Backfire fix

[Mi_ka]

On the lambda sensor diode mod, I do not expect this to work.
Most probably the ECU will through some check engine code related to bad sensor.

The narrowband zirconium sensor gives out ca. 0,9 Volts at 14,3 AFR, 0,15 Volts at 15.1 AFR and 0,450 Volts at a very narrow range around 14,7 AFR
(nice related thread here: http://www.explorerforum.com/forums/showthread.php?t=149917&page=2)
Most ECUs do not aim for this very narrow 0,450 V at 14,7 AFR but instead trim their injection values at about 0.5-1Hz in order to get rich-lean-rich-lean signal from the sensor in essence hovering their closed loop operation at the range 14,5-to-15,0 back and forth all the time.

There are low barrier voltage Schottky diodes at 0,2 Volts or plain silicone diodes at 0,7 volts ("knee" voltage) always depending on current passing and triggering the barrier phenomena - these values are lower for very small currents in the microamp region.
Putting a diode in series, I expect the maximum sensor voltage be trimmed to ca. 0,75 or 0,25 Volts depending on selected diode and will never achieve the specified voltage value. So the ECU will suppose the sensor is toast.

 

[kiwi dave]

What about using a simple resistive potential divider to lower the Lamda voltage? Currents should be minimal, so I wouldn't expect the arrangement to load the sensor or upset the ECU.

 

[Mi_ka]

What I post here is theoretical knowledge. I may be wrong and it may work as posted by Micha.

However, if a divider is used (no idea on proper resistance or current values not to mess with sensor operation), the problem remains the same: most ECUs look for a rich condition close to 0,9Volts and a lean condition close to 0,15 Volts.
They do not look for the very narrow range of 0,450 Volts. As far as I have read around that is...

 

[PP-1753]

If you want to play with the Lambda Sensor - just go get a FatDuc - all the work's done for you...

 

[DanPez]

For the DIY .......
The lambda sensor has an infinite impedance and is a generator of oscilating voltage as per Mi_ka explained.

If electronic components are to be used to modify the O2 sensor signal going to the ECU .......the sensor can be replaced by:


The LED limits the voltage to 0.7v and you have a variable oscilating circuit by adjusting R1 between CW ground (lambda signal 0.0v) and CCW 12v (lambda signal 0.7v)
Found this circuit at http://mechdb.com/index.php/O2_sensor_simulator

Cheers

 

[micha]

I just noticed a 2010 Norge has two O2 sensors, left and right. I wonder if injectors are timed individually, probably so.
Yes we can go buy a Fatduc but what fun would that be? We want to exercise our pumpkins not wallets mate!
I used to live in Vienna in early 90s, prostitution was legal but I always taught a proper girl freebee with some brain effort was way more fun. :woohoo: Actually I have a brand new Fatduc in a box for someone but I don't want to reverse engineer someone's work.
What are AF voltages on a wideband? Does the voltage drop with richer or leaner mixture? A local VW scrap yard owes me some money I might go and pick one up for the experiment.

 

[PP-1753]

Don't do it - this is to simulate the second O2 sensor in a twin O2 sensor circuit so as not to get engine errors IN A BLOODY CAR!

Our Guzzis don't have this feature. A FatDuc allows you to slightly trim the O2 feedback to the ECU to correct leanness at low throttle openings - this circuit will give you one reading and that's all - it may work at one rev and one load on the engine. It needs to work at various throttle openings up to around 1/4 throttle.

A FatDuc works on the standard ECU. It is not needed if you have Todd's GuzziTech ECU as it allows this to be varied through his system.

Sorry to be a wet blanket but putting this circuit in will be dissapointing at best and may actually do damage to the fuelling of your Guzzi - or maybe something worse!

Just get a FatDuc or go the whole hog and go visit Todd...

 

[Mi_ka]

What are AF voltages on a wideband? Does the voltage drop with richer or leaner mixture? A local VW scrap yard owes me some money I might go and pick one up for the experiment.

The wideband sensor is a different apple requiring sophisticated and much more expensive control.

In essence it is two zirconium narrowbands sensor elements properly molded together in porous material:
One narrowband element chamber, referenced to air, is being fed by the second element which is pumping oxygen ions from/to the exhaust gases in order to keep the first element happy at a steady 0,450 Volts which means almost no free oxygen ions in its measuring chamber.

The wideband controller adjusts the pumping electric current of the second cell properly and decides the AFR of the gases based on this current. It usually reports its measurement in the form of 0-5 Volts signal to the ECU. AFR range is sually from 10.0 upwards.
Total time lag for the measurement is in the range of 0,10-0,15 seconds.

(Please not that the nice Innovate wideband controllers follow a different PWM approach which is said to be much faster, around 50-70 msecs. That is why they need free air calibration. It is not the intended Bosch use of the sensor and maybe it affects longetivity, maybe not).

