Page 1 of 1

Alpha n/ itb's on m50b25/dme 403

Posted: Tue Jul 18, 2017 1:45 pm
by Evil
Hi all,
My car ls a bmw e36 with a m50b25, 403 dme.
This week I put a set of individual throttle bodies from a 1000 gsxr.
I run with no maf (unplugged).
I have some problems of lean AFR in fast throttle : AFR reaches at 16/17, and then go down to 11 (I run E85, my wide band controller is a Zeitronix ZT-3, so AFR values I give are the displayed ones).
I think I think that the problem come from my TPS (I use the GSXR one), who gives incorrect values at low throttle and at idle: 1.2v instead of 0.5v). Does someone knows the correct range for these values, at Idle and WOT?
Because of that, I also have no Idle bevause I think that the ecu dont see thé Idle state because of thé wrong values from TPS.
I think I need To modify throttle définition table, I think I have find It in 6E4D, (31x1). I have converted thé table values To airflow %, x axis in throttle % (I think in reality it is TPS voltage but dont know thé conversion factor). See on this picture:

Can someone tell me if thats thé good one and give le some advice about my setup?
Thank b'y advance
Charlie from France

Posted: Wed Jul 19, 2017 3:22 am
by Hairyscreech
I would think you have 2 issues going on, whipping the throttles open quickly is resulting in the air getting to the chambers faster than it did with the single throttle and plenum set up, this is going to throw out your acceleration enrichment.

Second thing the acceleration enrichment is based on the change in TPS value, with the higher TPS voltage map sites getting less fuel.

Not using a M5x TPS sensor is really going to bugger things up as well, the ECU is expecting the TPS to be ~0.3-0.4v at idle and 4.2v+ at WOT, anything else is not going to play nice without having to change quite a bit of code.
In this case the correct M50 TPS is cheap and adapting one to the throttles should not be too hard, especially as you are relying on the TPS to give all the load data to the ECU.

You should be able to use a TPS from an M52, M54 or M60 as well, many BMW TPS sensors are the same electrically.

Posted: Wed Jul 19, 2017 4:41 am
by Evil
Hi Hairyscreech, and thank you for your answer!
I see what you mean, I'll try To adapt the M50 TPS. Do you think that tuning the TPS définition table, To tell the ecu the higher airflow in low throttle opening, will solve the acceleration enrichment problem?

Posted: Thu Jul 20, 2017 3:03 am
by Hairyscreech
100% you will need to adjust the acceleration enrichment table, If your fuel tables were ok before adding the throttles then that may not be too tricky, otherwise it could become a bit of a circular task.

It is not the TPS transfer map you would need to adjust but the RPM vs change in TPS voltage map.
I can't see the correct map on your XDF or the 402,4 or 5 XDFs on here.

Take a look at the 506 ecu XDF and BIN, you will see there is an acceleration table that dictates the extra fuel the ECU needs to throw into the port to cope with the sudden gulp of air on a snap throttle opening.

I would expect there to be a similar map in the 402,3, or 5 software somewhere.

Posted: Thu Jul 20, 2017 5:33 am
by Evil
Ok its a good idea, i had found the tps definition by looking at a 413 xdf and bin To recognize the same table in the 403.Thanks for thé idea!

I replaced the TPS from gsxr by the m50 one, the range is better but I have again less range than with the oem throttle. I think I will definitely have To change thé TPS definition but I dont know box To change axis.

Posted: Thu Jul 20, 2017 9:24 am
by Evil
So I've take a look at the 506 xdf/bin. I've try to found the same table in my 403 bin but found nothing with the 3d viewer... I've try To do install the winols test version To take a look but my PC dont install it, nothing happens when I click on the .exe file...
I will try again tommorow to find it with Tunerpro on the 3D viewer, or must be a 4x6 table.
With the new tps its a little better, i've still problems in low rpm/low throttle opening but at faster rpm and fast opening AFR stay under 15, thats not optimal but a little better.

