Won't Accelerate

[AVC]

Indeed they are easy to replace, and there are numerous aftermarket sources, but it seems most use the same Xinese "new" or "refurbs", which often have visible molding defects and I imagine not great QA on the plate film sensor and other bits. I tried one aftermarket sensor and it was worse than the dodgy OEM. The cheapest OEM I found was around $140. Really, what I wanted to do was switch the "loaner" car MAF with my CX-5 for an hour and evaluate. Of course I didn't have both vehicles simulatenously away from the dealer.

 

[erhayes]

I would first start by removing the neg battery cable for 15 minutes or so. This will put the computers back into the factory default settings. That may solve your issue. Ed

 

[minnesotaart]

Really, what I wanted to do was switch the "loaner" car MAF with my CX-5 for an hour and evaluate. Of course I didn't have both vehicles simulatenously away from the dealer.

Maybe you could meet up with someone on here and do the one hour swap, perhaps for a six-pack as payment! I think it's already been noted that someone on here goes to the same dealership as you do. Just a thought.

 

[CX5Nut]

Since new, I've always switched to Sport in our 2018 CX5 if I wanted reliable quick acceleration. Otherwise it would not be quick enough for fast entry onto the highway and would seem to hesitate for 1 second when the accelerator was pushed.

But could the OPs issues be from carbon build up?

 

[bonediggitty]

could the OPs issues be from carbon build up?

I originally thought it sounded like a TPS issue, but after AVC described his similar issue and findings it appears to be a MAF issue.

I don’t think it could be a carbon issue at the MAF, because there aren’t any vacuum lines that far out on the intake inlet tube.

 

[AVC]

As I had an "extra" MAF, I went about modifying it; enlarged the sensor intake port with a Dremel (chipmunk with dull teeth), by tapering the 4 edges back into a slight funnel shape. No beauty content but just to test a scenario. While it didn't change the idle MAF output noticeably, it does appear to make the initial rising output curve a bit more aggressive (see below). Probably messes with other parts of the curve in undesirable ways, but this wasn't intended as a final solution.

It made a significant improvement in the off-idle stumbling, "u-turn" performance and eliminated the low-speed high load surging.

 

[Buzzman12]

Is it possible for you to post a picture of your "modified" MAF?

 

[AVC]

Is it possible for you to post a picture of your "modified" MAF?

Edited prior post.

 

[AVC]

I'll post some plots captured in ForScan and processed in Excel, but the conclusion is that these SKyactiv engines are VERY fussy about the MAF condition. I believe this is because MAF has a designed non-linear response curve, and the engineers manipulate the intake and exhaust timely very heavily and therefore depend on super accurate air mass input for Skyactiv to provide good drivability, throttle response, transmission shift points and performance.

Because of the number of drivability complaints, without CEL codes, that people post, I have reason to believe a number of them may be due to MAF performance. Not sure whether the OEM supplier of these MAF struggles with quality control, or whether the MAF output drifts out of calibration with exposure to heat over time, contamination, or distortion to the hot film sensor with cleaning solvents. Anyway, my factory MAF was quite far out of tolerance by 30k miles. I found at least 3 aftermarket MAF sensors to be seriously out of calibration (sent them back), and a 4th aftermarket sensor to be closer to (assumed) nominal factory MAF.

I further noticed that the "deep waaaahh" sound (engine running in high valve overlap Atkinson cycle mode) that forum members have reported in light to moderate throttle increases when cruising and u-turns, which I've also experienced, is DIRECTLY related to the state of the MAF. Basically, the further the MAF departs from the "ideal" output curve, the more likely the PCM is to interpret light load conditions and hold the current gear and remain in Atkinson--thus the boggy, low power feeling and the waaah sound. Since fuel trims and O2 sensors are still within closed loop control ranges, the PCM never flags the condition.

