NA Engine Control Unit tuning and Diagnostics

[longassname]

I'm very close to releasing a full blown tuning application to be run on a laptop packaged with a usb cable and interface to plug into the select monitor port. The software uses a definitions file to configure it for a particular control unit. It can check and clear diagnostic codes, monitor engine operating parameters, sensor data, switches etc, download the firmware off the control unit for backup and modification, edit the downloaded firmware to make modifications to the tuning or engine management hardware, and save the modified firmware for creating a "performance chip."

It will work with any OBD1 Subaru control unit for which there is a definitions file; I have already created definitions files for your OBD1 Naturally aspirated Subaru Legacy. These definition files can be expanded to include any new and exciting things any gurus may figure out in the future. They are already quite extensive taking your naturally aspirated OBD1 legacy vehicles from non tuneable to one of the most tunable vehicles in existance. Of course you don't have to use all of this power. The software is worth having just to adjust sensors, check and clear trouble codes, etc.

For the 90 to 91 na legacy the following parameters are already defined:
Load %
Throttle V
Injector Milliseconds
Idle solenoid %
ALPHA 1
O2 Sensor V
RTRD
CPCD %
BARO. P
K constant
Injector latency correction per volt
Injector latency
Temp based idle fuel enrichment %
Q for Air meter V/Q table
RPM for fuel tables
Load for fuel tables
AFR for primary fuel revision table
Advance for fallback ignition table
AFR for starting fuel table
RPM for ignition tables
Load for ignition tables
Advance for primary ignition table
Rev limit 2 off threshold
Rev limit 2 on threshold
Rev limit 1 off threshold
Rev limit 1 on threshold
Fan Vsp K/h upper threshold
Fan Vsp K/h lower threshold
Fan Degrees C upper threshold
Fan Degrees C lower threshold
TPS voltage for airflow fallback table
RPM for airflow fallback table
Q for Airflow fallback table
Ignition Advance revision


The following tables are also defined:
Temp based idle fuel enrichment
Mass Air Meter Translation
RPM scale for fuel tables
Load scale for fuel tables
Primary Fuel
Fallback ignition timing table
Starting fuel
temp based enrichment 2
RPM scale for ignition tables
Load scale for ignition tables
Primary Ignition timing
TPS voltage scale for Mass Air failure table
RPM scale for Mass Air failure table
Mass Air failure fallback table
Ignition Revision Table

All trouble codes are defined:
41 AF Control
42 Idle Switch
43 Throttle Switch
44 Wastegate Duty Sol
45 Atmos Presure/Pressure Sol
49 CO Resistor
51 Neutral Switch
52 Parking Switch
21 Temperature Sensor
22 Knock Sensor
23 MAF Sensor
24 Idle Solenoid
31 Throttle Sensor
32 O2 Sensor
33 Speed Sensor
35 Purge Solenoid
11 Crank Sensor
12 Starter Motor
13 Cam Sensor
14 Injector 1
15 Injector 2
16 Injector 3
17 Injector 4

[longassname]

For the 92+ NA legacy the following have already been defined:

parameters:
Battery Voltage
Vehicle Speed M/H
Vehicle Speed KM/H
Engine RPM
Coolant Temp F
Coolant Temp C
Ignition Advance
Airflow Sensor V
Load %
Throttle V
Injector Milliseconds
Idle solenoid %
ALPHA 1
O2 Sensor V
O2 max V
O2 min V
RTRD
BARO. P
K constant
Temp based idle fuel enrichment %
Q for Air meter V/Q table
RPM for fuel tables
Load for fuel tables
AFR for primary fuel revision table
RPM for ignition tables
Load for ignition tables
Advance for primary ignition table
Fan Vsp K/h upper threshold
Fan Vsp K/h lower threshold
Fan Degrees C upper threshold
Fan Degrees C lower threshold
Ignition Advance revision
AFR for high octane fuel revision table
TPS voltage for airflow fallback table
RPM for airflow fallback table
Q for Airflow fallback table
Ignition Advance revision for high octane table


