AFR install (AFR=Air Fuel Ratio)

Hello,

I am considering adding an AFR sensor to my panel, although not entirely sure how or what I would looking for in the AFR numbers. More to learn.

I have an IO 360 with mechanical fuel injection (bendix) and dual (legacy) EFII electronic ignition. I have been running this setup for years and it has been great. I have read a little about using AFR to better lean and set max power, etc....using AFR.

Does anyone have experience and the PLX system, or recommend another, for a AFR sensor?

https://www.amazon.com/PLX-Devices-SM-AFR-Wideband-Controller/dp/B00U1J1MCK

Never knew anything like this existed.

Reading the description on the Amazon link you posted, it says "O2 sensor harness 10ft." My concern would be how does the O2 sensor work with leaded fuel that we use in our airplanes?

I am under the impression that the leaded fuel will ruin the O2 sensor.

AFR

See ?Kitplanes March 2019?. Article on how to add AFR to your airplane. It?s definitely doable with today?s sensors.

Works Well - But...

The PLX sensors are designed to work in an un-leaded fuel environment - auto gas.
If you install the sensors according to PLX instructions you will get maybe 50 hours (some get much less) of operation out of them before they stop working in our leaded av-gas world.
But, I'm told that if you mount the sensor up in the exhaust flame (about 18 inches or so) it keeps the lead from depositing on the sensor and you get a very long life from them.

I'll not get into the technical details of the Stoic... sensing and the benefit, as that has been discussed in other threads here. I'll simply say that if you want to get the most accurate efficiency from your fuel burn, using AFR would be greatly encouraged.
Also, strongly encourage you to chat with Robert Paisley - since you have the EFII ignition, he will be an excellent source of technical advice.

Each engine / sensor installation will give different readings. What you should do is operate your engine as you normally do & see what your indicator values are at each power setting. Full Rich, Best Power, Peak, your Lean of peak setting. From that you can develop leaning guide lines referring AFR as an additional reference to your RPM, EGT, F/Flow tools you already have.

As a (rough) example 10.5 to 1 might be Rich, 11.5 thru 13 might be Best Power, Stoik around 14.7 (which means nothing), your Lean might be anywhere between 14 to 18, what ever your comfort level.

Would be interesting to know what YOUR numbers would be.

Isaac said:

See ?Kitplanes March 2019?. Article on how to add AFR to your airplane. It?s definitely doable with today?s sensors.

Click to expand...

I have installed the Ballenger AFR500v2 unit detailed in this article, with the NTK sensor which is said to be more tolerant of leaded fuel than the older Bosch sensors. It's still early in my testing, and I'm currently working on a couple of enhancements.

https://www.bmotorsports.com/shop/product_info.php/products_id/2337

Simplest is a smaller remote-mount, dimmable display. As mentioned in the Kitplanes article the built-in LED display is large, very bright, and non-dimmable. Note that if you have an EFIS (Garmin, Dynon, etc.) that accepts a 0-5v analog input, you can have the AFR (or lambda) value displayed & logged by the EFIS, and can mount the AFR box out of sight behind the panel without the need for a separate remote display.

Second & slightly more ambitious is a closed-loop mixture controller using a linear servo connected to the mixture knob with the same quick-disconnect linkage used in this product: http://flightenhancements.com/auto-lean.html This is ready for flight testing, likely starting this weekend. Set a target lambda/AFR value, and the controller will maintain it throughout climb, cruise, and descent. You can enter the target manually, or use preset targets for different phases of flight.

If you do this, aside from the lead concerns, try to find a wide-band O2 sensor. The narrow band commonly available sensors are actually only sensitive to a not very useful window of A:F Ratios.

Wide band sensors are commonly found on "lean burn" cars, older Honda Civics and such. Maybe more cars these days than when I was into cars.

rdrcrmatt said:

If you do this, aside from the lead concerns, try to find a wide-band O2 sensor. The narrow band commonly available sensors are actually only sensitive to a not very useful window of A:F Ratios.

Wide band sensors are commonly found on "lean burn" cars, older Honda Civics and such. Maybe more cars these days than when I was into cars.

Click to expand...

You would be hard pressed to find a narrow band aftermarket O2 setup. They've all moved to the wideband. Most use the Bosch LSU 4.2 or 4.9 and a couple use the NTK.


