Closed Loop Fuel Enricher

What do you guys think about this product?

http://www.splitsec.com/products/Enricher.htm

Seems like open loop fuel systems were originally designed around the limitations of the narrow band O2 sensors. The computers only had feedback from a 1.0 lambda sensor and 1.0 does not work during high loads and high RPM ranges so open loop programming was a necessity.

With the accuracy of the current wideband sensors why not stay in closed loop all the time and let the ECU make realtime adjustments to achieve the specified AFR for the specified conditions. I believe some new vehicles are doing this.

The problem they're solving does not exist with 05-09 LGTs or 02-12 turbo Imprezas. Also note that they're really aiming this product at customers who are putting turbos on non-turbo cars:

When an aftermarket turbo or supercharger is fitted to an engine, it is desirable to run at a much richer fuel mixture when the engine is in boost. Efforts to add fuel in boost are quickly overridden by the fuel trim that occurs in closed-loop. The Enricher alters the target fuel mixture in boost so that the stock ECU targets a user-adjustable richer mixture.

That said, staying in closed loop at all times might make them a bit more resilient to unexpected conditions (like boost leaks or weird fuel), so the "stay in closed loop" idea isn't necessarily a bad one. It's something I've been looking into, and the biggest problem (that I know of, so far) is that the ECU likes to vary the target AFR up and down by a couple tenths, a couple times per second. This is good for emissions in cruise, but probably not helpful at WOT. I might have stumbled on a solution to that a few weeks ago, but I haven't really dug into it yet.

The problem they're solving does not exist with 05-09 LGTs or 02-12 turbo Imprezas. Also note that they're really aiming this product at customers who are putting turbos on non-turbo cars:

 

We have enough control over our ECUs to make them run whatever AFR we want.

 

 

 

That said, staying in closed loop at all times might make them a bit more resilient to unexpected conditions (like boost leaks or weird fuel), so the "stay in closed loop" idea isn't necessarily a bad one. It's something I've been looking into, and the biggest problem (that I know of, so far) is that the ECU likes to vary the target AFR up and down by a couple tenths, a couple times per second. This is good for emissions in cruise, but probably not helpful at WOT.

Do our ECU's have the ability to target multiple AFRs while in closed loop? I was not aware of that. My thoughts were that the enricher could give the ECU a normal 14.7 target, and also a WOT AFR target. It seems that the Enricher allows the transition to the richer target to be user specified using boost and one other signal.

I might have stumbled on a solution to that a few weeks ago, but I haven't really dug into it yet.

Very interesting. If you don't mind, please share what you have discovered.

One of the problems with turbo cars is that O2 sensors do not read accurately under increased pressure. The stock location for the front O2 sensor in the header will not give a useful reading while in boost. Even a WBO2 sensor in that location cannot be used for fueling under boost.

True. Easy fix to move the sensor though.

True. Easy fix to move the sensor though.

Move it where?

Move it where?

To the downpipe, upstream of the cat. For example:

http://i193.photobucket.com/albums/z151/Legacy_NSFW/Exhaust/StereoWidebandBellmouth.jpg

It reads within a couple percent of my PLX wideband, but it only reads down to about 11.1 or 11.2:1. I'm sure it would work fine for closed-loop at about 11.5, maybe slightly richer.

Do our ECU's have the ability to target multiple AFRs while in closed loop? I was not aware of that. My thoughts were that the enricher could give the ECU a normal 14.7 target, and also a WOT AFR target. It seems that the Enricher allows the transition to the richer target to be user specified using boost and one other signal.

 

Very interesting. If you don't mind, please share what you have discovered.

Yes, there's a table in the ECU that adjusts the closed-loop target, based on RPM and load. By default it just adds a little bit of extra fuel when RPM and load increase beyond the 'cruise' area. I suspect that you could shape it just like the open loop fuel table, and tweak a few other tables to prevent the ECU from ever switching out of closed loop, but I haven't actually tried that.

