LTFT on eManage Blue

[shnazzle]

Does it make sense to have LTFT readings consistently around -15 when running the eManage?
I'm assuming the eManage is upping fuel, which the ECU doesn't like in closed loop.
BUt...when it closed loop it;s not using emanage...so it makes no sense.

If not, I seem to be running very rich. STFTs are perfect. So not sure how I;m getting LTFTs of -14ish

[lamcote]

Does it make sense to have LTFT readings consistently around -15 when running the eManage?
I'm assuming the eManage is upping fuel, which the ECU doesn't like in closed loop.
BUt...when it closed loop it;s not using emanage...so it makes no sense.

If not, I seem to be running very rich. STFTs are perfect. So not sure how I;m getting LTFTs of -14ish

I think you need to re read the thread we had when you first talked about fitting this because I reckon what you are now describing is showing that the conclusion we reached back then was correct.

To recap:
Yes it makes sense to have LTFT readings at -15 if the eManage is putting more fuel in than the standard setting because that is the standard ECU trying to compensate for the  extra fuel the eManage map is providing. As you say, "the ECU doesn't like" the added fuel and is trying to take it away.

The eManage is being used in closed loop, it works all the time, what is important is whether the tuner adjusted any settings in closed loop. It looks like they did. This means that the standard ECU is seeing those changes in the closed loop and therefore assuming the engine is running rich everywhere (including open loop) so it then sets the LTFT to compensate in both closed and open loop.

I think the only way to avoid this is for the tuner to make sure they only change any fuel settings in open loop because then the standard ECU wouldn't be able to see these.

It is absolutely correct that the STFT is good while the LTFT is -15 because the only reason the LTFT exists is to ensure that the STFT returns to 0, that is why it is there. I would think if you reset the ECU you would see STFT running at -15 and LTFT running at 0 for a few days, then when the LTFT takes over, the STFT will go back to 0 and the LTFT will go to -15. The only time you would have the STFT and LTFT running changes at the same time is when the amount of change the LTFT needs to apply is so big that it is beyond its capability so it also needs to add STFT to make the change necessary, you really don't want to be in that situation!!

This does suggest that the eManage needs to be set up very carefully by someone who knows exactly how to avoid making any changes in closed loop, otherwise the standard ECU just brings everything back to standard over time.

As I say, re read the last thread as I am sure I will have missed something here.

[shnazzle]

.....
The eManage is being used in closed loop, it works all the time, what is important is whether the tuner adjusted any settings in closed loop. It looks like they did. This means that the standard ECU is seeing those changes in the closed loop and therefore assuming the engine is running rich everywhere (including open loop) so it then sets the LTFT to compensate in both closed and open loop.

I think the only way to avoid this is for the tuner to make sure they only change any fuel settings in open loop because then the standard ECU wouldn't be able to see these.
.

That's what I was after. I wasn't sure whether the the eManage even operated in closed loop. If it didn't, it would make it very odd that the ECU was compensating. But as you pointed out, it does work in closed loop, so it all makes sense  s:) s:)

It does indeed verify the previous thread. I remember what was discussed, but the key I missed was that eManage is active all the time and "working against" the ECU.

So it seems the map Noble did on this is full range, so the only way to get the best out of it is to reset the fuel trim table frequently.
WHich leads me to the next question; is there a way to reset the table without disconnecting the battery? If so, i'd be very tempted to start looking into writing an arduino application to reset the ECU every time I shut down the car.

[lamcote]

I reckon the trouble with all this is that the effect is the same whether the standard ECU is using the STFT or the LTFT to compensate, so even if you reset the ECU every time you drive it, the STFT will presumably still apply the changes within a few minutes of starting the engine. All you would be doing is avoiding the ECU ever changing the LTFT but that doesn't actually help because the STFT will still be undoing all your eManage mapping!

It always comes back to the fact that the remap must avoid changes in closed loop settings.

