What do we like for wideband O2 sensors?

I've only installed one.  I put an Innovate LC-2 into my Miata and used it to tune Megasquirt.  It was fine I guess... I didn't like that it only had one ground connection for both signal and power, which introduced noise on the signal.  It stopped working after about 2 years and just read wide open 22.1, which I attributed to a bad O2 sensor but I sold the car before I got around to fixing it.

So... what do we all like?  Seems like in a reasonable price range it comes down to Innovate or AEM.  Pros and cons?

I have an AEM X-series, and it seems okay, but annoying that the sensor display and the reading in Megasquirt do not coincide, no matter what I do for ensuring ground planes are the same.  (I rewired half the car in an attempt to make the problem go away)

I will probably update the car to MS3 so I can use the CAN feature.

Interested, and also casting my vote against the innovate mtx-l.

Mine seems to eat a sensor every season, despite getting my exhaust redone and mounting it as close to recommended as possible.

Pete. (l33t FS) said:

I have an AEM X-series, and it seems okay, but annoying that the sensor display and the reading in Megasquirt do not coincide, no matter what I do for ensuring ground planes are the same.  (I rewired half the car in an attempt to make the problem go away)

 

I will probably update the car to MS3 so I can use the CAN feature.

I had the same issue with the Innovate, which I also attributed to ground issues tht couldn't be solved when signal and power share the same ground. I vaguely recall that some of the systems have distinct grounds, but I don't know which ones. 

I have installed Innovate widebands on my Exocet and my Miata - both have needed replacement sensors.  They seem to last a year, two max.  After some internet reading, I no longer let the sensor warm up before starting the car, maybe that will help.  The Miata also had the gauge/controller fail.  Not impressed.

This isn't helpful since I haven't actually used it yet, but the 14point7 kit is an easy build - took longer to weld in the bung.

I love love love my OBDII AEM widebands in the 350Z. They make nice stuff and support it well.

Tom Suddard said:

I love love love my OBDII AEM widebands in the 350Z. They make nice stuff and support it well.

Do you have a specific model that you like Tom?

Another vote against Inovate. I have a paper weight LC-1 that was dead out of the box, and outside of warranty.

I've installed & used Innovate & PLX. I prefer PLX.

ShinnyGroove (Forum Supporter) said:

Pete. (l33t FS) said:

I have an AEM X-series, and it seems okay, but annoying that the sensor display and the reading in Megasquirt do not coincide, no matter what I do for ensuring ground planes are the same.  (I rewired half the car in an attempt to make the problem go away)

 

I will probably update the car to MS3 so I can use the CAN feature.

I had the same issue with the Innovate, which I also attributed to ground issues tht couldn't be solved when signal and power share the same ground. I vaguely recall that some of the systems have distinct grounds, but I don't know which ones. 

Even the best analog wideband signal always has significant noise in it -- as Pete mentions the right answer is a digital readout like CAN bus (some systems use rs232 serial ports as well)

Good discussion here for reference: https://grassrootsmotorsports.com/forum/grm/wideband-o2-and-ms/158082/page1/

A lot of sensors fail because they're turned on immediately when the engine starts or key-on and the moisture in the cold engine out exhaust kills them.  It's best practice to incorporate a delay before turning on the sensor, this is how OEM O2 sensors work, some standalone ECUs allow this function but the wideband controller itself won't do this.

Ideally find a wideband controller that uses Bosch LSU 4.9 sensors, since they don't require free-air calibration, and are readily available.

EcoTron is the most cost effective option for a simple wideband that also has CAN output:

https://www.ecotrons.com/accurate_lambda_meter/alm-inline/

AEM X Series also has CAN output:

https://www.vividracing.com/aem-xseries-inline-wideband-uego-afr-controller-lsu49-p-152414916.html

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

I've had an Innovate LM-2 in the Jeep for years.  And it feeds data to the ECU as well, so it has to work.  The setup isn't perfect, and early on, it was pretty buggy and would periodically error out and need a reboot.  I also killed a couple of sensors early on.  However, once they updated the firmware to support LSU4.9 sensors instead of the older LSU4.2, I swapped over and it's worked great ever since.  No more lockups, no sensor failures, no more calibrating, etc.  And the sensors seem to do just fine with starting their warmup cycle just before engine start (it starts warming up immediately at key on, so it's usually 2 seconds into warmup when the thing starts).  And it warms up fast enough that going to closed loop 30 seconds after start isn't an issue, as it's outputting signal by then. 

No complaints with the 14point7 running in my truck, it's been flawless so far. 

Tom Suddard said:

I love love love my OBDII AEM widebands in the 350Z. They make nice stuff and support it well.

Another vote for the AEM - bonus for having a nice oem looking gauge.

alfadriver (Forum Supporter) said:

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

I assume the heater is controlled by the controller and not something the sensor does all by itself? So it's something that the manufacturer of the controller should be able to address in the controller's firmware?