I've been reverse-engineering the software in the ECU, and I think I may have found the stuff that causes the rich/lean oscillations during closed-loop cruise. But I won't know for sure until I start modifying the tables it uses, and I'm not ready to take that step yet. Figuring this stuff out is tedious work, and I have a few other things I want to work on first.

The table and figures are known as adaptive numerators, there are two different types. Short term and long term.

Short term fueling can adjust to a much greater scale/curve than a long term can.. but typically only makes the correction for a short period of time. If the ecu sees the need to hold short term fueling at the extreme of one end or the other, it starts adjusting the long term accordingly. Im not sure how many base values/ 'boxes' Subaru uses for standard close loop fueling... but a typical factory map across all auto makers uses 16-28 ( depending on engine necessities and of course, emissions.) The more valued the ecu can handle and tweak, the better the car will run.

Should also add.. each value/box corresponds to a certain rpm and engine load.

Someone who has recently had a repair (say to an o2 sensor) and is having stumbling problems at 3500 rpm up a hill in third gear should look into doing a battery disconnect or resetting the adaptive.

In response to the op; id fear that with keeping the car in close loop all the time.. you may not allow your engine to really perform the way it has to in certain situations. Open loop has no set fueling decisions for the simple reason that everytime you punch the throttle or start your car cold the conditions are different.

Unless you can account for all those situations with an extensive closed loop program.. I don't see how it could be beneficial on a daily driver.

To the downpipe, upstream of the cat. For example:

Don't the usual WBO2 kit vendors recommend more distance between the turbo flange and sensor than that?

Don't the usual WBO2 kit vendors recommend more distance between the turbo flange and sensor than that?

Yeah, but since Subaru puts the turbo three feet from the cylinder head, I don't think the usual guidance applies.

The table and figures are known as adaptive numerators, there are two different types. Short term and long term.

 

Short term fueling can adjust to a much greater scale/curve than a long term can.. but typically only makes the correction for a short period of time. If the ecu sees the need to hold short term fueling at the extreme of one end or the other, it starts adjusting the long term accordingly. Im not sure how many base values/ 'boxes' Subaru uses for standard close loop fueling... but a typical factory map across all auto makers uses 16-28 ( depending on engine necessities and of course, emissions.) The more valued the ecu can handle and tweak, the better the car will run.

I've attached a picture of the table I'm talking about. This is how the table looks in a stock ROM.

Short-term / long-term refers to the corrections that the ECU makes to achieve the target AFR. This table determines what the target AFR should be in the first place. If you get the rest of the tuning right, the short-term and long-term trims should stay within a couple percent of zero, just as with regular closed-loop cruise.

Yeah, but since Subaru puts the turbo three feet from the cylinder head, I don't think the usual guidance applies.

... except the 5th gen, and that's the direction Subaru will likely go with future turbo models.

Well, I have seen some Spec C downpipes that had an O2 sensor in that location. Not sure if it was was used for primary fueling, as I don't remember if there was another in the manifold.

... except the 5th gen, and that's the direction Subaru will likely go with future turbo models.

 

Well, I have seen some Spec C downpipes that had an O2 sensor in that location. Not sure if it was was used for primary fueling, as I don't remember if there was another in the manifold.

I guess 5th-genners might want to move it further downstream.

I'm told that the twin-scroll Imprezas have the factory O2 sensor in the downpipe from the factory, but I don't know about the TS LGTs. I don't know if the ones with the sensors in the downpipe stay in CL all the time or not, but I would be surprised if that was the case.

I guess 5th-genners might want to move it further downstream.

Might be tough if they want to run a cat...

II'm told that the twin-scroll Imprezas have the factory O2 sensor in the downpipe from the factory, but I don't know about the TS LGTs. I don't know if the ones with the sensors in the downpipe stay in CL all the time or not, but I would be surprised if that was the case.

The Spec C certainly has a WBO2 in the bellmouth, but I also very much doubt if they always run in CL. A look at a Spec C ROM will tell you, if it's been defined...

I've attached a picture of the table I'm talking about. This is how the table looks in a stock ROM.

