NGK AFX vs. AEM 30-0300 CAN Comparison

[schpenxel]

Indeed. You can use inj PW, inj flow rate and a WB reading to basically create a VE or "GMVE" table and/or MAF tables. Basically if you know how much fuel you're using and know what lambda reading you're getting out.. then you can calculate how much air must be going through the engine. That + some info on # of cylinders, RPM, etc. and you can calculate VE.

Gen 4's have cylinder volume and some sort of gas constant also factored into their VE values (i.e. "GMVE"), but that's obviously simple enough to factor in.

I've tried doing it in the past but results weren't so perfect. I never spent the time to figure out why, though it very well could have been from WB errors (or me screwing up the math, which is at least equally as likely)

[GHuggins]

The gauge is , now, an official part from AEM. Part# 30-0333 is a 30-0300 w/ the firmware update and a pre-wired splitter cable. So, you should be able to call AEM, or one of their distributors, to get hold of one.

AFAIK, they have not done the non-gauge "inline" WB, yet. I think they are waiting to see how popular the gauges are.
If it is going to be a while, I MIGHT offer them, with the splitter cables, until AEM has them ready.



That is a problem with the Bosch sensors. And, more so with the 4.9 than the older 4.2. They require VERY tight control of the sensor heater to maintain a constant temperature under changing EGTs. Several controller brands do not do a very good job of this. The AEM gauge actually uses a DC power-supply to run the heater, rather than the normal switched 12v PWM signal. That may help some. I have not heard of any coming back for this.

So no sensor problems reported so far then? That's good to know...... I've killed bosch after bosch sensor - several thousand spent on them - main reason I went to the very reliable ECM product....

I've got a flathead V8 I'm fixing for a customer in between tuning jobs that in essence is using 4 carbs (dual 97's) - really wish I had the ability to tune it with individual widebands instead of just getting bank to bank the same and then having to read the plugs for cylinder to cylinder... That would make life SO much easier - of course I also just found out that this thing has a "prototyped" European carb base plate setup that won't work with multiple carbs - explained all the problems I was having (would stay at wot fueling air fuel ratio after doing a pull until you shut it off or adjusted the carbs due to some primary, secondary circuitry scavenging screw up) - at least it can and will be fixed with new base plates

[10_SS]

Little do some know. Using an AEM 30-0310 the following was recorded on an LS-2 with a broken header bolt. Slight exhaust leak, sucking air on scavenge.

Attachment 59094

Because the AEM wideband has an actual latency of around 10ms or less, it was able to track the single cylinder lean events ( ~40ms ). With the oscillations synchronized to RPM/2.

Latency makes a huge difference. if your t63 response time is around 100ms; then, after 100ms, your reading is still 37% off. No need to go looking for that 4th digit.

A lot happens in an engine in 100ms.

I believe can see low RPM, single bank cylinder combustion events with the AEM 30-0300, since it tends to smooth out as RPM increases (IN WOT PE mode).

[10_SS]

Holy f'ing cr@p Mike! Here is a quote from the OP's very first post in this thread...




With that last statement you just posted, you clearly succeeded in answering the OP's original question, which was the whole point of this thread. You brought us all full circle back to the beginning. And you did it with facts and in a rational, adult, logical and respectful manner. Even when questioned and doubted, you kept posting more information that backed up your opinion. An opinion that was NOT based on emotion or brand bias, but based on your vast experience and knowledge in the actual real world of lab testing and validation of the exact product(s) being discussed. This is obviously your job, your expertise and why the companies that actually produce these products pay you for this data.

For me, the OP's question has been answered definitively. In HIS particular situation, he should trust the AEM. Especially if he's using histogram data only, which I'm betting 90% of "tuners" do. Like you've said, copy and pasting bad data just makes bad tunes. That's not his fault, again, I'm guessing 90% of the guys out there don't realize this can be a problem. I truly believe "tuning" IS an art, and I'm so amazed and happy that you said that. It was a form of validation for me, and I'm sure many others out there that truly understand how this stuff really works. You said it perfectly, if all you have to do is copy histogram data and paste it into your VE or MAF tables, anyone could tune. BINGO! You've just explained why a "tuner" can build a tune, and someone else can look at it and how the engine is actually running and shake his head at what a "hot mess" it is. (OP's exact words btw.)