The following tables have been defined:
Temp based idle fuel enrichment
Mass Air Meter Translation
RPM scale for fuel tables
Load scale for fuel tables
Primary Fuel
RPM scale for ignition tables
Load scale for ignition tables
Primary Ignition timing
Ignition revision table
High octane fuel table
TPS voltage scale for Mass Air failure table
RPM scale for Mass Air failure table
Mass Air failure fallback table
High octane ignition revision table


All trouble codes have been defined:
38 Engine Torque Control
49 CO Resistor
51 Neutral Switch
21 Temperature Sensor
22 Knock Sensor
23 MAF Sensor
24 Idle Solenoid
31 Throttle Sensor
32 O2 Sensor
33 Speed Sensor
35 Purge Solenoid
11 Crank Sensor
12 Starter Motor
13 Cam Sensor
14 Injector 1
15 Injector 2
16 Injector 3
17 Injector 4

[longassname]

Interface is accomplished through a usb cable which plugs into a custom interface board I have made for select monitor coms. In this picture the output connect is looped back beacuse this is a test unit hooked up to the workstation I am writing the software on. It actually comes with a select monitor plug already wired to the 4 wire screw terminal strip so all you have to do is plug the usb cable into your lap top and plug the select monitor plug into the select monitor port of your car.

[longassname]

The software has a tabbed user interface which is hopefully intuitive and easy to use. I'm suposedly pretty good at that and I've included pretty extensive text directing the user on what things do to help you along.

This tab is the setup tab. Most users won't need to do anything on the setup tab besides choose the definitions file to match their car and select the communications device to use. Everything else will be configured by the definitions file and the only communications device that will appear as available is an official ECUtune interface cable. For more advanced users, controls are available on this screen to allow communication with a control unit for which there is no definitions file yet so they can download from it and poke and prod at it to figure out how it works and create a definitions file. The definitions file is a microsoft excel spreadsheet which can of course be edited in microsoft excel as well as in my software. Mid level users who will be editing their firmware to create their own custom tunes will also use the load bin file feature to load the firmware they wish to modify.

[longassname]

The definitions editor tab allows advanced users to expand, modify, or create their own definitions file while in the ECUtuner software. Each definitions file has multiple worksheets--AddressSpace, Coms, Parameters, Bitmaps, and Tables. Slect the worksheet to edit and you can edit it/add rows to it just like you would in excel. Again only advanced users will actually do this but making it available to them is what makes the software a tool for cracking new control units and what encourages them to figure out new definitions for more stuff for everyone else to play with.

[longassname]

The monitor parameters tab is used for monitoring sensor data/calculated engine management parameters. Just having the ability to monitor your tps voltage in order to adjust your tps correct makes the software an almost essential tool for anyone who does their own work on their legacy. The ability to monitor sensors is also essential for home mechanics to diagnose intermittent sensors or sensors with poor response. Of course monitoring and logging data is critical to anyone wanting to tune their own engine control unit. 4 parameters can be monitored, graphed and logged at a time.

[longassname]

The monitor bitmaps tab lets you check your trouble codes, clear your trouble codes and learned data, and monitor your switches.

[longassname]

The transfer firmware tab is for getting firmware off of and if you have emulators into your control unit so you can modify it. Emulation enables real time tuning so you can edit the data live while the car is running and being driven be it on the street or even better on a loading dyno. Downloading firmware takes an hour and half to two hours so you only want to do it once. Once downloaded the software can calculate what the checksum should be and compare it to what the recorded checksum is to ensure it is error free.

That's where I'm at now. I took a little break from programming to create the definitions files for you so as soon as I have the software done it will be fully functional for naturally aspirated legacy owners. Now I'll get back to work on programming to finish the emulator coms and the firmware editing features.

[longassname]

I will also offer off the shelf performance chips for both the Engine control unit and transmission control unit.

