View Single Post
Old 07-24-2003, 02:43 PM   #36
Scooby Newbie
Member#: 40416
Join Date: Jul 2003
Location: CA
68 Porsche VW Bus


Hi Jon,

Here are some answers:

that info was derived from their well-written user's manual.
Thank you.

Until now, I'd never read about any electrical property or relationship of either the sensor cells or their heater that varied with a change in gas pressure but didn't vary with a change in partial oxygen pressure or gas temperature.
The properties that change with gas pressure are sensor current (raw lambda) and sensor response time. They do not change proportionally. So from the difference pressure compensation data can be derived when comparing both to the free-air current and response time (free air assumed at normalized sea-level pressure).

Although there appear to be some problems getting this technique to work well with the Bosch sensor, I take it that LM-1's pressure compensation technique DOES work well with the NTK sensor?
No, for time and money reasons we have not implemented it for the NTK. To realize the pressure compensation we need to make a lot of tests with the sensors at different pressures at different lambdas. Because we are shipping with the Bosch, we put the majority of our development time there (we have been a non-profit organisation for too long ;-) ).
The temperature regulation for the Bosch actually does very well compensate for the higher temperature dependence of the Bosch sensor. It actually works much better than the NTK temperature compensation. The NTK is spec'd for a constant htr voltage of 10.5V, typically using the positive temperature coeff. of the heater to self-regulate. During initial heater calibration we run the NTK at a constant 10.5V after warmup for 20 seconds, then measure the heater resistance to ~5 mOhm precision (quantization accuracy, not absolute accuracy). This value is stored and used to regulate the heater of the NTK to a constant resistance using a PID algorithm. Because there can be a temperature gradient between heater and pump cell, the NTK will not work as precise as the Bosch implementation.
We actually considered dropping the NTK support completely.
With the limited number of NTK sensors tested (5) we found a minimum of about 0.67 Lambda (9.85 AFR) for the NTK, assuming the max. negative current of the NTK is the same as the worst case free-air current. We have much less manuf. data for the NTK than for the Bosch, that's the second reason we did not spend too much time on it. As I understand, Horriba has the rights to use the NTK sensor for AFR measuring in the US. I assume that's why there is so little data available. My hat off to the original designers of the DIY-WB. It must have been difficult to derive a working design from so little data.

* Registered users of the site do not see these ads.
Klaus is offline   Reply With Quote