I'm happy and grateful that you spent the time to clear things up. I'm disappointed that you HAD to spend the time to clear things up. lol. This should have been a discussion and not an emotion based dogpile. Emotion has no place in the work that we all do professionally. If you can't be objective and unbiased, you can't diagnose or test effectively. I have no doubt that the reasons you're good at what you do are those exact traits. I've seen literally hundreds of techs come and go over my 33 years of being a lead dealership tech/foreman/manager. A VERY small percentage can be objective and unbiased and keep emotion out of the data acquisition and final assessment. Thank you VERY much for being unbiased and objective, and pulling it back into an excellent discussion! MUCH respect from this old guy here!

this was me seeing the difference, however I think it's common knowledge that Wideband readings can be delayed and is why your supposed to run WOT fueling in a tall gear, like 4th or higher so RPM climbs slower and your not creating a bad Histogram/chart, so this is why I bought the AEM 30-0300 so help speed things up, and with the Dr. Mike CAN update it's just awesome since it was always hard matching the analog outputs to the actual signal... but the NGK would go low then high at power on, so made things easier... but still not perfect.

So anyway, Now I made 1 correction after a few lower RPM WOT pulls using the AEM 30-0300 and first thought WOW, one correction and my error is so small now... which is seems to still be, however I'm getting some inconsistency in the Error at the first WOT pull, at least that's what it seems.. check it out. I compared injector pulse vs. RPM at the point where the AFR leans out between WOT run 1 (leans out at 1600rpm) and WOT run 2 (looks great), and they appear to be identical.. ---UPDATED - added fuel pressure to one of the pics... it is rock solid----... but what could this be? BTW, yes, I know, the Narrow Bands dip on the leanout and not something just the AEM 30-0300 is doing, so yes it appears to be real... (I figured gmtech16450yz would like that example)
Camaro.Layout.xml
125e 1st WOT LEAN.hpl
125e 2nd WOT OK.hpl
Attachment 59154
125e 2nd WOT OK.PNG

[dr.mike]

So no sensor problems reported so far then? That's good to know...... I've killed bosch after bosch sensor - several thousand spent on them - main reason I went to the very reliable ECM product....

That's the trade-off with Bosch vs. NTK sensors. The NTK ceramic sensor elements have a much larger thermal mass and are more solidly coupled to the metal sensor housing than the Bosch elements. The little holes in the metal shielding that the exhaust gas flows through are smaller as well. This makes the ceramic element heat up and cool down more slowly,due to EGT swings, as compared to the Bosch elements.

The sensors all have built in heater elements that are supposed to keep the ceramic element at a constant temperature ( usually, around 800C ). But, to do that, the controller needs to constantly measure the element temperature and adjust the power applied to the heater.

The old Bosch LSU4.2 sensors required the heater power to be updated twice per second, to maintain proper operating temperature. While the new LSU4.9 sensors require the heater control to update at a rate of 100 times per second. This is why controllers that were originally designed for the LSU4.2 don't always do so well, when modified to run the LSU4.9 . New generation controllers that are designed, from the ground up, for the LSU4.9 should do better.

If you don't keep the sensor element at a near constant temperature, the ceramic will, eventually, fracture.

By being bigger/heavier and less exposed to the exhaust gas, the NTK sensor is more durable; especially when the controller's temperature regulation is not so great.

But, the NTK sensor pays a price for being big/heavy and having reduced exhaust gas exposure. And, that is the much slower response times, when AFR changes rapidly.

[LSxpwrdZ]

I've been spoiled with a loaded dyno. The response greatly helps transients and quickly changing conditions like you would see street tuning but I have the ability to slow everything down on the dyno (load) and really get a good population of data. When you have a 1200+hp car and the pull still lasts 8-10secs rather than something like a DJ where it takes 2-3secs you can gather some pretty good data!

[dr.mike]

I've been spoiled with a loaded dyno.

Yep. a load-controlled dyno helps a lot in slowing things down enough for the wb to catch up.

But, you still cant tune the actual transient events, like fuel-cutoff recovery or power-shifts.

