Timing advance for better fuel mileage using Link V3

[92nolatist]

If this is in the wrong section, I appologize, I know its a topic beaten to death but also has to do with tuning. Move or delete as needed.

Hey everyone, Ive been fighting with my car to get good gas mileage for years now and have done some extensive research about using timing advance for better MPGs, but havent found any solid facts yet. Now, obviously I have the capabilities to adjust timing for part throttle cruising via the link DA timing table, but Im not sure about any ill effects it may have and if it's even a good idea. Or is there something else I could be looking at? Im only getting like 220~ to a tank (12 gallons used) and averaging around 19 to 21 mpg.

This is something I would really like to know about because it interests me alot, maybe even a few others too, not to mention I drive my car about 60 miles everyday and fillups happen twice a week (or more sometimes).

Thanks

 

[ENICKB]

21mpg sounds right. But think of the timing advance this way.. With the same tune (same ammount of fuel used) if you tune your car really "retarded" and it runs like poop, then you use say X ammount of fuel but get hardly knowhere. If you advance the time until the choking stops (back to stock values) youw ill get there quicker and still using X ammount of fuel under throttle. But if you start advancing the timing, you will get there even quicker, using the same X ammount of fuel under throttle. Dam man, what am I trying to say... I think advancing your timing WILL in fact get you the best possible gas mileage... But arent we supposed to max out our timing as DSM-ers right under the point of knock, as it is?? I think we should all be advancing our timing as much as safely possible for everybodys car! I myself am a huge fan of my timing maps on dsmlink

Yes I take into consideration that the engine will be spinning faster and using fuel faster, but you'll also be GOING much faster too. Think about it..

 

[goodbye]

21mpg sucks. I used to get 28mpg all the time on the highway with my original 20G, VPC setup. I didn't mess with timing though, it was all in leaning out the mixture at the cruising rpm.

 

[Imback]

^^ Ding Ding Ding

 

[kelvinb]

running 1450s/v3/wideband im right at 25 on the highway. my car doesnt have cruise control so its alittle harder to get it right. im never really in the throttle at all on my 50 mile drive, but ive got it between 14-16afr between 60-75mph. one thing ive noticed is that my bov is not recirculated, and between shifts it will get rich, and letting off the gas it will get rich enough to shoot fireballs. not very fuel economy friendly haha.

 

[TexasTurbo]

I'd recommend a pyrometer if you plan to tune to the ragged edge of good mpg's using timing advance.

I had my car running ~16AFR in cruise and 11.3-11.5AFR in boost. I was getting ~28mpg regularly and that was driving with a heavy foot every time i got in the car.

I'd be more concerned with the cruising at rich AFR's as the others above me mentioned than advancing the timing to get better fuel economy. The only way I can see timing advance improving your MPG is if all the fuel is not being burnt as a result of the timing not being advanced enough.

It is also important to consider that whatever fuel you use while in boost is a bit negligible because the percentage of time out of the total time that the motor is spent operating in boost. It's a small percentage for most everyone that is DD'ing their car so the "cruise" tune is a lot more at fault for bad mpg's than what's going on in boost.

Just some food for thought.

 

[92nolatist]

I've already leaned my cruise mixture out a little bit but could probly lean it out a tad more. I only ask about timing because doing 70 on the highway, which has a bunch of hills in my area, my timing is in the mid to high 20s to very low 30s. And I've read about people cruising with almost 40 degrees of timing, so that's why I inquired.

I also have my bov vented but am running a blow thru gm MAF, however, my AFRs go very rich if I let off the gas with cruising; I think that is the majority of my bad mpgs due to heavy traffic during rush hour. I can't for the life of me figure how to correct that rich issue with link. I have no boost leaks as well and my Evap canister is hooked up.

 

[92nolatist]

Alright, I have a problem. When I first put my engine in last year, I would get some knock at part throttle everytime I changed gears. Went into link and adjust my knock buttons to after 2200 rpm and after 40% throttle. Never had it again.

Tonight, I undid all that and I got 8* knock going into 5th gear ??? Is this phantom knock or do I have another issue on hand? Also while cruising, I get like .5* knock every so often.

This is a bother to me since Im trying to advance my timing while cruising to see if I net part throttle power gains along with some mpgs (just experimenting)

 

[snooopy365]

I actually tuned my car for good gas mileage.
But I'm using E85, so my values are not good to use with normal gas.

