How Do You Achieve Zero Fuel Consumption On Deceleration?

[neil]

Any ideas?
I'm trying to get the fuel flow to virtually drop to zero when the vehicle decelerates.
No matter what I adjust seems to have any effect on the fuel usage on deceleration.

Regards,

Neil.

[Biggsy]

Have you played with the DFCO settings? It will cut fuel on decel when its conditions are met.

Cheers,

[neil]

I've tried modifying the DFCO, Throttle Cracker and other areas but they only drop the fuel consumption to about 4.5ltrs/100kms then it starts to increase to about 12ltrs/100kms before the vehicle completely stops. The fuel consumption then drops away to about 2ltrs/100klms at idle.
You can also cause surging to occur if you go too far.
After releasing the throttle the vehicle goes into cruise control slowly dropping in speed.
There should be away of instantly reverting to the idle state after releasing the throttle.
Does anyone have any examples of the required table changes that will achieve this result?

Thanks,

Neil.

[Tordne]

The injectors will never TOTALLY shut off. The pulse width that will be commanded when in DFCO is the value in the Injector Voltage Offset table. If reading thi you have the temptation to change this value to 0 DON'T!!!

I assume you are monitoring the fuel consumption on the LCD display in your particular vehicle. I'm not sure how accurate that will be in this case. I can assure you the the minimum pulse without DFCO active would e in the region of around 1.7ms which will flow a lot more fule than the 0.52ms of pulse in DFCO!

Give up on this one and move on to the next tuning item on your list ;)

[ringram]

I think you might be able to tweak dfco more. Change the map points too. Mine hangs onto cutoff until almost idle then opens out again with a bit of a jerk but thats cool.
You can change the ramp rates to make it cut off faster too. There are quite a few dfco settings to play with.
You can drop your enable speed as well as raise the enable map point (depending on how your car is)
Look down under spark for more dfco settings and obviously kill the 2 sec enable delay.

[limited cv8r]

do you have a manual or auto? Auto's don't really activate DFCO that often when decelarating in 4th as there usually isn't an converter lock up so all the parameters aren't met to activate it. Try locking it in 3rd or 2nd and try if it is an auto. Mine usually indicates 0.4 on the instant fuel when DFCO is active.

[neil]

I've got an auto but I don't think DFCO must be working properly.
It seems to want to go into cruise and slow down very slowly.

Regards,

Neil.

[Biggsy]

Get your car up to 70-80 kmph, manually pull it into second gear and let go of the throttle. Does it engage then?
(Adjust to a lower speed if you are running 4.11's or you will over rev!!!)

Cheers,

[limited cv8r]

post up your tune so maybe we can help you out.

[N0DIH]

I suspect my LT1 (1994 5.7L Gen II Small Block) DOES completely shut down the injectors under decel when DFCO comes in. When monitoring the O2 sensor voltages both drop to 0.004v fixed nearly immediately when the DFCO engages. You can then feel the deceleration kick in. As soon as the DFCO disengages, the O2's returned to normal operation (fluxuation).

I have heard people talk about agressive DFCO being a very bad thing, like piston breakage. Can someone elaborate on why?

My settings are not the same as EFILive has, but might shed some light on a good DFCO setting (note this is very close to factory GM settings).

This is NOT a terribly agressive DFCO, but it works pretty well. I can make it run a little harder, but I don't feel the need, I did try it and returned back to these settings and perfer them over being more agressive.




The injectors will never TOTALLY shut off. The pulse width that will be commanded when in DFCO is the value in the Injector Voltage Offset table. If reading thi you have the temptation to change this value to 0 DON'T!!!

I assume you are monitoring the fuel consumption on the LCD display in your particular vehicle. I'm not sure how accurate that will be in this case. I can assure you the the minimum pulse without DFCO active would e in the region of around 1.7ms which will flow a lot more fule than the 0.52ms of pulse in DFCO!

