Does anyone have the scaling for grams/rev, instead of Load units?

The value Load/95, some people use doesn't seem accurate.

 

How does this work? I found it in 21810021.

<scaling name="Airflow16" units="g/s" toexpr="x*600/65536" frexpr="x*65536/600" format="%.1f" min="0" max="600" inc="1" storagetype="uint16" endian="big" />

 

That is grams/second. Used more with MAF stuff.

The 2G DSM Boost Cut is supposedly 2.55 g/rev, and in the ROM is 160(0xA0).
This gives a conversion of Load*(51/4000) = g/rev.

Which is ~ Load/78, instead of the Load/95 I found posted.

I have been trying to find some screenshots of the old version of DSMLink when it had factory map axis, to see if my axis matches. but have not been able to find any pics.

Edit:
The Load used for Axis scaling is (x*10/32).
After searching a little I've found this scaling in g/rev is (x/128) in old ecus.
Just need to find a way to confirm this.

 

Unrelated sidepoint, but the load limit in the STi is 4.0 g/rev which if you were to multiply by 95 is 380 which is coincidentally what the Evo is.

Ceddy on the 2G (at least 2GB) where we know the load limit I would think we'd be able to reverse and see. I am curious where the 2.55 value came from? Was that from Thomas or Jeff?

 

The 2.55 g/rev is from the dsm-ecu yahoo group, but might be from Tom or Jeff second hand. I'm trying to find the Timing Load axis in the 95 eprom DSM to see if it matches the 2GB, but the old ecus don't have map and axis headers so it makes things harder to find.

 

Ceddy, I'm interested in this also, as I am currently working on evo3 stuff and which I believe uses the same scaling.
Andy Fox may have some good answers on this from his dissassembly work.

 

I have to wonder if it is really even meant to be g/rev.

The load used is different between fuel and timing. If you did them both by the same scaling, they would actually reflect two different things. The timing would be based on something proportional to g/rev. However the fuel would be proportional to volume/rev since it does not account for baro correction (temperature too?).

 

I think if you follow PV=nRT in the code, you come up with Liters of Air/Rev, but this is converted to Grams of Air/Rev because its easier to understand.

Most of the previous mitsu ems used g/rev instead of the Load we use. You just need to find the proper scaling to convert between them. If your MAF is properly adjusted g/rev is accurate.

 

Here are the g/rev scalings:

g/rev 16-bit (x/256 & x*256)
Used in table axis.

g/rev 8-bit (x/128 & x*128)

g/rev boost (x/64 & x*64)
Used for BoostLimit units.

These haven't been confirmed through logging yet.
I would be interested in how these compare to MrFred's Airflow Calcs.

I believe these are the units Mitsu used in their code, because of the power of two scalings.

Does anyone know where the Load units we use now came from? And if they have any real world equivalent?

Comparison of Load and g/rev from my 2G Ignition axis:


To convert Load to grams/rev, you would use the formula g/rev = Load/80.
I'm not sure where the Load/95 came from, but the ~18% difference might be explained by drivetrain loss if it was figured from HP.


To find lb/min: lb/min = g/rev * rpm * 0.002204623


Here are some Ignition Tables with lb/min figured out for every cell:
(This tells you the AirFlow at each cell.)

2GB DSM Stock Ignition Map


Evo8 Stock Ignition Map


88590715 Tephra Big Map
106 lb/min Max!

 

What is the last load value in the 2G map? Your screenshot shows a jump from 180 to 260, but I'm guessing that is your own value?

From what I remember from DSMLink and Tom/Dave, the last load column represented approx 2.1 g/rev. That wouldn't coincide with load/80.

Edit: BTW, I think it was me who came up with the rough estimate of load/95 based on mass airflow logging from OBD and then volumetric airflow logging from MUT. I was trying (back then) to come up with an equation for Hz to lb/min with baro and temp adjustment.

 

The 2.1 g/rev max is from the 1G map. The 1G 2.1 g/rev max is very low, you are off the map at like 12 psi or something. I think they did something so the map sizes are the same in the 1G and 2G DSMLink versions.

I found the scalings from the 1G maps axis and MMCD DataLogger and knowing the fuel cut g/rev.


That is the stock map. I think they entered 260 instead of 200 as the last value so you won't go off the map.

The 3000GT VR4 has an even bigger jump, it goes from 160 to 400. They only do this on the Ignition Maps.

 

I see a similar jump in the 3g eclipse maps, as it goes from 100 to 400.