I'm running as much as 50° in the low load range.
On normal cruising I see between 35-50° BDC.
I also leaned it out to lambda 1.2 (17.5 AFR) while crusing.
In this combination you NEED more timing to burn the fuel, the lean mix does ignited slower than a normal mix, so running more timing is a good thing while running very lean.

The only thing that I'm not able to get is the fact that the car used less fuel with this settings and a stock (epromed) Ecu and more fuel while running the Evo8 Ecu.

Anyway with stock (evo8) timing and fuelmap the car got 16.8mpg, after disabling the o2 Sensor and fueltrims + setting it to run lambda 1.2 (checked via WBO2) and tuning the timingmap it gets 20mpg.
(with the old Ecu it was getting 21.4mph)

Just what your knocksensor while tuning and on gasoline I would not go leaner that 1.1 and not give more than 45° timing.

If you'r wondering why my mpg still sucks: I'm in Germany, traffic isn't that relaxed here. Also here you don't have this much high distance drives than in the US.
On long distance travels with speeds (mostly) limited to 75mph I get about 23.5mpg with E85.
With normal fuel that would be ~29-30mph.

Long distance trip (just cruising!) with speeds between 81-120mph I will get 18-20mpg

 

[92nolatist]

Thanks, that makes somewhat sense haha. So in order to get the best mpgs, I need to lean out the mixture in addition to adding more timing. Gotcha.

Now what about the knock issue?? Ive read of this happening to other people but there was never a definitive answer on how to combat that.

 

[snooopy365]

I'm not sure about the amount of knock you are getting. Never worked with DSM-Link before and in Evoscan I would read 0-70 units of knock.
Everything below 10 units isn't something to worry (as long as it isn't constantly doing that amount).
10-20 means there have to be something wrong. 20+ is: turn of the damn car and get it maintained

If you see light to middle know values at part throttle between 2300-3500rpm it is most defiantly not some real knock.
There is some sound of a dsm engines at this rpm that seems to trigger the knock sensor, it is just knocking until you create more or less load.
I had this kind of knocking, tried lowering AFR, nothing changed, used stock timings: still knocking.
After a few month if figured out that it is a specific load and rpm range.
I have a slight hill here, where I can reproduce the knocking every time.

I used some earplugs this a tube connected to the engine, to acctualy hear the engine, it wasn't knocking.
I just adjusted the sensor to ignore the sound (mostly because I had a flash on knock CE light active and it was driving me crazy....)

Changed the HG about a month ago and there is also no sign of knocking.

It is very hard to get any details on tuning for fuel efficiency, most dynoguys in germany just know how to make power.
My best guess is: you have to find the "sweetspot" leaning it out and adjusting the timing.
It should not knock even if you have (a little) to much timing, it will just get worse mpg (from my understanding!)

The problem might be the higher "low load" range.
If you are running gasoline and you set to much timing/to less fuel the combustion temperature can rise above any safety level.
Than you will get real knock, or real melted pistons.

That is one fact to love E85, it is nearly impossible to get the values so wrong that it will reach critical combustion temperatures in this load range.

So I your case I would guess it is phantom knock.

 

[92nolatist]

Wow, thank you so much. That is what Im expierencing at the moment. Only sometimes while going up a hill or going into the next gear will it knock; it will continue to do so until I let off the throttle. I went back into link and changed the settings so the ECU will ignore knock until 2200 and 40% throttle.

I tried leaning out my DA fuel tables more and I am getting no change in my AFRatio Estimate. It gets worse actually, plus airflow is WAY off at idle and I maxed out the airflow sliders and still nothing .
So I attempted to clean my gm maf and somehow broke one of the little wire things inside it. So now my car doesnt run anymore, Im pissed off, and Im thinking about going speed density.

 

[evileagleawd]

Another idea for better fuel milage is go with a taller 5th gear. The lower RPMs you turn traveling at a hight speed equals better fuel mileage. Could look into a Evo 5th gear set if you are realy trying to cram out every extra MPG you can!

 

[Calan]

I'm not sure about the amount of knock you are getting. Never worked with DSM-Link before and in Evoscan I would read 0-70 units of knock.
Everything below 10 units isn't something to worry (as long as it isn't constantly doing that amount).
10-20 means there have to be something wrong. 20+ is: turn of the damn car and get it maintained

ECMLink shows knock as degrees of timing pulled...not counts. So any consistent knock above 2-3 degrees is cause for some concern.

I tried leaning out my DA fuel tables more and I am getting no change in my AFRatio Estimate. It gets worse actually, plus airflow is WAY off at idle and I maxed out the airflow sliders and still nothing .

