The m20 Motronic m1.3 to m3.3.1 thread - Fully working now!

Wed Nov 11, 2015 4:30 pm

The increased timing on the m20 and reduced fueling makes sence with it being a older 2valve per cylinder head engine, so having less efficiant combustion and volumetric efficiency.

Thu Nov 12, 2015 12:22 am

Most definitely, it will be a result of poorer airflow and lower compression/offset spark plug. With improvements in all 3 areas I would expect them to get closer together.

Another interesting thing is the M1.3 173 ECU has a "low part throttle" map and a "high part throttle" map giving a total of 14x16 for the part throttle map.
The M3.x ECU only has a single part throttle map of 12x16. Meaning the older 173 actually has more resolution to its map that the newer ECU.
I am sure the M3.x ECU more than makes up for this with greater processing power and better interpolation of the maps though.

Interestingly the M1.3 seems quite a fully featured ECU for its age, seems like it was a bit of a flagship job for the time.

Thu Nov 12, 2015 12:42 am

also having more accurate load inputs, variable tps, and a maf would give the ecu more precise information on load so that may also account for not needing such large tables. i think once you have managed to get the new ecu mapped accordingly you should notice a good difference, my car felt like it had a bit more pull down low once i got mine running fully sequential fuel and ignition on ms.

Thu Nov 12, 2015 10:47 am

The transient throttle should be a lot better, yeah. Hoping for some MPG as well but I think that won't come until the 2.8 with better breathing, higher compression and improved BSFC.

What generation MS did you go for?

I had an older system (MS1 V3) on a zetec turbo and found it a bit lacking, Doing it with the available tech today I would be using a modified Honda or Nissan ECU converted to make it function like stand alone. The OEM ECU hacking is producing some really good bits of tech now vs what used to be available.
The Nismotronic is pretty fully functioned and I think it would be tempting if I were to build another turbo lump. http://www.moates.net/nemunismotronic-t ... l?cPath=68

Are the newer MS units better?

My Ostrich

should have been despatched now, I put a map together by copying the load and RPM points that matched and manually interpolating the areas between with a bit of an Err towards rich and retarded, it should be a lot better match than the M50 map but not ideal, Thinking its better to be rich and retarded and pull some fuel and add timing on the dyno later.
12.5:1 and 2 degrees retarded has rarely killed an engine, same values the other way.....

The biggest surprise has been the throttle response, I have only driven it a few feet but the change is huge. A little rev for 2k to shuffle it around on the drive has it jumping to 3.5k. Feels more like a race engine than a road engine, with a light flywheel it would be silly.
Keen to see what it is like with tweeked fuelling and timing.

Thu Nov 12, 2015 11:32 am

i have a ms3 with the ms3x card so pretty much all i could want and then some. ms2 is miles past the ms1 days and ms3 is way past ms2 so it has come along a good way. i take it you have a wide band so you can keep an eye on afr's?

Thu Nov 12, 2015 1:19 pm

Interesting to hear, sounds like it has stepped its game up a lot then.

Wideband? at the moment no, likely to be the next purchase after this months £450 of tyres and emulator stop causing ring sting.

Thu Nov 12, 2015 8:36 pm

gromgsxr wrote:also having more accurate load inputs, variable tps, and a maf would give the ecu more precise information on load so that may also account for not needing such large tables. i think once you have managed to get the new ecu mapped accordingly you should notice a good difference, my car felt like it had a bit more pull down low once i got mine running fully sequential fuel and ignition on ms.

when i changed to wasted spark and sequential fuel there was no noticeable improvement. Maybe it idles a little smoother but that's about it.

Thu Nov 12, 2015 8:42 pm

maybe its all in the mind but when i changed from un-timed batch injection to full sequential i'm sure the car drove smoother and felt a bit stronger down low in the revs.

Thu Nov 12, 2015 8:53 pm

improved response is possible ive often perceived this as more torque but the rollers proved me wrong often.

mine now puts down less power than what it did but a fair bit of water has flown under the bridge that i cant for certain attribute it to that.

Thu Nov 12, 2015 9:08 pm

amazing thread mate.

Fri Nov 13, 2015 10:02 pm

Right, finally got around to booting the other computer up and grabbing the maps I have been working on. (yeah I'm lazy, sue me).

First off, a comparison shot of the 413 and 173 definitions list, each one of those headers is something that can be tuned in the ECU with tuner pro at the moment.

Something to notice is the the 173 ECU has multiples of some maps, they seem pretty much the same and the ECU is just pointed to the different maps depending on the vehicle setup.

For the fuel and ignition maps I have tried to reduce the table size keeping the trends and rough values, I will do some more explaining on an future edit, for now, pictures below, it's Friday and the ciders open.

Wed Nov 18, 2015 1:33 pm

Had a bit of a break through with the way the ECU uses its maps and calculates some of its functions.
It seems my assumptions were incorrect with the fuel maps, they are actually target lambda values, using the 506 DME .bin from the US S50 M3 gives a greater understanding of the maps, the two ECUs are actually the same and maps/chips can be swapped between red and silver label as well as 506 and 413 ECUs.

My Emulator should be here this week so will be able to do some more with this.