Need Help With Remap Article

[fr0st]

I've now got my hands (well getting hands on... it hasn't arived yet) a spare RB20DET ECU to conduct some mods on. I've compiled a heap of info on remapping that I'm going to put into an article for the tutorial section once I have some pics to go with it.

I've got all the info for things like removing R&R, the speed limiter and allowing for larger injectors and AFM. What I need is a method or information on modifying the fuel and ignition maps.

The way I see it is an engine is a closed system. The AFM measures the air going into the system which is also the air leaving the system (ignoring the extra intake during spooling). Since you can measure the exact amount of air leaving the engine through the cylinders the ECU puts in fuel for a 14.7 A/F ratio plus the offset in the fuel map. Obviously this isn't the case in real life. What I'd like to know is what mods cause a change in the A/F ratio and by how much (approximately).

For example (and probably not true), putting in a less restrictive exhaust creates less back pressure, hence less exhaust gas is left in the cylinder after the exhaust stroke so theres more fresh air that can be can be combusted creating a leaner A/F ratio.

Ignition I'm kind of lost since I can't real time tune like a PFC that has a knock reading. Can anyone provide info on 'safe' ignition timing that would be a better compromise than the stock timing?

The original maps have several retarded positions under load (probably the retard out of R&R) that could be changed to improve them, but I'm looking for a little more.

Also, if anyone has a dyno graphs with AFR's for RB20DET's using the stock computer it would be very much appreciated if you can post it up with a mods list. Hopefully I can gain an idea on what mods affect AFR and by how much by looking at real world results.

Thanks for the help in advance

-Matt

[Warpspeed]

Forget it Matt.

You will never be able to figure out where in many different maps the engine is actually operating. And even if you knew where all the maps were, and what variables each map applied to, there is still no way to intelligently change the data.

Suppose you begin with the airflow meter signal, that is modified by air temperature, which is modified by engine Rpm, which is modified by water temperature, which is modified by rate of engine acceleration/deceleration, which is modifed by throttle position.

There may be six or more maps, and the location on each map is constantly changing.

So maybe your engine pops and bangs under light throttle acceleration at 3,200 Rpm, but only on a very cold day. What would you change to fix that ???

There will be many hidden maps with no way of knowing which is which, or if the values stored there should be increased or decreased, or even where on all six maps the engine is actually operating. There will be thousands of stored numbers with no clue as to what any of them do.

Just changing things around blindly will get you absolutely nowhere.

[paulr33]

it cant be that hard. you just need to be a bit gung ho and take a stab at it. and of course be prepared to face failure if it occurs. i did my fc on the street and it turned out ace.

i did ign and inj and havent had any problems. for timing you simply increase slowly each of the cells until it knocks (on the FC anyway) once you see knocking say past 10 or 15 then thats about the highest u can go. with inj you simply do it in real time on the street with a wideband and aim for 15s on light cruise, mid 12s on max load and smooth it around the middle. not sure how u would go about remapping in real time however

[Adriano]

Dont be detered by others, its really not hard at all, and i think you end up with as good a result as a pfc. I have the software the deencripts the code(you need an eprom burner)and puts it into the relevant maps(low and high octane fuel maps and low and high octane ign maps) They are also graphical and the numbers you tune in are direct air/fuel ratios ie 14.7.. but you have to convert those numbers into hexidecimal which is a bit tedious.You can select where closed loop is(you add a set amount to the hex value to turn on that function) We did it very successfully on a r32 gtr and z32 300zx.PS if you get a consault interface then you can see what cells you use in a power run in map trace. Live editiing is also possible with an eprom emulator. Hope this helps. PS shoot us you email and i will find the software(i think there are 3 different programs we used.

[rob82]

If your willing to spend lots of time you could get it somewhere close, what you've got to realise is that the PFC is just like a standard ECU only with half the sensitivity being that MAF transfer table is half the size. The thing to remember with both is that in using an AFM you are measuring the mass of air entering the chamber (grams/sec, or grams/cylinder) at that instant so techically all you need to do is command an AFR and that's what you should get at the tailpipe.

But in order for the maths of the equation to be right you need the injector sizing to be correct along with the MAF transfer table or "MAF ramp tables". These values are depedant upon each individual sensor and it age/wear. For dynamic injector operation the injector latency must be tuned accordingly.

Your really limited to how good the tune is by not being able to emulate. With our emulator and racelogic software(no longer available to anybody) we are able to watch the addresses being accessed at that moment in time and edit that point. It would take me about 6 hours I reckon to get a tune somewhere close to a 1 hour emulated tune.

[fr0st]

You can actually trace the fuel map based on AFM voltage and RPM.

The AFM voltage is stored in an 4 bit number i.e 0 to 64. This is looked up on the VQ Table to get a VQ value (16 bits)

(VQ x K Value / CAS Value) / Number of Cylinders

CAS = rpm / 50 x 256

Say we have an AFM value of 3.68v, this has a VQ Value of about 45000 (not actual value). The K value for a stock ECU is 257. If we had the revs at say 5000rpm we have a CAS value of 25600. So our TP is

[(45000 * 257) / 25600] / 6 = 75.29

The actual value looked up is 75 since the ecu only works in whole numbers. The fuel and ignition maps have RPM and the Theoretical pulsewidth (TP) as there axis's. Now we have the TP you can use that and the rpm to trace the fuel and ignition maps.

