Memory dump of ECU

[mbtoloczko]

Originally Posted by b3lha

OK. So simplistically, if the MAF is measuring the volume of air per minute and you are dividing that by the Revolutions Per Minute, then Load can therefore be thought of as the volume of air per revolution? Or even per combustion if the constant multipler takes into account the number of cylinders.

Correct. In its most simple interpretation, its the amount of air per rev going into the motor. It doesn't account for how efficiently the air is used though. If there is aggressive cam overlap, then at certain rpm ranges, air can get sucked in the intake valve and back out the slightly open exhaust valve. Evo's have electronically adjustable cam phasing, and I can cause an increase in measured load without increasing HP/TQ simply by increasing the overlap.

[TomsSVX]

ok now this is the discussion I have been waiting for. My issues I feel is that my car pulls timing once I go off boost to full boost (the supercharger is good for that). My feelings are that when this happens the ECU either is not programmed to accept this load increase or it is too slow to react(which is why a slower boost incline from a softer bypass valve may help). I also feel the ECU is programmed to run far too lean and needs more fuel in just about all areas. Let me know if this sounds right to you guys.

Tom

[mbtoloczko]

Originally Posted by TomsSVX

ok now this is the discussion I have been waiting for. My issues I feel is that my car pulls timing once I go off boost to full boost (the supercharger is good for that). My feelings are that when this happens the ECU either is not programmed to accept this load increase or it is too slow to react(which is why a slower boost incline from a softer bypass valve may help). I also feel the ECU is programmed to run far too lean and needs more fuel in just about all areas. Let me know if this sounds right to you guys.

Tom

Tom, how to deal with load is very important in your setup. Running FI on the stock ECU can definitely cause problems because the load will far exceed the load range in the fuel/timing/knock tables. However, I wouldn't think that LAN would overlook rescaling the load formulation. I believe that since LAN is using the Z32 MAF, he has probably rescaled the load formulation and also rescaled all the tables so that the load never exceeds the load range in the tables.

[Dessertrunner]

the select monitor instruction in the SVX manual page 97 says the Load Data uses Air Flow Sensor, Engine Speed & Crank angle sensor, I get the rest but don't understand the last. Any suggestions. Also of interest is "Ignition Timing" says to do the tests with the car in Netural so as it was suggested by somebody may be having the netural floating does screw timing when clutching in.
Phil if you drawn me a diagram I will see what I can do next time I get it.
Tony

[Calum]

Originally Posted by mbtoloczko

Tom, how to deal with load is very important in your setup. Running FI on the stock ECU can definitely cause problems because the load will far exceed the load range in the fuel/timing/knock tables.

Depending on how the ecu works, thats not really a bad thing and you can take advantage of it. Some ecus work with values in the fuel table directly being pulsewidths, and in that case if you run off the edge of the map you really are screwed. Nissan (and from the sounds of it Subaru) ecus either use a target AFR, or a multiplier for a calculated base injection time, so in that case running off the edge of the map isn't bad, your just operating on the last column. Because the value in the map is a multiplier or a target value its not like the car is going to lean out like it would if it was a fixed injection time. This makes for an easy way to tune the car. I purposely scale the table down so that when the car hits some low boost value (like 4psi) its on the last column. I then set boost to whatever the car is going to operate at, and now when I tune I only have to work with the last column. you can do the same for an NA car, just make sure under WOT your on the last column.

[SVXRide]

Originally Posted by b3lha

I agree with you entirely, except on one small point:

The copyright on the ROM belongs to Subaru not LAN. You can't take a book, change a couple of pages and claim copyright over it. I think technically both LAN and I are breaking Subaru's copyright by distributing their ROM, but I don't think they would care too much.

Just to nitpick a little: I know some Americans seem to believe that everybody in the world should follow American laws, but the internet is a global medium, and laws differ across the world. You should not assume that everyone reading this is bound by the same laws as you. However, such discussion belongs in the political forum. I don't want to hijack this thread.

Anyway, I can't speak for the other contributors to this thread, but the work I am doing is not intended to compete with LAN. I do not have the tuning expertise to produce a B3lha-stage-1. Anyone who want's off the shelf performance chips should go to LAN. He deserves to get back the money he's invested. I'm just hoping to shed enough light on how the ECU works to make it easier for people who want to have a go at tuning it themselves.

