I've been fighting with my TBI setup for a couple of weeks now (took a while to find some wires that were mixed up/cut wrong/hidden under the miles of electrical tape and it took a few days to diagnose a bad ignition module -Doh!) and been reading threads for longer. I know a lot more, but I feel like I know even less than when I started. Does anyone have all of the information about how tables and constants interact consolidated into a spreadsheet? Or even just the math listed on a website? Stuff like, VE1 + VE2 = Fuel delivered. Specifically I'm looking for 7747 ECM information using 42.xdf. What I'm thinking is, it would be nice if I could query certain conditions like coolant temperature and RPM and have specific IAC Step, Spark advance, Fuel delivery amount and MAP come up. I think it would give a good picture of what's going on and allow me to chase my tail less burning chips and then checking vacuum, spark and RPM while the engine is warming up and the garage is filling with fumes (It's an AMC 360 in a jeep and I haven't got it running well enough to remove it from the garage yet, and no, I don't have a way to log data yet). Anybody? Anybody? Bueller?

Nevermind... Another couple of DAYS of reading and searching and eyes glazing over from information overload and I found this site: Linky

 

That site and a lot of other TBI information is linked in the sticky above.

We have put almost all worthy links/posts/articles concerning any and all, including swapped in later version thirdgen ecm's in the sticky "Tuning Guidebook" at the top of the forum.

 

Yeah, I've read most of it. Some is relevant to what I'm doing, a lot isn't. I'm on major information overload looking here, the Scout forums and getting tid-bits from other Full-Size Jeep guys who have converted to TBI. That link has some of what I need, but again, a lot of it is irrelevant to what I'm doing and the ECM I'm working with. Fortunately I finally found a couple more needles in the haystack (eliminating ESC and finally finding a bin that *sorta* works) and have the engine running well enough that the jeep "should" be able to back out of the garage for the first time in months...

 

As far as the VE and VE adder tables go, those are added together to get the total VE. The VE adder table only exists because of the extreme limitations in the TBI ECMs, expecially memory. In lieu of a full VE table, they put the crap half sized table in with an adder table. Since the first table ends early, there is no way to taper off the VE as the RPMs increase, so to prevent the engine from becoming real rich out top, they use the VE adder table to taper off the VE's. Do some searches on this, there is alot of info on the adder tables.

You might even consider going to EBL down the road. While its not a perfect solution (hardware in the TBI ECMs totally blows), it does overcome alot of the software flaws that are in the stock computers, and also comes with documentation and better datalogging software to boot.

 

What I was really hoping for was some kind of "virtual tuning" ability so that I could enter certain conditions and have the program show at least the total fuel and spark numbers. If I can gather up all the information (I'm working on it) and put it all together in Excel (I'm working on it) I should be able to make something like this happen. It seems (to me) that something like that would be useful in tuning because you could see the numeric effect of adjusting one table or another before burning dino.

And since it's a full-size Jeep, I'm looking for a little performance to pull the 7000lb behemoth around, but I also want to be able to crawl down a trail without overheating.

 

It's a great idea, IMO, but would take a bit of software work. Maybe matlab could do it with a lot of work. I don't know if there's a cheap software emulator available that can run GM code in a motorolla whatever chip.

You could try to simplify things and just write a few equations in excel. Make an excel template (basically does everything that the XDF does) that reads and converts a BIN correctly, and then a place to input the current state such as RPM, MAP, TPS, ECT, IAT, VSS, etc...

Best option for most is an ECM test bench. Just uses a pulse generator, several potentiometers for sensors, switches for switched inputs, some displays (LEDs, digital tach, voltmeters), some dummy loads for injectors and coils and relays, and maybe mounted IAC. Then scope whatever data you desire. A bit more expensive than writing the excel stuff, but it's more accurate (it uses an actual ECU with the actual code), and less mental stress.

 

If only I had all the interaction stuff. I wrote an excel add-in (less math, more programming) for the reliability department at a big company I used to work for. Thus my thinking of doing this in Excel, I guess.

 

If you didn't want to do the Excel thing. The Megasquirt Stimulator ($40) has most of the signals that you need. Combine it with a 1 Amp wal-wart, a $25 ECM and some wiring and you would be up and running. That setup has saved me countless gallons of gas and days/weeks of frustration.

