Honda UEGO Sensor...

[orlin steven jared]

> converter with linearization done by the micro.  The calibration table could 
> be stored in non-volatile EEPROM.  The parts cost should be less than $50 
> ($5 for the uC, $1 for EEPROM, $5 for display/driver, $10 for ADC, $3 for 
> voltage reference, etc.).  If Honda is putting the electronics in their 
> cars, then the interface cost has got to be reasonable.

This would be great.  I hope if you pull it off you will make the plans
for building the system known to the group :)

> A few questions:
> 1.  Is each sensor calibrated with each interface box?  For example, with 
> the NTK or Horiba boxes, when you replace the UEGO sensor, do you need to 
> recalibrate the unit?  Hopefully the answer is NO.

Sorry, the answer is yes. :(   Accuracy will be off if you do not
calibrate the box after you replace the sensor.  I have used the Horiba
Mexa 101 and 110 units.   For one of these, the sensor is supposed to
COME with a factory calibration sheet unique to that sensor.  And you
use those values in the readout box or something.

But, I believe (can't totally remember, it was a while ago) that the
other unit requires no calibration sheet, and can be calibrated with
the box.

Also, there is a quick and dirty way of calibrating these things for
o.k. accuracy, but the *real* way of calibrating a UEGO is with test
gases, which cost $$$$.   I'm pretty sure this is why each Horiba sensor
comes with the calibration sheet, so you don't need to calibrate
with the gases.  But, if for some reason you should ever change sensors,
and install a used one, then you would need to recalibrate.

Since each UEGO is hand made, there will be variations.

> 2.  What kind of accuracy do the manufacturers claim for the sensors?  And 
> what accuracy do the interface box manufacturers claim?  This project would 
> require the initial accuracy to be decent, without having to calibrate 
> (tweak) each unit against an expensive lab standard.

Unfortunately, I don't think anyone can give you a really good quote
on accuracy without calibration using the test gases.  When I tried
to calibrate one of these sensors in the lab, I called Horiba, and
talked with a technician for a while as he walked me through the
'quick and dirty' procedure.  I asked him about the accuracy, and he
said probably 5-10%, but he was really hesitant about it.   It is
a hard thing to say, unless you talk to the test engineers who are
testing these UEGOs in the NGK/Horiba labs.


Steve