NGK AFX vs. AEM 30-0300 CAN Comparison

[10_SS]

I decided to update my NGK AFX which has been wired in perfectly with a relay at the battery for power and all grounds (HPTuners and NGK) wired right at the Batt Negative terminal... Finally got it tuned in solid over the last 3yrs... LOG attached of that (123 NGK AFX WOTs.hpl)

Now updated to the latest and greatest fast AEM 30-0300 with the great Dr. Mike CAN update... which is AWESOME by the way.. so easy... I actually laughed when I saw the "new" CAN address show up in the list of PID's...

But now this "ultra accurate" WB reading is showing me my fueling is no longer anywhere near accurate... and of course now Im wondering, since the NGK AFX had to deal with extra grounding to make the analog input signal correct, then what does something like WOT full load with a MSD BAP do to your AFX/MPVI PRO and really you only check it a few times unloaded (hard to compare WOT display vs. actual HPTuners reading).. with the AEM 30-0300 CAN, the signal is supposed to be 100% spot on and seems to be...

So... after thinking about this for awhile.. which data log would you trust?

If I want a finished tune, I say NGK AFX.... but then now this much faster "more accurate" AEM 30-0300 is showing my tune to be a hot mess...
Camaro_AFX.Layout.xml
Camaro.Layout.xml

Edit: It took me 3yrs to get my fueling right with the NGK AFX. It took me 2 weeks to get it right with the AEM 30-0300. Im not sure which one is "right", but I can tell you which one makes more sense.

[DSteck]

I'd trust the wideband made by ECM...

[Ben Charles]

Never liked anything from AEM

[gmtech16450yz]

Lol. You guys are funny. How about looking at the logs and seeing what the actual data is telling you instead of making personal assumptions?

10_SS, I took a quick look and what I saw was your LTFT+STFT numbers are telling you the engine is running too rich. The AEM seems to be telling you that too? I'll look at it more but at first glance I'd be trusting the AEM.


And yeah, I'm betting there are a TON of tunes out there that are a "hot mess" after someone has done all the tuning based on a wideband that isn't reading correctly.

[schpenxel]

What's the reaction time of the NGK vs. AEM? The NGK readings look so smooth compared to the AEM, and I don't mean that in a good way

Fuel trims are still pretty far negative so that makes me think something really might be off.

A lot of the little ups and downs with throttle changes are just from transient fueling not being right. Good luck getting that right... the AEM is just fast enough to see them where the NGK isn't.

As to which one is actually more accurate? I have no idea.. without a lab to test them I don't think anyone else knows either. I've not really been a fan of AEM widebands until their new CAN bus one came out that was supposed to be awesome. I never got around to buying one though, so I have no direct experience with it, and I know some of the others posting have a lot of experience with the NGK one.

Lol. You guys are funny. How about looking at the logs and seeing what the actual data is telling you instead of making personal assumptions?

10_SS, I took a quick look and what I saw was your LTFT+STFT numbers are telling you the engine is running too rich. The AEM seems to be telling you that too? I'll look at it more but at first glance I'd be trusting the AEM.


And yeah, I'm betting there are a TON of tunes out there that are a "hot mess" after someone has done all the tuning based on a wideband that isn't reading correctly.

It's running lean while in PE... fuel trims are negative most everywhere else though

I agree about the hot mess comment though. No doubt about that.

[gmtech16450yz]

Fuel trims are still pretty far negative so that makes me think something really might be off.

As to which one is actually more accurate? I have no idea.. without a lab to test them I don't think anyone else knows either.

I don't agree with that one, that's the whole point of reading the logs. You saw it also, the fuel trims are WAY off. If the wideband was working properly the factory fuel trims wouldn't be that far off. Isn't that the whole point of tuning fuel trims anyway? If you're running closed loop you want your fuel trims to do as little corrections as possible, not more corrections. If you tune an engine based on wideband readings and factory side of your fueling is then off, I'd say that's a pretty good sign that your wideband isn't reading properly.

And the problem with lab testing is that it doesn't mean that they'll read the same or be as accurate in a car. If the analog conversion is off, the readings are off. MY opinion is at least 50% of widebands when they're installed and operating in real world installations are NOT reading properly.