Posted: Thu Jul 20, 2017 1:44 pm
by aboutton
0x6F84 - 4 X 6 Acceleration Enrichment


Posted: Thu Jul 20, 2017 5:21 pm
by Evil
Thanks Aboutton, I'll try it as soon as im back from work 😉

Posted: Fri Jul 21, 2017 3:04 am
by Evil
Huge thanks To Aboutton for the enrichment table, now I will be able To do some work on this ITB setup.
The enrichment table is already at top value at low opening, but I have To solve the TPS problem and Perfect my fueling before doing anything else!

Posted: Sat Jul 22, 2017 2:15 pm
by Evil
Little update:
My TPS had a little clearance with the throttles shaft, solved with a better adaptatif with no clearance. Now voltage changes as soon as throttles move, and response on acceleration is better.
I'll receive my Ostrich 2.0 soon, more update when it will be there.
I'm thinking about wiring the TPS on the maf plug, because the alpha n table is only 6x7... with that done, and with the maf table modified I think I could tune the partial throttle tables for idle/PT and wot, the idle and wot tables unused because of the reduced TPS range.
Am I right with with it? If someone has ideas about it he's welcomed 😉

Posted: Mon Jul 24, 2017 2:57 am
by Hairyscreech
From what I know of these ECUs (sure someone can correct me if I am wrong) you will need to give the ECU a normal TPS input as well as giving the MAF input the TPS voltage.

Both are being used by the ECU so giving it the TPS voltage on the MAF input and running no TPS is also going to put it into limp mode.

I suspect you could get some joy by giving the MAF input a signal from the TPS in parallel with the normal TPS signal.

The MAF table would need to be setup specific to that engine as the signal would not be a measure of airflow and just throttle angle.
Two ways of solving this I guess, 1 would be to get the throttles on a flow bench and find the flow at various angels and derive a transfer function from the measured flow.

Other option is calculate the theoretical flow mathematically which would not be strictly correct to real life and then tune around it using the fuel map.

Is this a road car? If so then an air box with a MAF on the end like the M3 uses would be a lot better long term.

Posted: Mon Jul 24, 2017 6:48 am
by Evil
You're absolutely right, thats what I meaned by wiring the TPS signal on thé MAF plug: TPS signal will be use for both TPS and MAF in parrallel.
For the airbox + MAF, maybe I could do it in the future, but I have to build the airbox, it will ask a bit of work so I want first To make this setup work fine 😉
I'm also reflecting about speed/density (MAP sensor instead of MAF), with a "vacuum manifold". This solution is apparently a little complicated with ITB's, I will see what's better.

Posted: Tue Jul 25, 2017 8:55 pm
by aboutton
1- Did you make sure that the TPS voltage is ascending not descending when you step the throttle ? cause on M50 with M3 throttle bodies it will be flipped so you need to flip the positive and negative wires together to fix that issue.

2- Its true that the AlfaN table is 6 x 7 but it is not a fuel table, the ECU uses this table to calculate load so you can use the table to enrich fuel on ITB opening and you can still use the part throttle table to tune your fuel.

3- If you are gonna use the ITB's with no air-box you will loose low end torque.

4- If you want to use a MAP sensor instead of the MAF you need to alter only the MAF table and make it linear cause that is the only difference between the 2 signals, but i don't recommend MAP sensor on ITB's its not gonna work good because you will loose 90% vacuum on 10% throttle.

contact me for questions. good luck

Posted: Tue Jul 25, 2017 10:20 pm
by Evil
Hi Aboutton 😉
1. Yes I have switched wires 1 and 3 on the TPS plug .

2. Ok I see what you mean.

3. I know that no airbox will affect a low end torque, but thats an option that I dont neglect, maybe I will come To it with a bigger MAF when the actual setup will work fine.'re right, it seems that MAP with itb will only work fine on idle and very low loads. With all the renseignments I had from all of you I think that things will gonna be fine. Thanks for your help!

Posted: Tue Aug 01, 2017 12:20 pm
by Evil
Last Two days I made some tests with the ostrich.
WOT is now good, PT are better (not perfect but I have To work on it again), but I have some problems with idle which is verry erratum (sometimes rich, sometimes lean), but I think that's because injectors are biggers (270cc).
I had To lean a lot at wot (was under 10), but in PT and idle i had To enrich (I was at 17 in very low load).
I dont understand how to tune acceleration enrichment, it seems that nothing i do on this table has effect, i always have big hesitations when pressing the throttle... the y descriptor of this table is D3 but no idea of what it is on a 403 dme.
If someone has ideas about all that he's welcomed 😉

Posted: Wed Aug 02, 2017 3:23 am
by Hairyscreech
I know the Axis descriptors are a little different from the 413 ECU but the enrichment table is TPS volts vs RPM, 6 wide 4 high.