 

But I'm so cheap! And that all costs money! I'm getting closer to having this thing moving and then *hopefully* I can log some good data and tune. Tuning will be the easy part compared to what it has taken to get to this point.

 

Sitting in the house and tweaking tables while watching the PW move will pay for itself in fuel saved so quickly it's hard to believe.
Not to mention all the time it will save.

A WB and a bench will save an incredible amount of time and fuel.

 

Well... back to my original post... If I had all the interaction stuff I could easily make a little Excel program to do "virtual tuning". I'm still working on it, but I'm kind of stalling out in a few places and I'm trying to do it in parallel with my attempts to get the jeep out of the garage for the first time in months.

 

How about a "holy crap he's serioius" post... Or maybe it's a post...

What else do I need to include here:

 

I made this excel sheet a while back, its specifically taylored for the $58 Syclone/Typhoon code, but you could change it around to your needs. Also, I think this is an old version, I'll have to upload the newer version early part of next week, maybe sooner.

(NOTE: If it asks for a password click cancel. If it appears to have frozen, the password box is being hidden by another window).
https://jshare.johnshopkins.edu/asch...VE%20table.xls

 

Can you read assembly? If so, just open a source code for whatever you want, and chug away. Many are hacked already, but the devil is in the details. Excel can do math much better than the chip in the ECU, which may hide a real problem (remember the old Displacement and Injector size limit?)

For your interface, you'll want all sensors, and switch inputs. Then you'll have to sort out all of the special modes that an ECU can run in. Then in more sophisticated versions, you'll want a revolutions count slider, a time slider, etc.. Then when you get really serious, you'll have an engine model that responds to input conditions and loads.

But by the time you get that far, you'll probably realize it would have been easier to just get a load box and simulator (to get your methods set, learn what affects what, and look at how the tables and constants REALLY work), and then begin tuning on your car and learning the "old fashioned" way of guess and check + simple experiments + notes.

Actually, if you're not making code changes, you can jump right into a car with a chip burner or emulator, and a wideband, and tweak away. Ask questions here when you get stumped.

 

I know there are probably easier ways, but it seemed like the Excel idea might be useful. It would be neat to have inputs for vehicle weight, axle ratio, transmission ratios, tire size, and a lot of other variables too.

For now I'm fighting with a cold-idle problem. I *finally* have data, so that's a start. The problem is the engine still wants to cycle through high RPM/die/high RPM/die a couple times and then finally just quits. Here's some of the data if anybody is interested (I don't know what else might be useful since there's no ESC, EGR or VSS and nothing else seems to be as active as these fields):

SENSOR:IAC SENSOR:Coolant Temp SENSOR:MAP SENSOR:RPM SENSOR:TPS
145______________38______________35.8_________1100________-0.3
96 ______________38 ______________34 _________975 ________-0.3
96 ______________38 ______________35.4 _________950 ________-0.3
96 ______________38 ______________36.6 _________900 ________-0.3
103______________38 ______________66.4 _________350 ________-0.3
114______________38 ______________83 _________600 ________-0.3
99 ______________38 ______________32.5 _________1075 ________-0.3
90 ______________38 ______________31 _________1025 ________-0.3
84 ______________38 ______________32.5 _________925 ________-0.3
82 ______________38 ______________43.2 _________625 ________-0.3
98 ______________38 ______________87.5 _________325 ________-0.3
93 ______________38 ______________32.9 _________1025 ________-0.3
87 ______________38 ______________31.8 _________1000 ________-0.3
85 ______________38 ______________32.5 _________975 ________-0.3
84 ______________38 ______________35.8 _________850 ________-0.3
84 ______________38 ______________35.8 _________850 ________-0.3

 

Increase the fuel delivery at 38* and 30-40 KPA MAP. Surging is usually an indicator that the engine is starved for fuel.

 

Ive never really found simulations to be useful for tuning. The real power of simulation is when your writing new code and doing debugging. There its a must to have some way to run the code in some sort of simulator to test it and make sure there arnt any bugs, logic flaws, math errors, freeze-ups, stack overflows, etc.

Simulation can help show table interactions, but it can also lead to false conclusions. If you can make anything out of assembly, then looking at the actual hack is better.