ShinnyGroove (Forum Supporter) said:

Tom Suddard said:

I love love love my OBDII AEM widebands in the 350Z. They make nice stuff and support it well.

Do you have a specific model that you like Tom?

I've tried most of the brands mentioned here and these are by far my favorite: https://www.aemelectronics.com/products/gauges/digital-gauges/x-series-wideband-uego-obdii

alfadriver (Forum Supporter) said:

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

Good points. It's hard to tell what aftermarket controllers don't have this problem so adding a delay doesn't hurt, unless you regularly need cold start AFR data on an older tuner car (older meaning with only a factory narrowband O2).

BoxheadTim (Forum Supporter) said:

alfadriver (Forum Supporter) said:

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

I assume the heater is controlled by the controller and not something the sensor does all by itself? So it's something that the manufacturer of the controller should be able to address in the controller's firmware?

Yes, the controller does all of the work.  And while the physics are very much the same, the different sensors run very differently to each other- NTK vs Bosch.  Hardware is very specific to the brand of sensor.

Over the years, I've run ECM Lambda Pros and the Bosch (can't remember the controller) and now most of what we do is with small Bosch boxes and the on board sensor on the ecu.  

If I were to harbor a guess at the problem, Innovate may use a pretty old controller with newer sensors that have much more powerful heaters on them.  So they are much likely to crack the heater with old controls.

engiekev said:

alfadriver (Forum Supporter) said:

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

Good points. It's hard to tell what aftermarket controllers don't have this problem so adding a delay doesn't hurt, unless you regularly need cold start AFR data on an older tuner car (older meaning with only a factory narrowband O2).

Or not even an O2 sensor, ever.  If I were to convert one of my old cars to EFI, I would record every single drive in it for a very long time- to get the starts as perfect as I can get them.  As well as perfect the transient fuel.

alfadriver (Forum Supporter) said:

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

I never had an issue with my LC1, with the 4.2 sensor, warming up the sensor before starts.

The only issue would be when oil would climb up the oil filler tube, get into the crankcase vent, and flood the engine with oil.  Rapidly ingesting a quart of oil wou!d reliably kill the sensor, along with the local mosquito population.

I only stopped using it because of a wiring accident.  The unit slipped and rested on the header and damaged the wiring...

In order of my personal preference tuning lots of different setups, keeping a reasonable budget in mind:

Ballenger AFR500v2 with NGK sensor/NGK AFX

(big gap)

AEM X series, 14.7 spartan2/ADV, PLX, Zeitronix

(small gap)

Innovate LC2/MTXL, Glowshift, older AEM, other off brands

alfadriver (Forum Supporter) said:

BoxheadTim (Forum Supporter) said:

alfadriver (Forum Supporter) said:

In reply to engiekev :

FWIW, I've been using sensors to record cold starts (down to -20F) for over 20 years now, and I've still not broken a sensor.  It's been thousands and thousands of starts.  In the last 5 years, I've even started to run production sensors before the key on, and still have not broken one.

Which tells me that it's the controller that's the problem, not the sensor. 

We have actually determined in the last few years that what actually breaks the sensor isn't the drop- it's the heater reacting to a sudden drop in temp.  Which means that we now have production systems that are capable of being on at crank.  And within 4 years, every new car will have the O2 sensor alive within 5 seconds of start.

I assume the heater is controlled by the controller and not something the sensor does all by itself? So it's something that the manufacturer of the controller should be able to address in the controller's firmware?

Yes, the controller does all of the work.  And while the physics are very much the same, the different sensors run very differently to each other- NTK vs Bosch.  Hardware is very specific to the brand of sensor.

Over the years, I've run ECM Lambda Pros and the Bosch (can't remember the controller) and now most of what we do is with small Bosch boxes and the on board sensor on the ecu.  

If I were to harbor a guess at the problem, Innovate may use a pretty old controller with newer sensors that have much more powerful heaters on them.  So they are much likely to crack the heater with old controls.

Going off Paul_VR6's suggestion as well, I also only use the NGK AFX. It's been proven to be the most accurate controller on the market by a wide margin.

The NGK (Ballenger) AFX are manufactured by ECM, and V2 now supports both Bosch LSU and NGK NTK sensors

http://afxmonitor.com/

If you go on ECM's website, you'll see the same display the NGK AFX uses (different controller, but this just proves it really is made by ECM):

Note this doesn't offer CAN or digital output, only analog.  But unless you already have a vehicle with CAN bus (and are planning to add sensors to the existing bus) or have a Standalone ECU that has CAN Bus, that's not really a deal breaker.

In reply to Paul_VR6 (Forum Supporter) :

My mind was blown that the FAST XFi's onboard NTK controller reacted so fast that you could tune transients not by looking where the fuel dipped lean or rich, but by looking at where the fuel trim was adding or subtracting fuel.  Because the computer would run desired ratio rock steady.