[shnazzle]

I can't imagine thats very easy as it spends most of its time in closed loop. THe odd thing is that I've had the emanage on for a while now without resetting the ecu, and it's still noticeably nippier.
So in open loop it's definitely using eManage. But what I've noticed as well (will log this as a separate parameter) the fueling status doesn't change when I expect it to. I've had TOrque running for a couple days now and it's interesting to see when the car goes into open loop. It's certainly not only over 4k rpm. Deceleration obviously as well, but I've seen it in open loop under 3.5k rpm.

I could do with knowing the exact parameters for determining when it goes open

[lamcote]

Yes, that's the problem, I've never seen any information on these parameters but I think it is a combination of revs and/or throttle position.

I am totally guessing but I would think open loop may kick in as follows:

Full throttle at any revs
Over c.4,500rpm at any throttle position
Over c.2,500rpm and more than 50% throttle
etc etc

[shnazzle]

Yes, that's the problem, I've never seen any information on these parameters but I think it is a combination of revs and/or throttle position.

I am totally guessing but I would think open loop may kick in as follows:

Full throttle at any revs
Over c.4,500rpm at any throttle position
Over c.2,500rpm and more than 50% throttle
etc etc

Indeed. Guessing is the game. Maybe even water temperature.

Another assumption here but I reckon Noble is pretty in-the-know. So I assume that they were able to somehow force the car into open loop in order to be able to map the closed loop areas of the map without the ECU constantly sticking its oar in. Otherwise whatever values they put in closed loop territory are totally dependent on the parameters that day. Humidity, temperature,pressure, etc etc

[lamcote]

Yes, good point, water temp must be one.

I think the remapping process just adds/subtracts a percentage value to the standard settings rather than specifying an absolute value. This means the standard ECU can still adjust for humidity/temp/pressure as it does normally and the remap then just adds/subtracts from that value, so this gets around that issue.

Remember, the standard ECU and the eManage work in series, the standard ECU sends its output to the eManage which then modifies the value (based on the remap) it receives and passes the adjusted value on to the engine.

I would agree that Noble should know what they are doing, BUT how else do you explain the LTFT you are seeing which is the reason why you have asked this question......?

[shnazzle]

And this is why I wanted to tap into the map to see what they've done  s:) s:)
So many unknowns. And as far as I can tell, the Boomslang harness I have does have the extra injector wiring, so they've clearly done all they can do.

[lamcote]

Seeing the remap would be interesting but unless you know the parameters when open loop kicks in, it may not help.

You sound like a bit of a tech wiz, couldn't you write something to keep a log of some relevant settings each time the engine goes into open loop?

[shnazzle]

Seeing the remap would be interesting but unless you know the parameters when open loop kicks in, it may not help.

You sound like a bit of a tech wiz, couldn't you write something to keep a log of some relevant settings each time the engine goes into open loop?

ooo tech wiz   s:) s:)   Praise indeed.
Not a bad shout. I will see if torque lets me log the fueling state, clear the ECU and drive around on as many loads and rpm ranges as I can

[lamcote]

Sorry, hope that wasn't patronising???

[lamcote]

I will see if torque lets me log the fueling state, clear the ECU and drive around on as many loads and rpm ranges as I can

If you can do this, please would you share your results?

[shnazzle]

Sorry, hope that wasn't patronising???

haha no not patronising at all! JUst having a laugh. I am decently techy to be fair...I was called a geek at work yesterday.

I will see if torque lets me log the fueling state, clear the ECU and drive around on as many loads and rpm ranges as I can

If you can do this, please would you share your results?

Absolutely. I might roll this all up in to the earlier post I had, as it's all around the same topic

[lamcote]

That would be a really useful reference because this topic does come up fairly regularly and no-one seems to have any data about these parameters.

[lamcote]

Just been out for a walk to clear the head and I've had a thought.

Do you know for sure that the LTFT you have shown in the original post is definitely being applied in open loop? It's possible that it may only be trimming the closed loop settings, this would explain the presence of that LTFT along with your impression that the car still feels faster. Is there any way of finding out whether there are separate trims for closed loop and open loop? Indeed there may be no trim for open loop at all.......