It has a noticeable effect on the 413 ECU, reducing the enrichment does drop the initial AFR when you change throttle position.

As you say you have changed injectors do you have the injector latency set right.

I just posted the M50 and M20 values on the other thread (

I have noticed the motronic is very sensitive to changes of injector without the lag being set right. Until I adjusted mine it was all over the place.

Did you need to go up to the 270cc injectors? Quite a bit bigger than the stock ones if all you did was add throttle bodies.

Posted: Wed Aug 02, 2017 6:24 am
by Evil
Yes 270cc are a bit too much, but I couldnt use the m50 one because I run e85 now .
I tried To adjust the latency but I dont know what need these injectors, it seems To be better with higher values. I will try more on it .
For the accel enrichment, descriptors are rpm and D3, but TPS is CB, maybe is there a calculation? Maybe D3 is throttle change rate but I dont know 🤔
Thanks for the precision because I was thinking that high value was higher enrichment, so I think I was doing it wrong 😅 I will look at that too, thank you very much!

Posted: Wed Aug 02, 2017 6:56 am
by Hairyscreech
It seems to behave as an additional enrichment on top of the values in the VE table.

Makes sense as the VE table is already working off load, which increases with acceleration...

The acceleration enrichment is intended to compensate for the wetting of the walls of the inlet port and the delay in the time it takes between the air beginning to flow into the cylinder and the flow increase being detected by the ECU.
In theory if it was set up ideally then you should see the AFR hit the required value as soon as you open the throttle and stay steady after that, too little and it will go lean for an instant, too much and a glut of fuel will be added and an instant of over rich will happen.

It is a fine adjustment on all ECUs but one that life would suck without.
Think the feeling you get from a carb with a broken accelerator pump function. The accelerator pump is doing the same job, dumping a bit of fuel in instantly to keep up with the air flow change until the depression at the carb increases enough to up the fuel flow.

You might be able to find the injector lag values on the web, whats the part number for them?

Posted: Wed Aug 02, 2017 8:59 am
by Evil
Thanks for thé explanation, I knew the principe about accel enrichment but the matter was I thought I had To increase values to enrich more. I have tried by lowering and its better but there's still a "lean hole". The enrichment seems To be "retarded": when I Press throttle it goes lean, then too rich and then runs normal afr. Also when I releases throttle, AFR goes very rich for a second and then injection cuts as it should. Its like there was a delay between throttle action and enrichment or cutoff.
Anyway, its quite more driveable than before and I begin To ses the benefits of itb's, little decrease of low end torque but a full engine until 7200rpm. Maybe the runners are a little too shorts, will try To increase lenght and see what it does

Posted: Wed Aug 02, 2017 9:01 am
by Evil
For thé 270cc injectors: there is no ref on them, they are from the gsxr, I'll search more To see if I find their latency.

Posted: Thu Aug 03, 2017 1:45 am
by Hairyscreech
The table should be higher number for more enrichment, certainly that it the response I get from the M3.3.1 ECU.

When I say reducing the numbers lowers the AFR I should have been more specific, Reducing the numbers should reduce the fuel, which means the numerical value of the AFR would increase.
Should probably stick to talking in lambda :x

I won't profess to be an expert in these ECUs and I am sure there are some on here that could give a lot more details.

Posted: Thu Aug 03, 2017 3:39 am
by Evil
No problem hairyscreech 😉
Now its clear: higher the value = higher the enrichment.
This confirms there is a problem 😅 with all values at 255 enrichment is not better... I'll check if I dont have fuel pressure problems, I dont think because my fuel pump is a 255l/h and its far enough for 270cc injectors.

Posted: Thu Aug 03, 2017 4:44 am
by Hairyscreech
I would start by either finding those injector dead times or swapping in a known set, these ECUs really need to know how fast the injectors open to work right.