BTW, yes, I know, the Narrow Bands dip on the leanout and not something just the AEM 30-0300 is doing, so yes it appears to be real... (I figured gmtech16450yz would like that example)

Yeah. when I 1st started logging with the 30-0300 I saw a lot of things that turned out to be "real" that were, 1st, assumed to be anomalies. That trace of the header leak was one of them
1st I thought "hey! this thing is unstable! and it's oscillating". But, then I noticed the RPM/"oscillation" correlation. Sure enough... broken header bolt on #2.

Same with some funky looking traces, when recovering from fuel cut-off.

[gmtech16450yz]

this was me seeing the difference, however I think it's common knowledge that Wideband readings can be delayed and is why your supposed to run WOT fueling in a tall gear, like 4th or higher so RPM climbs slower and your not creating a bad Histogram/chart, so this is why I bought the AEM 30-0300 so help speed things up, and with the Dr. Mike CAN update it's just awesome since it was always hard matching the analog outputs to the actual signal... but the NGK would go low then high at power on, so made things easier... but still not perfect.

So anyway, Now I made 1 correction after a few lower RPM WOT pulls using the AEM 30-0300 and first thought WOW, one correction and my error is so small now... which is seems to still be, however I'm getting some inconsistency in the Error at the first WOT pull, at least that's what it seems.. check it out. I compared injector pulse vs. RPM at the point where the AFR leans out between WOT run 1 (leans out at 1600rpm) and WOT run 2 (looks great), and they appear to be identical.. ---UPDATED - added fuel pressure to one of the pics... it is rock solid----... but what could this be? BTW, yes, I know, the Narrow Bands dip on the leanout and not something just the AEM 30-0300 is doing, so yes it appears to be real... (I figured gmtech16450yz would like that example)
Camaro.Layout.xml
125e 1st WOT LEAN.hpl
125e 2nd WOT OK.hpl
Attachment 59154
125e 2nd WOT OK.PNG

Holy cr@p I love looking at logs!!!!!! lol. Yeah I know I'm weird.

10_SS this is SO COOL! So wait, first of all, are you trying to say that the narrowband reading changing along with the wideband reading verified that the wideband data was correct? WTF ARE YOU TRYING TO IMPLY??? That narrowbands can be used as a "sanity check" to widebands? Now you're just crazy talking!!!!!

Sorry, I couldn't resist. haha.

Ok so I don't know exactly why you're getting that lean spot, but I can tell you where to look. This is a PERFECT example of logging AND charting the right stuff. I opened your log in my layout and immediately saw what very well could be the cause. Then I opened the same log in your layout, and guess what? I didn't see the problem! Good thing you were logging the data though, all you'll have to do is add it to your chart. Look at this screenshot...





What I saw was that voltage drop right at the lean event. You didn't have ign voltage in your chart so you couldn't see it. What didn't look right was the fact that it was in the middle of the event. Then I looked at the polling speed for that ign voltage, it looks like it's updating at 1 second intervals. So that .2v drop doesn't look like much does it? But that's in the MIDDLE of the drop, it's an average of what that voltage did over a full 1 second period. Increase your polling speed for ignition voltage as fast as it will go. I'd like to know what that voltage was doing during that whole event. The voltage updated right before the start of the lean event, so it literally missed a full second of the voltage drop!

The next thing I would want to see IF the voltage drop is what's causing the lean event is coolant fan operation. It looks like the fan might have passed through the 48-52% area of desired fan speed. Where are your fan settings? And does the ZL1 have relay controlled fans like the V's? If so, did the fan just go from low speed to high and caused a voltage drop large enough to effect the injector drives? I'm guessing here because there's a bunch of info still missing, but that's where I'd start looking.

[GHuggins]

Yep. a load-controlled dyno helps a lot in slowing things down enough for the wb to catch up.

But, you still cant tune the actual transient events, like fuel-cutoff recovery or power-shifts.




Yeah. when I 1st started logging with the 30-0300 I saw a lot of things that turned out to be "real" that were, 1st, assumed to be anomalies. That trace of the header leak was one of them
1st I thought "hey! this thing is unstable! and it's oscillating". But, then I noticed the RPM/"oscillation" correlation. Sure enough... broken header bolt on #2.

Same with some funky looking traces, when recovering from fuel cut-off.

I'm curious Mike, what are the delay times for the lambda pro unit? Only reason I ask is because I do power shift and transient fuel tuning with it on the dyno all the time. Of course I'm also using load profiles built into the dyno that simulates vehicle weight and wind drag coefficients. You would be very surprised how much that can change your air fuel tuning......

