Home |
Products |
Syclone 1178 |
_____________ |
Tech Articles |
Engine |
Transmission |
Chipping |
Electrical |
Misc |
_____________ |
Links |
JS Mfg. |
TurboTime |
CanadianSyTy |
ChitownSyTy |
SyTyArchives |
SyTyAuthority |
SyTy.org |
SyTy.net |
_____________ |
EFI Tuning Tips (especially GM EFI)INTRODUCTION This purpose of this document is to assist someone in the tuning stage of an Electronic Fuel Injection (EFI) project. It contains discussions about the monitoring equipment required and the methods and strategies for the actual tuning. The information is based on the experiences of tuning a GM EFI setup on a small block. However, the methods should be applicable to most any engine. This paper doesn't reveal the location of the maps and tables in the ecm software but does discuss their functions. The maps and tables for some ecm's can be found on the Internet. For example, the Syclone/Typhoon and the Buick T-Type fanatics have found most, if not all, the needed information to alter the EFI systems for the extreme modifications on their cars. Look out, here comes a warning. You can really mess up an engine if you don't understand what is going on and are careful and methodical about tuning your EFI setup. This is not lawyer gibberish, this is for real. You can turn that expensive, close tolerance machine into a pool of molten crap real quick. I personally know of a kid who had too much money and not enough smarts. In the early days of EFI, he bought an O2 sensor monitor and a device to alter the O2 feed back signal to the ECM. He didn't hook up the O2 monitor properly, started twisting the knob on the signal modifier, leaned the fuel mixture too much, and melted the pistons. It can be done! After all the time and effort you have put into achieving this stage, it would be a shame to ruin something because your enthusiasm carried you away and you tried to tune too fast. WHAT IS BEING MODIFIED When tuning the EFI setup, the data controlling the fuel mixture and spark timing is being altered. Each engine type and setup will require a unique set of data to achieve it's ultimate power and efficiency setup. The camshaft, intake manifold, head design, and exhaust system all influence the torque curve of the engine and therefore the fuel requirements. The spark advance plays into this also. The following section discusses the different tables found in some ecm's. LIST OF TABLES Tables, Maps, Switches seen or heard of: Main Fuel (VE or LV8); Fuel Enrichment WOT; Main Spark (rpm vs Kpa or gms/sec); Spark added WOT Idle Speed; IAC steps (by cool temp); WOT enable TPS %; TCC lock/unlock by road speed; TCC disable by TPS %; TPS Enrichment (sensitivity, quantity, duration); Injector constant; AFR % enrichment vs Coolant temp; EGR temp enable; EGR enable % TPS; EGR timing correction table; EGR fuel correction table; Speed limited MPH; RPM Limiter; Degrees retard by knock sensor by MAP; Degrees retard by knock by rpm (WOT); Turbo Boost Control; All diagnostic malfunction disables. Not all ecms use all these tables!! DISCUSSION OF TABLES The following tables are for a 3020 lb. car: Some details are not mentioned (gotta keep an edge). Table 1 Spark Table (RPM vs Vacuum Signal in kpa) 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100 Table 2 Power Enrich Spark (RPM vs Added Spark Advance) 800 0 Table 3 Volumetric Efficiency (MAP vs RPM) 20 30 40 50 60 70 80 90 100 Table 4 Volumetric Efficiency WOT Enable (RPM vs Additional fuel factor) 0 10 Table 5 %Throttle Position Sensor for Wide Open Throttle (RPM vs % factor) 0 35 These tables are very different
from the OEM setup. But, for this engine combination it works.
The items to
note are the lean/retard overrun (high vacuum values when you lift
your foot off the accelerator pedals at speed
) areas. Glance at the spark table to see when the engine wakes up.