The NTK is said to be more lead compatible but they also want it to be recalibrated every hour with lead.

FWIW, Bosch recommends moving the sensor away from the exhaust port when using leaded fuel. They recommend 36".

Recalibrated every hour eh?

So when they die from lead poisoning, they drift out of calibration over a period of operating hours?

DanH said:

Recalibrated every hour eh?

So when they die from lead poisoning, they drift out of calibration over a period of operating hours?

Click to expand...

Honestly that's just what I read in the instructions (page 11). I haven't used the NTK sensor.

I do have experience with the AEM UEGO running the Bosch LSU4.2. It survived a tank or two of 100LL and 120 race gas but mostly lived on pump gas + NOx. It was in the middle of the collector on long tube headers. It never had to be recalibrated.

These sensors are invaluable street tuning a car, especially with a carb. On an airplane? I don't know. EGT gives you a lot of info but you don't get the actual AFR. EGT is also slower to react but again does it matter?

With the LSU4.2 I've read numbers of something like 80k km service life with pump gas and 10k km with 0.60g/l lead. Which is like 10% more than 100LL has. You could maybe extrapolate that out at like 80km/h average and call that ~125 hours on lead?

Wideband O2 sensors in airplanes

Wide band O2 sensors are used in most vehicles these days.
The information (air fuel ratio - AFR) from a wideband sensor is the best fuel tuning information. Most of our customers (over 600 EFII installations now) use a wideband O2 sensor on their aircraft engine.

An O2 sensor needs an O2 sensor amplifier to work. The amplifiers might have a built in display, or they might put out a voltage that can be interpreted by a separate display. In any case, you need a sensor, an amplifier, and some way to display the AFR information. With System32, the AFR is displayed on our Cockpit Controller - no additional display is required. We have many customers who use the wideband kit from PLX Devices - this works fine. There are many other brands available. The Bosch LSU 4.9 sensor is used in almost all of these systems - this works fine - no need to shop further.

In a car application (unleaded fuel only), the sensor is typically located downstream in the exhaust to it can see combustion results from multiple (or all) cylinders.
This doesn't work when leaded fuel is present.

What does work is the following:
(for a 4 cyl Lyc)
Mount the sensor in the #4 exhaust pipe, 3 inches down from the cylinder.
Sensor sitting in a horizontal fashion sticking aft off the pipe.

This puts the sensor in the flame front coming out of the cylinder which will keep it purged of lead. I have over 400 hours on one sensor using 100LL exclusively.

Why #4 cyl (#6 cyl on 540 engines)? At this location, the exhaust pipe is the rear pipe on the cylinder and there is better room for the sensor mounting.
The sensor could be mounted on any other cylinder if you wish.

Some people will report that they have experienced short O2 sensor life on their planes. This is typically due to:
1. The sensor is located too far down in the exhaust system.
2. Some area of their engine operation is extremely rich - this causes lots of lead and carbon build up in the exhaust system.

That's pretty much the whole picture. AFR is very good info. Adjust your fuelling to get 11.5 to 12.0 AFR for max take off power (on the rich side) and let it range up to 15 (leaner) at light cruise power and you will have a very happy engine.

Robert

rcpaisley said:

That's pretty much the whole picture.

Click to expand...

Except for the question. When they die from lead poisoning, they drift out of calibration over a period of operating hours?

lead poisoning

DanH said:

Except for the question. When they die from lead poisoning, they drift out of calibration over a period of operating hours?

Click to expand...

That problem is usually from issue #1 in the post.
There is one more factor that occasionally comes up:
Avgas has a lead scavenging chemical added to it called ethylenedibromide (EDB).

EDB doesn't do it's job well unless CHTs are in the 360F to 390F range.
When engines are operated at very low power cruise levels, the CHTs may go below this range and lead may start to appear everywhere it can including in the exhaust, in the engine oil (makes milky green colored oil), and on all interior engine surfaces where oil can get. This effect can cause sensor fouling.

Robert

So again, when they die from lead poisoning, they drift out of calibration over a period of operating hours?

DanH said:

So again, when they die from lead poisoning, they drift out of calibration over a period of operating hours?

Click to expand...