[gmtech16450yz]

I'm not going to beat this dead horse anymore after this, but I couldn't resist showing this screenshot from 10_SS's log. THIS was my whole point about comparing narrowbands to widebands.
This is real world. This is valuable data. This is data that any "tuner" would be ignorant to ignore or deem "unusable".

What I've done here is to "zoom in" on the narrowband data. I don't really care what the actual values are, although the low 9's reading does also give me an idea that it's too lean. I've set my chart to show a range of .85v to .95v, just enough range to cover this small event in the chart window. Obviously you wouldn't want to run this range normally, but the power to change ranges is there, why wouldn't you use it? It's extremely powerful. This example shows that perfectly.

So look at the narrowband (red trace) vs. the wideband (white trace). Pretty cool huh? The narrowband IS telling you something, sometimes actually MORE than the wideband. Don't poo poo this source of information, it can help you tune cars. And don't try to kill the messenger of what may sound like a crazy idea. It may actually have some merit.

'Nuff said, I won't harp on that point anymore. Just keep an open mind guys. It's what separates the smart people who come up with new ideas from the ignorant ones that just repeat what everyone else is saying.




Here's the difference in reaction speed between that wideband and the stock narrowband. I'm measuring from when they both just started to indicate a difference in direction on the mixtures. The narrowband saw it first, the wideband saw it .23 seconds later. Dr. Mike, you can tell me if I'm doing that correctly. I know it's a barbaric way to see the reaction differences, but we're talking real world here, not lab equipment in a controlled environment. To me, I see the wideband reacting ~230ms behind the narrowband. Make sense? Like I said, it's kind of crude because you kind of have to interpret what constitutes a significant start of change trend, but it surely shows that the narrowband "saw" this event happen BEFORE the wideband. How much obviously can be debated, but between 100ms and 300ms would be my ballpark guess.

[dr.mike]

Really hard to say, with a slope that flat.

[10_SS]

I'm curious Mike, what are the delay times for the lambda pro unit? Only reason I ask is because I do power shift and transient fuel tuning with it on the dyno all the time. Of course I'm also using load profiles built into the dyno that simulates vehicle weight and wind drag coefficients. You would be very surprised how much that can change your air fuel tuning......

I will say that I see what seems like full resolution data between shifts, at WOT tip in, and after WOT close throttle events clear with the AEM 30-0300, the NGK was just kinda smooth lined and very misleading in a way, but since you know not to use it for that anyway, no problem. Just nice to see more real AFR data go along with what you "feel", things start to make allot more sense.

gmtech16450yz :
Yea I honestly do not remember people saying ignore your Narrowbands completely, just said don't base your PE fueling off of them. Maybe they did I don't really care, allot of stuff like this goes without saying. And thanks for finding something tied to the AFR leanout, hopefully that's "all" it is, based on the slow steady recovery after it went lean it would make sense it was tied to voltage as V ramps up so does the fueling come back in line. I'll play with it more... thanks again for looking at it

[dr.mike]

I'm curious Mike, what are the delay times for the lambda pro unit?

IIRC the ECM lambda pro uses the NTK LHA sensor. They quote it as " Less than 150ms ". So, I am guessing that it is around 120ms, like the AFX/Powerdex, which they also made.

This makes it a poor choice for transient events. I have a log with a 30-0300 and a slower WB, reacting to a fuel-cutoff recovery. The slower unit ( around 80ms response time ) completely missed the over-rich event ( rich spike ) followed by a lean spike..

Capture.JPG

In this case, the older LSU4.2 based WB (red) is actually about a foot closer to the engine than the 30-0300 (green). At the begining of the log is a fuel cut-off recovery event. The detail is completely lost on the 80ms WB. But, pretty clear on the 10ms WB. The throttle blip reactions , on the right side, were, almost, completely missed. And that ~80ms LSU4.2 WB is still considerable faster than the ~120ms NTK based WBs.

Trick is. You don't know you are missing it, until you can see it.