Note that the VE table values can add up
to over a 100%. Also, because the values in the
%TPS for full throttle table are low, and with the existing VE
in the first table, the system will change the air fuel ratio to rich
instantly. Might also note the funny timing under TUNING TOOLS In order to know if your wonder toy is about to melt down on one end off the air fuel ratio spectrum or put a smile on an Arab Sheik's face on the other end of the spectrum, you need to monitor the fuel injection system and some of it's sensors. Here's a list of tools: Digital Volt Meter (DVM) - (10 meg ohm impedance) for monitoring the Throttle Position Sensor Vacuum Gage - For idle and part throttle analysis; Tachometer - If you don't know, if you've got a lot of learning to do; Performance analyzer - To give you an unbiased reference of the car's performance changes other than the uncalibrated seat of your pants; Scan Tool - To know what values the ecm is using internally; Oxygen Sensor Monitor - Usually a 10 element LED Display. The 10 LED AFR Display, and Knock detector schematics are at the GN T-type tech area. O2 Sensors and Indicators As clarification, this O2 discussion is about/for conventional V-8s, Small Block Chevy, Small Block Ford, Big Block Chevy, Big Block Ford. While there are probably numerous others, there are a bunch of VTEC types that this may not be accurate for. Monitoring this sensor will keep you from melting your engine. From some personnel notes from
the early-mid 80's there were many 2 gas analyzers that went to
market for
the then pending inspections for CA. Here are some conclusions drawn
from some GM training center notes
and a Radio Shack engine book. First from a slip of paper, circa '83,
that lists O2 sensor volts relative to AFR
Air Fuel Ratio, and one from 2-3 years ago from Cyberdyne, who manufacturers
a 10 led AFR Monitor. The
voltages match i.e.: 1v=17:1; .3v=16; .5v=14.9
; .7v=13.8:1; .9v=12.7:1. Also, going by the RS book are some
differences in response time vs temp, and differences in sensor voltage
depending on load. So yes there are variables, Hopefully, when starting with any custom ecm or drive train combination, you'll have something to start with, and better yet at least an eprom from the same engine series. If not, then we'll cover both anyway, again with the GM OEM, because that's what was experimented with. Regardless of what your working on you have to pick a spot to start with. This involves pulling numbers out of the hat to begin with. The more you do, the smaller the hat gets, because you can rule out some things. Idle A suggestion is start at idle. For a SBC the initial timing for the distributor is 6 degrees before BTDC. How do you setup 6d without a timing light? On GM style distributor there is a set of little arrows inside the distributor. On a V-8, 8 on the reluctor, and 8 on the sensor ring. Just line the marks up and tighten the distributor clamp. The engine fires when the arrows
are ever so slightly past each other in the direction of rotation.
The second you get
to about 400rpm is when the engine starts running off your ecm control.
Less than that and on a cold day the engine
will windmill when it's cranking. Fill in the 400-800 RPM areas of
the timing map with 16d (Table 1). Also, fill in
the lower rpm VE tables (Table 3) with 30, and try to start the engine
without grinding the starter too much. The
engine will warm up enough to wake up the O2 sensor, look at the scanner,
and see where the BL is, and IAC Now you have some idea of where the mixture is, and what the Kpa is from reading the scan tool. From experience you can't run the same numbers everywhere at idle. On the high side of idle Kpa go richer, ie; with an engine that idles at 60 Kpa if it is 30 there, at 65-70 it would be 32, and at 50-55 28. Yes, the 30 is a lot but its easy to go lean, and clean up the idle, then having stall all the time, and having to wait for the prime shot to burn out. Next, lean the VE values (Table 3) down till the O2 is swinging (value moving back and forth through .45 volts) , the IAC is less than 10, and you have a min idle speed. For the initial butterfly opening I cut a strip of paper a 1/4" wide and set the throttle idle setting so when the paper is pulled through the butterfly opening with a minimum of effort. If it tears the paper open it up. So now the engine starts and will idle. This is a good start. Light Cruise This section is on how to do light throttle cruising set ups, using GM ECMs. This is for an engine using a slightly hotter than stock cam, healthy ignition and assume some roughing out of the fuel timing tables have already been done. The objectives are: Advance timing as much as possible consistent with minimum throttle opening; and use just enough fuel to run 128 or slightly lower BL numbers. For monitoring the engine, a
digital volt meter (DVM) is attached to the Throttle Position Sensor
(TPS), a
vacuum is gage is hooked up to the manifold but this is almost redundant
of the TPS, a scan tool is hooked up
for looking at the Block Learn (BL) numbers, and a 10 element LED display
is attached to the oxygen sensor.
