LSM11

[garfield at pilgrimhouse.com]

On Wed, 25 Mar 1998 22:22:46 EST, PNefdt <PNefdt at aol.com> wrote:

>  I've worked on the O2 system for about 2 years. I'm fortunate to work for a
>race dyno company, so have facilities available. Unfortunately all this work
>was done in company time, and I'd be hanged if I disclosed the 3d map,
>however, I can explain exactly how to go about doing it;

Well, that would be welcome; I've read about applying PID to temp
correction of O2 sensors, but I've ALSO been told by someone directly in
Bosch that in order to do that, you *must* have a t-couple ON the
element, not just measure the ambient EGT. So some light shed in this
area would be useful. If this is too nitty-gritty, we can take it
off-line, but I suspect there'll be enough interested to justify
on-group. Either way, I'm game.

>first of all, the facts.
>The LSM11 was designed to be temperature independant above lambda = 1.
>Below 1, it's just like any other old sensor, except very repeatable between
>any two sensors (which is the important part in production assembly of course)
>There are two totally different methods which can be used
>1. Stabilise the sensor at a known internal temperature, utilising a PID
>controller, to control the temp to 750degC (method used by a upmarket($5k)
>instrument)  
>2. Compensate for internal temperature (the method I followed)
>
>Method 1 is the easiest.
>I, of course, being a gutton for punishment, chose the more difficult way, and
>which I will explain.
>You will need to be conversant in Motorola asm for the 68hc11, if you would
>like to see/follow my code.
>
>I will prepare a mail attachment, which I can post to the list, or to you,
>whichever?

Or put it up on the ftp site for posterity, so when the topic comes
around again (trust me, it will, like everything else we discuss here!),
you can just point to the file. Whichever, I dunno fer sure what's best.
But knowing this group, I'm sure some who have a strong preference will
speak up. B)  If you could give a sorta step-by-step but still kinda
brief rundown of BOTH methods you mention above, that would be great.
The code would be good once the concepts are clearly understood.

I had a hunch your involvement was commercial; this is really sad we
can't share stuff like this, but I understand your employment situation.
One more vote for the NTK, I guess, eh? Cuz the horrendous price point
everyone wants for these meters is really shutting the individual
eXperimenter outta some of the fun stuff. Phoooeey.

Course, if we teach each other "how-tos", we can go catch our own fish
and still dine in style.

Garfield

P.S. I'm sufficiently interested in this topic that if we can achieve
critical mass on how the above two methods are done in practice, I & I'm
sure others can help to whip circuits together for play-time. Besides,
PID is an old analog trick from waaaaaaayback. Just don't see how to
apply it without some key hints. The more "modern" method I'd love to
understand as well.