Give up on this one and move on to the next tuning item on your list ;)

[Tordne]

The feel of engine breaking is the spark being withdrawn... Love the avatar :)

[N0DIH]

If you look at my tables, I don't change the spark on decel. The factory setting did, I leave it alone so it uses the Closed Throttle Spark table instead. I was having some anomoly that when it would go into DFCO, and the timing dropped, it would never come back up and my timing would be in the dumpster until I restarted. Puzzled me until I figured out it was DFCO triggering it. If I didn't allow DFCO, no issue. So DFCO Timing Retard is the amount of timing that is pulled out when DFCO is active. So whatever value in the Closed Throttle Spark table - DFCO Spark Retard = Total DFCO timing. So when I set DFCO Spark Retard to 0, then I leave my timing alone to the Closed Throttle Spark Table.

There is a significant difference in decel without the fuel vs with, and even factory you can feel the 45-48 mph point where fuel comes back on. I have tried and tried to get it significantly lower than 45 mph and although I can, the transition from fuel off to fuel on is too abrupt and my limited PCM wouldn't let me transition smoother.


The feel of engine breaking is the spark being withdrawn... Love the avatar :)

[Biggsy]

I havent touched my DFCO in my tune, so I decided to give mine a test today.
My car is VX Commodore LS1 Auto.
I can make the dash display 0L/100 if I go downhill and do a manual downshift to second at speed.

Cheers,

[dfe1]

When DFCO comes in on pulse width typically drops to somewhere between .200 ms and .350 ms (depending on the vehicle) and O2 voltage goes to .004 mv. With the stock cals I've looked at, the transition to DFCO is very slow because of a slow rate of spark adjustment. DFCO won't become active until ignition timing hits the value in B3336. (The description is not correct. This isn't the value to which timing will be adjusted when DFCO is active-- it's the value that activates DFCO. Spark will drop until it reaches the value in B5915.

Make sure that the value in B5919 isn't higher than the values in B3336 or you'll never hit the timing enabler for DFCO. Try values between .06 and .09 in B3334 to achieve a ramp in rate you're happy with. Values in B3335 (Ramp Out Rate) will be in the .4 to .7 range. Don't forget to adjust the other enablers (rpm, MAP, speed and throttle). If all else fails, send me your tun file.

[N0DIH]

Sorry for the n00b question, but I haven't got EFILive yet. What do you mean by the values in B5919 or B3336? Is this something in the datalog/scanner that is a specific location?

Love the sig! 1+1=10..... And FWIW the systems I used to work on 7+1=10.....

[dfe1]

I was begining to wonder if anybody got it. Sig used to be 7+1=10, then F+1=10. Figured it needed another change, so I went to the binary version. I used to work on octal systems as well. One of the best stories is the time we were installing an electronic cash register system in a department store and some electricians were on site. We were counting cable pairs coming into the system and you should have seen their faces when we counted 1,2,3,4,5,6,7,10, then went on to 16, 17, 20. They couldn't understand what happened to 8, 9, 18 and 19. Ever try to explain an octal or hex system to someone?

As for your EFILive questions, every table in the tuning software is numbered. B3334 is the DFCO Spark Ramp In Rate table, B3335 is the Ramp Out table. It's a little confusing at first, but all you have to do is enter the table number in the navigator window and hit search. The system will display he table for you. It might not be a bad idea to download the software from the EFILive web site and begin looking through it before you get your system.

[joecar]

Sorry for the n00b question, but I haven't got EFILive yet. What do you mean by the values in B5919 or B3336? Is this something in the datalog/scanner that is a specific location?

Love the sig! 1+1=10..... And FWIW the systems I used to work on 7+1=10.....N0DIH,

No problem, feel free to ask questions all you like.

Download and install the software from here: http://www.efilive.com/download.aspx

Without a cable it still runs... you can view logs and view/edit tunes... you just can't log or flash;
real sample logs and tunes are included as are tutorial PDF's (see the Auto VE tutorial, for example).

