Car Analogy

DIY for installing replacement APP sensor in DBW S2000 (will likely cure intermittent dbw hesitation issue)

Tools needed:

10mm socket and extension
12mm open end wrench
Flat blade screwdriver
T20 Torx bit socket
Cutoff wheel or dremel
OBD scanner

Replacement sensor. I got mine from Amazon seller NewYall, $25.


Remove black throttle assembly cover. Under hood, driver side (left hand drive) fender, just aft of shock tower. The cover the throttle cable goes to.



Unplug electrical connector from sensor.

Remove throttle cable. Loosen throttle cable lock nut, enough to remve cable from bracket. Manually advance throttle so you can remove cable end from throttle rotating assembly.

Remove throttle assembly from car, by removing 2 10mm bolts.



Now the fun part. Removing old APP sensor. Its secured by two screws, but there is no head. They just look like rivets.

Use cutoff wheel or dremel to carefully cut a groove in each head, so you can use screwdriver to loosen screws. The heat from grinding will help break them loose, so loosen each one as soon as done grinding groove. Make a decently deep groove. It'll just make it more difficult if you don't go low enough then strip the groove if it can't handle torque needed.

Once the screws are out, old part just comes off. There is a copper gasket, reuse it.

Now install new sensor. You'll see there is a slot in rotating throttle assembly, and a corresponding tab on sensor. Sensor tab needs to fit in that groove. Make sure it does.

Now secure the new sensor using the provided torx screws. Tighten just enough so you can still rotate sensor in its slotted mounting holes, but tight enough so it stays where you set it.

Sensor alignment:

Plug the electrical connector back into sensor. Note, if you turn the key on with sensor unplugged, it will throw a code. Two codes actually, codes are for both APP sensor circuits low. If you accidentally turn on ignition with sensor unplugged, just clear the codes and continue.

With sensor loosly tightened and plugged in, rotate sensor as far clockwise as ot will go. Throttle assembly is not bolted in place yet. Only sensor is loosely bolted in place and plugged in.



Turn on ignition, but DO NOT START ENGINE. Press pedal ALL the way to the floor. Make sure no code is thrown.

Now we're going to find the spot where sensor alignment just throws a code, then back off a hair.

Rotate sensor counter clockwise by as small an increment as you can. Press throttle to floor again. Make sure no code. Repeat this process until it throws a code (code will be P2123, APP circuit D too high).

Now back the sensor down clockwise by as small an increment as you can. Tighten into place, being careful it doesn't rotate at all.

Clear code. Then try throttle to floor again, make sure it does not throw code.

Reassemble everything, go for a drive. Make sure to floor it during drive in something higher than first gear, to make sure no code is thrown (it often wouldn't throw the code in first when misaligned). Keep floored well up into vtec. Make sure its FULLY floored.

Drink a beer, you just saved about $200 (how much more it would cost to buy a whole new assembly with sensor already installed and aligned).

My car had noticeably better throttle response and dbw hesitation has yet to reoccur (its so intermittent can't tell yet if its cured for good).

Zer0ne

Replaced my whole assembly when mine gave me issues last year. I still have the old one and will be doing this. Thanks for the write up!

Car Analogy

An observation about the replacement sensor. I measured resistance of oem sensor and replacement sensor, over the full sweep of its range, from fully closed to fully open.

The new sensor has a larger range. While there are two circuits, one which sweeps low to high, other high to low (presumably as a safety redundancy, to make sure both always jive, to reduce potential for unintended acceleration), I'll just talk about low to high to help explain the point.

The oem sensor low to high circuit sweeps from about .05 ohms to about 342 ohms. While replacement sweeps from .07-420.

So for each degree increase of physical pedal movement, the new sensor will read a comparatively higher resistance change compared to oem sensor.

That means with the new sensor in place, engine response feels snappier. Its more responsive to the slightest touch of pedal.

It also means its more difficult to drive smoothly in parking lots and stop and go traffic. I think my custom Flashpro throttle table is making it worse. I'd modified it to make throttle more responsive. More linear, more like direct cable throttle. I think the new sensor does much of that all on its own. Combined, the new sensor and modified map are making it too sensitive sometimes. But with a stock throttle map, its probably going to be great right out of the box. Even if you have a Gernby tune, you still probably have a stock throttle table, as he doesn't usually modify that.

You may be wondering if the new sensor is more responsive, what happens at wot? Won't the throttle body be at wot even before the pedal is at wot? That is true, but guess what, its also true with oem sensor and oem throttle table.

In the stock throttle table, 100% wot throttle body is always achieved prior to 100% pedal to floor, and how much pedal is required to get wot at tb varies with factors like rpm.