Phil,
agreed...I probably went a little overboard with my statement, only with the intent of recognizing LAN's efforts to date.
-Bill

[SVXRide]

Originally Posted by mbtoloczko

Tom, how to deal with load is very important in your setup. Running FI on the stock ECU can definitely cause problems because the load will far exceed the load range in the fuel/timing/knock tables. However, I wouldn't think that LAN would overlook rescaling the load formulation. I believe that since LAN is using the Z32 MAF, he has probably rescaled the load formulation and also rescaled all the tables so that the load never exceeds the load range in the tables.

Mychailo,
I believe that when LAN came out with the 2v line of chips using the z32 MAF he did post how he had rescaled the tables.
-Bill

[Dessertrunner]

Does anyone know what the commuication protcol is used for the RS232 connector, for example would it be ASCII In/Out the reason I ask is that to do what we need it may be quicker to plug in a PLC as programing them can be fast.
Tony

[mbtoloczko]

Originally Posted by Calum

Depending on how the ecu works, thats not really a bad thing and you can take advantage of it. Some ecus work with values in the fuel table directly being pulsewidths, and in that case if you run off the edge of the map you really are screwed. Nissan (and from the sounds of it Subaru) ecus either use a target AFR, or a multiplier for a calculated base injection time, so in that case running off the edge of the map isn't bad, your just operating on the last column. Because the value in the map is a multiplier or a target value its not like the car is going to lean out like it would if it was a fixed injection time. This makes for an easy way to tune the car. I purposely scale the table down so that when the car hits some low boost value (like 4psi) its on the last column. I then set boost to whatever the car is going to operate at, and now when I tune I only have to work with the last column. you can do the same for an NA car, just make sure under WOT your on the last column.

Evo's will also use the value in the last column. Seems like this could be ok for a setup that only mildly exceeds the last column, but for a major increase in throughput, I would think that blowing past the last column would prevent optimizing the timing in the out-of-bound load values and leave HP on the table.

[mbtoloczko]

Originally Posted by SVXRide

Mychailo,
I believe that when LAN came out with the 2v line of chips using the z32 MAF he did post how he had rescaled the tables.
-Bill

That's my recollection as well.

[mbtoloczko]

Originally Posted by SVXRide

Phil,
agreed...I probably went a little overboard with my statement, only with the intent of recognizing LAN's efforts to date.
-Bill

The issue of showing vendor maps has been batted around quite a bit on EvoM as well. Showing vendor maps is not permitted. However, once the ECU was unlocked and free tuning software hit the streets, everyone started tuning and posting their results. It only took about a year for the community to establish the optimal range of fuel/timing/etc maps for the Evo.

[Calum]

Originally Posted by mbtoloczko

Evo's will also use the value in the last column. Seems like this could be ok for a setup that only mildly exceeds the last column, but for a major increase in throughput, I would think that blowing past the last column would prevent optimizing the timing in the out-of-bound load values and leave HP on the table.

No, it doesn't prevent anything if the car always runs at one boost setting. If you always run the car at 10 psi, and tune the last column with that boost setting, your good to go.

For cars that are using something like a Greddy Profec-B with a hi and low boost setting what I do is map trace or datalog the load value with the high value, then do the same for the low value. I then make the second to last column the average load value for the low boost setting, and the last column the high boost setting. Now you might think that limits your tuning to just two boost settings, but doing the above is actually pretty slick and here's why: the ecu interpolates between the columns (and rpms) in a linear fashion (probably, Nissans do and I bet the Subbie does too). So what you've really made isn't just a tune for a hi-low boost seting, but boost dependant timing retard and fuel enrichment based on the parameters of your two columns. Neat.

[Hocrest]

I got my USB cable, I downloaded JDash and vwrx. What do I do now?? I've tried to follow this thread and others on the subject. I've read back through, and am now even more confused.

I'm guessing there is someplace that I change settings, but I can't really find where to do that?? I also found someone mentioning Phil editing the VWRX .ini file, but I can't find the details.

If you want, I can be IM's at hocrest1

Thanks

[Nomake Wan]

Desertrunner: It's binary, not ASCII. 1953 baud rate, 8 data bits, 1 stop bit, even parity.

Hocrest: JDash I don't think is going to help you right now, but VWRX will work.

First run the SelectDump program in the VWRX software. It'll give you an ECU ID if your connection is working correctly.

Then get Phil's modified SelectMonitor.ini file from here and replace the one that comes with VWRX with it:

http://www.alcyone.org.uk/ssm/ecu/SelectMonitor.ini

...find the entry for "1994 UK SVX" and change its ECU ID value to yours. Then save the file, and run the SelectMonitor. It should be fully functional now.

[Hocrest]

I get nothing when I run VWRX, so I guess I have connection issues...