With this wired up it is easy to now start changing the timing of the
40 kpa areas of the fuel map by rpm for
cruise. On a flat road with the car in top gear, find the cruise speed
at 40 kpa and note the DVM signal from the
TPS. Add timing and repeat the road test at the same TPS value. Observe
whether the vacuum value is above From the 40-50 kpa vacuum value, again depending on your car, drop off as little timing going toward 100 kpa (full throttle), as possible. Run enough fuel in the VE table so that when pulling slight grades the motor doesn't chug (love that technical language). Also use a minimum amount of timing for the WOT spark value. A suggestion is stay on the low
side of the BL 128 at cruise so the engine is instantly smooth
and pulling
when the gas is nudged. There is no difference in MPG between this
and 128. If you have the capability for
pulse width modulation (PWM) of the EGR I'd turn this off during the
initial tinkering so that you don't chase
too many items at one time. Just doing the spark and fuel can be an
exercise in patience. Also, with a big cam
just off idle and just as you go to light cruise at say 30-35 mph,
you may have a dodo area (in this area of the
prom maps, BLs are lousy regardless of timing afr), What is happening
is the engine is starting to wake up and Perform the light throttle cruise initial tuning in whatever gear is 1:1, or less (overdrive). Doing the tuning while giving the engine the advantage of gearing might leave you having to do a major retune. On some of the eproms you can lower the TPS enable settings. If your running an auto tranny watch it. If you have a throttle value (TV) cable sensing transmission you must have enough line pressure increase to keep from smoking the tranny to handle the extra HP at low TPS settings. WOT General Concepts Remember, this is from personal experience, and the example vehicle is the vehicle used in the light cruise tuning. DON'T DO THIS WITHOUT ADULT SUPERVISION. In other words, be careful, patient and methodical. To review, starting with a GM ECM with initial programming set to 10-14 degrees WOT maximum retard. Set the decay rates slow. Set the total timing amounts to 30 degrees. Set the WOT AFR to 11.5. Hook up a scan tool, LED O2 sensor display, and performance analyzer. Now adjust the F/P so that you can drive at WOT. Watch the O2, as you drive (OK glance at it). After the pass is over, look at the knock counts. If the O2 sensor was plenty rich (all ten LEDs Lit), and no knock counts advance all timing 2 degrees and lean the WOT A/F mixture to 11.7 Repeat driving test. A few words about the last item, analyzing performance requires something self triggering, and easy to use. A home grown performance analyzer is described in Appendix A. Now adjust the F/P so that you can drive at WOT. Watch the O2, as you drive (OK glance at it). When the pass is over, look at the knock counts. If plenty rich (all ten LEDs Lit), and no knock counts, advance all timing step 2 degrees and lean to 11.7 Repeat driving test. Plenty of LED, and no knock counts, add 2 degrees more timing and lean the mixture to 11.9 . Somewhere now you'll start either slowing down, or tripping the knock sensor. Usually the knock starts triggering between 2000-2600 rpm. The WOT AFR stored in the prom can be a wide range of values depending on your injector sizing, and PSI. Remember this is an ecm calculation and has nothing to do with reality. During all of this do some plug cuts, and learn how to read the removed plugs. If the F/P is changed then the fuel VE values at part throttle in the prom tables will have to be adjusted. Some juggling back and forth between the values in the VE table and the WOT AFR will be expected. Now, the ecm has
been told a huge lie. This means we didn't know where the injector
constant lives. The
TPS
Enrichment (EFI talk for carb accelerator pump) location was unknown,
so for this application huge injectors
were used. The engine is running a 90#/hr injector, with an injector
constant in the eprom size of 65. This is
not
the best way to do this. But, by creative use of Fuel Pressure Regulators,
and time consuming recalibration of the
VE tables, the system was made it to work, but some of the constants