So now with this sensor it will just get to 100% a little sooner. Like I said before, I modified the stock table to make things more linear, which included something closer to wot only at larger pedal positions.

Its important to understand in all this that wcu doesn't read pedal % from 0%-100%. On the live data display on my odb scanner, reading pedal percentage, with new sensor in place idle shows as 13.2%, and if I floor it, it goes to 79.2%. So in tbe throttle table, tb has to be wot by 79.2% pedal or I'll never be fully wot.

This lack of full % range seems to be linked to the fact that the APP sensor can't reach its full sweep range once installed into throttle assembly. The full range of throttle assembly is less than what the sensor can achieve.

Thus no matter where you place sensor in its slotted mounting holes, it can never hit its lowest low or highest high within its potential range. This seems deliberate in the design, and makes sense.

Thus the full range of readable pedal percentage is always something less than 0-100.

It seems confusing how the sensor wouldn't need absolutely precise calibration to work. Calibrated so its idle resistance perfectly matches what ecu expects idle to be. I was able to install both stock sensor and replacement all along theor potential range, and with one exception car still idled at normal rpm. My theory is that whatever resistance ecu reads from APP sensor at startup, within a certain acceptable range, is considered the new APP sensor value for idle. That way, replace sensor and car still idles. Move sensor calibration (within reason, within that acceptable range) and car still idles.

The only reason replacement sensor even needs calibration is because it has a faster rate of change per degree of pedal motion. That means its capable of exceeding the resistance value ecu thinks is too high no matter what. No matter what the current idle calibration value is. Outside the acceptable range for wot.

So with this new more sensitive sensor, with a broader resistance range, calibration is a delicate balance of finding a position that won't ever exceed acceptable range resistance at wot, yet is still within the acceptable range of initial value for idle when sensor is at rest.

The calibration process described in first post is one way to achieve that balance.

If you buy an aftermarket APP sensor, understand it will make the throttle more responsive, and require a more delicate touch in parking lots and stop and go. And require calibration.

shind3

^I tuned my DBW tables to emulate a more linear torque response instead. This is because even at half throttle at low to mid RPM, I was getting full engine torque leaving the latter half of the pedal being useless. Engine torque is proportional to MAP. Mine isn't perfectly 1:1 pedal vs torque but a significant shift towards that. I find it easier to control the car during a slide on the track like this. YMMV.

KONcept

I purchased the Neyall sensor and have it installed. I find it almost impossible to get it not to throw a code. Either too low, or too high when the pedal is fully depressed. Is there any other modification needed? perhaps adjusting the throttle cable?

Thank you.

Car Analogy

Nope, just the finesse of Goldilocks position.

Since finding the spot, mine has not thrown another code. I drive it often enough.

So set and forget, once you get past challenge of setting it.

KONcept

Do you remember when you installed yours that setting the sensor fully clockwise would still throw a "input too low error"?
Is it necessary to unhook the battery to fully clear or re learn?

Should I just calibrate for the high end and let the car have a low input CEL? Will it relearn a new lower range?

Car Analogy

Shouldn't have to reset ecu to clear previous codes.

Definitely don't live with cel, too low or too high.

When I set it too far one way, got low cel, other direction high cel. Exact Goldilocks position, no cel.

Maybe try ohmeter measure bare sensor (not installed into assembly, since it limits range of motion), at full off and full on. Lets compare to readings mine had when new. Maybe there is quality control variance, and yours is just enough different than mine that there is no Goldilocks.

KONcept

So with 6 pins, where do i put the negative and positive leads? Lest say with the sensor wheel facing up, and the pins facing you?

I didn't get any readings 0.07-420 ohms. Either i have a cheap reader or i was doing something wrong. I think i had readings in kohm or something.

I got a 2nd sensor from neyall today, same thing happend, i've got to be doing something wrong. I tried on the OEM sensor just for kicks, calibrated in like 2 tries, so much easier. I felt this is insane i'm getting both high and low errors at the same time...

Is the goldilocks spot essentially 1 perfect spot in the range? The first post made it sound like there is a bit of wiggle room as they are trying to calibrate it as close to the top as possible. How small exactly are the adjustments? seems almost less than a mm of change going from high error to low error.

Thanks for helping.

Car Analogy

This weekend I'll get my meter and get some pics of what wires to measure, etc.

The Goldilocks position is a narrow window. I had some wiggle room, but not much. Yes, its sub mm adjustment.

Since mine is working, my measured high and low resistance as installed should be within acceptable range for ecu.

Occurs to me another way to calibrate is measure resistance as you adjust, and set so low isn't lower than mine, high not higher.