Cheers
Joe
:cheers:

[joecar]

I was begining to wonder if anybody got it. Sig used to be 7+1=10, then F+1=10. Figured it needed another change, so I went to the binary version. I used to work on octal systems as well. One of the best stories is the time we were installing an electronic cash register system in a department store and some electricians were on site. We were counting cable pairs coming into the system and you should have seen their faces when we counted 1,2,3,4,5,6,7,10, then went on to 16, 17, 20. They couldn't understand what happened to 8, 9, 18 and 19. Ever try to explain an octal or hex system to someone?

As for your EFILive questions, every table in the tuning software is numbered. B3334 is the DFCO Spark Ramp In Rate table, B3335 is the Ramp Out table. It's a little confusing at first, but all you have to do is enter the table number in the navigator window and hit search. The system will display he table for you. It might not be a bad idea to download the software from the EFILive web site and begin looking through it before you get your system.lol...

How about base 36 (...I think IBM did use base 36 a long time ago...)...

Base 36 uses the digits 0-9 and A-Z, so... Z + 1 = 10

:D

[ringram]

Looks like you still have enable delay there!

[neil]

Hi dfe1,
I would appreciate if you could have a look at this tune and advise on the correct setting up of DFCO.
It doesn't seem to work no matter what I do.
I would really like for the engine to virtually cut all the fuel on deceleration.
It's just a standard non modified auto engine
Any help that you could offer would be greatly appreciated.
Thanks.

Regards,

Neil.

[limited cv8r]

looking at your tune your map B3320 settings were to high.try this tune.

[dfe1]

Try this. Some of the settings may not be what you're expecting-- don't change them. DFCO should become active as soon as you take your foot off the gas pedal. Let me know how it works.

[mistermike]

Try this. Some of the settings may not be what you're expecting-- don't change them. DFCO should become active as soon as you take your foot off the gas pedal. Let me know how it works.
Did you remove the attached file for a reason?

[dfe1]

Did you remove the attached file for a reason?
I wasn't aware it had been removed. I'll repost it as soon as I can find it.

[dfe1]

Images reposted

[mistermike]

I wasn't aware it had been removed. I'll repost it as soon as I can find it.
Thanks.

[N0DIH]

I am looking to buy EFILive soon, but don't have any experience with it, what are these tables here, like B3336 and such? What are the values? How do I view them?

[joecar]

Download/install the software from www.efilive.com/downloads (http://www.efilive.com/downloads)

Then from the tunetool, open a sample bin file (say 2001_Corvette_Man.tun or 2002_Camaro_Auto.tun for example)
located in the folder My Documents\EFILive\V7\bin.

Then in the tunetool's search dialog box (upper left, under "Navigator:"), type B3336 (uppercase B) and press Search...
this takes you directly to table B3336...

Each table shows:
- description of what the table does,
- links to related tables,
- each axis's units and max/min values.

Press F1 to bring up the tunetool user manual pdf for more info.

[ZL1Killa]

Images reposted
so these settings are working for ppl? I have been told (seen on here) from several that cutting DFCO off makes it worse...well from these tables several of the enablers are set to off..that doesn't entirely cut it off correct? just eliminates that enabler?

how many ppl have had succes stories from this? I'm interested in just stopping my car from going rich on decel...i don't want to try and CUT fuel, just get it away from going rich

[dfe1]

I've used the data (or something very close) in the tables I posted in a variety of vehicles with excellent results. The values are fairly aggressive, so you may want to soften the spark ramping a bit. There won't be any doubt when DFCO becomes active-- wide band AFR readings will go to 19:1 and you'll feel a change in the rate of deceleration. If you're having problems, I'd be happy to look at your file and make some suggestions.

Why don't you want to cut fuel?

[mistermike]

dfe1's settings work great on my GTO. There's an abrubt transition when fueling comes back on. I haven't played with ramp rates yet.

[dfe1]

dfe1's settings work great on my GTO. There's an abrubt transition when fueling comes back on. I haven't played with ramp rates yet.
Must have something to do with the blower. In the vehicles I've done the transitions in and out are undetectable except for the change in the exhaust sound and increased engine braking during decel. You may need to alter the MAP settings to account for differences in manifold pressure transition time. It sounds like you have the air and spark back well before fuel is back to normal.