10_SS... I took another look. Your fueling is pretty far off. If it was done with the AFX, I think you already have your answer on which one you should trust. From what I see, you're correct in thinking your tune's fueling is a hot mess. Try starting from scratch with the AEM and see what happens. I know from my experience using the AEM 30-0300 is that the factory side of the fuel trims come in DAMN NEAR PERFECTLY if you simply tune by copy/pasteing lambda error into your maf calibration. Using the data from the 30-0300 has made my fueling closer than it's ever been in 15k miles and about 1000 logs and tunes worth of tuning. With bad data you can chase your tail forever. With accurate data everything lines up the way it should.

Only "trusting" hardware from one brand, or not "liking" another brand is a very close minded way of operating. Opening your mind has a weird way of making you smarter. Assume nothing, you may learn something. I looked at the OP's logs without bias or prejudice towards ANY brand. THAT'S how you evaluate properly, by looking at facts and not relying on emotions to base conclusions. Just sayin'.

[schpenxel]

No doubt the tune needs some work

I believe you're assuming that he did the tune in open loop and then when he went into open loop got these negative fuel trims values though.. I'm not sure that is the case. If he used fuel trims to tune closed loop to begin with then which wideband was being used at the time makes no difference (in those areas, obviously PE is a different story)

I like doing a quick sanity check of what they're reading in closed loop.. which should be right around 14.7 on both of them (regardless of fuel trims since they'd be corrected to lambda=1), and you get something like the graphs below. I didn't do any real analysis on this but just a quick eye balling it makes me think the AEM is closer to 14.7 than the NGK. The AEM seems to be reading a little leaner everywhere.. the NGK a little richer.

To me that's the only real reference point we have to work with. It would be really interesting to see these two logged side by side on this car though.

AEM:
AEM.PNG


NGK:
NGK.PNG

[gmtech16450yz]

No doubt the tune needs some work

I believe you're assuming that he did the tune in open loop and then when he went into open loop got these negative fuel trims values though.. I'm not sure that is the case. If he used fuel trims to tune closed loop to begin with then which wideband was being used at the time makes no difference (in those areas, obviously PE is a different story)

I like doing a quick sanity check of what they're reading in closed loop.. which should be right around 14.7 on both of them (regardless of fuel trims since they'd be corrected to lambda=1), and you get something like the graphs below. I didn't do any real analysis on this but just a quick eye balling it makes me think the AEM is closer to 14.7 than the NGK. The AEM seems to be reading a little leaner everywhere.. the NGK a little richer.

To me that's the only real reference point we have to work with. It would be really interesting to see these two logged side by side on this car though.

AEM:
AEM.PNG


NGK:
NGK.PNG

Making your comparison doesn't work on two different logs made in two different conditions. The ONLY way that could be of value is if both widebands were installed and those comparisons were made on the SAME log. Especially when the fuel trims are off and constantly correcting. How on earth can you rule out the variables of what exactly the LTFT's were in both logs at any particular moment? Those two logs are so very different those kind of comparisons can't be made.

[gmtech16450yz]

10_SS- I'd be a little concerned about this too...

[DSteck]

Most widebands are accurate at stoich. Off of that is when they begin to deviate bad.

The AFX is made by ECM who makes lab grade equipment. Studies have been done on what's accurate. ECM stuff is.

[gmtech16450yz]

Most widebands are accurate at stoich. Off of that is when they begin to deviate bad.

The AFX is made by ECM who makes lab grade equipment. Studies have been done on what's accurate. ECM stuff is.

Again, you're making an assessment based on what? Have you even looked at his logs? This was the OP's question...

after thinking about this for awhile.. which data log would you trust?

The OP didn't ask what the "Studies" have shown, he wants to know which one of the two widebands he has should he trust. Is this what you'd tell a customer, "brand X is the best so disregard brand Y's information? Wow.

And having the voltage offsets wrong would still make the greatest wideband in the world accurate, correct? Ummm no.

You are completely disregarding the possibility of the logging of the wideband in real world situations can most definitely be inaccurate. You can have the most accurate wideband ever made and if it's not converted properly it WILL NOT read properly.


Supposedly this tune was dialed in using the AFX sensor(?) Does this look like it was properly modified with accurate data? What you're saying is analogous to somebody saying "Ruger makes excellent guns. There's no way one could be responsible for somebody shooting themselves in the foot while using it incorrectly. It clearly had to have been a Glock that shot the person in his foot". Does that make any sense? I'm trying to look in the person's hand to see what brand gun shot him in the foot. I don't give a d@mn which one did better in "studies". Studies are meaningless if the hardware is not used or setup properly. Again, MY opinion is a huge percentage of ALL brands of widebands that have to go through a analog voltage conversion are setup incorrectly.

