Widebands That You Would or Wouldn't Buy

[HelmutVonAutobahn]

It looks like the extra ground ( heat shrink ) is for the sensor cable shield. It is marked DRN ( drain ). Probably an attempt to isolate the sensor signals from external interference. I'd have to check to see what it is actually connected to. The actual sensors don't have a drain/shield.

[DelSolid]

DelSolid:

Do you have accuracy and and noise values for the various controllers?

Sorry, not that I can share. AEM does not post competitor test data unless it is performed by a 3rd party testing house and the testing you refer to was performed internally. So there is no way they would give me authorization to post that stuff.

But it's easy to get. The test gasses I posted earlier in the thread will give a repeatable value for error testing and the noise can be determined by placing the sensor in a constant known gas and determining the standard deviation of the analog output signal over some test time when it is in the fixed (unchanging) gas. Since the gas is unchanging the analog signal should be unchanging as well and the constant changes of the signal over time is the error caused by signal noise.

[HelmutVonAutobahn]

Sorry, not that I can share. AEM does not post competitor test data unless it is performed by a 3rd party testing house and the testing you refer to was performed internally. So there is no way they would give me authorization to post that stuff.

I figured that you guys would not have posted that table without some backup. It's almost begging for a response from "The world's fastest wideband".
It would be one thing if you just edged them out. But, you have them in the bottom 1/3 of the field.

You could post the data for your own units ?

[DelSolid]

Sorry, not that I can share. AEM does not post competitor test data unless it is performed by a 3rd party testing house and the testing you refer to was performed internally. So there is no way they would give me authorization to post that stuff.

I figured that you guys would not have posted that table without some backup. It's almost begging for a response from "The world's fastest wideband".
It would be one thing if you just edged them out. But, you have them in the bottom 1/3 of the field.

You could post the data for your own units ?

Oh, there is definitely plenty of backup, I can assure you. You really have to have all your ducks in a row before you explicitly call anybodies product out by name & P/N. Referring to "The world's fastest wideband", we made sure that both the MTX-L and the LC-2 were there in their latest forms (with the 4.9), but also the discontinued LC-1 was included since we knew that would be the fastest one and the one most people refer to. We also made sure the test house verified and explicitly confirmed that the test units were on the fastest available analog output speeds.

I will check to see if they are cool with the posting of data on our units but in all honesty the data alone without other units to compare them with is pretty much a non-event. Saying our unit has a snr of XXX and a std dev of YYY is pretty useless without context. We generally prefer 3rd parties to do all the tests and posts their results.

[Fred]

After a quick perusal, it looks like a "signal ground" could be picked up off of, what appears to be, the programming header ( marked JP3 ). FFrom PIN1 ( square ). If you have a differential input, using that signal as the negative side will likely provide a cleaner signal. i.e. the heater noise SHOULD cancel out, via common mode.

The black harness wire looks to go through a shunt resistor; then, straight to the MOSFET. I would cut this wire as close as possible to the gauge. Then use HEAVY ( 14g or better ) wire to ground it to something solid, like the battery terminal. Make the run is as short as possible.

That's about the best you can do.

Sounds like wiring up a MegaSquirt :-D

[fuzzysig]

why no 14point7 on that list?

[DelSolid]

why no 14point7 on that list?

With any list, there will be many that are not on it. I tried to get as many mass market devices as possible and based most of my shopping on amazon and ebay sales data from data collection services. Additional units were selected based on my own subjective judgement regarding the number of units in the field and the monthly sales of the units. I wanted to get the best coverage for the least $ spent, remembering that not only were all the devices purchased new for the testing (except I believe, the Motec PLM, the LC-1 and the Powerdex AFX) but consider there was a per unit charge for the 3rd party testing. I could have easily found 40 units to test if I really wanted to but tried (subjectively) to limit it to the most common units used in the aftermarket.

However, I have reached out multiple times in this very thread to get a unit from 14point7 to test (unofficially) and will share the results here. The offer still stands.

[DelSolid]

DelSolid: Do you have accuracy and and noise values for the various controllers?

Here is some data from my output noise tests. Please note that the following data was gathered by me personally and is not in any way considered to be an official statement or information from AEM. It is not to say one unit is better, faster, higher, prettier or smarter than any other. It was simply my observation on a particular day and is in all likelihood flawed and totally wrong. Use at your own risk. Your mileage may vary.

The Signal Noise as used above is defined as the +/- value of one std deviation of the analog (unless otherwise noted) signal output with the sensor in 14.6:1 and 13.6:1 AFR test gases. In English this means that ~68% of the samples taken in a steady state test gas situation fell within this band and thus ~32% of the discrete measurement points fell outside this band. A smaller number is better.

You will see a coupe more items listed here than in the response table, that is because this was an informal test and I ran a few additional configurations just for my enjoyment. The CAN bus version of the 30-0300 was included to isolate the true analog noise on our unit to the same signal received over the CAN bus. The Bosch LT4 is a very high end unit that is not in any way the same class of device as the others but it a useful benchmark for internal testing for CJ125/LSU4.9 based devices as it is a true Bosch implementation of their technology.

I did my best to connect the test units exactly as outlined in their respective manual. I powered the units with a 20A BK Precision power supply parallel driving a 660 CCA 12V deep cycle marine battery.

In my mind, the data falls into two very distinct groups; the ones that are acceptable and the ones that are not. I leave it up to you to decide which is which but it does sort of jump out at you.

[HelmutVonAutobahn]

Does the PLX measurement include the little filtering capacitor that they include in the kit ?

[DelSolid]

Does the PLX measurement include the little filtering capacitor that they include in the kit ?

No, not for this test and come to think of it, I don't believe the test unit came with one and the online instructions don't mention it (I just checked). I vaguely remember an earlier one having it but maybe that's just an implanted memory prompted by your question. I have not played with the PLX for something like 6 months so it certainly may have.