Vanilla Ice

I'm looking around but finding it difficult to find any answers about this setup. All links seem to be 404, or never answered.

I also am not receiving registration email to be able to get onto code59.org

Can SAUJP handle a t5 now?
Can $59?

I have written some pretty awesome code for the turbo subaru world so I'm not against disassembly and code writing. Do I need to start on an AXXD hack?
If I do this I'm going to attempt to bring as much into it as I find worth doing.

I think my ethanol sensing code could work depending on space and inputs for flex fuel. I've only got experience in 32 bit OBD2 cars so it will take considerable time for the learning curve.

Are the code writers still around and want some new eyes and ideas? I'd love to piggyback and get some new age features possible for older cars. Maybe an SAUJP V5.

Traxion may already know my work...

Orr89RocZ

iTrader: (20)

$59 can. Well i assume so, i ran it in a full manual auto.

There are some vette bins for manuals you can try to use and tune for the 305 but the aujp stuff doesnt support boost

Robs EBL does

I have a base 305 lb9 tpi with single turbo $59 bin with 36 lb injectors
But with water injection so it will be leaner than without

Vanilla Ice

What I'm seeing now is basically $8d is the starting bin and these axxd and aujp are just calibrations of the same setup. So really aujp should be able to be calibrated for manuals somewhat easily. And unless some code is dependent on the trans such as shift light in SAUJP it should work too.

Anybody in the know able to tell me if I'm heading in the right direction? I don't have my tuning laptop so I'm unable to download these and check them out currently.

Vanilla Ice

Damn. I was under the impression SAUJP handled boost.

Can the $8d simply have its MAP sensor voltage calibration altered to read a 2-3 bar?

Orr89RocZ

iTrader: (20)

Guy on here with user name starting with junk made a code but hasnt let anyone play

Aujp is just the 350 auto bin. Its 8d. Theres a vette manual 350 one but i forget what it is. I know kevin91z and dynodon have used it on some of the cali tpi manual cars and it worked better for their manuals than the automatic bin did

Vanilla Ice

That sounds like an easy swap too. Would just need to change a couple parameters, but I don't see the point if it can't handle boost. I might as well tackle this myself and open source it for the masses.

Anybody want to throw me some disassembly info? I guess I can try to google the processor manuals.

RBob

iTrader: (1)

Will want the MC68HC11 'pink' book from Motorola. Basically only the instruction set section will match the uProc's that GM used. Currently can be found here:

http://www.nxp.com/files/microcontro.../M68HC11RM.pdf

There is a pdf file of the '90 Vette auto $8D, ANHT out there (attached). This is fairly well commented and can be converted to assembly to assemble. There is a copy of AUJP that has been disassembled, but is hardly commented.

Note that the code base varies among the various $8D BCC's. Such as AUJP having a hard coded high idle RPM limit, while the Vette BINs do not have this.

You can look at the $58 code to see how GM handled a 2-bar MAP sensor. First thing is to convert the 2-bar MAP value into what I call a 1-bar Speed Density value. This is then used for all of hte current code that uses this value.

Create a modified value for use in the injector calculation.

Can make a spark retard table and a fuel multiplier table for use when in boost. Or, as I found is easier to deal with when tuning, a spark advance table based on MAP and RPM for use in boost.

RBob.

Vanilla Ice

Cool, I found those earlier. Means I'm already on the right track. Got a Disassebly going in IDA. Using the XDF to search and verify the AUJP similarities.

Definitely need to get used to non 32 bit code. It's foreign to me as I'm a relative noob.

Vanilla Ice

How much free space is in the RAM?
Is there a section that is free to use that you know the address range of?

Thinking ahead for when I start to write up some code. Might want to make some new Ram Variables if I can.

Really what I'm seeing short term is that I just need to fix the code where it compares the MAP to Baro (if that's even necessary). Then adjust everything that references MAP for a 2 or 3 bar. Haven't run into anything else yet and I may be off base, but I'm still rooting around in the ROM.

OR If I can use some RAM space then I just write up a subroutine that will take the Raw MAP value and adjust the voltage scalar and have MAP referenced through the new RAM variable for calculations...

Speed Density will throw a wrench in my stuff I can tell already.

RBob

iTrader: (1)

There isn't much free RAM available in stock form. Just some bytes here and there. If you get rid of mode 4 commands can free up a block of RAM, may also be able to cut into the stack area some.

