel, the 39lb injector will inject 0.00086667lb of fuel. Doesn't seem like a lot, but it actually is.

Anyway- to know how much time one needs to hold an individual event on for-

(Mass of air /(air/fuel desired))/flow rate- So lets say that at 1500rpm, the air calculation says the engine is using 4lb/min.

4 lb/min/ 1500 rpm = 0.002667lb air/ revolution.

It's a 4 cyl engine, so for every revolution, 2 cylinders are pulling in air.

(4lb/min)/(1500rev/min) = 0.002667lb/rev / 2cyl/rev= each cyl is getting 0.001333lb air.

If I want an a/f ratio of 14:1, I need to inject 1/14 the mass of that, or 0.00019048 lb of fuel.

How much time is that then?

0.00019408lb/0.0108333lb/sec = 0.01758 seconds. there's your pulse width.

Since we know at 1500rpm I have 0.08 seconds to inject that fuel- that's a pretty easy job to do.

If you have a car that uses a RPM-MAP table to output pulse width- you will need to roughly go through that calculation for every speed/load point to get started.

Most aftermarket system batch fire pairs of injectors- and many allow them to fire at the same time. But I was running a TECII that did NOT allow the two banks to fire at the same time. So at 1500 rpm, the allowed time was only 0.04 seconds. It was a problem to deal with that.

All in all- I like systems where I tell the system the injector flow rate, and it calculates how much on time on it's own. It's much easier to change injectors that way.

So a tuning note here- most will go an bench test injectors. And then trust it. If you can do that, DO IT. No, the mass of fuel isn't going to be perfect under all conditions, but using the bench flow to be correct is one good thing to trust. It's better than other variations. You can either test it on your own bench, or send it out, but given all of the errors- the fuel injection bench calibration is one of the best ones to trust. Set it, leave it alone (unless you see some real oddities in the air calculation). Doing that makes the air calibration a lot easier.

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/19/15 2:48 p.m.

More later- spark is next.

Dusterbd13
Dusterbd13 UberDork
11/19/15 5:52 p.m.

In reply to alfadriver:

Im going to have to read and understand what you already wrote, so take your time. First read through confused the hell out of me, but hopefully it will make more sense after a couple hours of sleep.

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/19/15 6:07 p.m.

In reply to Dusterbd13:

Please ask questions.

My point of view is as an engineer who has done this for a long time- so the way I put it can easily be ineffective.

Strike_Zero
Strike_Zero UltraDork
11/19/15 8:15 p.m.

In reply to alfadriver:

This is GREAT!!

Fuel made sense . . . I have to reread the air part.

Ranger50
Ranger50 PowerDork
11/19/15 8:53 p.m.

In reply to Strike_Zero:

Air is just a given volume over a certain time at a given pressure. It's about as close to a contast in the whole formula....

patgizz
patgizz PowerDork
11/19/15 9:00 p.m.

are you going to be using gmecm or one of the other free obd1 softwares? i might have an ALDL cable around here.

Gearheadotaku
Gearheadotaku GRM+ Memberand PowerDork
11/19/15 9:16 p.m.

This is of intrest to me too. Also, The '91 ECM you are using is slow, try to get a 93 or 94. Its much more powerful and gives more options. I beleive it reads the O2 sensors much more often for a better tune as well. The TBI board at thirdgen.org has more info.

SVreX
SVreX MegaDork
11/19/15 9:20 p.m.

In reply to alfadriver:

Good stuff, Eric.

Thanks!

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/20/15 6:58 a.m.

I'm waiting to move on to answer questions.

Basically because spark is really hard to fine tune- so it's more- make sure you are getting what you ask for- from there....

06HHR
06HHR GRM+ Memberand HalfDork
11/20/15 8:30 a.m.

In reply to alfadriver:

Those are the most informative forum posts I've ever had the opportunity to read. I think you missed your calling, learned more from those two forum posts than my last two years of undergrad in college..

alfadriver
alfadriver GRM+ Memberand UltimaDork
11/20/15 8:39 a.m.

In reply to 06HHR:

I have thought about going back to teach once I retire.

But I do ECU tuning for a living. So...

edit- full disclosure- I've been a product development engineer for 22 years, most of that in powertrain calibration. And the last 15 of that really getting to know emissions calibration and the physics/chemistry around that. If anyone was wondering.