[gmtech16450yz]

Here's an example of what happens when you're tuning with accurate wideband data. This is a 2 minute screenshot. Anyone that's compared widebands to narrowbands knows that anything off of stoich and the narrowbands are either going to go full rich or full lean. It's very hard to get a narrowband to actually read a steady mid-range value, they inherently want to fall towards rich or lean, that's how they work.

What this is showing you is that the AEM 30-0300 in this log is pretty much locked at 1.00 lambda, or stoich obviously. The mixtures ARE actually so close to stoich, and staying there consistently that the narrowbands will actually stay somewhat steady in their midrange. (I realize .450 is the actual center point, but expecting the narrowband to sit EXACTLY in it's center is asking a little too much.) This is telling me that the wideband reading 1.00 lambda HAS to be pretty darn close to actual stoich otherwise the narrowbands would hang on the rich or lean ends. (This engine is running in open loop so there is no O2/mixture switching. That's a different situation and of course the narrowband won't sit in the midrange if switching was on.) For you guys that are trying to learn something from this, switch off closed loop and O2 switching and see how your wideband compares to your narrowbands. Most likely even if your wideband is reading stoich, your narrowbands are going to be hanging around on either the rich or lean ends, not perfectly in the middle like this log shows. Dave, maybe you can show me an AFX wideband log that shows this same correlation? Again, if the wideband is even slightly off, the narrowbands WILL NOT sit steady like this. Getting narrowbands to sit steady at the crossover point is like balancing a beach ball on a seals nose. This screenshot shows a beach ball balanced on a seals nose for a full 2 minutes.

Schpenxel- if you want a sanity check for your wideband vs. narrowband, do the open loop test like this. When your wideband is reading stoich, your narrowbands will either be full rich or full lean. At least that will give you an idea of which way your wideband is biased. If you get lucky and your narrowbands hover near the midpoint like this screenshot, then your wideband is pretty darn accurate, at least in relation to your narrowbands. (Which are usually pretty darn accurate when working properly.)

[matty b]

So you are saying that you should sanity check the ACCURACY of a given wideband against the stock narrowband sensors? WTF? That is the worst scientific control I've ever heard of. Comparing them against reference gases is the only way to determine how accurate the HEGO is, not that.

You want a basic control? How about powering the HEGO controllers from a regulated power supply? Yes that means powering both from a proper power supply that maintains the same voltage at all times. From there, at least you can take out wiring variables which exist between both of them. If they read differently, then you use a control gas to verify the accuracy of the actual sensor (and the electronics of the controller). The scientific method does not start with using a variable control, thats why we have the hockey stick Global Warming model.

[DSteck]

Jesus Christ. You just don't get it.

[mowton]

Man this thread has taken a step into the dark side....I for one don't understand why "lab quality" accuracy is needed when tuning our personal projects or even customers cars. There are units available which while not in the same league, can provide a good platform for most tuning projects.

This is just my personal opinion on the subject and may not agree with yours. I understand the views relating to accuracy so here is mine on whats available....here goes :-)

In my observations and tests I have read/seen, the difference in power from 12.5 to 13.5 AFR is very very close, certainly not worth the risk of tuning on the lean side to produce those extra 5 ponies or so. So to that, I am not walking a fine line needing any more accuracy than the "consumer" widebands available today will provide. Certainly ECM makes a great product including reliability and repeatability but at a cost of $1500-$4000 dollars which is prohibitive for the average tuner. Somehow we need to make the available wideband units work for a large portion of HP Tuners customers.

Secondly, I am also of the opinion that GM engineering did a pretty good job in controlling fueling through their Narrowband O2 sensors/programming and to me establishes a pretty good baseline to be able to "calibrate" your wideband setup at through the Closed Loop limitations. Driving down the road in 4th gear at 40 mph w2ill provide a pretty good test scenario to align you wideband with for Stoich. At a minimum to null out analog developed offsets through ground loops and voltage variances. Again, I am going with the safe tune approach and am not pushing the limits of the accuracy limits of the standard, back yard mechanic equipment available to us today. Of cours all of this depends on a well operating narrowband system so visually verifying the switching rate and levels is a key prerequisite,

I can appreciate the desire to maximize the speed and accuracy in our wideband but there is a trade off based on our ultimate goal and think acceptable limits should be established and discussed based on these goals ad objectives...not just the quest to push accuracy levels further to the right of the decimal place......what is our return on investment if you will.

