Tuning and Troubleshooting: This page will detail tuning the TPI system as it is installed on a different vehicle. The Suburban is significantly heavier then a Firebird GTA. Also, although the engines are (were) both stock for their particular application, the Suburban engine was designed for TBI system and emphasizes a lower torque curve, and demands a different spark table (much lower settings than a stock Firebird).
DIY (Do-It-Yourself) Fuel Injection Tuning the VE (Volumetric Efficiency) Tables Tuning the SA (Spark Advance) Tables
DIY (Do-It-Yourself) Fuel Injection: In the process of troubleshooting my conversion, I was exposed to many new terms in electronic fuel injection. Using my Auto-X-Ray scanner, I was able to read real-time data and use this in the first steps of understanding what was happening and what may need to be adjusted to fit my particular application. Other than buying a custom chip, I had little choice in fine tuning this TPI system for my particular application. It should be noted that in order to do a real "custom" chip, the vehicle would need to be taken to a shop with a chassis dyno that could simulate real-time data, under real operating conditions. There are too many variables to make a custom chip over the phone or fax. Doing it yourself gives you as much time as you need to slowly refine and make changes. And no matter how long it takes, it is a one time investment.
After reading an article titled "Chips Know it All" (1MB - pdf file, click here for a high resolution copy 3.8MB) by Shawn M. West, which originally appeared in the October 2004 issue of Chevy High Performance magazine, I decided that perhaps I should attempt the DIY-PROM route. Before attempting to do any DIY programming, it is imperative that you read up and study this subject in great detail.
I will not go into any detail on the definitions of MEMCAL, EPROM, binary or "bin", mask, etc. Read the attached article for it is the best summary I have found to date. In fact, read it two or three times, whatever it takes to start understanding the differences and their meanings. Then, download and read Holley's manual for their Commander 950 ECM. This is a very informative and detailed manual that defines many common EFI terms. Many of the tables, constants and other items associated with the 950 system are similar to the items you will find when working with the GM system. Last but not least, visit www.thirdgen.org and use the search feature. Thirdgen.org has a number of good articles written by some of the senior members. There is a DIY EPROM section in the forum that is second to none when it comes to finding information on questions you may have. You will have to join as a member, but I assure you it is one of the best moves you can make when attempting DIY-EFI. If you cannot find an answer, post the question, and chances are, you will get information that will help you solve the problem(s).
Here are some articles/threads from various sources that I found extremely helpful:
Programming and Reading Data: For the purposes of my tuning, I am using a #727 ECM that uses an $8D mask. The #727 ECM is an under the hood model that came in speed-density Corvettes. The $8D mask is the same that is used in #730 ECM that came in speed-density Camaros/Firebirds (1990-1992). If you read the articles I linked above, you will understand this. If you are using a different ECM, then you will need a different mask, bin & ECU.
Based on my research (above), this is how I programmed my own chips:
HARDWARE: I decided to go the "flash" chip method, so I purchased* the following items from www.moates.net:
After getting the TPI system tuned for the stock TBI engine (or reasonably close), I ordered and installed a G1 Memory Adapter using a AT29C256 chip. The G1 adapter could be used in lieu of the GP1, but removing and reinstalling chips would take more time and care. The G1 adapter will permit the ECM lid to close without interference, whereas the GP1 is too high. You can see the difference in the pictures below. If I were working with an interior mounted ECM and I was planning on a lot of tuning, I would probably not worry about closing the lid, and just use the GP1 adapter. In fact, if a high performance application is what you are doing, you could consider the G3 Bank Switching Adapter [Model: G3] ($55) or the GP3 Switcher Adapter with ZIF and Chip [Model: GP3] ($75).
SOFTWARE: Download TunerProRT at http://tunerpro.markmansur.com/. This program is free, but a $30.00 donation is suggested. Considering the alternatives and the power and versatility of this program, it is worth every penny. Read about all its features here. It can also be downloaded from Moates' site.
Before using TunerProRT, do yourself a favor and read the entire "Help" section. Mark Mansur has taken a complex subject and streamlined it into sections that guide you through reading and burning chips. His explanation does not go into how to tune your vehicle, but how to use his program.
