AEM or Motec?

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,749
Location
Calgary, Alberta, Canada
The biggest problem with the T51 replicas is the stub exhaust makes it difficult to install a wideband sensor due to reversion air contaminating the sensor. The Honda uses a more convoluted exhaust collector prior to the stacks than the Suzuki V6 which may offer enough length to negate contamination from the atmosphere during valve overlap but you'd want to confirm on a dyno the readings are accurate. If you build an airtight shroud around the exhausts to shield the sensor, this might affect how the engine runs and is not practical to fit on the aircraft for flying, maybe just for ground runs.

Our experience with widebands and leaded fuel has been mixed. I have seen them last as little as 3 hours and as long as 100, depending on TEL concentrations. Long enough to complete a tune with our system but you may not be able to rely on readings hours down the road. The latest PLX wideband kit allows you to test the sensor health which is a great idea. If you think you are getting good readings but are actually not, bad things can happen if you chase the tuning the wrong direction.
 

TXFlyGuy

Well-Known Member
Joined
Apr 25, 2012
Messages
1,589
Location
Republic of Texas
The biggest problem with the T51 replicas is the stub exhaust makes it difficult to install a wideband sensor due to reversion air contaminating the sensor. The Honda uses a more convoluted exhaust collector prior to the stacks than the Suzuki V6 which may offer enough length to negate contamination from the atmosphere during valve overlap but you'd want to confirm on a dyno the readings are accurate. If you build an airtight shroud around the exhausts to shield the sensor, this might affect how the engine runs and is not practical to fit on the aircraft for flying, maybe just for ground runs.

Our experience with widebands and leaded fuel has been mixed. I have seen them last as little as 3 hours and as long as 100, depending on TEL concentrations. Long enough to complete a tune with our system but you may not be able to rely on readings hours down the road. The latest PLX wideband kit allows you to test the sensor health which is a great idea. If you think you are getting good readings but are actually not, bad things can happen if you chase the tuning the wrong direction.
This is a valid observation, to be sure. The suggestion in from my exhaust fabricator is to mount the sensor within 4 inches of the manifold using a bung to protect from the heat. At 3,500 to 4,000 rpm, my AEM tuner is confident that there will not be any mixing of ambient air into the exhaust stack. As we will have a multi-port exhaust (J35A4), there will not be a "log" or "collector" where the exhaust will exit from. We will have a series of one-into-two exhaust pipes. I estimate that the total length of each pipe from flange to exit point will be 12 inches (+/-).

The plan is to monitor two cylinders, one on each side. We will do the initial map tuning using a factory type collector exhaust, then fine tune with a short stack exhaust. We can program to run sans O2 sensor if we need to.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,749
Location
Calgary, Alberta, Canada
This is a valid observation, to be sure. The suggestion in from my exhaust fabricator is to mount the sensor within 4 inches of the manifold using a bung to protect from the heat. At 3,500 to 4,000 rpm, my AEM tuner is confident that there will not be any mixing of ambient air into the exhaust stack. As we will have a multi-port exhaust (J35A4), there will not be a "log" or "collector" where the exhaust will exit from. We will have a series of one-into-two exhaust pipes. I estimate that the total length of each pipe from flange to exit point will be 12 inches (+/-).

The plan is to monitor two cylinders, one on each side. We will do the initial map tuning using a factory type collector exhaust, then fine tune with a short stack exhaust. We can program to run sans O2 sensor if we need to.

12 inches from sensor to atmosphere should work ok once the rpms are up.

We never run at high power using O2 sensor feedback. If the WB dies, usually they indicate rich and the ECU will trim fuel out until the engine blows up.

Will be interesting to see the running/ tuning differences between the OEM collector exhaust and the stacks. On the Suzuki, there are definite peaks and valleys in the curve as the stacks pipe in and out with rpm changes, altering the VE.
 

TXFlyGuy

Well-Known Member
Joined
Apr 25, 2012
Messages
1,589
Location
Republic of Texas
12 inches from sensor to atmosphere should work ok once the rpms are up.

We never run at high power using O2 sensor feedback. If the WB dies, usually they indicate rich and the ECU will trim fuel out until the engine blows up.

Will be interesting to see the running/ tuning differences between the OEM collector exhaust and the stacks. On the Suzuki, there are definite peaks and valleys in the curve as the stacks pipe in and out with rpm changes, altering the VE.
It is my understanding that we can program default values so that if the sensor dies, or suffers from erratic data, the ECU will continue at a pre-specified value with regards to air/fuel mixture. Our program will have a very small window, so that if an exceedance occurs, we will resort to default mode operation.

This also applies to most of the other sensor readings as well with regards to failure modes and default value settings.
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,749
Location
Calgary, Alberta, Canada
It is my understanding that we can program default values so that if the sensor dies, or suffers from erratic data, the ECU will continue at a pre-specified value with regards to air/fuel mixture. Our program will be have a very small window, so that if an exceedance occurs, we will resort to default mode operation.

This also applies to most of the other sensor readings as well with regards to failure modes and default value settings.
You can do defaults for most of the other sensors but usually not O2 being used in a closed loop mode since you don't have any idea if the data is valid or not. There is a fine line between living and dying with regards to AFRs and pistons.

With coolant temp, air temp and MAP you can default the open and shorted ends to something to keep the engine running. Even getting as lean as 13 to 1 at full power may eventually cause piston failure. If the software is very smart, it could check by adding some trim at anything leaner than 13 and observing the AFR for the expected % change. If it did not see it, it would default to open loop mode.

One other thing to watch for with widebands is if you get a misfire, it shows lean when in fact, the AFRs could be normal or rich. If it is not smart, it will keep trimming richer until you actually get a rich misfire and eventually could shut the engine right off from over-fueling.

Lots of things to consider here and many people in the past have not- resulting in forced landings or worse.
 

TXFlyGuy

Well-Known Member
Joined
Apr 25, 2012
Messages
1,589
Location
Republic of Texas
Another ECU we are looking at is LINK. I know of extremely good performance with this in aviation applications. Another motor-head engine builder guru suggested using Delphi, as many race types use this with their LS Series V8 engines. Completely programmable, with lots of options for data logging.

And, SDS has not been completely ruled out.
 
Group Builder
Top