I just talked to a couple people about tuning their cars today, and it seems a lot of people think an S-AFC is a great part to buy for your car. I'm not sure how common this knowledge is, but apparantly it's not common enough, so I'd like to share a couple of my views.

First off, some very basic information about how your ECU controls fuel. I'm mainly talking about Subarus, but most every fuel injected car works on the same principles. In closed loop fueling (low throttle), your ECU will constantly monitor the oxygen sensor, and adjust fueling to maintain a stoichometric AFR. Basically, the MAF signal doesn't really matter -- if you're leaner than 14.7, the ECU will add fuel, regardless of the MAF. So the fuel computer won't affect it. In open loop fueling, rather than adjusting to hit a specific AFR, your ECU uses look up tables to determine the amount of fuel to use. On Subarus, these tables' axiis are RPM and engine load. Engine load is the mass of the air per cylinder charge, or, more simply, the MAF. See the table below.

http://www.enginuity.org/files/jojofuelnew_130.jpg
Edit: That's load (g/sec) on x axis, RPM on y

All piggy back fueling computers (S-AFC, e-manage...) basically work the same -- they receive input from your mass air flow and throttle position sensor, and spit an altered MAF voltage back out to the ECU. What does this do? It tricks the ECU not just in to using a different cell from the lookup table (it does not directly add more fuel).

Now, you're saying, "big deal? it works, right?" Yes, it does -- kind of. The problem is that the fuel table isn't the only table relying on the MAF sensor. I can't recall all of them off the top of my head, but the most important is the timing table. See below.

http://www.enginuity.org/files/enginuity20a_5_903.jpg

When you use a piggy back fuel computer to add fuel from the ECU's base map, you're also usually decreasing timing advance, and affecting who knows how many other tables (there are over 100) in unpredictable ways.

That's really the major problem with piggy backs. I've seen people mention things about the ECU "learning" around them, but I'm haven't seen any evidence of that. The other drawbacks are pretty straight forward and hopefully obvious. I don't think there are as many people using fuel computers on Subarus as other cars, but I wanted to make sure. ;) Good luck.

good deal. that is why openecu is a good thing ;) along with enginuity :)

Good points, all true.

On turbo Honda's it's even worse because you put in larger injectors and scale them back, which advances the timing to crazy numbers.

They only really work with older DSM's and cars that have a vacuum advance, or where the the MAP signal has no bearing on timing.

The map sensor in subarus is used only for fuel cut. Timing is based off rpm and load sent by the mass air sensor.

The map sensor in subarus is used only for fuel cut. Timing is based off rpm and load sent by the mass air sensor.

Right. I should have typed MAS instead of MAP. I'm not used to typing MAS, :D

ok, studied A LITTLE of how cars are 'tuned', if Emanage/Ultimate can't adequately address ALL the different affected/important fuel/ignition maps, what's the solution for those seeking 1. knowledge & 2.dependable-safe amateur tuning?
guaranteed, I'll READ any suggested threads... thanks, Justin

The timing issue can be addressed with piggy backs that have timing control/adjustment.

However I agree, piggybacks are not a perfect solution. I've run into that problem with the first gen legacy. There's pretty much nothing out there other then piggy backs or stand alones.

I never understood why people would want to use piggyback solutions on a vehicle with a flashable ECU. On a vehicle without that capability... sure, it may be the only solution, but I'm still amazed on how people want to implement these even on new-age Subarus. Oh well... to each his own I guess.

I never understood why people would want to use piggyback solutions on a vehicle with a flashable ECU. On a vehicle without that capability... sure, it may be the only solution, but I'm still amazed on how people want to implement these even on new-age Subarus. Oh well... to each his own I guess.Because, until recently, the only way to reflash your car was to see an EcuTeK tuner/dealer or run an off-the-shelf Cobb AP map.

Because, until recently, the only way to reflash your car was to see an EcuTeK tuner/dealer or run an off-the-shelf Cobb AP map.

Until recently? If you're refering to the OpenECU stuff, ok, but EcuTek's been arround since the early WRX days 3-4 years ago for the US market. The way I look at it... if your ecu has the ability to be tuned by reflashing the ECU then that's the best EM solution you need to implement on your car. You don't see VW people buying isht fuel controllers do you? The only people that should be buying them are people without flashable ECUs from the factory... because it's the only route they have with the exception of replacing the ECU with a standalone of some kind. On the Civic I ran every FMU/Fuel computer combo known to man to try and get it to run respectably and finally gave up on using the OEM ECU. I switched to bigger injectors and a PowerFC and haven't looked back since. When I got the WRX the first thing I did in terms of mods was reflash the ECU. It's almost a given...

