Need advice on managing my BLMs and INTs

Printable View

03-07-2017
Danspeed1

Need advice on managing my BLMs and INTs

Hey Guys,

I believe this is my first post although I have been lurking on here since last summer. Great site... I've been able to learn a ton and get my moates setup and tunerpro rt working like a charm; unfortunately I've finally hit a little road block on the 1990 Corvette I am working on. Just a little background on the car. Its has a 350 with a Miniram and long tubes with a heated O2 sensor. I am using the $8D file. I have limited experience using Hptuners to tune newer cars but this is the first OBD1 Car I've attempted. I am trying to use the same logic in tuning it but I seem to be having some trouble managing my long term fuel trim and short term fuel trim. I found a calculator spreadsheet somewhere on here for the $8D file i've been using. DEFCO and EGR off, in closed loop, with no power enrichment, yet I can't seem to get my Lower BLM table to regulate out. After doing a little reading on here again tonight I am going to be changing my history table sample from 10 to 100, but my main issue is that no matter how many times I seem to calculate out the cells ranging from 114 to 134, I can never seem to close in on 128. I guess what I am asking is what is the best way to approach this? On my upper table I managed to get it worked out so that my cells range from 127-129, but it was my understanding that since the BLM is made up of a bunch of INTs I should be able to get all the BLMs to 128... is this a correct way of thinking about this? Also, how much should my INT's fluctuate and do I need to specifically tune for them as well? Ive read a bunch of thread on what BLMs and INTs are but noting concrete on how to properly arrive at a properly tuned car. Using HPtuners on the newer stuff, we just try and get the long term within about 3%-5% and we are set... I feel like there is more to this.

Any help is greatly appreciated!

Thanks,

Dan
03-07-2017
1project2many

You should be tuning to achieve BLM as close to 128 as possible. The sensors, computer, and algorithm that determine air density and required fuel are not perfect. The INT is expected to make up for inaccuracies up to 10%. With BLM at or near 128 then INT shouldn't stray too far from home.
03-07-2017
dave w

Have you tried using Log File Analyzer? http://www.gearhead-efi.com/Fuel-Inj...-File-Analyzer Intermediate to advanced Excel skills could be required. I've had excellent results using Log File Analyzer for tuning $8D.

dave w
03-08-2017
Danspeed1

Quote:

Originally Posted by dave w

Have you tried using Log File Analyzer? http://www.gearhead-efi.com/Fuel-Inj...-File-Analyzer Intermediate to advanced Excel skills could be required. I've had excellent results using Log File Analyzer for tuning $8D.

dave w

Hey Dave,

I appreciate the link and I will certainly give it a thorough read, however I think you lost me at least for the moment. I have been using the two VE tables and a similar spreadsheet to calculate the BLM's and only the BLM's. I have the majority of the upper table close... between 127-129, but the lower table is still a mess after 25-30 chip tunes. I must be doing something wrong, just not sure how to get it dialed in any closer. I see my INT's jumping around as I suspect similarly short term fuel trims do on a newer car, but I also have been reading of people tuning INT's as well. I have no idea how to do that or if it really even needs to be done. Also, using TTS datamaster I am able to see the actual blocks.... but as far as Tunerpro RT is concerned I haven't been able to find anything that signifies those blocks.

Thanks for the tips so far... I'll keep trolling on here and see if I can find any additional info on how to properly attack this.

Dan
03-08-2017
1project2many

I will sometimes lock the BLM and INT at 128 then tune for the way the car feels and the readings on the O2. With all the corrections locked a lean engine will show up very quickly and a rich mixture is easily seen on the O2 sensor. Your feedback from the vehicle is much less damped that way and it can help you if you're having troubles due to cam behavior.
03-08-2017
Danspeed1

Quote:

Originally Posted by 1project2many

I will sometimes lock the BLM and INT at 128 then tune for the way the car feels and the readings on the O2. With all the corrections locked a lean engine will show up very quickly and a rich mixture is easily seen on the O2 sensor. Your feedback from the vehicle is much less damped that way and it can help you if you're having troubles due to cam behavior.