 

Short-term / long-term refers to the corrections that the ECU makes to achieve the target AFR. This table determines what the target AFR should be in the first place. If you get the rest of the tuning right, the short-term and long-term trims should stay within a couple percent of zero, just as with regular closed-loop cruise.

That's exactly what im talking about.

To the downpipe, upstream of the cat. For example:

 

 

 

 

It reads within a couple percent of my PLX wideband, but it only reads down to about 11.1 or 11.2:1. I'm sure it would work fine for closed-loop at about 11.5, maybe slightly richer.

Obviously you leave your factory sensor in the bellmouth all the time. Do you also leave your wideband in the bellmouth constantly? Have you had any reliability issues with either sensor?

I have read that the sensor may wear out quickly when it is close to the turbo. I don't understand why, because the temperatures should be much cooler after the turbo as opposed to the stock location. I am considering moving my stock O2 sensor to the bellmouth.

Our stock sensor normally lives in the exhaust manifold, at the bottom of the up-pipe, so moving downstream means less heat and less pressure, so I don't expect any trouble.

The other sensor came with the PLX SM-AFR, I think it's the same Bosch used in most aftermarket widebands. It's been in the position pictured for two years as of tomorrow, and it works just fine. Like I said, most wideband instructions suggest mounting further downstream of the turbo. But they also suggest about a foot between the exhaust ports and the sensor in a non-turbo car. Most wideband instruction authors are assuming that the turbo is just inches from the exhaust ports. Putting those two things together, I think it won't be a problem to put the sensor in the bellmouth when the turbo is a couple feet away from the exhaust ports.

Our stock sensor normally lives in the exhaust manifold, at the bottom of the up-pipe, so moving downstream means less heat and less pressure, so I don't expect any trouble.

 

The other sensor came with the PLX SM-AFR, I think it's the same Bosch used in most aftermarket widebands. It's been in the position pictured for two years as of tomorrow, and it works just fine.

In your opinion, would you find your stock O2 sensor located in the bellmouth adequate for tuning the car or do you believe that an aftermarket wideband is necessary for tuning. I am just talking about for logging and tuning. I realize the benefits of a separate AFR gauge for monitoring during daily driving.

In your opinion, would you find your stock O2 sensor located in the bellmouth adequate for tuning the car or do you believe that an aftermarket wideband is necessary for tuning. I am just talking about for logging and tuning. I realize the benefits of a separate AFR gauge for monitoring during daily driving.

Actually, this discussion has me thinking about moving the stock O2 sensor to the bellmouth.

It's not a true wideband in that IIRC it only has 4 wires, and the scaling isn't linear like you'd expect from a true WBO2. I would also guess that since Subaru ignore it's readings in OL, it either wouldn't be well scaled past perhaps 13.5 or 14:1 AFR, or is deliberately scaled to read rich for safety.

Just thinking out loud, would it make any sense to relocate it, and scale it against a reference WBO2. Would be a very stealthy and clean mod since the lead for the stocker will certainly reach back there if some cable clips are released, just don't know if it would be of any use for tuning...

EDIT: I see from some old RR discussion that NSFW has his stock O2 sensor in the bellmouth. I was wondering if you were running two widebands in there for some reason. How is this working out now, and are you running either or flattened load correction or revisions in tables that create the AFR 'hunting' behavior in CL?

With the stock scaling, and relocating to the downpipe, the stock sensor reads within about 3% of my PLX. I'd need a third sensor to know for sure which one is more accurate. It just doesn't read lower than about 11.2.

I think you could tune with it. I'm going to try that approach on a mild stage-2 setup in the next couple/few weeks. Stock TMIC, catted DP, stock O2 sensor in the DP, no other mods.

I'm not sure what you mean by load correction.

At idle, my AFR is steady. In cruise, it oscillates up and down a little bit, but that's because the ECU asks it to - you can log the CL target and watch that bounce up and down, so that's by design. You can also log AF Correction #3 (not #1), it's related.

I'm not aware of any ill effects from moving it to the downpipe. I get a little bit of stumble around 2000 RPM, but that still happens if I force the ECU to stay out of closed loop, so it's not related to moving the sensor.