Others may know the answer to this?

[shnazzle]

One thing i did notice that in open loop, STFT figures go to 0.0, but LTFT figures still show

[lamcote]

What figures for STFT and LTFT show in closed loop? Are they different to open loop?

Edit: Hmm, thinking about it that probably shows that the situation may be as follows:

1. Closed loop creates a STFT which eventually gets transferred to the LTFT, therefore you can get a report of both STFT and LTFT in closed loop
2. Open loop doesn't have any STFT, that makes sense because there is no mechanism to create one, therefore STFT would report as 0 or provide no report at all
3. The LTFT from closed loop does get applied to both closed loop and open loop (but if crucially the STFT doesn't, see edit 2), therefore a LTFT would be reported in open loop, but there would be no STFT as in 2. above

That may be bad because it means the standard ECU is able to pretty much override the eManage, (subject to edit 2 below). I suppose you may still be getting some benefit from the remap under the following theoretical scenario:

Based on the dyno run to achieve maximum power the mapper increases fuelling by 10% on average in the closed loop settings and by 15% in open loop. The standard ECU notes that you have 10% "too much" fuel in closed loop. The standard ECU therefore reduces fuelling by a corresponding 10% but it applies this to both closed and open loop. This means that you are getting no benefit at all from the remap in closed loop but you are still getting 5% more fuel (ie 15%-10%) in open loop so the car is still faster, but you are not getting the 15% more fuel that you really need for maximum power.

Sorry if I sound like a broken record but.....it must be possible to work round this situation BUT only if the mapper knows (or can control) exactly when the engine switches to open loop. Therefore I can only conclude that ideally you should always ask any mapper if they know this information before agreeing to use them to tune a piggyback ECU.

Edit 2
N.B. However if the above scenario is correct and only the LTFT gets applied to open loop, then resetting the ECU regularly would actually help because; if it is only the LTFT that ever gets applied in open loop, and you can prevent the standard ECU from ever arriving at a LTFT by regular resets, the standard ECU won't then ever be able to apply any override in open loop. Hooray.... but for how long? Booooo. (Read on).

[shnazzle]

What figures for STFT and LTFT show in closed loop? Are they different to open loop?

Edit: Hmm, thinking about it that probably shows that the situation is as follows:

1. Closed loop creates a STFT which eventually gets transferred to the LTFT, therefore you can get a report of both STFT and LTFT in closed loop
2. Open loop doesn't have any STFT, that makes sense because there is no mechanism to create one, therefore STFT would report as 0 or provide no report at all
3. The LTFT from closed loop does get applied to both closed loop and open loop, therefore a LTFT would be reported in open loop, but there would be no STFT as in 2. above

That is bad because it means the standard ECU is able to completely override the eManage.

I've just mapped it against o2 voltage. Obviously 0 volt = open loop, >0 volt = closed loop.
No more clarity...
From what I can tell there are is one very clear trigger; deceleration. Letting the car coast in gear puts it into open loop. ANd obviously no STFT (0 reading) in periods when o2 is reading 0 volt.
LTFT however does not return to 0 or even return no reading. It returns an adjusted value, but it doesn't change in the period of time that the o2 sensors are returning 0 volt.

So, the ECU does apply trim even in open loop... but that makes no sense because then I would never be able to feel any change. The only thing I can imagine is that the ECU applies max trim, but the signal comes from eManage, so really you never see what it's actually fueling. WHat'd be interesting is if I could overlay the map from eManage onto a run.

[lamcote]

That is valuable information, see my last post which I have edited to reflect this. My edits have crossed with your last post.

I agree, you may not be able disentangle the different outputs of the standard ECU and the eManage. That's tricky.

[shnazzle]

So we've come full circle haha.

...wonder if I can build something to autoamtically clear the fuel trims on car stop  s:) s:)

[lamcote]

We have if my guesses about what might happen are correct, (and it will still mean you are getting no benefit in closed loop). Can anyone actually confirm or deny?

