It is really not that hard. Here is the method:

Step 1:
test the sensors as they are against a caliberated gas

Step 2:
run the sensors for one hour on a test engine with race gas to simulate wear and tear.

Step 3:
re-test the sensors against the caliberated gas once step 2 was over.

Put all the data together and present it in the table.

 

Sorry bro, that doesn't really tell you much

Was the test gas moving? what was the diameter of the pipe using (if at all)? how long was the measurement taken? Was the +/- they recorded an average reading or just split second +/- data points?

See what I'm getting at? The test setup was good...but they left out a lot of critical data that may give a much better/larger picture of how these sensors actually perform

 

Yeah, we posted at about the same time. We are eye-to-eye on that.

It does seem like they just held the sensor there, which I do not think is the best way. But, it seems to me that they were trying to hold it at an angle for full exposure to the gas. No, it's not perfect, but it probably gives an accurate reading. It just seems like a HUGE oversight ... someone would have suggest a rig of some type.

I also think that at the angle in the photo it's not an accurate representation of an actual install, which, as you mentioned, would be at 90* to the gas flow. That is probably the biggest flaw in the experiment.

Do I think that the 2 flaws noted above can skew the test results? Yes.

Do I think it would cause a sensor to read .5 AFR off? No. Maybe .2 or so maximum. But, I'm not an expert either.

Do I think this test is more accurate than the test TTP posted? Yes.

 

Considering how exhaust leaks can throw off wb readings quite a bit, just holding it next to the stream like that could be extremely inaccurate

 

You are correct.

 

""har har har...

I'm a mechanical engineer, work on electro-chemical systems, I know a thing or two about bench marking sensors and using calibration fluids/constants

The article and the "data" it presents sucks ***. It would have been better to use a single W/B as a standard (you could compare it to the calibration gas), then use that as a bench mark as to compare all of the other meters.

Not only that, but the first article doesnt provide a single real data point, just rough estimations and silly full and half-full circles. Where the sensors off by through out the entire test? Just during start up? What were the avg afrs recorded? They went through such lengths to get the calibration gases and such, yet couldnt find the room in the article to post a single real data point?""

Use what single wide-band as a standard and why? If a wide-band can't get 2 points correct what's the point?

 

If all the sensors were exposed to the gas using the same method as pictured (If that was the way it was done), then they will suffer the same inaccuracy. If innaccuracy entered the test, by the method of exposing the sensor to the gas, then ALL the sensors were effected in the same way. We really do not know how the sensors were exposed to the gas. You are speculating as I am.

The above said, this article is FAR AND AWAY better than the one posted by TTP in many ways:

1. It is more recent. His is JULY 2004 and this is JUNE 2007. A lot of technology has developed between then and now. For one, Innovate only had the LM-1 back in 04. Now they have the LC-1. AEM's wideband did not measure below 10 AFR in 04, now it does.

2. It is not tampered with. The one posted by TTP was tampered with. The original did not have the "Horiba" text typed onto it. Now that I know that he has tampered with the original article, I am suspicious of it. I wonder what else did he tamper with? I tried finding the article online, but could not. It is in the Import Tuner July 2004 magazine.

3. TTP is also being selective with this article. In this same July 04 article, the magazine declared the LM-1 to be the best for track tuning. TTP does not even mention that. He selectively presents the article W/O a mention of the LM-1's track tuning score.

4. The Import Tuner 04 article does not set a benchmark based on caliberated gases. They simpley decide that the NTK (Horiba according to TTP) is the benchmark since it is the most expensive. Why? Is it because it is expensive.

So despite your SPECULATION about the shortcomings of the article, it is far better than the article that TTP posted.

Looking objectively at both articles, can you say otherwise?

 

 

True, yet un true. If it really was conductd this way, there is going to be a run to run change depending on where the tester placed the sensor. From my professional experience, for some sensors, being just 1mm off can dramatically throw off readings. So no, I cannot agree with you that all of the sensors were subject to the same error, since user error is everything but consistant



Well, as far as I know, the actual sensors being used is Identical, the lm-1 and lc-1 use the same sensor/logic, and none of the tests were done below 10:1, so the aems lack of resolution was not even tested

OK...so he puts horiba there a few times, big deal? Does it really invalidate the rest of the article?