I am getting the impression there are quite a few different ways to get this done. Sadly as I mentioned before I "feel" like I nailed the upper VE table, but the lower one is still a mess. It drives OK but I have a slight shudder around 1600. I was hoping there was some fine tuning that could be done in just that area. Also thinking that any long term fuel trim variance could lead to poor gas mileage. Just trying to nail it. I have friends that have had good drivability success tuning with a wide band but realistically the car's that drive the best are tuned by narrow band using a wide band only to check wot.
03-08-2017
dave w

5 Attachment(s)

Quote:

Originally Posted by Danspeed1

Hey Dave,

I appreciate the link and I will certainly give it a thorough read, however I think you lost me at least for the moment. I have been using the two VE tables and a similar spreadsheet to calculate the BLM's and only the BLM's. I have the majority of the upper table close... between 127-129, but the lower table is still a mess after 25-30 chip tunes. I must be doing something wrong, just not sure how to get it dialed in any closer. I see my INT's jumping around as I suspect similarly short term fuel trims do on a newer car, but I also have been reading of people tuning INT's as well. I have no idea how to do that or if it really even needs to be done. Also, using TTS datamaster I am able to see the actual blocks.... but as far as Tunerpro RT is concerned I haven't been able to find anything that signifies those blocks.

Thanks for the tips so far... I'll keep trolling on here and see if I can find any additional info on how to properly attack this.

Dan

Results from a remote tune I did last year using the Analyzer Spreadsheet to tune with. The vehicle owner was up to speed with using a TunerPro RT and emailed me .xdl files. I crunched the .xdl and sent an updated .bin file for the vehicle owner to burn a chip with.

The Analyzer separates lower tables and upper tables.

dave w
03-08-2017
1project2many

Quote:

I am getting the impression there are quite a few different ways to get this done.

Yes, there are many ways to get this done. I learned tuning before EFI and sometimes I still rely heavily on how the dang thing drives. Maybe it's a dead art but when the ecm is trying to be smarter than I am, when there are competing corrections that need adjustment, I find it can be just as easy to take as much of the ecm's smarts out of the equation as possible.

When looking at logs, understanding what you're reading can be tricky. Sure, the values can be easy to see but how to interpret them and where to make a change can require some thought. If you can find a nice stretch of road with an incline then you can work on some testing under fairly steady state conditions. Target a MAP and RPM and drive with the engine at those points for enough time that the BLM / INT can make their corrections. You can pick cells for the target and set adjacent cells close to the target cell value. That process can build a reasonably close VE table. If you're testing under acceleration with MAP and / or RPM changing it's a tougher job. Determining where AE does and doesn't affect the readings is important. AE should not be set to cover for improper VE, it should add fuel to make up for suddenly increasing air density in the manifold. Knock or lean fueling when rpm is increasing can be due to too much advance or too little fuel several hundred rpm earlier than the sensors record it. The same is true for rich readings under deceleration. It's very helpful to get as much of the VE table right before trying to evaluate these transient problems. Dave's tool and my old school methods are two ways to try and do this.
03-08-2017
Danspeed1

Quote:

Originally Posted by dave w

Results from a remote tune I did last year using the Analyzer Spreadsheet to tune with. The vehicle owner was up to speed with using a TunerPro RT and emailed me .xdl files. I crunched the .xdl and sent an updated .bin file for the vehicle owner to burn a chip with.

The Analyzer separates lower tables and upper tables.

dave w

Thank you for posting those logs. If i am analyzing what you posted correctly it took you 18 tries to get it ironed out properly? If so I think i'm headed in the right direction. Your tables are starting to look VERY similar to mine. I think what i have been trying to justify in my head this entire time is whether or not every BLM box on the table has to be 128. You see, on the newer cars you have a percentage margin of error, 3%-5%. And that is evident looking directly at the Long Term Fuel Trims... The Short term fuel trims are never used for adjustment purposes and they always move around to try and keep your long terms in that 3-5% range. This feels different. My thinking was that I have a box known as a BLM that needs to be at 128.... and within that box are how ever many boxes that make up INT's which need to be some where in the range of 128 in order for the car to be tuned correctly. I was only harping on this because i see people tuning INT's which never really gets done on the newer stuff. So if I am understanding what you are posting correct... SIMPLY... I can just think of my BLMs as my LTFT and the INTs as my STFT and never worry about the INTs... am i correct?

Thanks for everything you posted so far. This site is amazing, i've learned so much just reading random threads on here!