You are locking the ECU in open loop...right?

If you are still using a front O2 sensor (or simulated NB) and the car is entering closed loop like normal, changing the DA AFR tables in the low-load cruise and idle areas won't have any effect.

*****

The "proper" way to get the best power and mpg is to run what's known as a spark hook test on a load dyno. This basically involves sweeping the timing at a set load and RPM and measuring the output torque, under as many operational conditions as possible. The idea is to find MBT (Maximum Brake Torque) - the amount of timing advance that yields the most torque (or the point at which increases in torque are no longer linear with increases in timing) at each cell in the timing map. Then back off a degree or two for safety. If knock is encountered at any point, MBT goes out the window... avoiding knock trumps all else.

Maximum torque comes from a combination of AFR and timing. If the AFR changes, so will the amount of timing advance needed to provide maximum torque at the new AFR.

 

[92nolatist]

You are locking the ECU in open loop...right?

If you are still using a front O2 sensor (or simulated NB) and the car is entering closed loop like normal, changing the DA AFR tables in the low-load cruise and idle areas won't have any effect.

No it operates as it normally should. I didnt realize thats how that worked, thanks craig!

The testing for this is haulted right now because of broken MAF, so Im gonna take a time out on this but I will still be doing some research in the meantime. I think Im gonna go SD as well, because the maf has given me problems ever since I put it in back in 09...didnt come with the honeycombs, just a mesh screen.

Oh, and a taller 5th is in my future but thats going along with a built trans with evo3 gearset (need to save for that one)

 

[Calan]

No it operates as it normally should. I didnt realize thats how that worked, thanks craig!

Yep. As long as the car is in closed loop, the ECU is watching for a switching voltage from the front O2 sensor (or a simulated signal from a wideband). It uses this signal as feedback to maintain a stoich AFR of 14.7:1.

When the ECU drops into open loop, it looks up a target AFR from the DA tables based on load and RPM, and then injects whatever fuel it thinks is needed to hit that AFR, based on what it "knows" about your fuel system and MAF calibration.

Since a factory NB O2 sensor is designed to switch at 14.7:1 AFR, the only way to change the AFR in closed loop is by changing the NB voltage that the ECU switches at...via an offset simulated NB signal from the wideband or within ECMLink.

 

[perks563]

Semi-revival...

In you're opinion, which is better - locking the ecu in open loop or using an o2 sim to change the switching point while remaining in closed loop? I've been reading a lot lately about tuning for better mpg and these seem to be the two popular methods to do it, but not sure which is the better one to use.

 

[turboglenn]

ON a side note to add to Craigs imput (which it can get price to tune the entire spark map on a load bearing dyno) The simplest way and proper way if you don't access a dyno is to lean out a tad first and add a few degrees of timing, the sometimes after adding timing you can go back and lean it out a little more (and just go back and forth between the two without making too drastic of a change in either fuel or spark at one given adjustment before tweaking the other)

The change in timing will often effect the AFR's because of the change in time for the burn cycle to complete.. (but in DSMlink where you are actually tageting an AFR for fuel tuning i'mnot sure if your ECU will make changes on it's own, i'm usually tuning full stand alones where you have to enter the actual pulse width in milliseconds at each RPM vs. Manifold Pressure cell (which i have 512 cells available to cover the entire RPM and boost ranges of -14.7 to +32psi)

I have a few pointers outside of tuning as welll.. One thing i've found that's HUGE is the speed at which you cruise.. I recently rented a Kia Rio to drive to the east coast in and on the way there following my family that was in another car i got 42MPG and better at 60-65 MPH, on the way back i averaged 80-85 MPG and over the course of that drive home i droppedthe fuel economy down into the 30-32mpg range (even had it at 29 when i was doing a tad over 90mph through a large stretch of iowa) The car simply burns a lot more fuel at the higher speeds due to not only the higher RPM and more injections taking place per mile and also at a highe pulse width (more fuel used poer injection) but you start running into variables like wind resistance because there's a lot more fuel burned in fighting the wind to keep the speeds up

Most cars on the market are tuned from the factory to provide the best milage in the areas between 45 and 65mph (and most milage claims by manufacturers don't take into account the wind resistance or other variables of the sort) I would say that no matter the tune i'm running this is the single most important factor in determining my fuel economy on the interstate period!