looks really strange, i.e. WOT AFR of To achieve the eprom files for this setup, burning 400 chips was not required to get it right, 400 chips were burned to get it PERFECT. More than likely 98% of the world would never notice the difference in the last 100+, and the first couple of 100 was figuring out how to do the above. PS When doing the above settle for the least amount of timing consistent with the best performance. Trying to run maximum timing results in hitting the knock sensor and having varying ETs. IF you run too much timing you may have to disable the knock sensor diagnostics because the car is so happy with yesterdays tuning, and today the ECM is using the OEM specs and can't run the diagnostics, and the fault codes get pinged. (This setup is that way) Mild Performance Engines, Engine Swaps Now things take a step up, tools are still the same, but now it would be nice to have a scan tool, With probably a different cam the operating range of the sweet spot of the engine has changed, or the weight of the vehicle has. This weight thing is not to be taken lightly (HOHO HAHA), just had to sneek a little humor in. A light car and/or one with higher numerically final drive gears will want a lot more fuel earlier, or more timing taken out sooner. So now you'll have to have some control of the timing. But still if you're using a distributor, you can get by with just timing at cruise or WOT. If it's a mild cam (and if your doing this for the first time I would highly suggest you stay conservative). Now use your scan tool and look for the magic 128, try adding/subtracting 2-4d here and there do the 4d to get a feel for things, and do not use any WOT testing when going 4 at a time. Real Fire Breathing Hot Rods (Things that go lump, lump, rump, rump at idle, non lock up converters, spin tires in any gear less than overdrive.) Now ya gotta have all the tools. Table changes will include: Main fuel table; WOT enrichments; Main Spark table; WOT spark adders; TPS Enrichments (sensitivity duration quantity); Enable WOT TPS % settings; Cooling fan operation; Idle speed. This engine will be able to do your basic China Syndrome Routine!!!!!!!!!, and it don't take long at WOT. Once ya get to this stage, ya have ta have the scanner, DVM, vacuum guage, full time fuel pressure guage (AND ISOLATOR), and 10 led AFR Display (full time), wouldn't hurt to have a audible knock detector Ok higher compression, no big shakes Small chamber/AL Heads generally make a difference, for either less total timing, at WOT Al Heads need more timing during light throttle cruise. Non-Lockup converter, more initial enrichment (less drivetrain inertia). There are several other reasons for the non lockup, being you'll start needing more fuel earlier ya gotta lower the % TPS WOT enable. This will beat the dodo out the TCC linings.. (I bought 4, before I figured this out), and the linings ruined the tranny internals causing 2 overhauls. Now I'll be the first to admit to being hard on things, (I'm a severe duty application). At this level they require constant attention, to prevent damage. With EFI, and knock protection usually there is protection from things going wrong. But when things go wrong ya can't forecast what the problem will be. Also at this level you might consider running a fuel separator, which is basically a large tank for the sediment to separate in. So there is no pressure drop due to a restricted fuel filter. GENERAL TUNING TIPS ( Abbreviations: WOT, wide open throttle; DVM, Digitial Volt Meter; AFR, air fuel ratio; FP, fuel pressure; VE LV8 mixture; TPS, throttle position sensor, TCC; torque converter clutch) No matter what, save your original data!!.. Tinker with STOCK ENGINES, to get your feet wet. One of the easiest errors is to use too much air filter. Actually, the problem is getting the air to organize to flow through the throttle body. Having several hundred square inches of filtering surface area is fine, but having a huge volume of air sitting on top of the butterflies, can have it's down sides. I.E. when the throttles snap open, it takes/can take a huge accelerator pump shoot to cover the delay in this column of air to get moving. If your using an open