Need to be careful with baro as the ECM does pseudo baro reads. That feature needs to be disabled with a boosted engine.

Note that the '7730 ECMs have a provision to add a RAM chip.

RBob.

Vanilla Ice

Interesting though I wouldn't know how or even how to reference it once added.

Thanks, I'll look into that too.

Vanilla Ice

So I calculated 2/3 bar values to change in the .bin to get it working pretty much perfect.
Next time to trace all MAP values and possibly bypass some routines to keep the ECU happy with new values.

These ECUs are so simple this is so easy to do.

3 bar values:
scalar: 210 = 0x00D2
Offset: 921 = 0x0399

2 bar values:
scalar: 329 = 0x0149
offset: 2253 = 0x08CD

These values get changed in the subroutine at LB178. They are in there twice for pseudoBaro and the MAP.
also at LB5FD. LB632. CEC0.

Vacuum calc:
B658: = 0xF6 Change to 0xF7

Diag Error Values (2 Bar Sensor):
Error 33:
8283: 0xDA Change to 0x8A (assuming 14.7Baro = 12psi Overboost... for now)
Error 34:
8288: 0x26 Change to 0x17 (same as OEM vacuum)

Saving this here for posterity.

Orr89RocZ

iTrader: (20)

Interesting. Curious to see how it all works

Vanilla Ice

There's not much to it after this.

Checking how the pseudo MAP is used then checking the ratio calculation and error code values. Then all that's left is making new extended tables for larger map values so more resolution isn't lost. You lose some with the new sensor but that's impossible to fix since it's stuck with 256 values and 5 volts no matter how many KPA it reads. (1/2/3 BAR)

RBob

iTrader: (1)

Run a 1-bar along with the 2 or 3 bar MAP sensor. The '7730 has two MAP inputs to the ADC. In testing I was able to feel the loss of resolution just going from the 1-bar to a 2-bar sensor. It is ever so slight, but it's there. Using a 3-bar in place of the 1-bar and it was obvious.

RBob.

Vanilla Ice

That's a good idea, but I want this to be as easy as possible. No rewire, no special needs.

Resolution:
1 bar is .04 psi per step per .35kpa.
2 bar is .11 psi per step per .8kpa.
3 bar is .17 psi per step per 1.2kpa.

At the resolution of the stock MAP axis LU tables I think stock like driveability can be obtained. Hell I may be wrong but I'm going to try.

Vanilla Ice

Just came across the NVSRAM module on moates.

In order to get this installed must I burn the eprom prior to install or can I install then upload the altered axxc rom?

Also installing the module does not require soldering or does it? Haven't found any information on physical install of the module.
Pretty sure I've answered this myself, requires 3 wires soldered. Why isn't this stuff documented better?

Want to get equipment going so I can start implementing and testing.

I may get this and use the RAM space for some new variables... Probably use it to create the mapswitching or Virtual Flex Fuel Sensor code.






Here's another question for you, I can't get to code59.org since they are clearly not accepting new members. So I am totally blind to the whole setup.

But if $59 can be run on a 730 ecm to run boost, then the A/D channel is still 8bit. That means the resolution for $59 is the same as what I will have. With that in mind, stock driveability should be easy with my setup.

Am I missing something?

Vanilla Ice

Wrote some code to rework how the Baro is calculated.
Have it JSR to this subroutine instead of a compare and load.

It checks if it's zero, if so Loads Baro from MAP value.
If not Zero:
Check RPM to see if engine running.
If running Baro doesn't change.
If 0 RPM Load Baro from MAP value.
Return from Subroutine.

ROM:F800 ; ---------------------------------------------------------------------------
ROM:F800 psha ; Push accumulator A into stack
ROM:F801 clra ; Clear
ROM:F802 cmpa pMAP_PseudoBARO ; Compare accumulator with memory
ROM:F804 bne loc_F80A ; Branch if not equal
ROM:F806 pula ; Pull accumulator A from stack
ROM:F807 staa pMAP_PseudoBARO ; Store accumulator in memory
ROM:F809 rts ; Return from subroutine
ROM:F80A ; ---------------------------------------------------------------------------
ROM:F80A
ROM:F80A loc_F80A: ; CODE XREF: ROM:F804j
ROM:F80A cmpa pRPM_VAR ; Compare accumulator with memory
ROM:F80C bne loc_F812 ; Branch if not equal
ROM:F80E pula ; Pull accumulator A from stack
ROM:F80F staa pMAP_PseudoBARO ; Store accumulator in memory
ROM:F811 rts ; Return from subroutine
ROM:F812 ; ---------------------------------------------------------------------------
ROM:F812
ROM:F812 loc_F812: ; CODE XREF: ROM:F80Cj
ROM:F812 pula ; Pull accumulator A from stack
ROM:F813 rts ; Return from subroutine
ROM:F813 ; ---------------------------------------------------------------------------