It looks to me like revtronix has put quite a bit of work into supporting the turbo legacy. It also looks to me like retronix probably has an individual much like myself behind it. As a hard working individual with my own (scary amount of) student loans I'm sensitive to how awful it would be for that individual if I came along and competed against him in the tiny market he has supported and just as bad if my software was used to steal his tunes. For that reason I have no intention of creating definitions files for or creating chips for the hitachi turbo ecu's (I do have a definitions file for the jdm sti and ra sti).

[ericem]

great to know others are out there making products. A transmission chip changing lockup and shift points would be phenominal.

[longassname]

Thanks,

I went ahead and started a thred for people to discuss what they want for tcu chips.

http://bbs.legacycentral.org/viewtopic.php?t=42271

great to know others are out there making products. A transmission chip changing lockup and shift points would be phenominal.

[Lunatech]

Seriously not trying to be a smart ass or anything.


I know jack about tuning, well new school tuning anyhow. So, what does this get you in the way of control, over say, ignition timing or fuel flow?

It sounds very interesting to me, but not sure of it's applicable use. It seems that is is either too good to be true or too wild to fathom.

[longassname]

Complete control. Anything that is being controled by the control unit can be changed by you or a tuner by modifying the data in the control unit. You just have to know what you are doing. Chances are there are very few people to nobody in the community that does yet. That will change. Given the tools that are about to be available the forum will likely add a control unit tuning section and those who know the most will start helping those who know the least. Other cars for which this kind of software is availabe have entire forums many times the size of this one dedicated to ecu tuning.

Here's how fuel works:

The mass air meter is a sensor which puts out a voltage according to the mass of air which flows through it. It is used to measure the amount of air being ingested by the engine. Meters more or less all output a voltage range from 1 to 5 volts but the amount of voltage corresponding to a given amount of air is different for different meters. One meter may max out at 100hp worth of air for which it puts out 5 volts. Another meter may max out at 500hp worth of air for which it puts out 5 volts. Another meter may also max out at 500hp worth of air for which it puts out 5 volts but its voltage output for 200hp worth of air may be radically different than the first meter. You may note one of the tables I have defined is the mass air meter translation table. This is a very acurate 16 bit table specific to the mass air meter being used. The ECU uses the curve in this table to translate the non linear voltage response of the meter into an airflow measurement.

Here's a comparison plot between the nissan z32 air flow meter and a subaru SVX airflow meter. Not only are the shapes of the translation tables very different but the airflow they represent where values are the same is very different and so there is a constant which the looked up values are multiplied by we will need to change in order to swap meters as well. I'll get to that in a bit. For now just know that the max reading on the svx meter is about 250hp worth of air and the max reading on the z32 meter is getting close to 500hp worth of air and these plots are of of the translation tables * constant ratio to show the true difference in airflow.




Ok, so say you have a naturally aspirated legacy and you want to install a turbo charger or supercharger on it that is going to force more air down it's throat than the stock mass air meter can measure. You could swap in another meter..maybe an svx meter...not big enough?..ok a z32 air flow meter. Now you have a meter that can measure close to 500hp worth of air. Then you copy and paste in the translation table for the z32 air flow meter to replace the translation table for the old meter that was too small for you. Now the value the ecu looks up from the table is multiplied by a constant which need to be changed as well. This is frequently called the k constant and that's what I called it in the definitions file. Lets say the max flow of your old meter was 150hp now the max flow of your new meter is..well let's call it 450. That's 3 times more. So we multiply the K constant by 3. (problem is that value is too big to fit in a 16 bit location but that's ok...we're going to need to divide it when we put big injectors in next).

Ok, so now we can measure 500hp worth of air for our 2 liter engine....
We need to be able to inject a lot more fuel too. What are the stock injectors? Let's say they are 250cc. Let's say we decide we want to use nismo 740cc fuel injectors (which would drop righ in by the way). So we install our 740cc injectors which are 3 times as big as our old injectors and divide our k constant by 3...what do you know using the hypothetical numbers I just made up we are right back to the k constant we started with. 3 times as much air requires 3 times as much fuel. If we used a meter that could only measure twice as much air and fuel injectors that could inject 3 times as much fuel then we'd be multiplying k by 2 and dividing it by 3 and k would still fit in it's 16 bits and everything would be fine. The only thing we can't do is go putting in a mass air meter that is huge compared to our injectors and there is no reason to.