Another thing is not leaving the line too fast which is obvious but also taking off too easily can burn a lot of fuel because it takes much more time to get up to speeds where you plane out and start only using enoughfuel to "maintane" the speeds you're at... (first learned of this when i bought my first datsun 280zx, it was in the owners manual )

ON e85 in the city with a mix of stop and go as well as interstate (beltway) driving in omaha, just using those two rules can mean the difference between me getting 12-14MPG or me getting 16-20mpg) Just get up to speed relatively quickly and keep your interstate driving speeds in the 60-65mph range.. (for acceleration i use the switch of from closed to open loop to determine how fast i take off.. I have the ECU (Full Standalone) set to keep me in closed loop anytime i'm under 32% throttle. (i have less adjustment on this than you DSMlink guys who can vary it by load and rpm at different points, mine is simply aboce or below a set TPS and RPM range which i also shut off closed loop t 4500RPM but if i'm above 4500 i'm usually wide open or at least 70% or more TPS.

 

[perks563]

That's great info Glenn, thanks for the post. What I'm not sure about is which is a better way to change the AFR. If you lock the ECU in open loop mode, then you would change the targets just like you would tuning a WOT pull. Another way is to use narrowband simulation and change the switch point voltage, so that instead of targeting 14.7, it could target 15.5 say. This way you would retain closed loop operation.

 

[96ETT]

To add to Glenn's great post, just remember, the equation for wind resistance has a velocity squared term in it. That means you double your speed and have 4 times the wind resistance.

Also the reason you don't want to take off too slow or too fast is because the engine has different efficiencies at different speeds and loads. So too slow and you may be at a low rpm, but also a low efficiency, too fast and you may be at a better efficiency point, but the higher rpm needs more fuel.

 

[johnovision27]

Very interesting thread. I've been trying to tackle this MPG issue on my DSM for a while now, ever since I did a major overhaul (2.3, 16G, 650s, etc) Used to get around 28-34 highway, now its closer to 22...
I think try leaning and advancing the timing bit by bit. I should be able to get pretty lean (16.5-17) at cruise as I'm currently running a 50/50 mix of 91 and 100LL (Avgas)

Great info from all parties.

 

[Calan]

Just as an FYI -

Something a bit unexpected happens in ECMLink when you change the NB switching point.

When the switching point is set to 14.7:1, CombinedFT and WBError will pretty much track each other...they both should be around 0% when everything is dialed in. But when you raise the switching point to something like 16.0:1, your wideband error and CombinedFT will start to deviate from each other. CombinedFT will stay around 0%, while WBError goes positive.

This is because of how the ECU is deriving the simulated signal, and the fact that it is internally hardcoded for an AFR of 14.7:1 while running in closed loop. When you raise the NB switching point, the ECU code still thinks it's switching around 14.7:1; you're just changing the sensor side of things. As ECMLink compensates for this with the fuel trims (and after some minor airflow adjustment), CombinedFT will fall back to 0%...but the wideband error is still comparing your wideband reading (which is now 16.0:1) to an internal value of 14.7:1, so it goes positive.

It's not a big deal...just something to be aware of.

 

[perks563]

Calan, that's great info, thanks. I did not know that, but is exactly why I am asking which method is better. What are the drawbacks to locking the ecu in open loop? Which method do you use?

 

[Calan]

What are the drawbacks to locking the ecu in open loop? Which method do you use?

A little of both, mostly because I'm always experimenting and trying to learn new stuff. If I am right on top of my tune (know exactly what fuel mix I have, etc), then I sometimes will lock the car in open loop to check the airflow calibration across the board, or maybe tweak the VE table a bit (I'm a perfectionist...it sucks. ). But I normally keep it in closed loop, so that the ECU can handle any unexpected things that come up like weird gas, a sudden vac leak, etc.

For typical daily driving, I don't raise the NB switching point much (if any) at all. I've found that mileage is so dependent on driving habits and speed (as Glen mentioned), that I just don't see many gains by raising the switching point... except on long trips with lots of 65-70mph highway cruising.

You'll probably get more gains by optimizing your timing for best torque at those cruise points, rather than trying to wring a few more HP out of a couple points of AFR.

YMMV of course.

 

[perks563]

except on long trips with lots of 65-70mph highway cruising.

Funny you say that, because the other-other way I was reading about was to define steady state parameters, like those you would encounter cruising on the highway, and when those specific conditions are met, the ecu would go into open loop. I think I like this idea best. I would keep all closed loop functionality like you said, but maybe get a slight gain in mpg.

I'm really just starting to wrap my head around Link and tuning in general and it is truly amazing what you can do!