Going to change this to check RPM first. That way if engine is off it updates baro every time.

Vanilla Ice

Ran through diagnostic values and searched the rom for other requirements. Still have a little looking around to do.

Need to look at vacuum one more time, have to decide where ROM space is that j want to utilize then really it's edit the ROM officially and it will be ready to run a 2 Bar MAP on a non boosted setup. Create and implement larger tables for higher KPA and its ready for Boost. If I had the hardware I could have it released this week.

Street Lethal

iTrader: (16)

There is support for that code on the Sy/Ty forum if you can't get in Code59.org...

http://www.syty.net/forums/

JP86SS

iTrader: (1)

WhaaaaaaT?
Been there for 10 years, AUJP_JP2.zip on Moates Fileman on Gearhead site.

The AUJP file is commented and assembles to match the stock bin.
Is also relocatable as variables have been changed to allow changes to the code without it becoming corrupted.
The ANHT disassembly will not assemble due to missing info.
Great piece of work and was used to create the commented AUJP files but was incomplete.

In any case, $58/59 is going to be your best starting point if you want to go boosted. Several have modded the $8D but as stated, nothing in the wild for use.
There is a copy of some old $59 files also on the Moates fileman that may help.

Jp

Vanilla Ice

That's going to change. I've already posted some info on howto.
Thanks, I'll look for that.

Street Lethal

iTrader: (16)

I think the only member who had access to it was Joe, and the way he explained it the code was very limited the way it was written, but that was awhile ago, and maybe it was preliminary, as junkcltr might have changed something after that, I am not really sure. He gets quiet whenever someone brings it back up...

The code RBob put together for his version of a boosted $8D is far superior, but of course I am talking about the EBL-P4. There have been many threads about a drop in $8D prom with boost, but nothing was ever finalized. I looked at $8F and wanted to incorporate it into the '7730 prior to the EBL-P4, but was getting a headache from the XDF, and some of the comments seemed off. Apparently Robert Saar wrote something for it up to 3-Bar that I never got a chance to look at, and I think there was something written for a manual $8F code as well, I am not sure. Other than Saar and Mars, not too sure who else to source for that particular code. As for $59, I believe Six Shooter wrote some code for it, so he is another one to source if you need information. I'm curious to see how the OP finishes his version and then compare it to what junkcltr did. I'm sure there may be a road block or two, but it is still very refreshing to get a new perspective on a boosted $8D...

JP86SS

iTrader: (1)

There was the $60 code as well...Grumpy's own with DIS IIRC.
Never went far as nobody was using a setup similar to his monster.

I'm still looking to add boost one day, would be nice to have 730 code even if it was crude.
Jp

Vanilla Ice

This won't be crude.

Already wrote the algorithm for map switching on the fly with stock ECU. Plan on locating a free byte of RAM for it. Then hijack the code for some custom subroutines to alter what I need.

Don't know my end purpose for it other than just for fun but I'll probably make it a Valet Mode and kill timing and lower max RPM.

Vanilla Ice

Been working on more disassembly of the code.

Found some IO addresses that help. Got a basic code to pull from the f14? pin and store to a free ram spot. Can take that and have the ecu read my wideband directly.

Currently thinking about the Drive Mode Switch.
Needs to be in NVRam to work correctly so I am planning on Hijacking an unused couple bit as a simple 00,01,11,10 four mode switch. Should work perfectly... as long as I don't hijack the wrong bits...

Problem now is finding enough NVram bits to come close to Flex Fuel work... But I should be able to...

Is the stack part of the nonvolatile ram? Makes sense that it is, but I can't find an answer to this.

Thinking about hijacking the CCP routine and use it for an EBCS routine. This way I can have a valet mode that drops the turbo down to wastegate pressure too. No big deal if I can't. Haven't looked into it to see if even possible, but I believe they work off the same hz range so who knows.