Now the ecu has the correct data to match the new hardware and can calculate how much air is being ingested and how much fuel to inject in order to reach a stoichiometric air to fuel ratio of 14.7. When you are plodding along at light cruise that is great but when you are accelerating you want to run a lower afr. The ECU uses a fuel revision table to specify what afr to run according to rpm and load. Load is qty of air per revolution. Load is the main and just about the sole thing determining what afr and how much ignition advance you should be running at any given time. RPM is the 2nd biggest factor which is why it is the other axis of the fuel and timing tables. Now, since we've correctly modified the data to match our new injectors and our new mass air meter....we actually don't have to edit our fuel table if we want the car to run the same as it did before but what we do need to edit is the load scale for our fuel table (and our timing tables but no need to cover that right now). The load which the load scales of our important tables for fuel and ignition and ignition revision is calculated using the K constant. We didn't change the k constant but what it is representing is different. Say we have a load of 50 now. Well that 50 is 3 times as much air and fuel as it was before we changed our air flow meter and injectors. So we divide all of our table load scales by 3. Whenever you change your fuel injectors you have to change your load scales to match. The air flow meter doesn't play into it. Beginners often have a problem wrapping their head around that one but you get it after you sit down and do the math. With that done we are back to our engine management working perfectly with the new injectors and air flow meter. In fact we haven't changed the way the car will run at all we've just given it the ability to manage a lot more power than it could with the old hardware.

To change the way the engine runs we change the data inside the tables. Tweak the afr here..tweak the timing there. That's called tuning. Using emulators you can adjust your afr and timing in real time while running the car on the dyno to see if power goes up or down in order to evaluate the changes as you are making them. Not using an emulator you can make changes you think will make improvements to try later. In either case when you are done you have to write your modified firmware to a piece of memory which you then install into one of my memory adaptors which you then install into your ecu. The final product is a "performance chip" just like you would buy from me but custom made by you.

If you weren't doing any custom modifications to your car and just wanted to improve performance of the stock car it would be faster cheaper and easier to buy an off the shelf chip from me. If you wanted to change your components to match the components I chose for an off the shelf chip it would be faster, cheaper, and easier to buy an off the shelf chip from me. If you want to be able to use whatever fuel injectors you want, or whatever air flow meter you want, or want to be able to change your tune yourself then you are in the category of people who will need the ECUtuner software to do so. End user tuning takes a lot of work and time and will more or less be a hoby for anyone that delves into it. There are also plenty of people who hire other people to do tuning for them. If you have the software and the memory adaptor you can take your car into a dyno tuning shop and pay them to use it to tune your car.

[Lunatech]

That sound awesome, just what I've been wanting. What is the price tag?

[93Leg-c]

Sounds exciting!

Yes, what is the cost of a chip? Will the '92+ chip work for a '98 2.2L engine or will you have to make another chip for it?

[Lunatech]

Bump

[longassname]

I haven't set prices yet. Off the shelf chips will be priced similarly to the SVX chips. The software will be a public download but the coms won't work without purchasing the cable which is used to license the software. The cable, interface, license bundle will be modestly priced. I want it to be cheap enough for people who don't want to do their own tuning to purchase it for its select monitor abilities.

[Lunatech]

That's super good, on my modest income a modest price works out pretty well. You'll let us know when it is ready, I trust.

[longassname]

Presumably you are talking about swapping a 98 2.2 engine into an earlier model legacy. To do that you would need the software so you could customize the firmware to match the engine. It's not a huge difference but you'd most likely want to adjust your load scales to match the difference in displacement.

Sounds exciting!

Yes, what is the cost of a chip? Will the '92+ chip work for a '98 2.2L engine or will you have to make another chip for it?

[93Leg-c]

Actually, I was asking about the '98 2.2L engine that is in my daily driver. It's a '98 Legacy L wagon with a 2.2L NA engine and auto tranny.

