How to setup a fuel MAP without a dyno

Give input on tuning as well as any tips and tricks you may have. Also feel free to share base mapping files for various engine types.
FrankCorrado
Posts: 120
Joined: Tue Jul 26, 2011 7:45 am
ECU Model: S80 Pro
Firmware Version: 76
Location: Montreal, Canada

How to setup a fuel MAP without a dyno

Post by FrankCorrado »

Well, I will try to gather my experience and knowledge from the other DTA forum (the one dead as of today on psycode) about how to setup a fuel map without using a dyno.

Where's my beer... ok I am ready. :)

With the huge help of eighty666 from the other DTA forum and the inputs of some others, I was able to understand it is possible to get yourself a certain basic fuel map by yourself. The trick is to use CL (close loop). So you need wideband.

So after you have some fuel numbers to startup your engine and idle correctly, and hopefully being able to drive around without engine stalling, you can set your CL to help you out on establishing a fuel map.

What you want to accomplish is have a CL which gradually hits your target without oscillating.

Before you start, disable Throttle Transients.

First, in Lambda menus, input some AFRs in the Target MAP. You can use either lambda values or AFR values as the units, lambda values are easier to calculate with when you think with % changes, but AFRs are usually easier for the eye to match with the fuel, depends what you prefer. Those numbers may be guesses, you may not know yet what the engine really likes, you may just think it may like 13.5 but in fact it may prefer 13.0. But to start with something you have to guess based on simple logic: you won't set 14.7 at 5000rpm under 200kpa. See what I mean. And you probably won't set 11.0 at 750rpm under 0% TPS (idle). :)

Whether you use TPS or MAP as load doesn't change a thing. Now in your Lambda Parms menu, please, do not use a sample rate too low. That is the first thing to change. If you use 2-3-4 and most probably numbers under 10, your CL will be way too slow to react for fuel changes and apply corrections. Sample rate of 2 means CL checks and adjusts fuel twice per second. Now that may be fine IF you already have a near-mint fuel map. But here I assume you got a crap map and want to smooth it out. :lol:

So you have to increase sample rate until you see your AFR oscillating all the time and almost never hitting target. Take time to check on that, make sure you remain in a stable cell and that AFR never hits target. Then drop one number and see if it works, it should if you increased one number from that number before. What I did is the opposite. I tried with 30 to start with. :lol: Then I dropped to 25, 20. At 20 it was not really oscillating. I tried 21, it was. Since I wasn't really sure about 20, I backed off to 19hz. VR6Turbo uses 15 I think and eigthy666 12 or 13, so above 10 you should be fine. If your wideband still oscillates at 10, drop lower. Like DTA says, if at 1 it oscillates again, you got a very slow sensor and must buy another one!

Then the PIDs... ah the great explanation eighty666 had on that. Can't remember all of it, I read it only once, but basically the PROP (P) is the first correction applied, the biggest. Then the INT (I) is a correction of the P. The DIFF (D) is a correction of the I­. All 3 add up to a certain extent, I think the smallest factor is the D, middle is I and biggest is P. How to find good values. It seems there is somekind of proportions to respect. You could use the method in the DTA manual. But if you start with small numbers you have something helpful. The goal, as eighty666 said, is to apply many small changes so that it hits target from a slope and not oscillating up and down and then on target. It also seems numbers like 750-200-50 is the exact same thing as 75-20-5. I would suggest starting with these numbers (75-20-5), they are from eighty666 and have proven being quite good for me. You could play with I and drop it to 15 as well. If you want to compare different PID setups, give some time for your parms to settle down first. If you don't want to alter your Fuel Correction Map (FCM. it's recorded on ECU, not lost, in the Lambda menu) while you test different PIDs, take a screenshot of it before you switch and if you come back to your old PIDs, edit the map using the numbers from your screenshot. Also it seems too low a P will not be very good, as it is slower to apply the correction, the factor is not high enough to start with. In my case, P60 was not as smooth as P75. I think I may not go under 70-75 in my personal case.