MaxpowerTA

If this works, It seems like it would be an awesome addition to integrate into the latest SAUJP code........

Vanilla Ice

If what works?

Any of this will need ported to the other ROMs. I won't do it unless I can get someone to share SAUJP code or disassembled. Got enough to do with axxd. But should easily be doable. I'm handwriting hex right now but will switch to coding and can easily share all I'm doing.
I just hope to fit enough in without needing the NVSRAM board. I don't foresee an issue with selectable modes on the fly. And boost will definitely happen. It's the e85 that will probably need extra nvram

Orr89RocZ

iTrader: (20)

Even if its axxd for 305 you can still tune it for any motor, just change the parameters around

Vanilla Ice

Right but to get to saujp or any other custom, you'd need to port it.

RBob

iTrader: (1)

Just grabbed your version, I haven't seen that one before. It is well commented. The one I was referring to was done is 1998 and is a basic dis-assembly.

RBob.

Vanilla Ice

Might utilize another PIN to read an ethanol sensor. That will save on the need for ecu ram addons but will require $70 or so in sensor. Plus it will substantially lower sram needs too.

Wrote the code for modeswitch. Requires 2 bits in nvram. 1 bit ram.
Use the mode to change anything needed after that by multiplier or adder to values. Done and done. Pretty useless until I get into writing other code in there. Probably just use it as a coded injector killswitch at first.

Flex Fuel code is so simple I'll probably do it just for fun.

Also have a great idea for a momentary switch to toggle modes. Or it can be used to trigger ethanol sensing if I go through with the "free" (after nvram addon) flex fuel version. Or it can be used to read out diagnostic codes or current Mode or anything else really.

Vanilla Ice

Pulled my AXXC Stock bin. Attached.

Compared it to the AXXD file I've been Disassembling. Very, very little difference.

Car is running on a eeprom AXXC now. Officially ready to start testing my code...

JP86SS

iTrader: (1)

Is there anything I can try to "port" into the AUJP code?

Vanilla Ice

Nothing yet as I just got the hardware nothing has been finalized or tested.

Much of it is in my head or coded for 32bit. Without all my ducks in a row I have nothing to share.

Are you the SAUJP master? I was looking around for a disassembly or source code for it. Had I been able to get a hold of that, I'd rather piggyback off that and make a version 6 or whatever it's on. Then you could look it over for your official release.

No sense in two $8D versions, but I am doing what I require of my car so if I can't piggyback, I'll simply continue on my path, no worries.


Right now I'm working on Datalogging new Ram Variables. New to XDF stuff. Took a while to teach myself that today.

Vanilla Ice

Thanks to a MAP sensor donation received from AmorgetRS I will be able to now test Boost code (driveability only) on my ROM.
I have found enough nvram to implement what I need for mode switching.
Got inputs taken care of for an eventual E85 flex fuel build and Wideband inputs.
Figured out the backend of datalogging for new variables. May still go to a more elegant option here... Haven't decided. I like to make the coding as complex as I can to allow for idiot level changes later.

First things I'll work on and release will be mode switching (injector killswitch only since I have no boost to do any better modes, more of proof of concept for now.) CEL flash for Knock and what should be a full implementation of Boost (no wastegate control for now) but by no means less than needed for fluidity.
Then Wideband support built for LC2, but I'll make sure Scalar and offset are easily altered for others.

That's all for now. Have family in town so it may not be this week but I'll probably be running this code next week sometime.

Edit: Got a significant amount of the 2bar Map alterations done today. Just need to look into the updater to the Pseudo Baro reading. See if I want to disable it all together or what. That's a great place to hook for my custom subroutines...

Now I'm moving on to table creation. Don't know table lookups and XDF editing yet.

Anybody want to clue me in to where the ECU knows the values of the Axes?
How does it know RPM goes every 200, TPS goes every 12.5% etc? I'll figure it out, but I'd love a rundown to save time.
Got it. Prior to the LU routine the variables are scaled with Logical shifts.

Vanilla Ice

Still Deep into ram variable tracing. Have written a few needed function subroutines into their permanent place in the ROM. Linked in the pseudo Baro code. Haven't had much time to think about baro update setup, but it's still there. Found a few more places I didn't work in new values for. A lot more tracing to do.