element filter than try this. Assemble the air cleaner housing, and element as it would sit in the air horn. Measure the distance from the sealing ring portion to the lid, any thing more than an inch is too much. Go to the hobby shop (say you had 3 1/4" dimension) Lets also say the air filter sealing ring was 5" in diameter. You'll need enough thin brass to make two rings, 3" wide and about 16" long. To make a tube 5" in diameter, and 2 1/2" tall. The reason I say two is you have to experiment, on the height, and that means going too far, which will ruin the first one. I haven't found getting more than 1/4" close as necessary. To get involved all you need is the locations of the WOT enrichments, or the VE tables (TBI), or LV8 (TPI) in the eprom tables. With the WOT enrichment/APR you can tune the FP till the cruise feels right, and then enrich WOT. If that doesn't work try leaning things down. Always remember, too rich and you buy spark plugs. Too lean and you buy head gaskets, exhaust valves, rings etc. Or tune the WOT, experimenting with the tables to even things out. Tuning cruise is just watching the DVM for the lowest TPS. On the Fuel tables changing things by 10% should make a noticeable difference. Again move just a couple at a time, till you get a feel for things. If you've only been driving high performance cars for a couple 100K miles, it might take some getting used to for what makes a change. If you've got a ton of time behind the wheel, and a lot of SCCA time, you might have a better sense of things, either way take your time.. Sometimes a FP adjustment can
make the engine respond perfectly. If the fuel pressure is increased,
and the
engine runs better, that means one thing, it was universally lean.
Eventually you'll reach a point when changing it
doesn't seem to make much of a difference, and then performance falls
off quickly. When addition fuel is added
(more than what is needed) the ecm just
leans things down, so the BL changes. When the engine goes to WOT
enrichment it glances over at the BL and if it's low (less than 128),
then it will reduce the injector open time. If I have not read anything that
says you can't run vacuum referenced fuel pressure regulators with
TBIs, but the
pressure output variance is usually too much. But, if you run a second
fuel pressure regulator, you can trim this
variance. This is the only way out of not being able to tune the accelerator
pump features of TBI, since no
commercial program I've heard of allows this to be done. However, once
you have things rather close you can try
other chips from other applications, and put your tables on them. From
the good old days, a big block always took Again rich you buy spark plugs, lean it's internal engine parts. For Spark, not enough timing, and you foul spark plugs, too much advance it's internal engine parts. (too long, too retarded, overheating) If TPS enrichments aren't changeable then try setting the VE high and low, (to try having different BLs), then this BL will start to work as an enrichment during TPS transitions. The reason this works is that the ECM will start moving the IAC counts around. When tuning remember the IAC is a vacuum
leak. Less is better. A NOTE ABOUT TIMING I have done a bunch of playing with timing especially at WOT, and lately have noticed that 2d less than triggering the knock sensor has had no effect on performance. In the past I'd spent lots of time doing dyno work and many times the customer would wind up taking 2-4d out from where we had it. Please note this is in the early 70's, and we were using one we made, and were no where as sophisticated as they are now. Might mention this was also getting the AFR to correct within .2 volts at each timing spec. Try adding timing to the spark table, and remove a similar amount from the WOT spark addition. The same with the fuel table, and WOT enrichment. When you start getting close, change only one thing at a time, one thing only. Move two and things might seem fine but then 2 moves later everything takes a turn for the worse and you won't know where to start. MAF SYSTEM DIFFERENCES On a 89
vette there aren't really the VE tables as shown before, my example
was for a TBI which is a Speed
Density.