Because, until recently, the only way to reflash your car was to see an EcuTeK tuner/dealer or run an off-the-shelf Cobb AP map.Until recently? If you're refering to the OpenECU stuff, ok, but EcuTek's been arround since the early WRX days 3-4 years ago for the US market.Didn't I just say that? ;)

Maybe it was an echo... :lol: What I meant to say is that US guys have always had the ability to run flashable EM solutions... they are just sometimes too freaking cheap to do the job correctly.

Yes, I think that's a better way to put it. At least now the cheapos can run OpenECU stuff and do the job right... the first time.

Maybe it was an echo... :lol: What I meant to say is that US guys have always had the ability to run flashable EM solutions... they are just sometimes too freaking cheap to do the job correctly.I'll admit... I personally *liked* the idea of running something that I could tune myself (e.g. the UTEC on my old '03 WRX) versus a reflash, because I was changing my car so much. If I had to go back to an EcuTeK tuner every time I changed something, I would have been broke. :lol:

Eventually, I did settle on an EcuTeK reflash that I tweaked myself (great relationship with local shop). ;)

While your description about the fueling change has merit, you statement about the ECU not learning is clearly inconclusive.

The largest, and most difficult factor in using an offset based piggy back device on a modern subaru is the dynamic self learning Timing system. The fuel and timing offset issues are at least constrained and linear. The dynamic system is something that is very hard to tune for, and just plain impossible if you do not have a datalogging solution to see what the stock ECU is doing.

Having removed many a piggy back from broken cars, the dynamic behavior has almost always been the source of the problem.

I wrote a bit about the fueling algorythms several years back here:

http://forums.nasioc.com/forums/showthread.php?t=773994&highlight=fuel+table

http://forums.nasioc.com/forums/showthread.php?t=468040&highlight=piggyback


Cheers,

Jeff Sponaugle
PDXTuning.com

While your description about the fueling change has merit, you statement about the ECU not learning is clearly inconclusive.

The largest, and most difficult factor in using an offset based piggy back device on a modern subaru is the dynamic self learning Timing system. The fuel and timing offset issues are at least constrained and linear. The dynamic system is something that is very hard to tune for, and just plain impossible if you do not have a datalogging solution to see what the stock ECU is doing.

Having removed many a piggy back from broken cars, the dynamic behavior has almost always been the source of the problem.

I wrote a bit about the fueling algorythms several years back here:

http://forums.nasioc.com/forums/showthread.php?t=773994&highlight=fuel+table

http://forums.nasioc.com/forums/showthread.php?t=468040&highlight=piggyback


Cheers,

Jeff Sponaugle
PDXTuning.com

Nice read Jeff.

It's too bad a lot of the software/hardware that's compatible with the newer ECU's are not backwards with the OBD1 ECU's. You're really limited...

While your description about the fueling change has merit, you statement about the ECU not learning is clearly inconclusive.

The largest, and most difficult factor in using an offset based piggy back device on a modern subaru is the dynamic self learning Timing system. The fuel and timing offset issues are at least constrained and linear. The dynamic system is something that is very hard to tune for, and just plain impossible if you do not have a datalogging solution to see what the stock ECU is doing.

Having removed many a piggy back from broken cars, the dynamic behavior has almost always been the source of the problem.

I wrote a bit about the fueling algorythms several years back here:

http://forums.nasioc.com/forums/showthread.php?t=773994&highlight=fuel+table

http://forums.nasioc.com/forums/showthread.php?t=468040&highlight=piggyback


Cheers,

Jeff Sponaugle
PDXTuning.com

Jeff,

I read through your threads above. However, you only mentioned OL Fueling in UTEC. Do you know how non-OL UTEC works and what it affects? I'm curious what's happening in the backgroud to timing since there's no MAF table to reference. In non-OL is it more like an SAFC like qoncept mentioned?

I'm currently running mine in UTEC non-OL fueling over a reflash that I had to spend a considerable amount of time on the MAF table. This was the only way that I could run the UTEC. Here's a ref to how my MAF table looks now http://forums.nasioc.com/forums/showthread.php?t=1024855

Thanks,

Todd

Jeff,

I read through your threads above. However, you only mentioned OL Fueling in UTEC. Do you know how non-OL UTEC works and what it affects? I'm curious what's happening in the backgroud to timing since there's no MAF table to reference. In non-OL is it more like an SAFC like qoncept mentioned?