Max % fuel increase/decrease, at first you may use 20-10.
Overrun TPS position, 2 is good.
Overrun RPM turn on, 1500
Overrun lock time, 1.5-2.0sec. That is the time the CL will not adjust after an overrun has started. It gives time for the engine to settle down during an overrun.
TPS Red to lock, I'd say from 10 to 50%. If you reduce TPS by x%, it will assume like an overrun and lock. For me 10% was not enough, the engine and AFRs felt better with 50.
Red Lock time, 1.5-2.0sec. That is how long it will lock, I think it's linked to the TPS Red to Lock.
Allowed RPM band for correct MAP, 100. But you may need something different, check your Main Map Column for RPMs and if you increase by 500 all the time, maybe a bigger number is ok, but you have 250rpm increases, 100-125 is better. If it's too high, it may change cell before it records a fuel correction to the map.
Allowed TPS Band, 1. I think for same reason as above.
Correction MAP filter, 1 at first. Once you smooth out a bit your crap map to get a better one, then you can increase this. Filter=1 will require a lot of passes in each cell in order to get a more accurate #, but that's what you want to cumulate as much data as possible.
EDIT: Lambda Filter time at the bottom, apparently it's better to start with 0.0sec. Then like transients maybe after you apply the fuel corrections from the map you can try to tweak a different number.

Now with this you should be in good shape. Then you drive for 1000s of miles. Well, maybe 2 weeks using the car regularly. After the 2 weeks you can check your FCM and apply the changes. Then go back OL (open loop) and try it out, it should at least be better. From that point you can continue using CL to smooth out even more your map.

It worked for me! I still have bad spots here and there, but overall the driving experience is smoother. So now it makes it easier to tweak. :)

EDIT: One thing to know, the Delay from Start. That is the number of seconds CL will wait before starting after cranking engine. This is basically to allow Startup values to be passed through if you crank above the Minimum Water Temp for CL to work (setup in Lambda Parms), otherwise CL would apply corrections to your startup values. You have to know that for most of the changes you apply in Lambda Parms while you drive, CL will reset itself and go through that Delay from Start again. So if you have 60sec set up and you change the sample rate while you drive, after hitting F4 CL will shut down, wait 60sec and then start up. You must remember that if you check your AFRs and after 30sec you don't see any change don't think it's not working, it's just that your CL has not started up yet! :) I got fooled for a while with that!
Last edited by FrankCorrado on Sat Jul 30, 2011 4:07 pm, edited 2 times in total.
Frank
VW Corrado VR6 Turbo GT3582R getting into a super light kit car
S80Pro
claggett
Posts: 20
Joined: Sat Jul 30, 2011 6:22 am
ECU Model: S40 Pro

Re: How to setup a fuel MAP without a dyno

Post by claggett »

Thank you.

When I get my car back out I will give it a go.


Darren
Jon K
Posts: 602
Joined: Fri Dec 04, 2009 1:45 am
ECU Model: S80 Pro
Distributor: DTA UK
Firmware Version: 73
Location: Pennsylvania, USA
Contact:

Re: How to setup a fuel MAP without a dyno

Post by Jon K »

Frank - thanks for the info. Going to try these PID settings when I get in the car this afternoon.
1992 700RWHP Pump Gas BMW
Image
FrankCorrado
Posts: 120
Joined: Tue Jul 26, 2011 7:45 am
ECU Model: S80 Pro
Firmware Version: 76
Location: Montreal, Canada

Re: How to setup a fuel MAP without a dyno

Post by FrankCorrado »

So Jon, did you try?

For everyone, I have now switched to MAP as load (previously TPS as load). There's been a long debate on the old dead forums about TPS or MAP as load. Basically, if you are happy with one, keep it. I was not that happy with TPS as I have seen in the log files situations it could not handle the way MAP would and I was having either too much fuel or not enough. I could see these situations in real life as well. Maybe a run on a dyno would mostly fix this, but since I like to try things and understand how it works and the differences, I decided to give a try to MAP as load. Also my TPS as load map is now almost top notch, there were no more fuel adjustments I was doing on the map. So I am losing nothing to try MAP as load. And I'll keep the one I feel works better.