I killed a number of them already (Bosch LSU 4.9). They start lagging when they get lead fouled but they are still accurate. After a while they take minutes to respond. When they are new the response rate is incredibly fast, faster than the EGTs.
I tried an aftermarket one recently that had a different sensing tip configuration, promising better lead tolerance and compatible with the original Bosch sensor. It died in less than 2 hours.

You can take your pick from the big 3 wideband controller makers- PLX, AEM, Innovate. They all work. The latest models use Bosch 4.9 sensors. We find AEM and Innovate are somewhat better than most other brands. No experience with the NTK based ones.

Bosch publishes guidance for sensor orientation and preferred EGT range here: https://imfsoft.com/files/Lambda_Sensor_LSU_49_Datasheet_51_en_2779147659pdf.pdf

Just to be clear, all WBs have a heater element to hold the array within a specific temperature range to be accurate. If memory serves me correctly, this is not below 1450F. The sensor can only heat itself, cannot cool. The sensors can withstand up to 1800F for very short periods but life will be shortened. Thermal shock is a leading cause of premature failure due to condensation forming in the exhaust on warmup. Follow Bosch's orientation advice. It's important to avoid this issue. OEM controllers ramp up heater current slowly to avoid thermal shock. I'm not aware of any of the low cost WB controllers doing this. One reason few aftermarket O2 sensors last anywhere close to OEM application ones.

The sensors do degrade slowly with age. Bosch also mentions that TEL (lead) affects sensor life substantially depending on concentration and is likely to be 1/10th as long as operating under the same conditions with unleaded fuel.

We've seen them last (with drift) as long as 350 hours and as little as 2 hours on 100LL. Drift on may reach .6 AFR at 2000 hours on unleaded fuel according to Bosch.

Running AFRs richer than about 11 to 1 for long periods will shorten life.

I've seen some independent tests on a bunch of aftermarket WBs and many were not all that accurate, despite manufacturer claims, when compared against lab quality equipment. Some were downright inaccurate.

We've had 15+ years of experience with these things on thousands of EFI systems. 6 different controller brands. When the sensors are new, they can offer a good guideline for AFRs and tuning. When using leaded fuel, degradation may be pretty quick (in some cases). If you have EGTs, you can check WB accuracy by leaning to peak. On clean (no ethanol) mogas, the AFR should be around 14.7, with avgas, around 15.2.

On aircraft, there is the question of accuracy with altitude, you can check this aspect as well by leaning to peak at a few different altitudes. Indicated AFR should be close in all cases. There are some technical reasons with O2 pressures as to why they may not track with good repetition if calibrated at a different altitude. This may not be so important in automotive use but certainly could be in aircraft.

In short, don't believe these are the gospel without question. The only thing worse than not knowing what the true AFR is is to have a gauge you believe to be correct and isn't, while you chase tuning. Good to question these things armed with some understanding and logic. Too many people believe digital instrumentation must be 100% accurate. There are lots of reasons why this may not be so.

Good Timing!

I'm at the end of my build and hope to do a first start next week. I have installed the EFII legacy system. I even bought the AFR system mentioned in the literature. However, I was too stupid to remember to have the sensor bung welded on before installing the pipes. That being said, I also read that EGT's should be in the "Low 1300's" at cruise. Since I'm sick of building, and want to get it in the air, I was planning on using those numbers as an adjustment tool.

After reading this thread, something popped into my head (I think it was a blood vessel). What if the proper AFR, and the proper EGT do not happen at the same setting? Which is more important/accurate? I would think the EGT. But I'm the guy who forgot to install something I bought!

AFR is AFR period. I'd argue that this is really want we want to know to tune an EFI system. However, see my comments above on actual accuracy displayed.

EGT doesn't indicate AFR directly and is influenced by ignition timing to a fair degree and AFR as well. If peak EGT is at 1500F, you can be LOP at 1450F, say 17 to 1 AFR and also ROP at 1450F, say 13.5 to 1 AFR for arguments sake. You always better know which side of peak you are if you're tuning EFI. The only time we know the rough AFR using EGT is at peak. Other things can affect the gross EGT values like compression ratio and probe placement but on a given engine, EGTs are relative at least.

On the bonus side, EGT thermocouples aren't compromised by lead and don't degrade in accuracy like WB O2 sensors do.

I consider these two things complimentary and as I mentioned before, EGTs can and should be used to cross check WB accuracy.