Also-- an item of interest. In my LL8 (Trailblazer 6 cyl), I noticed that pulse width during DFCO is reported at .000 ms. Obviously, that's not correct, but it apparently has something to do with the P10 ECM's data stream because I haven't seen zeros with any other system.

[mistermike]

It's not as bad as I made it sound. I've got several issues going on, so the DFCO settings are way down on the list of things to pick on. Bigger motor, bigger cam, bigger heads and idle transition is pretty shakey. It turns out my wideband readings in EFI Live were way off, so I did some bandaid STFT tuning in the interim.

Blower is in bypass mode, so it's probably not that. I'm running catless now so you can hear everything. When fueling comes back on, there's a distinct "blorp" but nothing severe or unnerving, and certainly no worse than it was on the factory DFCO maps.

[ZL1Killa]

ok so just make the transitions not as bad... i will check that out and change them a little

[stigmundfreud]

sorry to be late to the party but here goes my questions/observations.

Car = 04 Holden Monaro/Pontiac GTO with capa Long tubes, miltek sport cats, miltek cat system

I've gotten my VE dialed in and now gotten my MAF mapped out to 1.00 +/- 0.01

Using the original Holden DFCO settings if I enabled dfco I could hear some unburnt fuel going into the cat and popping in there, rather than just getting the burble you get out of the exhaust exit. I quickly pulled over and turned it off for now but I have just applied the settings shown in this thread to my map (but kept the enabler temp high to stop it working yet).

My questions are this:

1.) Do the M6 cells refer to the manual six speed of which my car is? If so are the settings listed above for the Auto? For now I have kept the M6 settings as they were.

2.) I am running open loop + maf, both stft and ltft are disabled. Will this effect to the detrement of the car, the effect of DFCO?

I think that is it, basically will those settings be ok to try with a six speed modified manual? I'm eager to try it out but want to make sure it should be ok first. Also could someone confirm if the M6 does indeed refer to manual six speed cars - if so in the pasted settings why do they appear off?

thanks

edited to add my tune: http://www.stigmundfreud.f2s.com/tuning/olmaf.tun
also the log just in case: http://www.stigmundfreud.f2s.com/tuning/logs/olmaf.efi

[Tordne]

The DFCO M6 is in additional to the general DFCO controls which are still valid on an M6 vehicle. The DFCO M6 becomes active (assuming the enables are hit) when the clutch is pressed in between gear changes.

I used to get some pretty spectacular popping in the exhaust with the default DFCO settings. I still do if i really nail it to beat a red light ;)

In my case this was markedly improved by increasing the values in the B3334 & B3335, basically reducing the ramp in/out times.

[stigmundfreud]

well going to give dfe1s settings a try - it was only under larger decels where the unburnt would go into the cat - ie 2nd or 3rd gear and letting it go down from 4000rpm.

Will see tomorrow, tbh I cant see dfco really saving me much but that lovely lovely sound it makes ;)

[stigmundfreud]

dont suppose you could whack up your m6 enablers could you tordne? I'm going to run with dfe1's settings this afternoon and so long as I get no fuel popping in the actual cats I'll look at the m6 ones too

[mistermike]

The DFCO M6 is in additional to the general DFCO controls which are still valid on an M6 vehicle. The DFCO M6 becomes active (assuming the enables are hit) when the clutch is pressed in between gear changes.

I used to get some pretty spectacular popping in the exhaust with the default DFCO settings. I still do if i really nail it to beat a red light ;)

In my case this was markedly improved by increasing the values in the B3334 & B3335, basically reducing the ramp in/out times.
You need to post this up over at the LS1GTO site. Despite the rantings of myself and others, there is a persistent belief that the best way to eliminate popping on decel is to defeat DFCO altogether. Hmmmm. Have the decel mixture so bloody rich it couldn't possibly ignite by itself. Brilliant.

[emarkay]

Of course none of this is available on US based LS1 cars, correct?