You've got to be 50% good to guess the numbers in the map. What I did is simple, after 0ing out the MAP Compensation table (ignition and fuel) I told Lambda Parms to use MAP as load and then I guessed basic fuel numbers to start with. Everything above 100kpa is easy, you just add what was in the MAP Comp table, usually 10% fuel per chunk of 10kpa (120kpa = 20%, 200 = 100%, etc.). So the boost side is done.
For the vac side, everything must be less than what you had in our TPS table. But this "less" varies with kpa of course. So to keep it VERY simple I only adjusted the IDLE cells portion and the 20-30kpa columns, as I wasn't very sure which TPS column would equal to which kpa. These 2 columns I know will require much less fuel, something like 70-80% less. I removed I think 50%. Then the idle cells, I checked my logs for the TOTAL injection time at idle and I used those numbers in, so I can at least start the engine and idle.
Now for the Lambda Parms, the fuel increase and reduction, I added 50% fuel and removed 75%, cuz I knew my vac columns would have way too much fuel. Also, I would change the Delay From Start to Control. Put an extreme number in. Depends what you want. If you want CL to control quickly after startup, put a number like 3sec. But if you want to play with fuel in OL and let the startup values go away, put a number like 120sec. I would say if your startup values in the column you start up are very small, have CL kick in as soon as possible. If the numbers are in the 15%+ range, I'd say wait until your startup go away, so that CL will NOT try to hit your AFR target during startup map.
Oh one more thing, make sure you adjust your TARGET Lambda values as well! They are for the other type (TPS if you came from TPS, MAP if you came from MAP) of load, so it may affect. May, but may not. In my case with TPS as load my fuel values are the same from 60% onwards cuz my TB is bigger than required and the turbo boosts the same pressure at 70% than at 100%. Obviously switching to MAP as load eliminates totally this misconception. Don't forget to think about this for your application.
And don't forget to do your ignition table as well!! I feel it's much easier with MAP as load, cuz the numbers you enter are the real ones. With TPS as load they were a basis point and on that was the MAP Comp table adding or removing timing.

Then all set to start up. I did a WARM engine startup, water was around 90-95, so that startup values would play less in the equation.

VROUMMM!!! It worked. So it means my numbers were not that stupid. loll Ok I was about 50% too much fuel at idle, but I forgot about the startup values so I waited a bit and it went down to 20-25%. I did a few idle rpm taps making sure the engine would not stall at idle while I rev up and down. I had to adjust a little bit, removing fuel in all cases. It turned out the idle injection time was not equal to the number I have seen in my logs with TPS as load, but anyway it was a start. Then once the engine is stable at idle and little rev ups, I did a slow launch, stopped, reversed. To make sure the engine load on launch in 1st gear was ok. It was not bad. So I decided to launch for real. I did a few 1st gear launches and a few 2nd gear shifts, then i checked the fuel correction map and saw some big numbers and many differences. I applied them blindly, cuz there was no way to know how accurate it was and anyway I was still off to a point I could not drive. It helped. So I was able to take the car around the block. CL was doing all the corrections quite nicely. I came back, check the corrections, applied them. It's not good cuz not enough passes through the cells, but the corrections seemed logic and I needed some to help CL and get a better OL map.

Again it worked fine. So I went off with that. After 30mins of driving, fuel corrections are all under 20% with most under 12%. Now it's time to reduce the CL fuel corrections to whatever number seems ok for your situation. The car is very driveable, in fact I feel more punch than with TPS as load. Might be the ignition table I set up differently, I don't know, but it does not ping under vac (and boost) so it's fine.

I am still off on my map, but I will now drive like this for 2 weeks and pass thousands of times on as many cells as possible and apply the fuel corrections then. It worked fine with my TPS map, it will certainly with this one too. :)

I did not change my CL PID numbers, so it means they are good enough to control fuel even when the map is quite bad, without generating situations like removing 75% of fuel and 0.1sec after adding 50%. It hits target with minimal oscillations or none at all. And that's what you need to accomplish in order to trust your CL and have it smooth out your map on the long run.


So that technique, using CL to smooth out your map, really works. Then for the fine tuning the dyno may help.
Last edited by FrankCorrado on Tue Aug 02, 2011 8:12 pm, edited 1 time in total.
Frank
VW Corrado VR6 Turbo GT3582R getting into a super light kit car
S80Pro
Jon K
Posts: 602
Joined: Fri Dec 04, 2009 1:45 am
ECU Model: S80 Pro
Distributor: DTA UK
Firmware Version: 73
Location: Pennsylvania, USA
Contact:

Re: How to setup a fuel MAP without a dyno

Post by Jon K »

Frank I did try your settings for PID. Seem to work well. My only issue right now is that when I let the car generate an EGO Correction table, it only applies it to MAP1. My car runs in MAP2 because I have my Aquamist unit setup so that the gauge/failsafe button controls what map my car runs in. So I haven't applied EGO correction yet.