On the winning Reno Sport class stuff, we use all data gathered to get to a place where the engine makes good power but stays together for the race. Sometimes power must be sacrificed for longevity as we are running at the thermal and structural limits of pistons, valves, turbos etc.

If you own a rebranded or SDS ECU with V14 (circa 2005) or newer software, you can display AFR in the programmer from the linear output wire on most WB controllers connected to pin 24 on the ECU. Be sure to select the correct WB setting in the programmer to match the WB output. There are several different scales. AEM alone uses 2 different scales depending on the vintage of the controller.

We offered an extended bung to keep the sensor element out of the direct lead stream at the expense in response time. The jury is still out if that extends sensor life when operating on leaded fuels. Some reports say so, some say not.

I hear Bosch has new sensors under development. Will see what they come up with next.

Solutions to the leaded fuel replacement are still ongoing and may be available to the Experimental fleet before the certified one: https://www.flyingmag.com/when-will-we-see-unleaded-av-gas/

Of course Swift Fuels has been sold at an increasing number of sites for the last 3 years or so. Looks like others may be joining soon.

One other very large company in TX, not mentioned in this article, is also pursuing their unleaded avgas blend. I've been talking to them for a couple years now and it's being tested on a turbocharged Lycoming at this time.

For the OP, I'm not sure how useful a WB will be but it would be interesting to have it I think.

rv6ejguy said:

AFR is AFR period. I'd argue that this is really want we want to know to tune an EFI system. However, see my comments above on actual accuracy displayed.

Click to expand...

Thank you Ross. As always, good information, without the BS.

As a rapid tuning aid, I absolutely agree. Given the available EFI systems require initial adjustment of controller settings, I understand why EFI vendors support them. Dialing in the settings using EGT data would be a cluster for the average builder.

Same (x10) for the racing application. In racing, nobody cares about long sensor life. The goal is rapid acquisition of tuning data, and of course EFI then allows rapid changes to fuel settings. The ability to make rapid changes at the races is one of EFI's little miracles. (Dear readers, if you don't believe it, you probably haven't been around long enough to have re-jetted carbs while standing in pit lane.)

That said, is a wideband O2 sensor a long term engine management instrument? Nope, not as long as we use leaded fuel.

When we move to unleaded, we'll keep the EGT probes. An unusual EGT split will still be the best indicator of an individual cylinder problem.

An AF ratio display will become unnecessary. EFI vendors will use O2 sensor output to vary mixture at the digital level, without pilot intervention. Then (and only then) will aircraft EFI be "just like your car".

Agree with your points and conclusions here here Dan.

We'll probably continue to display AFR in the programmer on the next gen ECU just because airplane guys like to have all data available.

judoka5051 said:

I'm at the end of my build and hope to do a first start next week. I have installed the EFII legacy system. I even bought the AFR system mentioned in the literature. However, I was too stupid to remember to have the sensor bung welded on before installing the pipes. That being said, I also read that EGT's should be in the "Low 1300's" at cruise. Since I'm sick of building, and want to get it in the air, I was planning on using those numbers as an adjustment tool.

Click to expand...

Lance, pull the pipe and have the bung installed. Not being critical here, but your comment about EGT suggests you're not the customer who should try do EFI setup without a wideband sensor.

Addressing the specific comment, "low 1300's" means absolutely nothing. What matters is temperature referenced to peak temperature, i.e degrees rich of peak or lean of peak. Peak may be (for example) 1550 F with Engine A, and 1425 F with Engine B, depending on probe, probe location, compression ratio, and ignition timing. Using "low 1300's" for both would result in different FA ratios. Using (again for example) "200 F rich of peak" would result in roughly the same FA ratio for both....but the displays would read 1350 and 1225 respectively.

Thank you gentlemen!

Alright, going to add a couple of hours (won't be the first time!). I understand the confusion with the 'low 1300's', thought a lot about that myself. The best I could figure was that it's on the rich side for the sake of engine longevity as he also mentions a 13:1 mixture during cruise. So , the plan was, emphasis on 'was' to run the engine ROP adjusted for the low 1300's, which seems to me a bit cold for best economy. But for the purposes of initial tuning looks like I'll be using that AFR after all.

Thanks again, Lance

Once you've established what your EGTs are for the AFRs you want to run at different power settings for your engine, you can just use EGT. Some people remove or ignore the wideband once past the initial tuning stage.