 

[AVC]

For my research into MAF information and my CX-5's throttle response and performance anomalies, I determined that a common Wide Open Throttle test commonly used by tuners and technicians provided the most informative data. Specifically, in ForScan, plotting MAF output and total fuel trim over the RPM rise. Then dumping that as a .csv, and into Excel. It was immensely helpful having a very low mileage 2019 Mazda 3, 2.5LNA available to benchmark for normal/nominal MAF and fuel trim response, and then comparing that to the MAF my CX-5 came with, and a number of Ebay and Amazon $30 replacements, rather than popping for $170 Mazda replacement.

The Mazda 3 curve shows that target geometric curve, with slowly varying fuel trim command of 0 to +6%.

The MAF that came on my CX-5 is next. Notice that the curve is pretty linear, which might seem great, except I know it should be geometric curved! Notice -4 to +9 fuel trim swing, which seems trivial, but the PCM can only swing that 13 points in the span of a several seconds and doesn’t plateau until about 60g/s. The idle trims of -4 means it’s caught flat-footed with an idle launch stumble as the fuel trim is far off of target and the )2 sensor is screaming too lean!

The last plot is the best of the aftermarkets so far; An BAIXINDE unit, that advertises output curve match to OEM of <2% variance. Eh, other than people like me, who has the means or desire to verify that? . While the curve is not as geometric, it’s fits the target curve well enough to keep the fuel trims from idle to 140 g/s in a fairly tight +8 ~ +12 range. That brief trim spike to +16% is probably a data anomaly. Overall good idle launch and light throttle, decent medium throttle with a very slight surging, and good WOT throttle response. I could probably live with this sensor, vs popping for a new $170 Mazda OEM. Anyway, it was a good reason to experiment with ForScan and explore/solve the reason for my CX-5 drivability complaints that did NOT set any codes. [shrug]

Also: the AMAZON 1 sensor produces a SIGNIFICANT reduction of the number of times, the duration and intensity of the Atkinson cycle “waahh” sound.

 

[AVC]

An update: I sprung for an OEM MAF sensor mid-2021. Actually two OEM's: one was significantly better than the other, which confirmed my suspicion that QC on these particular MAF is not great, which might explain the various drivability "peppiness" comments from various members. I will note that the vendor opted to let me keep the 1st sensor at no cost--wasn't worth it to them to ship back.

Even with the better of the two sensors, it was still more sluggish that I cared for at low speeds. 89, or even better 91 octane fuel improves low speed behavior GREATLY. I've settled on 89 for the cooler 9 months in North Texas and 91 for the hotter months. Rather than be annoyed all the time, it's worth the $.30 or $.60 a gallon price difference, to me.

As an aside: as an DIY diagnostician, in 2020 I built my own in-cylinder pressure test jig + a digital scope I had already to verify that my 2.5L didn't have a base timing issue (jumped chain, dodgy crank or cam sensors, cam phasers). Everything measured spot on to the service info I got off of AllData.

 

[181962]

An update: I sprung for an OEM MAF sensor mid-2021. Actually two OEM's: one was significantly better than the other, which confirmed my suspicion that QC on these particular MAF is not great, which might explain the various drivability "peppiness" comments from various members. I will note that the vendor opted to let me keep the 1st sensor at no cost--wasn't worth it to them to ship back.

Even with the better of the two sensors, it was still more sluggish that I cared for at low speeds. 89, or even better 91 octane fuel improves low speed behavior GREATLY. I've settled on 89 for the cooler 9 months in North Texas and 91 for the hotter months. Rather than be annoyed all the time, it's worth the $.30 or $.60 a gallon price difference, to me.

As an aside: as an DIY diagnostician, in 2020 I built my own in-cylinder pressure test jig + a digital scope I had already to verify that my 2.5L didn't have a base timing issue (jumped chain, dodgy crank or cam sensors, cam phasers). Everything measured spot on to the service info I got off of AllData.

That's most impressive bit of diagnostics I've seen in these pages. Dumb question from a non-DIYer since you still have issues--did you also clean the throttle body?

 

[AVC]

Yes