The only part I am not too happy with on my car is how it revs sometimes. I mean, its 95% perfect. But every now and then I come out of over run and it feels hesitant. I feel like accel enrichment doesn't work well if you have the new over run setting on (where it cuts all fuel).

Still trying to understand what Throttle Reduction to Lock does.
1992 700RWHP Pump Gas BMW
Image
FrankCorrado
Posts: 120
Joined: Tue Jul 26, 2011 7:45 am
ECU Model: S80 Pro
Firmware Version: 76
Location: Montreal, Canada

Re: How to setup a fuel MAP without a dyno

Post by FrankCorrado »

Jon K wrote:Frank I did try your settings for PID. Seem to work well. My only issue right now is that when I let the car generate an EGO Correction table, it only applies it to MAP1. My car runs in MAP2 because I have my Aquamist unit setup so that the gauge/failsafe button controls what map my car runs in. So I haven't applied EGO correction yet.
Oh you got 2 MAPs, that I don't know how it works, but you just experimented it. lol Then I guess you can apply manually the corrections to MAP2. Hopefully MAP2 does not need different corrections than MAP1, if yes maybe DTA (a new firmware/software) should take that into account. 2 MAPS is new stuff for me.

Jon K wrote:The only part I am not too happy with on my car is how it revs sometimes. I mean, its 95% perfect. But every now and then I come out of over run and it feels hesitant. I feel like accel enrichment doesn't work well if you have the new over run setting on (where it cuts all fuel).
Yeah, don't worry I have problems related to overrun-accel pump as well and Kevin also had it before he got everything right. Whether I run TPS or MAP as load, the accel pump situation is bad, it gets too lean in the first 0.5sec and too rich in the second 0.5sec, then it stabilizes after 1.0sec. I don't know if CL can fix this, so far it did not for me, but probably dyno can. Or knowing how to adjust the fuel-transients.

But when I overrun and tap the pedal to get out of overrun, I never have an hesitation there, even with a super bad map. Is that the situation you describe? I think yes if I re-read your sentence. But when you use CL to smooth out the map, you have to turn off transients, so they won't come into play and mess up the numbers. It may feel worse for the time you will use CL, but after you apply the corrections it should then be at least a bit better. Then transients could help smooth it out even better. I have never been able to make transients work correctly yet, there must be something I still don't understand with them. lol
Jon K wrote:Still trying to understand what Throttle Reduction to Lock does.
If I understood what eighty66 told on the dead forums, it means if you reduce throttle by that %, CL will ASSUME it is an overrun situation and will then stop making adjustments for the numbers of seconds you inputted in the parms (red to lock time or something like that). Maybe if you play with that number (throttle red to lock) it may help on your hesitation. Or just try OL, try your map OL and see how it comes out of overrun. My 3rd test would be to turn off the overrun function and use your own numbers on the left columns, both with and without CL working, just to see if there are any differences.
Frank
VW Corrado VR6 Turbo GT3582R getting into a super light kit car
S80Pro
Jon K
Posts: 602
Joined: Fri Dec 04, 2009 1:45 am
ECU Model: S80 Pro
Distributor: DTA UK
Firmware Version: 73
Location: Pennsylvania, USA
Contact:

Re: How to setup a fuel MAP without a dyno

Post by Jon K »

Yeah I run two maps - one is with water/methanol system on, one is without. So the AFR can vary greatly in boost.

Hmm reduction to lock sort of makes sense... sort of.

My car runs fairly well over all. I am not sure on open loop as I drive with CL on, but it drives "well". Its just sometimes transitioning into throttle from over run can feel lean or rich, I am not sure which sometimes. I need to spend more time with it. I know that since using the new over run fuel cut, it is a bit lean when tipping back in. I wish there was a time delay before cut on over run fuel cut. Like 2 seconds of engine braking before over run cuts fuel.
1992 700RWHP Pump Gas BMW
Image
FrankCorrado
Posts: 120
Joined: Tue Jul 26, 2011 7:45 am
ECU Model: S80 Pro
Firmware Version: 76
Location: Montreal, Canada

Re: How to setup a fuel MAP without a dyno

Post by FrankCorrado »