I'm currently running mine in UTEC non-OL fueling over a reflash that I had to spend a considerable amount of time on the MAF table. This was the only way that I could run the UTEC. Here's a ref to how my MAF table looks now http://forums.nasioc.com/forums/showthread.php?t=1024855

Thanks,

Todd
the UTEC whether running in OLF (open loop fueling) or CLF (closed loop fueling) mode always has OL ignition advance (timing).

the reason the UTEC was so nice on '02-03 WRX's was because the ignition advance was a static map (predictable) and the fueling (which was only CLF at the time) was easy because the values in the map just scaled the MAF readings to adjust the amount of fuel delivered.

the initial reason turboXS had to implement OLF was because of the issues that arose with the release of the '04+ stock ECUs. for emissions reasons (if i recall), these ECUs had a delay when switching from CL to OL fueling. this was extremely dangerous on modified cars because you could be caught in between a rock and a hard spot when your boost goes up but youre still running stoich AFRs for two (2!) seconds. so in the end, turboXS ended up enabling OLF on the UTECs which allowed the UTEC to drive the injectors itself...which means it didnt have to rely on the stock ECUs MAF values when switched over into OL mode.

so that was a long winded response to a relatively simple question-

in closed loop fueling mode, a UTEC is still far superior than an SAFC since it uses its own ignition advance table (not an offset of the sometimes seemingly unpredictable stock ignition advance), yet still allows a good control over the fuel.

HTH

the UTEC whether running in OLF (open loop fueling) or CLF (closed loop fueling) mode always has OL ignition advance (timing).

the reason the UTEC was so nice on '02-03 WRX's was because the ignition advance was a static map (predictable) and the fueling (which was only CLF at the time) was easy because the values in the map just scaled the MAF readings to adjust the amount of fuel delivered.

the initial reason turboXS had to implement OLF was because of the issues that arose with the release of the '04+ stock ECUs. for emissions reasons (if i recall), these ECUs had a delay when switching from CL to OL fueling. this was extremely dangerous on modified cars because you could be caught in between a rock and a hard spot when your boost goes up but youre still running stoich AFRs for two (2!) seconds. so in the end, turboXS ended up enabling OLF on the UTECs which allowed the UTEC to drive the injectors itself...which means it didnt have to rely on the stock ECUs MAF values when switched over into OL mode.

so that was a long winded response to a relatively simple question-

in closed loop fueling mode, a UTEC is still far superior than an SAFC since it uses its own ignition advance table (not an offset of the sometimes seemingly unpredictable stock ignition advance), yet still allows a good control over the fuel.

HTH

Thanks for the long winded response ;) but maybe I should have clarified a little better. In UTEC, anytime that you're in 0% load (no manifold pressure) then the UTEC is using stock ECU timing unless you decided to use fixed valves which probably no one would ever do (it's possible though). So with that said, what's the UTEC doing in non-OL fueling with the timing that the ECU is running? It would seem to me that the ECU, while in control of timing still suffers from the same fate that qoncept posted about "why piggybacks suck" if you've got that column modified. Since I've been tuning my ECU, I have some of my lower load columns where the ECU runs where most people have fixed UTEC numbers. Hopefully that made since :lol:

Jeff,

I read through your threads above. However, you only mentioned OL Fueling in UTEC. Do you know how non-OL UTEC works and what it affects? I'm curious what's happening in the backgroud to timing since there's no MAF table to reference. In non-OL is it more like an SAFC like qoncept mentioned?


do a search, noob!

http://forums.nasioc.com/forums/showpost.php?p=14142297&postcount=24

:)

do a search, noob!

http://forums.nasioc.com/forums/showpost.php?p=14142297&postcount=24

:)


Ken, Stop bluffing ;) :lol:

Thanks for the link.

Thanks for the long winded response ;) but maybe I should have clarified a little better. In UTEC, anytime that you're in 0% load (no manifold pressure) then the UTEC is using stock ECU timing unless you decided to use fixed valves which probably no one would ever do (it's possible though). So with that said, what's the UTEC doing in non-OL fueling with the timing that the ECU is running? It would seem to me that the ECU, while in control of timing still suffers from the same fate that qoncept posted about "why piggybacks suck" if you've got that column modified. Since I've been tuning my ECU, I have some of my lower load columns where the ECU runs where most people have fixed UTEC numbers. Hopefully that made since :lol:
the danger is still there, but it is easily controlled where you need it (just punch in the values)

ie- i could leave my whole ignition advance map to 'ECU' but that would defeat the purpose of using the UTEC, really. on my UTEC, i have 'ECU' in the 0,10, and 20 columns. when i am in that area of the map, i am at the mercy of the stock ecu, but i did it only to get rid of the hesitation. in CLF mode, youll want to run UTEC ignition advance all the way from the 10 column for optimal results, IMHO...especially if you run larger than stock injectors or a larger MAF for example.

. on my UTEC, i have 'ECU' in the 0,10, and 20 columns. when i am in that area of the map, i am at the mercy of the stock ecu, but i did it only to get rid of the hesitation. .

happ, i think that's a good idea regardless.

i only put utec timing #s where i either want to

1) run a static value for stability
2) make sure the utec is in control so that it can pull if knock occurs

... otherwise the oem ecu is better at running things. i have ecu timing in the entirety of the 0 column, up to 5500 in the 10, and up to 3500 in the 20.