[dfe1]

Based on my experiences, Tordne is absolutely correct, the M6 DFCO settings control fuel cut while shifting gears and the clutch is depressed. I use some fairly agressive settings with my 01 Corvette, so they are available on US-built vehicles. As for popping being caused by fuel entering the converters, I don't see how that can happen if you cut the fuel. I don't notice any popping when I back off the throttle. As for running the mixture so rich that it can't burn-- that's a good way to kill the converters.

[stigmundfreud]

well your settings seem to work nice but I might look on my logs to see if I could increase the map trigger points. Certainly no popping in the cats - on the default factory there was

[mistermike]

The GTO seems to produce a bit of "popping" on decel, mostly when long tube headers are installed. I can only assume this is due to the delay factors in the factory DFCO calibration allowing a certain amount of fuel to be "sucked out" of the injectors before DFCO engages. This mixture is naturally quite rich. Once DFCO kicks in, the mixture transitions to an A/F range where it can easily self ignite, hence the popping. I would guess this is occurring in the header tubes, as opposed to the cats. Completely defeating DFCO seems to eliminate this, so I can only assume the decel mixture remains too rich to self ignite in the headers, but rather is burned off in the cats, undoubtedly shortening their life and tossing fuel economy out the window.

[stigmundfreud]

well I suppose it is totally feasable to be in the headers rather than in the cats - since using dfe1s settings it doesnt appear to happen at all (perhaps once or twice) but at the same time its hard to tell if its kicking in! I think the ramp in ramp out rates have smoothed things a lot but the map pressure switch points I might play with - trying them back at stock settings but keeping the other settings dfe posted

[jamesbond2509]

Can someone look at my MAP settings for DFCO and let me know if it is optumized for a blower and 2bar MAP sensor. Please advise as I sm confused as to what the PSI is referenced to.

Thanks



10035

[limited cv8r]

Not sure what you are trying to achieve but the way you have b3308 to b3312 will not allow DFCO to activate.

[jamesbond2509]

I am trying to get DFCO to work effectively on my vehicle. B3308 to B3312 is for an M6. Any help will be nice. Thanks!

[phils01z]

When DFCO comes in on pulse width typically drops to somewhere between .200 ms and .350 ms (depending on the vehicle) and O2 voltage goes to .004 mv. With the stock cals I've looked at, the transition to DFCO is very slow because of a slow rate of spark adjustment. DFCO won't become active until ignition timing hits the value in B3336. (The description is not correct. This isn't the value to which timing will be adjusted when DFCO is active-- it's the value that activates DFCO. Spark will drop until it reaches the value in B5915.

Make sure that the value in B5919 isn't higher than the values in B3336 or you'll never hit the timing enabler for DFCO. Try values between .06 and .09 in B3334 to achieve a ramp in rate you're happy with. Values in B3335 (Ramp Out Rate) will be in the .4 to .7 range. Don't forget to adjust the other enablers (rpm, MAP, speed and throttle). If all else fails, send me your tun file.

dfe1,

I understand this is old, but if the descriptions of B3336 is incorrect, and we have data that proves that, why is it that the descriptions are never updated? I understand this is a daunting task to accomplish, but how does it ever get done? I am used to Ford tuning with BE and the biggest complaint I have about all of it, it is the lack of descriptions and clear documented logic diagrams of how calculations are actually made. I would think the entire community as a whole would benefit if there was a clear path. In the Ford market there is nothing. No clear route and if you want to search for it the task is nearly unachievable for a hobbyist.

The GM platform from the birds eye view is much better documented and defined, but I always question if this description is incorrect, what else is incorrect. I suppose that is the cornerstone of the argument to go standalone. Everything is known.

Philip

[GMPX]

dfe1,

I understand this is old, but if the descriptions of B3336 is incorrect, and we have data that proves that, why is it that the descriptions are never updated? I understand this is a daunting task to accomplish, but how does it ever get done? I am used to Ford tuning with BE and the biggest complaint I have about all of it, it is the lack of descriptions and clear documented logic diagrams of how calculations are actually made. I would think the entire community as a whole would benefit if there was a clear path. In the Ford market there is nothing. No clear route and if you want to search for it the task is nearly unachievable for a hobbyist.

