[Wbo2] Using the Bosch WB sensor with L1H1 circuit board

Tue Nov 29 13:41:48 UTC 2005

I was really hoping someone would reply to this email, but no one did, so I'll ask some follow-up questions.  Your percentage errors are calculated directly on AFR.  One example you gave is 13.2 - 15% = 11.3.  However, is the error really in AFR, or is it in Lambda, current/voltage?

Also, the typical AFR curve is not linear.  It is better approximated as a piecewise linear.  If it is just a percentage of AFR, then the actual AFR error depends on whether you are rich or lean.

What happens at and around stoich?  If you're at 15.1 with one sensor, then is -15% to another really going to give you 12.83???

What about the errors in other fuels?  Methanol will give you the same voltage output at stoich as gas, but the real AFR is around 7:1.  So do these errors drop also -15%?

My point here is I'm not positive you can throw +/-15% (or any percentage) directly at the AFR measurement.  But, I don't know where the percentages you are quoting came from, a datasheet, actual testing, pulled from another thread?  If any one has any comments on this, it would be appreciated.

Steve

-----Original Message-----
From: wbo2-bounces at diy-efi.org [mailto:wbo2-bounces at diy-efi.org]On
Behalf Of Ron Vinsant
Sent: Friday, November 25, 2005 12:29 PM
To: Andrew Robertson
Cc: wbo2 at diy-efi.org
Subject: Re: [Wbo2] Using the Bosch WB sensor with L1H1 circuit board

Andrew, Agreed. Repeatability is the issue in the direct sense, not 
calibration. The problem I have is that the sensors go "bad" meaning 
their results are incorrect, or at least different, after a time. This 
is a big issue if you happen to use leaded gas, even once, such as on a 
track day, Now things start to change in the engine response due to 
contamination of the sensor (in a closed loop control system). Easy you 
think, I'll just replace the sensor.You install a new sensor and the a/f 
is still off (as measured by looking the spark plugs or egt).  You do 
some additional experiments and find that the heater current is a little 
different, the offset voltage is a little different, etc.Thus even if 
the sensor is calibrated at the factory there are many other things that 
can make the meter read differently than before.

If the meters vary by a significant amount, say -15%, and my air fuel 
was initially 13.2 then I could be as low as 11.3! If the meters are a 
15% span (+/-7.5%) then my 13.2 could be 14.19 or 12.21. If we had a ?5% 
meter then we would have 12.5 to 13.9. a/f.

If can see that there is a reasonable calibration ?1.5% is great, ?5% is 
probably good enough, then I would not worry as much about changing 
sensors.

The other item that is missing here is how the sensors age. Do they 
maintain their initial accuracy for their entire lifetime and then, at 
end of life just quit? My experience says that this is not the case. 
They start getting off calibration as they age. The dyno shop I do my 
set up at changes the sensor on the Motec (NTK) after 40 runs as the 
tech does not trust them after that.

I too want consistency. I would also like to understand what consistency 
I should expect. That is the information that is hard to find.

Ron

Andrew Robertson wrote:

> As a DIY enthusiast that uses the TE 2AO kit with 6066 to dial in my 
> comp car, accuracy at this level is purely academic - most DIY tuners 
> who dial in on the actual road aim for an AFR thought to correspond to 
> max power without really knowing if their target figure is in fact the 
> one that gives max power (i.e. is it 12.6?, is it 13?etc).  I have a 
> turbo car so I aim for AFR of 12 to 1 (i.e. slightly fat) - a 5 % 
> error would be of little consequence to me let alone 1.5%.  
> Stability/consistency of operation is more of a concern to me.  Just 
> wanted to add some real world context for the home tuner.
>
> Cheers, Andrew Robertson, New Zealand
>
> ----- Original Message ----- From: "Ron Vinsant" <ron at poweracumen.com>
> To: <bcroe at juno.com>
> Cc: <J_Holland at btopenworld.com>; <wbo2 at diy-efi.org>
> Sent: Thursday, November 24, 2005 5:33 AM
> Subject: Re: [Wbo2] Using the Bosch WB sensor with L1H1 circuit board
>
>
>> All, I just checked the Motec site and looked at their Lambda meter.. 
>> They claim a +/- 1.5% accuracy (depending on sensor 
>
>
>
> _______________________________________________
> Wbo2 mailing list
> Wbo2 at diy-efi.org
> http://lists.diy-efi.org/mailman/listinfo/wbo2
>
>

_______________________________________________
Wbo2 mailing list
Wbo2 at diy-efi.org
http://lists.diy-efi.org/mailman/listinfo/wbo2