One or two oxygen sensors

[Claudi]

I'm in the planning stage for my 1983 Porsche 911 EFI conversion.  I intend on using the Fury.  As there are two cylinder banks; is it important, or are there advantages to, using two oxygen sensors?  One for each bank of cylinders?  As the Fury has one Wideband sensor input I'm assuming the second sensor would need to be wired through a CAN Lambda unit.

[Adamw]

If it has a common plenium and single throttle then air/fuel distribution will be decent enough that I would be happy with a single sensor.  If you are going to ITB's or a throttle each side then probes on each side using dual channel closed loop can improve small throttle/low speed driving where there is the largest side to side error.  

Yes for the Fury you would add a CAN lambda for the 2nd probe.  

[Claudi]

I'm planning on using the AT Power ITBs, so ya, I hear you saying I should control the banks separately.  Just to be clear (that is to say, I'm quite new to this), when you say "dual channel closed loop", you mean two CAN Lambda units on a common CAN Bus?  Or is there another configuration I'm missing?

Thanks for the help!

 

[Adamw]

Dual channel closed loop is just the software strategy - this means you assign which injectors and which oxy probe belongs to each bank then the ecu will trim the fuel on both sides independently to keep the air fuel ratio as close as possible to target.  So if one side is lean and one side is rich then it will add fuel to one side and remove fuel from the other.  

Single channel closed loop would be where you have only a single oxygen sensor in a common collector - or you have one each side but the ECU averages the reading from both, and applies the same fuel correction is applied to all cylinders. 

Since you have a Fury you would just need one CAN lambda connected for dual bank control.    

[Claudi]

So then this leads right into my next problem, question; is the termination resistor inside the Link Fury ECU selectable?  As it's a Porsche with the engine in the back, I intend to install the ECU under the passenger seat (where the original "computer" was installed), and I'd like to install an OBD port under the dash (and possibly a display, we'll see), the ECU would naturally be in the middle of the bus.  I would prefer to have everything on the same bus as I don't want to include a third connector in the permanent installation.

[Adamw]

3 hours ago, Claudi said:

is the termination resistor inside the Link Fury ECU selectable

No, it is fixed.  

3 hours ago, Claudi said:

the ECU would naturally be in the middle of the bus.

Doesnt need to be.  CAN bus is often run backwards and forwards in the same bundle when you have long stubs etc.  So something along these lines:

[Claudi]

Now that I rearranging my design ideas about oxygen sensors; is there a maximum length for the analogue oxygen wiring?  As I could wire at least one of them into the ECU which is inside the car, this run would be pretty long.  That said, I'm even tempted to put the CAN Lambda unit inside too, the Porsche engine bay can get pretty hot and I wouldn't want to shorten the life of this unit. 

[Adamw]

1 hour ago, Claudi said:

is there a maximum length for the analogue oxygen wiring?

Probably, but I dont think you will have an issue in a typical car.  My dyno wideband has 12M of cable between probe and controller.  

 

1 hour ago, Claudi said:

I'm even tempted to put the CAN Lambda unit inside too, the Porsche engine bay can get pretty hot and I wouldn't want to shorten the life of this unit. 

I wouldnt worry about that too much, all our components are AEQ100 or 200 grade 1 rated (-40-125°C), the CAN lambda is nearly always hanging next to the exhaust or turbo so I dont think yours would be much different than normal...  If you did extend the cable, be aware you must retain the factory connector on the sensor side as it has the calibration resistor inside.   

[Claudi]

I ordered a custom exhaust from Germany which arrived this week.  It looks amazing, only, they forgot to install the O2 sensor port in the right exhaust manifold.  Argh.  I'm in Canada so shipping it back is a huge hassle.  My welding skills aren't up to this task.  All that to say/ask; what are the consequences of only having one O2 sensor representing half of the exhaust flow?  Is there a setting for this in the ECU?  Is it huge bad?, or just not best practice?

[Adamw]

Its not the end of the world - I just find with ITB's on "2 bank" engines the throttle linkage will never be perfectly "in sync" side to side. At idle you can get them nicely balanced since they are sitting on the stop screws, but at partial throttle just above idle and light cruise you only need a very small difference in blade position from side to side to cause a large difference in airflow - which will mean some cylinders will be richer or leaner than the average.  Depending how much variation there is you can get noticeable flatspots and sometimes "spit back" etc especially when cold on the cylinders that are getting more air.  Obviously individual bank lambda control doesnt fix the air flow difference - but it does fix the mixture difference which usually solves any noticeable drivability effects.

  Having said that I have done an ITB 911 with a single common lambda and it drove pretty nice - I did spend a fair bit of time messing with the linkage system to get it as similar as poss on each side with minimum backlash etc.  It was a jenvey kit from memory and the linkage wasnt great design.