Dan
03-08-2017
Danspeed1

Quote:

Originally Posted by 1project2many

Yes, there are many ways to get this done. I learned tuning before EFI and sometimes I still rely heavily on how the dang thing drives. Maybe it's a dead art but when the ecm is trying to be smarter than I am, when there are competing corrections that need adjustment, I find it can be just as easy to take as much of the ecm's smarts out of the equation as possible.

When looking at logs, understanding what you're reading can be tricky. Sure, the values can be easy to see but how to interpret them and where to make a change can require some thought. If you can find a nice stretch of road with an incline then you can work on some testing under fairly steady state conditions. Target a MAP and RPM and drive with the engine at those points for enough time that the BLM / INT can make their corrections. You can pick cells for the target and set adjacent cells close to the target cell value. That process can build a reasonably close VE table. If you're testing under acceleration with MAP and / or RPM changing it's a tougher job. Determining where AE does and doesn't affect the readings is important. AE should not be set to cover for improper VE, it should add fuel to make up for suddenly increasing air density in the manifold. Knock or lean fueling when rpm is increasing can be due to too much advance or too little fuel several hundred rpm earlier than the sensors record it. The same is true for rich readings under deceleration. It's very helpful to get as much of the VE table right before trying to evaluate these transient problems. Dave's tool and my old school methods are two ways to try and do this.

I think what I am struggling with on both the OBDI and OBDII cars is the pursuit of perfection. There is really no "perfect" tune. Part of the reason why is because the atmosphere the car is in is constantly changing. But with all of these adjustments within the tune it makes me feel as though it can always be better. On my Carburated cars it was so much simpler... does it run, smell pig rich, and leave rubber strips? If yes, yes, and yes.... its perfect.

Dan
03-08-2017
Danspeed1

Quote:

Originally Posted by dave w

Have you tried using Log File Analyzer? http://www.gearhead-efi.com/Fuel-Inj...-File-Analyzer Intermediate to advanced Excel skills could be required. I've had excellent results using Log File Analyzer for tuning $8D.

dave w

Dave,

One more question if I could. I am currently using a different calculator which is to straighten out my tables. The one you posted above seems awesome. The only question I have, is I am not sure how to load my log file into it. Could you step me through it?

Thanks again,
Dan
03-08-2017
dave w

1 Attachment(s)

Quote:

Originally Posted by Danspeed1

Dave,

One more question if I could. I am currently using a different calculator which is to straighten out my tables. The one you posted above seems awesome. The only question I have, is I am not sure how to load my log file into it. Could you step me through it?

Thanks again,
Dan

The spreadsheet will process the .csv file exported from the data log. Have you looked at the "How to Use" document in the .zip folder?

dave w
03-08-2017
1project2many

Quote:

I can just think of my BLMs as my LTFT and the INTs as my STFT and never worry about the INTs... am i correct?

BLM and INT are LTFT and STFT. The OBDII standardization of terms changed GM's wording into "Everyone" terminology. As long as you're within about 10% in total correction (BLM +INT) you're in the target range. GM's target for OBDI is sometimes much wider, BTW. Look at the minimum and maximum limit for BLM before rich and lean codes are set for an example.

Quote:

I think what I am struggling with on both the OBDI and OBDII cars is the pursuit of perfection.

It's always a compromise. I know an individual who re-wrote GM's code and used additional sensors for calibration and eventually worked out a combination that was within 1% accuracy most times. When he showed it to a GM engineer who had been involved in coding , the engineer scoffed that the system would ever need to be that accurate. The engineer concluded they could save some processing power and codespace and end up with a 2% accurate system since OEM OBDI code can have greater than 10% inaccuracy.
03-10-2017
Danspeed1

So just a quick update... although I never got the file above working, I was able to use a combination of another table, plus taking the advice above on adding fuel and got the table dialed in perfectly. I can't explain the satisfaction I had knowing I finally got it. Unfortunately during the WOT pulls.... I ran out of fuel. I knew the injections were too small but against my better judgement I decided it would be close but i would give it a try. Unfortunately I'm at about 120% at 5000K which as we know is not going to work. Rather than bumping the fuel pressure Im just going to order another set of injectors. I am fairly confident I will be able to get it dialed in properly this time around.

The only other question I had was in regard to the spark tables. There is a main spark table and a spark adder table. It appears as though the adder table is always adding about 24 degrees... Is it better to zero out that table and work off the main, or just remove timing from the main table while in the 60 kpa plus range for WOT? Car drives better than stock around town!

Thanks for all the help so far.