Download the Super_8dm2.ecu file (for the $8D mask) - this free from Mansur's website as well. This is a very comprehensive ECU that will allow you to read and change most aspects of your $8D bin.
For a total investment of only $260.00*, you will have all the tools necessary for making your own "Custom" chips, and reading real-time effects of your changes (datalogging). My cost was only $230.00 because I used the ALDL cable from my Auto-X-Ray package. This plugs into the ALDU1 converter. * Plus applicable S&H costs. Think about it, for less than the cost of most "custom" chips, you can invest in the equipment read and make your own chips, over and over again, for as long as it takes to dial in that perfect tune.
Another "real-time" data logging software package that works with the $8D mask is DataMaster 8D from TTS Power Systems. This is program can be downloaded for free, but you only get 20 free runs, before having to purchase a key to use the program. Download it here.
Burning a "Custom" Flash Chip:
The first order of business was to read my original S&P chip and save it as a new bin file. I did this using the BURN1, HDR1 and TunerPro RT. Note: For an $8D mask, it is very important to specify the MEMCAL you are reading as a 27C256 (Read-Only). This will save the bin as a 32kb file.
I was then able to pull it up and review the VE tables, spark tables, etc. It is very similar to the original GTA bin (ANJF). For information, the GTA's ANJF bin is nearly identical to the very popular AUJP. The AUJP bin can be downloaded from numerous locations, including at moates.net.
For the most part, the S&P chip ran fine with the stock 22# injectors set at 43.5 psi (no vacuum). It does run slightly lean at idle, but this can be adjusted later in the VE table. However, because of the differences in vehicle weight, the S&P (AUJP) Spark Advance was set way too high. This needed a dramatic adjustment.
I made sure that my base timing as set by the distributor (EST disconnected) and as defined in the bin file were the same. For my case, both are set to 6* BTC. I decided that based on my knock counts, that the S&P spark table needed to be reduced pretty much throughout all RPM and load ranges. Before making and spark table changes, I read the original Suburban TBI EPROM (BJLH). But in order to do this, I had to download a $OD mask. If found this at moates.net, which again is free. Using TunerPro's "Right-click" option, I was then able to copy the entire spark table from the stock BJLH bin, paste it into Excel; then copy back into my new $8D bin (which I named "truck_2.bin"). But to do this properly, I had to edit out a few extra rows of RPM from the original BJLH spark table. Transplanting the original TBI spark table into the original S&P chip made significant improvements in spark knock (less occurred); but it still needs more adjustment, especially in the part-throttle range (70-90 kPa) in the 2000 +/- range.
I saved the new bin and burned a new chip using TunerPro's Tools>Moates PROM I/O option (which is the BURN1 process). NOTE: When saving a new $8D bin to a 27SF512 flash chip (comes with the GP1 adapter), you must specify the start address to 8000 hex. This is clearly stated at Moates' website, but be sure to do this or your changes will not take effect. I learned this the hard way spending about two weeks making changes and never once see the results when test driving the truck. Here is the link to the 27SF512 description. Below is a quote from that description, but the 727 is added for reference:
"Works fine with 1227165, 1227730, and other similar ECMs (including Corvette 727). If your car normally takes a 27C128, 27C256, or 27C512, these should work for you. Just adjust the 'Start of Device' value when programming. (For 27C128 original, set to C000 hex. For 27C256 original, set to 8000 hex.)
Using Moates' GP1 (picture here), I can quickly change flash chips whenever I make a change. Then I test drive and scan the real-time data to see if my change(s) had any effect. But more refinement is still necessary.
VSS - Is it really Necessary?:
For reasons described in the next paragraph, I started out running my TPI ECM without a VSS signal. I learned quickly that without a VSS signal, the ECM obviously does not know the vehicle speed and when it is decelerating. Functions such as EGR and CCP (charcoal canister purge) rely on MPH readings for operating parameters. But most importantly, the ECM cuts off fuel completely when decelerating. I noticed that without a VSS signal input to the ECM, my BLM readings would drop to a very rich 114-116 when coming to an abrupt stop or slow down. The idle would also start to hunt up and down when coasting to a stop, especially an abrupt stop or slow down (like coming off an interstate ramp). Not only would the idle be affected, but frequently the engine would just stall. I decided that it was imperative to have a proper VSS signal hooked up to the TPI's ECM.