Without getting that table set up right any other adjustment is going to be guess work or a work around.

Posted: Tue Aug 08, 2017 9:44 am
by Evil
After many research I didnt found dead time value of my injectors.
I have tried To adjust Injector constant and dead time but with no satisfying results... ls the injector constant just a factor, or is there an "interpretation" To read the value? In the original bin the value is 54953 . Original injectors are 180cc. So for 270cc injectors I put 36635? (54953*(180/270))? For e85 do I need To change this value to decrease AFR or just tune the fuel tables?
Unfortunately my acceleration enrichment problems are still here, It seems that values on the tables are not big enough ... I really dont know if the problem comes from my injectors caracteristics not tuned properly, or if its a problem of tuning the alpha n table...

I found some other maps with same descriptors than the accel enrichment table, maybe are they related, I'll do some screenshots and post it here

Posted: Tue Aug 08, 2017 10:26 am
by Evil
6x6 table, adress 7053
Descriptors CD (rpm) and D3 (dont know what it is, maybe related with CB/throttle signal), same that acceleration enrichment table

6x6 table, adress 7618, same descriptors

4x4 table, adress 7251, same descriptors.
All these maps are on a dme 403.

I dont know if they are related To acceleration or decceleration enrichment, the first table is not traced with the ostrich but the 2 others are traced, if someone knows anything about it he's welcomed.

Posted: Fri Aug 11, 2017 5:55 am
by Evil
After a bit of reflexion I think I have 2 problems that cause my lack of acceleration enrichment and also deccel enleanment:
-my "manifold " between head and throttle bodies is stainless steel, and I think it involves condensation because he's always cold when engine is hot.
-injectors are further from the intake valve than with the stock manifold (5cm further) because they are on the throttle bodies.

So I think that both causes wall wetting more important (more surface To wet + condensation), more fuel is used for that instead of going into cylinders, and the accel enrich table dont allow To reach a sufficient amount of fuel.
Same problem for deccel enleanment: the wall film is more important and after fuel cut-off in decceleration the wall film goes into cylinders and causes a drop of AFR, so the decceleration is not instantaneous, mostly at low rpm.

Maybe is there a factor somewhere that could permit To increase the possibilities of acceleration enrichment? I think there is also a deccel table so I will search about that, maybe is it one of the maps I have found. Anyway I think Iwill have To weld injector ports on the manifold closer of the valves, but It would be nice To be able To first solve the problem with tuning.

What do you guys think about it?

Posted: Sun Aug 27, 2017 2:19 pm
by Hairyscreech
Am I right in thinking you thread about making the tables bigger may have something to do with this thread?

Are you thinking of making the limp mode table bigger?

It is possible you are correct about wetting out the walls of your manifold.
Is your manifold a smooth polished surface inside? If so the trick of roughening the surface with ~300 grit may help, the V8 lot with wet manifolds have issues of fuel condensation on smooth polished manifolds.

It might explain your lean-super rich-normal. It is because most of the fuel is collecting on the walls making it lean, then trickling into the cylinder causing the over rich and then back to normal once this is resolved by greater air flow through the manifold. Exactly the same as the US V8 lot get with carbs and wet manifolds.
Do you have pictures of the setup you are running?

R.e. Injectors, we absolutely have to get those set up right. You are correct about the way to adjust the injector constant, however you also need to adjust the load values of all the tables as well to match that scaling.
Your injectors are 1.5x bigger but just dividing the injector constant by 1.5 will mean the ECUs base load calculation will also be 1.5x too low. You will never get past the bottom 2/3rds of the fuel table.
You have to move the tables load values down to match the new calculated loads which will be smaller to suit the smaller pulse widths needed from the new injectors.
I sound like a stuck record on here the last few visits but to do this right you have to understand the way motronic table headers work.
i just wrote a spread sheet to calculate the new table headers for any given injector scaling. I can do a bit more programming on it and post it up if your not sure about how to scale the table headers.

Posted: Mon Aug 28, 2017 12:53 am
by Evil
You're right about extending the limp mode table, thats exactly what I want To do. I've done another thread about it to make it easy To ses for other people.

I have some photos of the setup i will post them today.