Ok you have me here, but the ford muscle article would have been much better if they actually tested it on the track as well. Kudos to the LM-1 for coming out on top

Yeah for the most part, even the fordmuscle article you posted benchmarks them all against the "very expensive" unit used at that facility. When it comes to commercial grade products, you normally do get what you pay for

I think both were lacking, the import tuner article could have benefited from at least benchmarking the NTK sensor vs a known afr gas

The fordmuscle article could have benefited from posting real data (esp vs time, the import tuner article suffers this downfall as well) as well as some real world tests.

I'm a mechanical engineer like I said, so when it comes to bench marking any sensor, I'm extremely critical, as I know how much error can occur due to sensor noise, user error, sensor hysteresis, etc

 

Again, this is mere sepculation on your part. You are assuming from the picture that this is the way the sensors were exposed to the caliberated gas. We DO NOT know that for certain. It is POSSIBLE that the sensors were touching the gas tank when the the tank was opened. It is POSSIBLE that they used the same position for all the sensors. I really doubt that 1 mm will throw off the readings, but it is POSSIBLE.
The LM-1 and LC-1 are NOT identical. The LC1 is TWICE as fast as the LM1. They do use the same patented method, but the LC1 is faster. The AEM has gne through a second generation since the Import Tuner 2004 article.
Doctoring an article w/o telling the reader IS a big deal. How does he know that the benchmark NTK unit was a Horiba? It is not mentioned in the text of the article. How do you trust that the NTK unit is indeed a Horiba? I suspect that all the RED and BLUE text at the bottom of the scan near the table was ADDED by TTP to show how great the ZT2 is.
No that is not the only way this recent shootout benchmarks the sensors. There are TWO methods. First, the sensors are caliberated against a gas of known Lambda value. Second, the sensors are caliberated against a Westec Expensive AFR meter. So the recent shootout has TWO benchmarks the known gas AND the Westec AFR meter.
I know they are both lacking, but which article has a more methodical and scientific approach to testing?

 

I agree with you man, the fordmuscle article is good, my main gripe was the lack of detail in the presentation of the data, as well as how they actually went about testing each

 

The article claims the Zt-2 to be reading as much as 0.54 AFR off. Wonder how they got that number since the display resolution is 0.1 AFR? Where did they see "4"? Or is this all made up? They do not say in what kind of conditions or how exactly was the test performed. It only says they used calibrated gases. You can perform this test in many ways to get different results.There are many ways to introduce an offset to the wideband O2 sensor depends on measurement method. We contacted fordmusle to discuss the results of this "shoot at" and recieved no reponce.
Our tests with calibrated gases indicates much better accuracy than presented in this article.
FYI we have not provided any of our units for this test and since nobody told us about this result till it was published we had no chance to verify this measurement, inspect the measuring method, the unit or the O2 sensor. Seeing claimed in the article AFR discrepancy, if given a chance, we would definitely look at the measuring method and other variables. Yes, we found out about this top secret "shoot out" (or is it a shoot at?) just like you from a link above.

Zeitronix unit scored the lowest in this "test" in easy of use category and difficult software. It takes 2 wires connected to power and ground plus O2 sensor placed in the exhaust stream for the Zt-2 to read the AFR. It does not get easier than that. For most tuners it seams to be very easy, specially comparing to the "other" units which need to go through a procedure before using the O2 sensor and to maintain accuracy frequently manually calibrated. Also, our ZDL datalogging software is very difficult to use. Hit ENTER- it starts logging, hit ENTER it stops. It's not a "rocket surgery" but must have been too difficult for the reviewing body, after all there is a disclaimer "Based on tester's experience". Of course since this is a shoot at not a shootout the tester did not even mention Zt-2 capability of reading other critical to engine tuning values. Or perhaps those are too difficult to use? I seriously doubt that's the case. This is simply, gently said, very heavily biased article posted to put the Innovate product at the top.

Most tech articles on fordmuscle are possible to view in entirety by subscribers only. Wonder why articles on fordmuscle about Innovate products are free to view. Can you say, "Innovate sponsorship"?

Best regards
Zeitronix

 

Strange indeed that the noticed a .54AFR variation, when the resolution is only 0.1

 

Not that it has anything to do with your claims of bias, but it's pretty easy to arrive at a figure of "0.54 AFR off" if the reference has a resolution of 0.01 AFR. For example, reference of 10.04 AFR and display on ZT-2 of 10.0 (10.04 - 10.0 = 0.04 difference).

 

I agree with you. It's a simple math. But this is not how measurements are correctly performed.