Since my Suburban uses a high speed VSS output of 128,000 ppm, it was not a matter of just tapping my #727 ECM VSS lead to the existing VSS sending unit. The TPI type ECM uses a 4000 ppm VSS signal. I tried tapping into the buffered signals coming out of the truck's VSS buffer that output to the speedometer and cruise control, but I could not get this to work properly. So I ended up buying a Dakota Digital Model SGI-5 Universal Speedometer Calibration Box for $90.00 from Street & Performance. I mounted this box under the dash, just left of the steering column. I tapped into the signal wire (purple) coming out of the transmission tailshaft VSS plug and ran it to "Sig. In" terminal on the DD box. I then ran a wire from the "Out 4" terminal on the DD box and tapped it into the yellow wire coming out of the #727 ECM. The DD box also requires a ground and 12+ ignition lead. The DD box can also be fine tuned using the calibration buttons. Note: when running the #727 ECM VSS lead behind the engine, be sure to keep it secured and clear of all ignition sources, including the coil, distributor cap, plug wires and interconnecting wires. If the VSS wires picks up any interference from these sources, it will provide false speed signals and may even cause no signal at all, which again results in poor idle and frequent stalling when coming to a stop.
Tuning the VE Tables:
I am not going to go into the background of reading and using BLM measurements here. You can read plenty about that in the links I listed above.
Starting with the original Street & Performance VE tables, I set up an Excel spreadsheet to track the changes. This method gives me the ability to keep a chronological history of changes, as well as let the spreadsheet calculate the changes for me using simple formulas (i.e. Observed BLM reading of 132/128 x last VE number when datalog was recorded).
Using TunerPro, I pulled up the VE table and right clicked it. I then chose "Copy Table to Clipboard" and pasted it into the Excel spreadsheet. I set up three tables in my spreadsheet. The far right table is the original VE table (or the one I want to adjust), the center table is for my changes based on the DataMaster Histogram readings, and the left table is the revised table to copy and paste back into TunerPro. Note: you could also use TunerPro's datalog feature here as well, for you could develop your own Histogram by running through the log and manually recording BLM readings at the different load and RPM points.
Once I make my revisions, I then copy the left table, and again right click the VE table in TunerPro. But this time I chose "Paste Clipboard data into Table". I saved the table, and then saved the bin as a new file. I can keep inserting rows in my spreadsheet, copy all three tables up to the top, then copy the last adjusted table and paste (using Paste Special "Values") on the new right side table. As I start to zero in on my optimum VE tables, I decided to use TunerPro's 3-D graph feature to better blend my table. You can go point-by-point and raise or lower the listed value. This process has worked very well for me. I can then copy the revised table from TunerPro back into my spreadsheet as a new starting point for adjustments.
The above three tables are from the same page & file. DataMaster's Histogram can float and literally be positioned over the active spreadsheet. This is extremely convenient when reviewing BLM readings and making changes.
Tuning the SA (Spark Advance) Tables:
The amount of spark I had to take out of the S&P chip (AUPJ) was anywhere between 5*-15* in many different RPM and load points. The way I performed this step was to datalog and simply watch my Auto-X-Ray scanner for how much spark was being taken out by the knock sensor at various RPM and load points. The table on the left is the original AUJP SA table, the center table is the resultant SA for my application, and the far right table shows the difference. Some parts of the table needed little if any change because I never actually found (hit) those RPM and load points when testing. I used the 3-D graph feature to smooth out my specific SA table after doing many datalog and observation sessions.
Please note that I have emphasized "my application". If you use these tables (or any other data on my site) which is specific for my application, you do so at your own risk. Every engine and every application is different. The only way to determine what is best for your engine is to test (cautiously) and make changes as you hone in on what your engine "wants", not what you think it "needs".
Email me with any questions or comments: mail@edgesz28.com
|
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
){kind=link}