Also, the wideband sensor needs good heating control in order to keep the temperature above the working 600 Celsius and not too hot (under 800 Celsius I think) which means expensive and sophisticated heater control as well. Expensive is not really expensive but not just a 12Volt supply like on the 4 cable narrowband sensors - it requires electronic supervision.

All this adds up costs and hardware and makes the wideband system more sensitive.
Widebands have 5 cables. I think I have read about a 6 cable sensor somewhere but may be mistaking....

No, I am no guru at all, just read most here: http://wbo2.com/lsu/lsuworks.htm
and the Bosch datasheet here among other places: http://www.starmazda.com/competitor...Bosch Lambda Sensor/Y258K01005e03mar21eng.pdf

 

[micha]

Here is one with a 2.2k resistor installed, baffle removed, minor pop can still be heard. This bike used to pop like mad. Best part is the transition and steady throttle is smoother. Resistor fix works better on stock bikes though, I’ll try to make a before and after video on a stock bike and post it here.

[youtube]mBaKHhq_UUU[/youtube]

Here is one without the resistor but with baffle still in:

[youtube]k2HwIdAlayo[/youtube]

 

[micha]

So I went to test a spare fatduc today, plugged it in a 1200 Sport and rode it for about 10 minutes. What is that thing supposed to do? Didn't do a thing for me. Checked exhaust on a two gas machine and there was no change. I had to look into it further...that thing is variable resistor! How is a variable resistor supposed to alter voltage on a narrow band sensor's output? Well, it doesn't. There is no voltage drop trough it.
Maybe that thing does something on a Duc. Did anyone ever test this to see what it actually does on our bikes or did people just start plugging it in because the connector matches? :unsure:

 

[PP-1753]

It Does Work! On mine anyway. It only works from idle to 1/4 throttle and richens up the mixture to stop it from being so snatchy at low throttle openings. After 1/4 throttle it does diddly squat. Set it at mid point and adjust from there - the bike takes around 10-15 minutes to adjust to each setting of the FatDuc. you also need to stop the engine and re-start it after changes.

 

[micha]

Tony, if you have diagnostic software hook it up and observe the pulse width numbers. They don’t change with fatduc vs. without. Fatduc is a resistor, they don’t change voltages and ECU’s use voltages for AF references. If our bikes had a feedback system you’d need a resistor to modify a 02 signal. You can use a resistor on a temp sensor because that is a feedback circuit but not on an O2 circuit. These are basics of fuel injection.

Turning the key on/off activates short enrichment circuit on next start so what you are feeling after on/off is a slight enrichment built into the map, not fatduc. Check it with diagnostics or with exhaust gas analyzer. I have both on hand here. As far as I go fatduc is a waste of time at least on a Goose. Duc forums don’t think much different either. But in the end if it works for you more power to you!

Moving along…as Mi_ka’s theory confirmed to modify a 0.15->0.9 volt (lean->rich) analog signal of the lambda sensor we would possibly need to use a diode with a forward voltage drop of 0.2 or 0.5 V except in that case we are risking an ECU error (#21 Lambda or #55 self adaptability).

This remains to be tested in praxis. Anyone here ride a Harley with O2IED? I think that is what that is, voltage reducer except they have no heater on their sensors. I have some 3A diodes with a 0.5V drop and I’ll get us some OA90s so we can do some more tests.

I like DanPez link of a circuit board mimicking the 02 sensor but it is a bit too much for I had in mind.

 

[davem]

I am still waiting to get my B1200 onto a dyno with it's recently installed Fat Duc to confirm if the closed loop fuelling has been affected, which I suppose will settle any arguments about it's worth.
If it is doing nothing why are errors thrown up when set for rich running?
Has anyone stripped a Fat Duc to see what's in it? There may be more than a variable resistor inside the wrapping.

 

[Foxster]

Hi guys!

My brother is Davem and he asked my opinion on this really interesting discussion because he knows I have played around with the same stuff on my v-twin, a HD Fat Bob. As someone mentioned HDs in the thread he thought I might have something to bring to the discussion.

I do know of people that have tried the method of fooling the engine temp sensor on Harleys and in some circumstances it has been seen to work. I suppose the problem is that temp is such a variable thing; it will vary by 30°C between locations. So someone at, say, 20° might get some useful result but someone at 10° might find it’s now too rich.

The only similar commercial device for Harleys that I know of is the Xieds made by nightrider.com. This uses the other mechanism discussed here: altering the offset of the voltage from the narrowband O2 sensors by the use of a pair of resistors to create a voltage divider. It’s a simple circuit that looks like this:

The pair of resistors is tucked away in some shrink wrap. I assume that the Fat Duc device is something very similar but with a variable resistor in place of one of the resistors.

Here’s how it works...

If resistor A is half the resistance of resistor B then 2/3 of the voltage drops across resistor B i.e. you drop the signal voltage that the ECU sees to 66% of what it would have seen without the resistors.