Simple huh?

The consult interface sounds a little easier than datalogging the MAF sensor output and rpm to trace maps. The narrowband oxygen sensor is obviously going to be useless, so I'm probably going to be datalogging rpm and a wideband sensor anyway for A/F ratios.

Whether or not I can come up with an emulator depends on the speed of the eeprom and my microcontroller skills. Eeprom speed is likely to win there. Reading, writing and modifying the eeprom is now problem. I found all the software and already have the eeprom reader/writer.

Noone has a tomei or MINES rom dump they mind sharing do they? I wouldn't publish it, I'm just curious to see what they've changed in terms of timing and the fuel map.

Cheers

-Matt

[rob82]

You can actually trace the fuel map based on AFM voltage and RPM.

The AFM voltage is stored in an 4 bit number i.e 0 to 64. This is looked up on the VQ Table to get a VQ value (16 bits)

(VQ x K Value / CAS Value) / Number of Cylinders

CAS = rpm / 50 x 256

Say we have an AFM value of 3.68v, this has a VQ Value of about 45000 (not actual value). The K value for a stock ECU is 257. If we had the revs at say 5000rpm we have a CAS value of 25600. So our TP is

[(45000 * 257) / 25600] / 6 = 75.29

The actual value looked up is 75 since the ecu only works in whole numbers. The fuel and ignition maps have RPM and the Theoretical pulsewidth (TP) as there axis's. Now we have the TP you can use that and the rpm to trace the fuel and ignition maps.

Simple huh?

The consult interface sounds a little easier than datalogging the MAF sensor output and rpm to trace maps. The narrowband oxygen sensor is obviously going to be useless, so I'm probably going to be datalogging rpm and a wideband sensor anyway for A/F ratios.

Whether or not I can come up with an emulator depends on the speed of the eeprom and my microcontroller skills. Eeprom speed is likely to win there. Reading, writing and modifying the eeprom is now problem. I found all the software and already have the eeprom reader/writer.

Noone has a tomei or MINES rom dump they mind sharing do they? I wouldn't publish it, I'm just curious to see what they've changed in terms of timing and the fuel map.

Cheers

-Matt

Its all the other parameters apart form timing and fuel thats interesting.

[Warpspeed]

it cant be that hard. you just need to be a bit gung ho and take a stab at it. and of course be prepared to face failure if it occurs.

True, but failure could result in a severely damaged engine.

I did ign and inj and havent had any problems. for timing you simply increase slowly each of the cells until it knocks (on the FC anyway) once you see knocking say past 10 or 15 then thats about the highest u can go.

No that is not how you adjust ignition timing. You advance timing until there is no further increase in torque on the dyno, then just back it off so it is right on the edge of dropping off. In other words you run the minimum amout of timing for best power. Your method is to use the absolute maximum timing the engine can possibly stand, just below detonation. That is both foolish and dangerous.

With inj you simply do it in real time on the street with a wideband and aim for 15s on light cruise, mid 12s on max load and smooth it around the middle. not sure how u would go about remapping in real time however

What you say is true, but you absolutely MUST set the correct air fuel ratios before adjusting ignition timing. Air fuel ratios will definitely change burn rate and require different ignition timing. You suggest setting the ignition timing first, which is totally wrong.

You say it is not hard, but you have a power FC with the plug in control module. He has just a bare standard ECU with no keyboard or screen readout.

Try adjusting your FC without the hand controller, and then tell me how easy THAT would be.

[fr0st]

I just had an idea....

Since the consult cable can read things like MAF Voltage, RPM and oxygen sensor voltage at the same time, I could replace the standard 02 sensor with a wideband and add a little circuit to give a linear voltage output between 0 and 1v. As long as the ECU doesn't convert the o2 sensor to lambda or an AF ratio I should be able to log a reading without needing an external data logger.

No that is not how you adjust ignition timing. You advance timing until there is no further increase in torque on the dyno, then just back it off so it is right on the edge of dropping off. In other words you run the minimum amout of timing for best power. Your method is to use the absolute maximum timing the engine can possibly stand, just below detonation. That is both foolish and dangerous.

Athlough you are right, I dout that you can reach the minimum amount of advance for the most power at high rpm without knocking beforehand. Low RPM you certainly will though.

[paulr33]

rightly so but if he wants to have a shot it at and has the right tools it shouldnt be too hard. im not saying to remap it without any tools at all im saying to remap it with the right tools. thats a little different to programming an FC without a handcontroller, of course you can't do that.

[Warpspeed]

It all depends on how far below the detonation threshold you are running your engine.

If the engine is being tuned on a dyno well below full throttle, and if you keep advancing the ignition, the power will rise up to a peak, stay constant over a small range of ignition timing and then fall away without seeing any detonation at all.