Probably the biggest difference with my '98 Legacy and the first generation 2.2 NAs is that mine is OBD2 while the first generation had the OBD1 system. I had a '93 Legacy with a 2.2 NA engine and it had lots of power for an NA--it would snap a person's head back upon take off from stop with moderate acceleration.

The '98 is quite gutless--I'm not sure if could peel the skin off an overripe banana even if I floored it. I'm guessing that the OBD2 system plays a large part with the substantial lack of power (compared to my '93).

So, that's why I asked if a whole new chip would have to be made for the '98 2.2. I'm guessing "yes" but I figured you would know for sure.

Presumably you are talking about swapping a 98 2.2 engine into an earlier model legacy. To do that you would need the software so you could customize the firmware to match the engine. It's not a huge difference but you'd most likely want to adjust your load scales to match the difference in displacement.

Sounds exciting!

Yes, what is the cost of a chip? Will the '92+ chip work for a '98 2.2L engine or will you have to make another chip for it?

[longassname]

I haven't looked at it but I'm sure it won't take a chip. I haven't seen an obd2 subaru ecu with a socket. OBD2 ecu's are usually flashable. So if software is made to interface with it it can be customized/tuned without installing a chip.

Actually, I was asking about the '98 2.2L engine that is in my daily driver. It's a '98 Legacy L wagon with a 2.2L NA engine and auto tranny.

Probably the biggest difference with my '98 Legacy and the first generation 2.2 NAs is that mine is OBD2 while the first generation had the OBD1 system. I had a '93 Legacy with a 2.2 NA engine and it had lots of power for an NA--it would snap a person's head back upon take off from stop with moderate acceleration.

The '98 is quite gutless--I'm not sure if could peel the skin off an overripe banana even if I floored it. I'm guessing that the OBD2 system plays a large part with the substantial lack of power (compared to my '93).

So, that's why I asked if a whole new chip would have to be made for the '98 2.2. I'm guessing "yes" but I figured you would know for sure.

Presumably you are talking about swapping a 98 2.2 engine into an earlier model legacy. To do that you would need the software so you could customize the firmware to match the engine. It's not a huge difference but you'd most likely want to adjust your load scales to match the difference in displacement.

Sounds exciting!

Yes, what is the cost of a chip? Will the '92+ chip work for a '98 2.2L engine or will you have to make another chip for it?

[93Leg-c]

Thanks for the reply, Michael. I know you must be quite busy endeavoring to complete your work for the OBD1 NA software so I appreciate you taking time to respond to my question.

Oh, and one more BIG thing, THANK YOU for your great work in helping NA Legacy owners!

Even though your initial work won't directly benefit me, I'm really eager to see how your chips and software turn out. It'd be quite fascinating to see the difference between a stock NA and one with your chip in it.

[Buffman]

As for the monitoring abilities is this going to be able to correctly show data for us oddball 92 guys? I've experienced with two different 92 ECU's that when I used the monitoring software available, I get sporatic values (they're there, but then it might pop up with a 0 or no value at all)?

I'm definately interested. Granted the mods on my subaru are limited, I would like to be able to tweak things on it as I did with my RMS.

[longassname]

Yes, communications are rock solid with no errors. The problem with the hobyist interfaces you are familiar with is they treat ecu coms like they are ttl and they are not. The thresholds are different. The interface I use, the first picture in the thread, is identical to what the subaru select monitor uses; the heart of it is a gate with the correct thresholds to match the serial port of the ecu.

I have also handled the communications differently in my software. The short of it is coms are rock solid to the point that you can unplug the select monitor plug and it will just stop where it is at and then plug it back in and it will pick right back up without ever skipping a beat, missing a byte, or freezing up the software.

As for the monitoring abilities is this going to be able to correctly show data for us oddball 92 guys? I've experienced with two different 92 ECU's that when I used the monitoring software available, I get sporatic values (they're there, but then it might pop up with a 0 or no value at all)?

I'm definately interested. Granted the mods on my subaru are limited, I would like to be able to tweak things on it as I did with my RMS.

[Buffman]

Thanks. Definately interested depending on price