Jon K wrote:Yeah I run two maps - one is with water/methanol system on, one is without. So the AFR can vary greatly in boost.
Ok I see. Then if you are in MAP2 and CL applies corrections, it will log it in the Fuel Correction Map (FCM) without saying it's for MAP2. Then if you apply it will apply what are MAP2 changes to MAP1. Not good, you're right. Then maybe you have to 0 out every correction in the boost area and then hit apply.
Jon K wrote:My car runs fairly well over all. I am not sure on open loop as I drive with CL on, but it drives "well". Its just sometimes transitioning into throttle from over run can feel lean or rich, I am not sure which sometimes. I need to spend more time with it. I know that since using the new over run fuel cut, it is a bit lean when tipping back in. I wish there was a time delay before cut on over run fuel cut. Like 2 seconds of engine braking before over run cuts fuel.
In this situation, CL is not supposed to help, as it will be locked by the parms you inputted for overrun. What eighty666 explained is that when coming out of overrun the engine needs a certain time to settle down on AFR, so if CL tries to apply corrections too early, it's not good, hence the lock times. But I guess playing around with this and other settings after a while you find the right one for your engine and driving style. I haven't fully yet, so. :)
Frank
VW Corrado VR6 Turbo GT3582R getting into a super light kit car
S80Pro
VR6Turbo
Posts: 329
Joined: Wed Mar 17, 2010 9:24 am
ECU Model: S80 Pro
Firmware Version: 62

Re: How to setup a fuel MAP without a dyno

Post by VR6Turbo »

I still say MAF is the best way to measure engine load and MAP for boost control, but DTA doesn't support MAF.

Does anyone know what scale MAF sensors work to as they are clearly not linear. Is it logarithmic?

I've seen a really nice MAF from ProM Racing in the USA. http://www.promracing.com/mass-air-meters-c-2/pro80-p-5

As you can see, it can meter air for a good deal more power than an OEM meter and in a relatively small 80mm tube. There must be a way to get one of these to work on the DTA? :| :D

Frank and I discussed a typical lean condition we see when using TPS as load. The scenario is backing the car out of the garage and just brushing the throttle enough to reverse out slowly. This TPS movement is so small it doesn't go into the next column but just goes up the rows instead....straight into an over-run cell!! Obviously that little tap on the gas pedal represents a far bigger demand in air the ECU can't detect and subsequently doesn't fuel for. We've both tried cramming in loads of tip in fuel around that area but it doesn't help.

The same goes for this tip-in inadequacy when cruising on the freeway. Usual scenario. You're going along at a nice pace when someone pulls into your lane. You back off the gas pedal enough to slow down, but not enough to go into over-run. Then the person moves over and you tap the gas and it goes lean. That has been the bain of my life ever since using a DTA. I simply can't iron that one out.

I'm pretty sure I could with MAF :D Making injector duty directly proportional to air density makes so much more sense to me. TPS and MAP as load are pure guesswork as to how much air the engine is consuming it's no wonder we have all these tip in issues.

And overrun needs to be load based, not throttle based. I've told Allan that countless times but he just doesn't get it. OE ECUs go into over-run based on load + rpm + TPS. In fact OEs use 3D maps for everthing! For example my E34 525i and Polo 1.8T can both show over 10% TPS but if the air mass is less than X gramms/sec, it goes into overrrun and you don't feel the motor coming into and out it either! Having TPS do the over-run makes the motor snatchy as it comes out of it. It's crap quite frankly.
FrankCorrado
Posts: 120
Joined: Tue Jul 26, 2011 7:45 am
ECU Model: S80 Pro
Firmware Version: 76
Location: Montreal, Canada

Re: How to setup a fuel MAP without a dyno

Post by FrankCorrado »

Yeah well OEs are the best for sure and since they all use MAF it means it's really the best way to accurately measure air. But we cannot match OE with our aftermarket setups. :)

On that situation coming out of the garage, engine warm it seemed a bit smoother with MAP as load, but the real test is engine cold with a better map. So at least 2 weeks for me before I can test this. :) Seemed better cuz the map is more sensitive, but I still saw the leaner situation, though not as bad as with TPS.

I'm happy I don't have any issues with the overrun settings, though. Maybe the behavior can well be engine dependent?

What does rpm do in overrun for OEs? If it's less than 2000 it won't go into overrun, things like that?
Frank
VW Corrado VR6 Turbo GT3582R getting into a super light kit car
S80Pro
Post Reply