I polished the manifold up To 400, maybe its too much as you say.
But the bignproblem I think, ls that injectors spray directly on the wall (thats it with the original position on these throttle bodies), I will modify that by welding injector ports with better angle directly on the manifold, To make them spray on the valves.

For thé injectors constant: for the moment I run without change on the constant, I just tuned the injectors latency, and fueling is not so bad,, just had To enlean a lot at wot and enrich at idle . One things I have maybe not tell is that I run e85, and by decreasing a little fuel pressure I think I'm not bad for the constant.

The 2 problems now are acceleration enrichment and little variations of AFR at very low loads (first 3 columns at low rpm).
Thanks again for your help, I really ppreciate it.

Posted: Mon Aug 28, 2017 8:12 am
by Hairyscreech
Evil wrote:You're right about extending the limp mode table, thats exactly what I want To do. I've done another thread about it to make it easy To ses for other people.

I have some photos of the setup i will post them today.

I polished the manifold up To 400, maybe its too much as you say.
But the bignproblem I think, ls that injectors spray directly on the wall (thats it with the original position on these throttle bodies), I will modify that by welding injector ports with better angle directly on the manifold, To make them spray on the valves.

For thé injectors constant: for the moment I run without change on the constant, I just tuned the injectors latency, and fueling is not so bad,, just had To enlean a lot at wot and enrich at idle . One things I have maybe not tell is that I run e85, and by decreasing a little fuel pressure I think I'm not bad for the constant.

The 2 problems now are acceleration enrichment and little variations of AFR at very low loads (first 3 columns at low rpm).
Thanks again for your help, I really ppreciate it.
Yep, ethanol makes a difference as the energy density/different AFR for stoich will come into it.
Having done the math I think the 270cc injectors are not far off the stock injectors for size when you account for the extra liquid you need to inject.

Posted: Mon Aug 28, 2017 9:33 am
by Evil
Yeah thats what I found too after calculating.
In some cells i had To enlean for about 7-8% but I think my alpha n table need again some fine adjustments

Posted: Wed Aug 30, 2017 11:40 am
by Evil
Can someone confirm that D0 is IAT descriptor for a 403? I need to adjust a little the IAT compensation. I found several tables with this descriptor but I'd like To be sure these are the good ones.

Posted: Thu Nov 30, 2017 10:02 am
by Evil
Some updates here, for not "overloading" the 413/506 thread.

I have set to 0 both Idle valve position tables. Like I thought my VE table needed some adjustement, mostly at low throttle openings.

Still have some problems at overrun/idle transition at stop, AFR take some time to come at 13 where it should be, and stalls easily... At constant idle AFR is OK. It seems that with the overrun table set to 0 it's harder to reach quicly the correct AFR when switching to idle. Maybe this one can be like original to adjust fueling in overrun.

Throttle tip-in is a bit harder at very low rpm, don't know if there is some relation with the tables zeroed, but since my VE table is not perfect I can't tell.

Posted: Fri Dec 15, 2017 2:26 pm
by Evil
My VE table is now close to be finished (need some adjustements at high loads and wot to perfect fueling.

Enrichments are good, still have some trouble at throttle tip-in after overrun fuel cut, and the "overrun IAC position table" seems to have no effect to that. If I enrich more for throttle tip-in, then in low loads variations enrichment becomes very massive (all the way down to 9.6 before reaching target AFR.
I really need to add some fuel in overrun.

I see 2 solutions for that:
-find how to disable fuel cut in the code and adjust fueling with the smallest TPS value (would be the simplest but dont know where it is, probably something like the rev limiter)
-"cheating" with TPS to "disable" fuel cut and Idle switching, but that means re-tune the VE table :roll:


maybe that could be also solved with the "delta load transfer function" (what we thought was TPS transfer function), but like Olafu said, we don't know if it's used for something else or not, neither how it works exactly....

Posted: Sat Dec 16, 2017 6:17 am
by Mykk
I'm throwing this out there as an idea to investigate, I don't know for certain.

But...perhaps the answer to your engine overrun/decel/fuel cut is in those TPS threshold tables. We already know about the Idle to Part Throttle threshold and the Part Throttle to WOT threshold, but their appear to be several other threshold tables in that area..Maybe, just maybe one is TPS vs RPM Overrun to Fuel cut threshold?