The computer knows from various sensors what is going on but has to
calculate the air entering the engine, and then
fine tunes it by using the O2 sensor. So you have to "give it
a range" to estimate where to start these calculations,
and based on that a set of look ahead tables for it to consider. Now,
on an engine with a Mass Air Flow sensor the The listed items are so tightly wound together that I really don't know how to separate them for this document. And any one that has more to say on the subject I would invite to comment further. Again this is from my experiences on the subject matter, and my interuption of what others have told me, to correct difficulties that I've experienced. The ecm is a marvelous thing, but it has no self generating logic, it just follows instructions. So it can be lied to and will never know the difference, and generates signals we humans use as a language to understand what the computer is doing. In other words if the computer is told to generate a WOT AFR of 13.5:1 it will do this using the information that is feed it. So this 13.5 is a calculation, and may or may not be an actual real world reading. On a multiport the injectors are calibrated at 3 bar (bar = 1 atmosphere = 14.7 psi). Atmospheric pressure is 14.7 PSI. Absolute vacuum is 30 inches of mercury. A typical engine develops 15" of vacuum at idle referenced to a fuel pressure regulator referenced to the engine vacuum signal will change the FP approximately 7 PSI (30 in hg = 14.7 psi). So the FP changes about 16% (7 psi/43psi) due to this vacuum ref. Depending on it's initial flow characteristics the pressure regulator control can change some when it is vacuum referenced when installed in the car. Hence for calibration purposes the ecu value can be set for an injector size of 39 for an actual installed injector that is rated at 42. The less linear an injector is the greater this "error" will be. Also consider rising rate fuel pressure regulators, they will change the FP by a rate greater than the referenced change. IE on a turbo motor that generates 14 psi boost a raising rate could generate a change of 28. This will help explain why, when the inj size is increased by 10%, you may/may not wind up changing the base injector width by 10% across the VE tables. This would be a good start, then again tune as necessary. It is your decision to start your WOT ranges at 13.5 or 14.7. Another tuner has had good experience using 14.7. If there is no change possible to these settings, than changing the injector constant will have a very profound effect. To quote from the OTC (Ottawa Tool Company) Scanner Manual " LV8, Filtered Load Variable (LV8), In the filtered load variable mode, a number will be displayed that is proportional to the measurement of the amount of air being displaced per cylinder. The filtered load variable replaces the MAP sensor in some applications." So a MAF reading is grams/second, and a LV8 is a number that is a proportion. Therefore in a MAP set of calibration tables you have VE tables, and in a MAF you have LV8. Both have a table of TPS %s vs RPM to enable WOT. Both have a correction for WOT enable, either a VE adder or % increase in injector timing. ATTACHMENT A There are some nice $100-$150 performance analyzer units out there, and they don't have to be perfect, just reliable. What I did was use the VSS and a Radio Shack Stop Watch. The VSS was tapped at the ECM and that was feed to a LM339. The stopwatch was taken apart an wires added for the power supply and across the start and stop buttons. the power wires went to a LED that was the on indicator the LED voltage drop of 1.5 was enough to run the watch even though it had a 1.6v battery (use a 100mfd at the input) the timer runs off a 7805 volt reg.. wire the on + off wires to a 4066. Wire each side of a 556 to be a monostable on off pulse for the 4066. Going back to the 339 tie that to a divide by 10, set
the divide by 10 to six. There are 3 pulses per tire revolution So
this will take two tire revolutions to start the timer, which helps
reduce false triggering from wheel spin At the reset
connection tie that to a second divide by 10 and use that to trigger
the first of the 556 for the start 4066. Wire the
second divide by 10 to a third one, so that you have a total of divide
by 100, and that will be the stop 4066/556 pulse
so now you have a timer that ignores the first 2 revolutions of the
rear tire, and the counts 200 revolutions to stop the
timer figure 6' rolling diameter ya have a 600' portable drag strip
in car that eliminates error do to wheel spin (there is
something wrong in the math, but paced out was a little over 1/8 mile,
hell it was 7 years ago) just reread it last divide
by 10 may have been set at 8 Breadboard a 555 pulse generator, and
run burn it in, the watch should run to .001 ever
time, No I don't have a schematic, and it was a cobble together but
has been working for 7 years, and I not cutting it |