Could you run the car until a LTFT is reported and then do an ECU reset and see if it feels any faster after the reset, that may give some indication about what is going on?

Edit
Possible Options:
If it doesn't feel any faster then Noble have successfully avoided making changes in closed loop, but there is something else going on to create your reported LTFT
It is a bit faster which could mean you were getting the part benefit described in my pre edited post
It is a lot faster which could confirm the reset works as per my final edit

It would also be very useful to find out how quickly a LTFT is recreated after a reset, is it minutes, hours, days, as that will determine how often a reset is required and therefore how practical this solution would be?

Gosh this is fun.

Edit 2:

I've just done some general Google searches and it is quite possible that the LTFT is updated in a matter of seconds, (see quote below) in which case the reset strategy isn't going to be much good as the standard ECU will then be undoing the remap, even in open loop, within a couple of miles after each reset. Can anyone confirm this is a likely timescale on our cars? If so we have gone full circle but are back to square one!

"As a general rule, STFT can deviate as much as +/- 20%, but normally should be within +/- 10%. If it deviates by 10% or more for very long, LTFT will adjust the baseline to bring the STFT deviation back within the 10% range. This will then be reflected by a change (+/-) in the LTFT.

To illustrate: suppose the vehicle were operating at 0% correction in LTFT and +10% in STFT when an air leak occurred. STFT initially adjusts for this increase in oxygen by increasing fuel +20%. It does that in 0.6 millisecond. After 30 seconds, LTFT increases +10%. This brings the STFT down to +15%, which is still too much. So, after another 30 seconds the PCM moves LTFT +20%, and STFT drops to less than 10%.   "

[mrtoo]

i have heard that the LTFT is updated prety much straight away

[shnazzle]

Thread revival time!
So I've finally managed to get into my emanage and can answer another question;
Noble has no fuel adjustments for anything below 50% throttle. But adjustments at 50% and over for all rev ranges.
Ignition timing on the other hand has changes across the board.
This explains why when I first reset, I get a nice surge of extra power. But soon after, it fades on the low end. That;'s because all I've got left is a bit of airflow adjustment and ignition advance.

I know someone mentioned this on either this post or Lee's thread on his VTEC-controlled box of tricks, but it seems that since the ECU doesn't have a long term "learning" map for ignition timing, the EMB values are used regardless throughout the load/rev range and in open and closed loop.

Also, I have noticed that the front A.A.V. (airflow adjustments) dials on the front of the EMB have been set to 11 for up to 2000 rpm and 10 above that. I assume this is to compensate for the 57i K&N induction that Dave (robsonic) had on his car at the time of mapping. I should probably reduce these a bit as I have the stock MAF positioning and only a panel filter and markiii pipe. I'll dial these down to 8 and 9.

So here's my next challenge; mapping the fueling table for closed loop. As funny as that sounds.

I found a thread on newcelica.org about exactly this and it essentially says that if LTFT is negative for that load cell, I should remove fuel (can't do that on EMB) and if it's positive, to add fuel (this I can do!).
I wonder if EMB can reduce fueling at all. I cna't do this via the map. But what if I tell it I have 240cc injectors instead of 260cc. It should apply a correction of -7.69% across the maps. So if I then go into the current map and add 7.69 to everything, it also means I can reduce the fueling in the closed loop cells by adding any value less than 7.69 (not that it does 0.1 increments but work with me...).

Is that an area worth exploring or does the EMB simply not possess the power to reduce fueling in any way shape or form?

[SteveJ]

The simplest way to mess with the closed-loop fuelling is to get a wide-band O2 (such as the Innovate LC-2)with a simulated narrow band output. This can then be programmed to switch at different lambda's fooling the stock ECU into pulling (or adding) fuel using it's LTFT.

Couple this with the eManage playing with the MAF values being fed to the ECU and you should achieve a 0 LTFT