Hopefully, even at full throttle and full flat out boost, the engine power will peak well before any detonation occurs. If you keep advancing it further, it probably will detonate, but that is certainly not the optimum position to set the ignition timing.

But then the engine might be a horribly overboosted heap of crap, running with no intercooler on cheap unleaded fuel. It will probably then start to detonate fairly early with power still increasing. Only then might you use the dubious tuning technique of running it right up to detonation and backing it off 0.01 degrees.

If your engine is like that, good luck to you. It will blow up for sure, and probably fairly soon.

There are two very firm facts to keep in mind. Detonation destroys engines, and your engine will make more power when properly set up with correct ignition timing, than it will overboosted, with the timing backed off to stop the resulting detonation.

The other fact is, for best turbo response and fastest spooling, you need the highest possible exhaust gas temperature. Hot exhaust gas expands and drives the turbine, cold exhaust gas cannot do that quite so well.

If you want high exhaust gas temperatures, you run LEAN mixtures, and ignition as far RETARDED as you possibly can without actually losing engine power. The turbo will absolutely love it while it is trying to spool.

Those 12:1 maximum power air fuel ratios, and that super advanced ignition may be the hot tip around the pub on Saturday night, and an excellent way to tune so you end up having a lag monster. If you want your car to run really sharp, and be fast and responsive, those hot tuning tips will not get you there.

Anyone that thinks engine tuning is all really simple and basic, has a very great deal to learn.

Also anyone that says they can just look at a standard fuel or ignition table and decide how to "improve it" without exensive dyno testing with proper instrumentation are just kidding themselves.

[fr0st]

Whilst I get what your saying, the whole point of writing the article is to learn for myself aswell as inform others. At the moment I'm going to leave out fuel and ignition tuning and only have ECU modding and R&R, speed limiter, injector and AFM changes.

I have a few ideas on how to accomplish fuel map and ignition tuning. If you could give your opinion it would be very much appreciated

For fuel map tuning, in order to keep in the diy spirit I'm working on a custom circuit to read a 5-wire wideband sensor which replaces the narrowband sensor whilst your tuning. Since you can read the sensor through the consult interface you can also log it along with RPM, MAF voltage etc. The consult also does fuel map tracing so you can tell what cells need adjusting. This is all in the hope that the ECU doesn't scale the O2 sensor voltage when it gives the value to the consult cable. The datascan website shows screenshots with a 0 to 100 reading, so if its linear with would work perfectly. I pick up a consult cable later today so I'll be able to check soon.

For ignition I had another (perhaps) clever idea.

From what I understand, if you advance timing too much the combusted gas starts to push down the piston before it reachs the top of its stroke and hence creates a knock. The minimum best timing would be when the ignited gas pushs down on the piston exactly after it reachs the top of its stroke. Any more timing than this and knock should be present, even if its only small amounts. If I was too hook up the engines existing knock sensors to the microphone inputs on a laptop through a small amplifier I'd be able to record every small vibration in the engine. If I ran it through an FFT program to look at what frequencies are being generated I should be able to pick out anything irregular (like a small knock).

It sounds good in theory but I'm not sure if it would work particually well in practice.

Any suggestions?

[Rowdyr32]

Good to see some others having a go. I have just made my burner from dicksmiths and bought an eraser for $10 from Jaycar on sale. Looking at all the maps I have of 26's(around 10) they nearly all use the same ignition timing as stock down low with some either added or lowered up top in the load/recs. They seem to use all the factory settings with injectors. The fuel tables go the same way as the timing. Mainly adjusted up top.

Is there a link to a page that converts the hex inj sizing to cc's. Some bins have 00B4. Might be 550cc?

They say the factory ecu runs to rich higher in the revs with to much timing. Is this true?

Also do the factory knock sensors run off a voltage or frequency?

If some of you have some bin files could you please send them to me so I can have a look. Just messing around with the burner at the moment.

[fr0st]

Good to see some others having a go. I have just made my burner from dicksmiths and bought an eraser for $10 from Jaycar on sale. Looking at all the maps I have of 26's(around 10) they nearly all use the same ignition timing as stock down low with some either added or lowered up top in the load/recs. They seem to use all the factory settings with injectors. The fuel tables go the same way as the timing. Mainly adjusted up top.

Is there a link to a page that converts the hex inj sizing to cc's. Some bins have 00B4. Might be 550cc?

They say the factory ecu runs to rich higher in the revs with to much timing. Is this true?

Also do the factory knock sensors run off a voltage or frequency?

If some of you have some bin files could you please send them to me so I can have a look. Just messing around with the burner at the moment.

I havn't had time to download my ROM with the consult cable but the standard rom I have says the stock K value is 0101. Thats about a 40% decrease in fuel. If stock injectors are 270cc then the file would be for 380cc injectors.

I think the factory ecu just runs rich all the time except on closed loop cruise.

IIRC knock sensors are bits of piezo so they just act as microphones. So you'd get a frequency and a voltage. The higher the voltage, the higher the volume. I'd assume there somewhat tuned to pick up pinging so the ECU could just work off voltage and not worry about the frequency.

Apart from a stock bin, I've found nothing