[GHuggins]

I will say that I see what seems like full resolution data between shifts, at WOT tip in, and after WOT close throttle events clear with the AEM 30-0300, the NGK was just kinda smooth lined and very misleading in a way, but since you know not to use it for that anyway, no problem. Just nice to see more real AFR data go along with what you "feel", things start to make allot more sense.

gmtech16450yz :
Yea I honestly do not remember people saying ignore your Narrowbands completely, just said don't base your PE fueling off of them. Maybe they did I don't really care, allot of stuff like this goes without saying. And thanks for finding something tied to the AFR leanout, hopefully that's "all" it is, based on the slow steady recovery after it went lean it would make sense it was tied to voltage as V ramps up so does the fueling come back in line. I'll play with it more... thanks again for looking at it

Might be because your using a BAP? Heavily dependent on input voltage....

[gmtech16450yz]

Might be because your using a BAP? Heavily dependent on input voltage....

Good point. That voltage drop just lines up too well with his lean event to be a coincidence. I'd really be looking at what caused that drop, how large it really was and what effect it had.

It would be awesome if the MPVI had the capability to input 0-20v signals to log. It is possible to log a 12v signal like fuel pump feed voltage, I've done it before. You just need to build a simple voltage drop circuit coming from the fuel pump feed and then convert the dropped voltages back to their 12v equivalents in the hpt scanner. What I ended up with was being able to read my ~10volt to ~18volt fuel pump feed into the MPVI as a 0-5volt input, then converted back to equivalent 10-18 volt ranges in the scanner. (Talk about voltage offset errors! lol. I wasn't so concerned on the values as I was the signal and voltage changes.) It worked, and actually showed me that my idea of using a BAP in front of an Aeromotive fuel pump controller didn't work very well. At lower BAP gain settings it did increase the final voltage going to the pumps, but at high gain settings it actually made the output voltage worse. I always wondered if the same thing happened when guys put a BAP in front of a stock FPCM, but I never did any further testing to find out. Knowing EXACTLY what the voltage is going to your fuel pump(s) though is very valuable information. How many guys don't even log their fuel pressure though? Crazy.

[GHuggins]

Through my years as a drive-ability diagnostic, electrical and heavy line tech, you would be surprised at how many vehicles I got from other reputable dealers that resulted in something as simple as corrosion in the wiring causing volt drop issues under the right conditions These were vehicles that had been to multiple dealers and never could get them figured out - part after part thrown at them - I know all too well what voltage will do to fuel pump pressure or as far as that goes any sensor or motor operation characteristic on the vehicle... I know exactly what your talking about on that one

[10_SS]

Through my years as a drive-ability diagnostic, electrical and heavy line tech, you would be surprised at how many vehicles I got from other reputable dealers that resulted in something as simple as corrosion in the wiring causing volt drop issues under the right conditions These were vehicles that had been to multiple dealers and never could get them figured out - part after part thrown at them - I know all too well what voltage will do to fuel pump pressure or as far as that goes any sensor or motor operation characteristic on the vehicle... I know exactly what your talking about on that one

My BAP system is awesome... I get down to 30psi Delta pressure at 6500rpm. But, This voltage drop and leanout event did not happen at all today after 10-20 WOT passes while dialing in my higher RPM VE table.. I did check the FAN ON settings and they seemed to line right up to turn on high speed at 60%, which also happened right at the voltage drop... Here's a little snippet of my fuel pressure drop and injector pulse skyrocketing to compensate for your enjoyment. (this is untuned, so it is rich after 6000rpm)

I tell you what.. this fast AEM is really helping me get my AFR right, very fast! It only takes 1 or 2 corrections, max. Before it used to take me 5-10. Im using 2nd gear pulls too.. . thanks Dr. Mike!! BTW is it normal for VE Airflow to max out at 512g/sec while Dynamic Air keeps climbing?
125i WOTS on Harmon Rd 64F snippet.hpl
125i VE STFT with WOT VE changes from 125h.hpt

[schpenxel]

On mine VE and Dyn air both stop at 512 g/sec. Maybe dyn air keeps going up on newer ECM's?

[10_SS]

On mine VE and Dyn air both stop at 512 g/sec. Maybe dyn air keeps going up on newer ECM's?

if both of yours stop at 512 what is controlling fuel? I thought one of these two were what the ECM uses for fuel calc, guess not?

[schpenxel]

It's just a scanner quirk, fuel keeps behaving like it should. I found some old threads on it back when I first hit 512 that explained why but I don't remember