Thank you for listening and hope I didn't step on any toes.....I really enjoy tuning, training, and shooting the $hit with you guys :-)

Ed M

[DSteck]

Just because a controller has a reliable way to transmit the data it calculates (CAN vs. analog) doesn't mean it's collecting accurate data. I can accurately send falsified junk over the CAN bus if I want.

If the AFX was off and somebody wanted to claim the fact that the trims are messed up as reason the AFX is wrong... Look at the log in closed loop. The AFX tracks the commanded air fuel ratio along with the closed loop swings. But whatever.

[realcanuk]

I take the simple approach since I am not an expert. I tune using my wideband. At idle and steady cruise if its reading stoich, and my trims are within a few percent, I assume my wideband is accurate enough for my use.

Is it wrong to think this way ? I am no expert.

[gmtech16450yz]

I'm so sorry OP, it appears that only schpenxel and I are concerned with actually helping you by looking at your logs and answering your question. I stated MY opinion and tried to directly answer your question on which sensor it looks like you should trust when tuning your particular car based on your particular setups. I won't fill up your thread with any more wasted time arguing with others.

Good luck on your car and let us know what you end up doing/learning from your situation. There clearly is an opportunity for us ALL to learn here.

[mowton]

I'm so sorry OP, it appears that only schpenxel and I are concerned with actually helping you by looking at your logs and answering your question. I stated MY opinion and tried to directly answer your question on which sensor it looks like you should trust when tuning your particular car based on your particular setups. I won't fill up your thread with any more wasted time arguing with others.

Good luck on your car and let us know what you end up doing/learning from your situation. There clearly is an opportunity for us ALL to learn here.

Hey John, how is it going? I guess my response is more in line with how much a wide-band should cost versus absolute accuracy which as you state, is not what the OP was searching for :-)

But perhaps not stated as clearly as I would like, my response was suppose to iterate that using the GM closed loop functionality, you would be able to develop a trust and "fine tune" your setup using the offset adjust to dial in the WB setup at least at stoich. If you had to make large changes in the manufacturers Lambda/AFR voltage "transform" equations than there must be bigger problems that require addressing. No disrespect meant for the AEM and the great work dr Mike has done, it is a new approach and as such really hasn't had the "seat time" that most of the seasoned consumer wide-bands have had so it does represent a bit more of a variable in the overall scheme of things.

So yes, you are correct, I didn't sit and evaluate the details representing the two scenario's (but I do think getting the narrowband voltage to sit at stoich is pretty damn good :-)) but instead (right or wrong) chose to define the scenario for what I would determine accuracy, repeatability and which one to trust.

Ed M

[10_SS]

No doubt the tune needs some work

I believe you're assuming that he did the tune in open loop and then when he went into open loop got these negative fuel trims values though.. I'm not sure that is the case. If he used fuel trims to tune closed loop to begin with then which wideband was being used at the time makes no difference (in those areas, obviously PE is a different story)

I like doing a quick sanity check of what they're reading in closed loop.. which should be right around 14.7 on both of them (regardless of fuel trims since they'd be corrected to lambda=1), and you get something like the graphs below. I didn't do any real analysis on this but just a quick eye balling it makes me think the AEM is closer to 14.7 than the NGK. The AEM seems to be reading a little leaner everywhere.. the NGK a little richer.

To me that's the only real reference point we have to work with. It would be really interesting to see these two logged side by side on this car though.

AEM:
AEM.PNG


NGK:
NGK.PNG

I'm not real concerned about the LTFT fueling that's easy to fix, and I haven't touched that in 1-2yrs anyway but for some reason it appears to be drifting away over time. LTFT is not off due to the AFX... it's off due to something else and I simply havent corrected it in 1-2yrs is all. Plus, the NGK log was started with 48F coolant temps...

I posted due to the fact that the two widebands show such different numbers... I had my PE fueling pretty much good to go according to the NGK, then toss in the AEM hoping for similar numbers and now things are messed up pretty bad. Just don't know what one to trust, part of me thinks the NGK could be off simply due to the analog outputs and voltage swings, and the AEM seems be be getting great reviews an any accuracy tests I've seen, so I would tend to believe the AEM may be a bit more accurate just due to this. I was going to leave the NGK in there and just add the AEM, but though it wouldn't matter. Yes the NGK is showing much smoother since it's a much slower updating system and probably has some smoothing built in, the AEM doesn't.