Now, 0.500v is the voltage that the sensor puts out at stoic (AFR of 14.7:1) and this voltage increases with richness. So what we can do is make the ECU think that the engine is running leaner than it really is by dropping the voltage from the sensor. The sensor will put out 0.750V at 14.2:1, so with a 2-1 voltage divider this will drop to 0.500V. In other words the ECU will see a real AFR of 14.2:1 as stoic and run the engine at that AFR.

Some other ratio of resistors will give:
A 60% divider reduces .850v to .500v, emulating a 13.8:1 AFR
A 63% divider reduces .820v to .500v emulating a 14.0:1 AFR
A 66% divider reduces .750v to .500v emulating a 14.2:1 AFR
A 70% divider reduces .775v to .500v emulating a 14.1:1 AFR

Of course, this only works when the O2 sensors are in use i.e. during closed loop (lower revs and constant throttle). So it’s not a performance mod per se. What it helps to do on Harleys is drop engine temp, eliminate low speed surging and reduce popping on over-run.

However, the ECU on late model Harleys is a complex beast and has quite a mind of its own. If it finds that the closed loop requires a certain amount of fuel to achieve the target AFR then apparently this ‘bleeds over’ into the open loop operation. Personally I’m dubious of this but willing to be convinced; I haven’t seen this effect on a dyno but who knows?

It depends, I think, just what you are trying to achieve. If you want to tweak things because you have the issues I mentioned above then this is a pretty good method of going about it. If you want to play with the AFR in the open loop operation area for performance gains then it’s not for you; you want some form of fueling solution such as Power Commander, factory tuner or whatever else is available for Guzzis.

Anyway, that my 2C worth.

Cheers,

Rich

 

[kiwi dave]

An elaboration on what I was suggesting a few posts ago.

Whatever you try, I reckon it is always going to be a compromise, as it will be hard to get the system to track properly, whatever you consider "properly" is.

Personally, I prefer the GT-X ECU reflash with a PC5 with AutoTune option. It may be expensive, but it sure does work.

My $0.02 worth too. YMMV.

 

[PP-1753]

A narowband sensor simply supplies a set voltage for a given AFR. The FatDuc fiddles with this voltage to pretend that the Oxygen in the exhaust is different to what it programmes to the injectors. This then allows the ECU to richen up the mixture accordingly - not rocket science. From my understanding, there is not "pulsing" just a variable voltage output depending on the amount of Oxygen in the exhaust - it varies due to engine load and throttle position. The Oxygen sensor is ignored at throttle openings over 1/4.

As Foxter points out, a FatDuc is simply a variable resistor divider instead of the fixed divider as suggested in the diagram. This can be adjusted for best low throttle openings.

A word of warning - no two Guzzis are the same, therefore it may do diddly squat for your bike and work might fine on someone elses. My experience is that it works on mine...

 

[Roblatt]

A word of warning - no two Guzzis are the same, therefore it may do diddly squat for your bike and work might fine on someone elses. My experience is that it works on mine...

I've been following this thread in the vain hope of learning something. So...

Pardon my ignorance. I have heard this before. But how can 2 supposedly identical bikes have a different result to the same input? Modifications aside.

Robert

 

[Mi_ka]

But how can 2 supposedly identical bikes have a different result to the same input? Modifications aside.

Minor differences between cylinders/pistons/gasket thickness & assembly torque/piston rings etc (parts manufactured at the end of life of machinery tools vs made by brand new ones), minor differences at the various sensors and injector performance.

If one bike at the production line has gotten all parts made at one end of their tolerances while another one has gotten all parts at the other end of the tolerances spectrum then these two brand new bikes with the same fuel/tyres at the same place & time may have a 3% performance difference (maximum allowed on the production line I think in order to get the model certificate of conformity - I also think I once heard that it used to be 5% or so a decade ago).

3% may not sound much but this means a 100hp cerified bike can have 97hps or 103hps, which is a 6hp difference.

If both bikes are dynoed and the various dyno procedures tolerances are added then you may get a bike at 105hps and the other at 95hps and have a bunch of confused guys scratching their heads... :roll:

From my understanding, there is not "pulsing" just a variable voltage output depending on the amount of Oxygen in the exhaust

The transistion from lean to rich signal is too sudden and most ECUs do not look for the 0,450 Volts - just hover between just a tad rich to just a tad lean all the time. You see, just a few millivolts changed constantly & abruptly because of slack at eroded connectors could confuse the ECU if it tried to aim for the 0,450 Volts condition. So it plays safe and looks for >0,7 and <0,3 or so..
All this in theory of course - I have never hooked up an oscilloscope on a working ECU myself.

 

[PP-1753]

What Mi_ka says...

in the specs for the 2 valve Sport "more than 95 HP" would confirm the individuality of each of our bikes. Anyway, none of us purchased a Guzzi to be mainstream, now did we! We wanted to be individualistic - and that's what we've got.

Enough of this rambling, I need to have breakfast, jump in the shower, and go for a ride - shame it's to work, but better than not riding at all. At least I get to enjoy my 41km of individualism :woohoo:

 

[Roblatt]

Gentlemen,

I have learnt something.

thank you

Robert