Stop calling the stock front O2 sensor a narrowband...

[NSFW]

Stock sensor moved to bellmouth, vs PLX SM-AFR (Bosch LSU 4.2?) in bellmouth:

* Registered users of the site do not see these ads.

[Turn in Concepts]

Originally Posted by gabedude

I tuned a v8 twinscroll that had the stock 02 post turbo. It was a few years ago, so it would be interesting to compare my logs of my WB vs the stock one.

I was thinking the same thing. The JDM twinscroll setup puts the o2 sensor there I would assuem due to the isolation of exhaust gasses from the twinscroll headers.

Tony

[gabedude]

Originally Posted by Turn in Concepts

I was thinking the same thing. The JDM twinscroll setup puts the o2 sensor there I would assuem due to the isolation of exhaust gasses from the twinscroll headers.

Tony

I didn't notice anything different in the mapping as far as O/L C/L fueling goes. You might think that Subaru might have tried CL fueling into boost.

[NSFW]

Originally Posted by gabedude

You might think that Subaru might have tried CL fueling into boost.

They do that in 07 STIs don't they?

There's a CL enrichment table that seems like it ought to allow staying in closed loop at AFRs well below 14.7, has anyone messed with it yet? For my car it showed up in Merchgod's last set of definitions. (Others may have gotten it sooner, others might not have it.)

[69subaru360]

The stock sensor, moved to the downpipe is accurate. I tuned both of my 02 WRXs that way. I had an aftermarket wideband, it read comparably to the stock sensor moved post turbo. The aftermarket one broke, I never got another one, no need. On the 02 wrx at least, you can rescale it to read down to 10.5 to 1 if you want to.

[gabedude]

Originally Posted by NSFW

They do that in 07 STIs don't they?

Just 3/4 of a second into high loads is all at 14.7:1.

Originally Posted by NSFW

There's a CL enrichment table that seems like it ought to allow staying in closed loop at AFRs well below 14.7, has anyone messed with it yet? For my car it showed up in Merchgod's last set of definitions. (Others may have gotten it sooner, others might not have it.)

I have that table. I would limit the AFL and AFC corrections first though. You don't want a big AFR swing at high loads.

[van_dutch]

Originally Posted by NSFW

Stock sensor moved to bellmouth, vs PLX SM-AFR (Bosch LSU 4.2?) in bellmouth:

I cut out your images

Couple questions. First, what is the equation for the regression you did on the data? What is the R value? How did you handle the x intercept? Ideally this would be zero. I don't know how the sensors will react in that area of the graph, but it will have an affect on the regression.

[NSFW]

Beats me. I just logged both, made a scatter plot, and used Excel's polynomial fit thing. I played with the poly order until it looked about right. That's 4th-order.

Why do you ask?

[van_dutch]

Originally Posted by NSFW

Beats me. I just logged both, made a scatter plot, and used Excel's polynomial fit thing. I played with the poly order until it looked about right. That's 4th-order.

Why do you ask?

I'm an engineer and when I see data plotted without that kind of stuff, it rings bells. Playing around until it looks right doesn't really cut it. I'm just curious about the comparison. Without stuff like the equation and the R (really R^2) value, a comparison like that is pretty useless. You can draw numerous shapes that look right. Without an R^2 value, you have no way of judging how well a given curve fits. The equation is telling for a comparison of how they line up.

[Frank A]

Originally Posted by van_dutch

I'm an engineer and when I see data plotted without that kind of stuff, it rings bells. Playing around until it looks right doesn't really cut it. I'm just curious about the comparison. Without stuff like the equation and the R (really R^2) value, a comparison like that is pretty useless. You can draw numerous shapes that look right. Without an R^2 value, you have no way of judging how well a given curve fits. The equation is telling for a comparison of how they line up.

The comparison isn't useless, it's just statistically not sound. From eyeballing the graph (as an engineer, you still remember how to eyeball something? ), I would say the stock sensor tends to underreport the AFR by 0.2 to 0.6 units. This is pretty important of course in tuning, but I would think that after having seen that graph, you could just absorb a 0.4 unit offset in your tuning and get about the same level of precision in the tune as with using other sensors.

[arghx7]

The problem here is that you assume the aftermarket wideband must be the "true" value simply because it's aftermarket and because it's advertised as being more accurate. You really need to use a completely different instrument (5 gas analyzer) that isn't subjected to the same limitations of the normal wideband design.

FYI I've since been tuning with the factory wideband relocated post turbo. Seems to work well enough, and I will acquiesce on the location issue. I do trust it more when it is post turbo.

[arthurtoo]

Originally Posted by arghx7

The problem here is that you assume the aftermarket wideband must be the "true" value simply because it's aftermarket and because it's advertised as being more accurate. You really need to use a completely different instrument (5 gas analyzer) that isn't subjected to the same limitations of the normal wideband design.

FYI I've since been tuning with the factory wideband relocated post turbo. Seems to work well enough, and I will acquiesce on the location issue. I do trust it more when it is post turbo.

Hi there,

I have a JDM blob eye twinscroll. by default the a/f sensor is post turbo at the bellmouth of the downpipe. do i need to recalibrate the sensor voltage table before i use it to tune?

thanks buddy!