Dusterbd13
Dusterbd13 UberDork
11/20/15 12:21 p.m.
alfadriver wrote: So lets start with air. or some kind of measure of air... I'm going to start with the easy one to understand- a MAF sensor- which measures the Mass Air Flow. Since we mostly know what air fuel to run, if you measure the mass of air, you then know what the fuel mass you need to deliver. It is as simple as that. There are a lot of small calculations that change that down to an event basis, so that an individual injector will inject the right amount of fuel, as the same air mass at 1500 rpm vs. 3000rpm, the amount of fuel injected is twice as much as the other. When doing the individual event calculations, it's basic thermodynamics to estimate the relative amount of air in the cylinder vs. what an ideal charge would be. And the ratio of actual in cylinder/ideal in cylinder is pretty universally called load. That load information can be also used for spark advance- but more on that later. The thermo is pV = mRT, Pressure * Volume = mass if gas * gas constant * gas temperature (in absolute amounts). With that, you can guesstimate the ideal mass amount if pressure is atmospheric, volume is the total cylinder swept volume, and T is normally 100C/373K. And if you know what m should be based on the MAF and corrected down to an event basis, you can estimate P.... Anyway- the point is that on a MAF system, it's pretty easy to calculate the amount of fuel to inject. But many aftermarket systems use a MAP sensor- measuring Manifold Air Pressure. This is where it can get weird. The most simple injection systems will have a table of MAP vs. RPM that outputs a fuel pulse width. I've played with one of those, and while doing ONE calibration is easy- doing a lot, or modifying that for another engine, or changing hardware is a major PITA. Still, getting a desk top calibration is a spreadsheet of ideal gas law formulations away. That will get you close, and then you need to measure exhaust a/f somehow to correct. That also means you know the fuel flow of the injectors. The next most simple system just uses the ideal gas law to estimate the mass. Since pV = mRT, m = pV/RT MAP = pressure. And YES- you do need to measure the manifold gas temp. Once that calculation is done, there's a correction to it- which is kind of called volumetric efficiency. It's not used exactly how it's actually defined, but close enough to not matter. This is how the MS system works- you have a table of numbers of 0-1 (or higher, sometimes)- and the ideal gas law is corrected to the actual cylinder mass. I don't know the specific name of that table- but I will say that it's a pretty elegant solution. Once you have that correction, then you know how much fuel to inject. Done. Above that- you can use the MAP and the air temp to run a series of regressions of mass vs. MAP. That's what most OEM's do- as the actual in cylinder mass being really close to correct is really important. It's not a hard calculation, but not the easiest thing in the world to calibrate- and is normally done using a dyno. The calculation is mostly a linear one, but it's so easy to include a MAP^2 term in it, it's done. The equation coefficients change with engine speed AND cam timing- and even though that seems like a lot of stuff to look up- it's still less than having a bunch of tables correcting voleff. Again, the point of that is to calculate the mass of air, which gives you a simple calculation for the mass of fuel. The way the MS works is really nice- and hope that whatever you get is similar. That way, calibrations are quite portable. You can take a V8 calibration that has a similar cyl volume, and similar cam timings- and start with that as the base calibration for a 4 or a 6 cyl engine. Just getting it going does not require perfection- just something reasonably close. All of the systems are tuned via- are you getting what you are asking for? If you want 14.6:1 in the exhaust- are you getting that? If not, make a correction to the table where it's wrong. That does require that you trust your injection flow rates- but you have to trust part of it just to get going. So on the MAF system- if you should be getting 14:1, and it's either delivering 13:1 or correcting the same amount- then on the MAF transfer function, lower the air flow at that point where it's wrong. On the basic PW table- lower the pulse width at that Speed/MAP point to match the error. On the vol eff table like MS uses, lower the amount by 13/14 - (or let the auto-tune do that for you). While tuning the air calculation does take a lot of time- it's important to know that all of the tables should be smooth and continuous. There are no step functions (unless there is some flapper valve opening or closing). Even the point that a VCT system moves from A to B- it will be pretty smooth- just sudden.

I'll be working with a map system. So if im understanding this, map should be a pretty stable value across the board, directly related to manifold vacuum. Higher vacuum, less fuel. Lower vacuum, more fuel. And as long as I can find a tune that is roughly the same cam and displacement, it will be close enough to run. Then tune map and fuel for air/fuel ratio desired at idle, cruise, and wfo in that order. Am I understanding this correctly?

How do I figure out the map calibration if I can't find a similar setup? Say im starting with a stock tuned port vette tune, an engine that made maybe 200 horse at the wheels. Im going to change it to feed a 425 horse 370 small block. Obviously more volume and fuel per firing cycle, and obviously different vacuum values everywhere. At that point, do I just leave the map table alone and work on fuel?

Dusterbd13
Dusterbd13 UberDork
11/20/15 12:26 p.m.

And you have lost me on fuel. I don't think i understand how to measure the mass of the air. I know that you gave the formula, but its not making any sense.

1 2
Our Preferred Partners
QUQaSjXdNKLHAssR886OHj9i0Rw4U1F1XlRp2k13xKUumExhj39D9061cwdItN5a