Posted: Sat Dec 16, 2017 9:06 am
by Evil
I did try To set Idle/PT thresold To 0 with no success unfortunately... Switching occurs anyway,maybe is there an offset...?
Your idea wasnt bad, this could have been an easy solution 😉

I'm quite sure it is possible To change overrun RPM, i need To find the exact value and it sould be not so hard to find it in the code.
It seems To be near 1500rpm but my tacho isnt really accurate

Posted: Sat Dec 16, 2017 4:40 pm
by Mykk
I do agree I can watch mine switch out of fuel cut around 1500rpm and there very well might be a single byte or two that controls that RPM point.

If so I would imagine it's in the front of the code where the Rev limit, Speed limit, Injector Constant...etc. are located.

I have a feeling Olafu would be able to pinpoint exactly where and how.

Posted: Sat Dec 16, 2017 6:35 pm
by Evil
I think the same a bout it. I did found 2 "double bytes " which were like the rev limiter and with à value around 1500rpm, but change those bites didnt have any effect on overrun RPM.

But It can also be a single value in x*40.... I'm still searching but with thé exact value of overrun RPM i would know exactly what To search

Posted: Sat Dec 16, 2017 10:08 pm
by olafu
413/623 :
0xDC79 - 4x1 D7 - Some kind table for overrun fuel cut? Needs more clearing, but it seems like those values are RPM thresholds - Actual RPM below those values = fuel cut is disabled.?

If this is fuel cut table, there can be similar function in 403 too.

I wrote that in my notes at last summer: "255 in every cells = fuel cut disabled. 0 - fuel cutted and won't come back, engine stalls"

According motronic table structure, if this is overrun cut table, there can be more overrun fuel cut tables in near that 0xDC79 in "623" and in this case i think 403 ecu has they too.

I guess there are "load before fuel cut" and "delay for switch fuel off, when load threshold is reached" and those can be 2d or 3d tables...

Posted: Sat Dec 16, 2017 10:57 pm
by Evil
I have the same table in 403, scaling and data are not exactly the same but very similar, With X*40 it seems to be RPM values like you said!
I will try it today and tell you if it works :D

Posted: Sun Dec 17, 2017 5:03 am
by Mykk
I've isolated the table in my 404 as well, it's directly followed by a mysterious D2 axis identifier. This could shed some light on that axis identifier.

I won't be able to road test this table on my vehicle for a little while as mine isn't running at the moment.

Posted: Sun Dec 17, 2017 6:05 am
by Evil
It's also followed by a 4x1 unknown descriptor in 403.

Maybe this could be "engine deccel speed".

And to talk about unknown descriptors:
I'm quite sure there is also somewhere a dedcriptor for" engine acceleration speed", one for "Time" and one for "number of engines rotations".
For the 2 lasts, I don't know if these counts only after cranking event or if it's just a "counter" that can be triggered when ECU needs it...

In another subject:
There is a 4xIAT table that I suspect to be some kind of hot start correction, in 71F6 for 403 (DBA8 in 413), scaling for IAT is approximatively 20/40/60/80, with datas decreasing from 48 to 36 in dec. Maybe to give a correction for heat soak after a more or less long stop.

Posted: Sun Dec 17, 2017 10:13 am
by Evil
Table for fuel cut RPM works perfectly!
I can adjust fueling as I need with VE table and then with idle fuel table.
Throttle tip-in is now close to Perfect, and it has also solved my stalling problem at stops.
Many thanks olafu for these infos which have been very useful, and also To the others for these great discussions we have a bout our motronics and all the work done!

Posted: Sun Dec 17, 2017 1:31 pm
by Mykk
I just came here to say the same, tested it on mine this eve. It still seems to go lean but not completely into fuel cut. The exhaust has a cool sounding burble on decel now and throttle response seems a bit better. I set mine to all 254.

I think the other tables in play are still acting as a cut as the longer it's in overrun the leaner it gets.

Posted: Sun Dec 17, 2017 2:35 pm
by Evil
If you want richer overrun try To enrich with Idle fuel table 😉
It should work fine with or without maf (works for me in alpha-n)