The GM platform from the birds eye view is much better documented and defined, but I always question if this description is incorrect, what else is incorrect. I suppose that is the cornerstone of the argument to go standalone. Everything is known.

Philip
Philip, perhaps you have never taken a moment to think about how all aftermarket tuning packages for OEM ECM's are created. We aren't GM, or FORD, or Chrysler, we don't have access to the people that write the code for these computers, therefore there is a lot of assumptions made on how things work and sure there may be some parameters that don't behave as expected. But your options are really, live with these shortfalls and continue to use the OEM computer, install a Motec (or similar) where the company who designed then ECM openly provide info on how it works, take a .bin file from an OEM ECM and do a better job of figuring it out yourself.

The descriptions are updated as required (usually by people explaining to us things don't work as described), however, we don't regenerate all the .calz files for each new software release just for description changes, so in the case of the LS1 unfortunately there has been no need to regenerate the .calz files for years.

[phils01z]

GMPX,
I can clearly read you are the Admin of this site and I'm not sure if the goal is to look tough on this subject by attacking my lack of consideration. I started my paragraph by saying, "I understand this is a daunting task to accomplish, but how does it ever get done?". I would assume that would imply in just that instant alone I took the time to ponder how much work it would be. I thought about the scope of demanufacturing structured code written by someone else, who intends to not help you, and your only clues are likely the microcontroller processor calls themselves published by the processor manufacturer. You would look though seas of data looking for structured patterns based on the patterns in which they appear in other decompiled code. Then you would attempt to test that and see if you have hit paydirt. I also can appreciate considerable work goes into this to support the multitude of vehicles you offer BIN files to tune since you would have to do this process for each new processor you add to your list of supported models.

I made all those realizations based on the fact I am a programmer myself. Endless hours a week all over the world building research equipment. While we do nothing similar I am pretty sure I have a far deeper appreciation that most could. I am in no way trying to devalue your efforts by saying if my code stops lives are at risk. Its allot of work with relentless testing. We all have had 4am nights trying to figure out why a bit flips state and it shouldn't causing some sort of havoc. All I wanted to know is if there is a log of known information. DFE1 could be totally wrong, but maybe he is totally right. Only way I can find out is if I test his opinion, but maybe someone already did. Maybe those efforts are someplace and I don't know.. On that note I think my question is valid. Its a mountain of work but who keeps track?

Philip

[GMPX]

I made all those realizations based on the fact I am a programmer myself.
Then you also understand the 'legacy support' concept :hihi:
In reality (and this is by no means meant an insult to LS1 based tuner or cars/trucks), but these days it has to be considered an older PCM (legacy if you will), 1999 was when it came out. So relate this back to your computer programming job, how much time do R&D departments spend on Windows 98 enhancements or fixes vs Windows 7 or 8? EFILive is in the same boat, at this point unless there is a critical error that will brick the LS1 PCM there is probably not a lot going to be happening for it in the future. And I'm just saying this to try to put in to perspective why incorrect descriptions may slip by these days, nobody at EFILive is working on LS1 full time any more.

Unfortunately EFILive staff no longer have any LS1 based vehicles to test with either, we used to have three, but they've all been sold for later model vehicles (new R&D platforms I suppose). We can slot in to fix the descriptions, but, we can't verify the exact operation through vehicle testing. We'll see what we can figure out though based on what Dave said.

[phils01z]

" Then you also understand the 'legacy support' concept "

I do, but it the critical lab safety business I can't risk anything that runs on a windows machine. Everyone is an "armchair computer guy" and I like to give them as few buttons as possible ( else they will push them ). Everything I write is on the embedded level for that reason, I can't let people get thier hands in my honey. So, I'm a bit atypical. I have to support anything we have ever written almost indefinitely. Yes systems designed in 1986 still are expected to perform 10 years after EOL, just part of the GOV/MIL mindset.

If that is the end of the road for the LS1 type things I will have to learn to live with it I guess. I was secretly hoping that the GM stuff was going to liberate me from tons of research the Fords require. To be honest it is already far and away better in every way, but all of it still is based on the interpretation of the stock PCM.

Philip