I have a LOT more ideas that I want to implement into it. It's become to intense it's getting harder and harder to slip back in to coding every time I stop. Only had a few hours a week lately.

I'm not going to spill all the beans as far as ideas go until I get it all tested and running.

???

iTrader: (5)

very interesting.. e85 sensor. subscribed.


I'm still running the code 59 guys stuff, using a 3bar, it handled manual transmission well. since $58 it came from had a 4cyl manual option . and gm did have waste gate control in it, even if it's said to be a bit complicated to tune. I was only running a super charger. just a thought if you wanted something to look at as to home the oem handled wastegate.

Vanilla Ice

I think I have a ROM ready to test my 2Bar MAP finally. Need to double check everything first. Been so long I may have missed something I had only in my head. So a couple walk throughs and I'll test it out.

There's no way to run boost yet as there's no extended tables created yet. This is only to check driveability on a 2 bar.

Vanilla Ice

Edit: Ran the wrong Rom. It was a 1 bar bin.

Vanilla Ice

Ran the rom, car runs and drives now.
Edit: I'm an idiot.

Vanilla Ice

Done. Running a 2 bar.

didnt need to fix anything. Either time. First time running was an error where I loaded the wrong rom.
Second time was a bad ADX file.

Any other issues seem to be in my head and due to the 107 degree weather. Basically ran right out of he box the first time. How sad that I could be so poorly organized. It's annoying doing things on 2 different computers.

???

iTrader: (5)

that's awesome man. I truly wish I had a running car to test it on.


I hope some of the guys in the power adder forum will give it a shot before dropping mmoney on stand alones

The_Punisher454

so to summarize, what all does your code change do?

I have a twin turbo project in the works and haven't yet chosen which ecu I'll use.
I was kinda thinking code59 running on one of my 730's with a '427 piggybacked onto it for 4L80E control.

Vanilla Ice

The only thing it does now is run a 2 bar on a stock car. This is only a test rom. It has some hooks some custom log variables and some prep work hidden but has no purpose.
The plans are mostly laid out in this thread, but it will be a while before anything major gets underway as I have very little time.
The next part will probably be knock flashing, mode switching and injector kill switch.

The_Punisher454

Cool, I'll be watching this with great interest.

Vanilla Ice

Had some free time today. Wrote the subroutine into the ROM to trigger the CEL with knock.

Should be able to test this rom tomorrow.

Vanilla Ice

I added in VE map table lookup values up to 200 kpa.
So fueling is almost set for boost.

The cel flash works but needs work. It flashes the shift light at the same time and it also flashes cel when telling me to shift. So I need to separate those.
I have it setup to check for a map threshold before checking the KR threshold. If over both it lights up.

Edit: also went through and added in extended map and rpm timing tables.
Edit: Shiftlight is now independent, but knock flash still flashes shift light. Didn't have much time to delve into it but I caught that bug pretty quick.

Vanilla Ice

Seems to have oem drivability.
Running off custom timing tables up to 6400rpm and 200kpa.
Running off custom upper VE table up to 200kpa. Need to fix to go up to 6400 rpm.
Plan to work on a custom power enrichment table. I think that is probably needed soon.

Any other tables I need to alter for turbo?

Any turbo cars care to run this test rom for real world boost application?
Sign up here and I'll send your way of course I require feedback.

Vanilla Ice

Turbo Test bin is up on the power adder forum.

Added a MAT Timing table adder. Adds timing to value pulled from both lower RPM and upper RPM tables.
Added a Fueling table adder. Adds to VE value pulled from both upper and lower VE tables.

Not in the posted bin:

Added in a Drivemode switch (no true function yet other than mode 0 will cut fuel), killswitch for injectors based off "Mode 0" Flashes CEL when Changing modes. Once I get back into the CEL logic, I'll make it flash the mode number. The modeswitch ram variable is protected at key off so it will stay in the mode you put it in.

Plan is to have a valet mode, limiting upper timing table, cutting fuel at a lower rpm, and will also effect the Turbo spool through WGDC once a working wastegate function is made.
Then a low performance low boost (WGDC) lean mpg mode.
Then performance mode.
Then High performance Race mode. (anti lag launches and such).

These modes should not eat up too much ROM space to allow for blending with Flex fuel one day, but who knows.

It's starting to dawn on me that writing this in C would be so much easier at this point...