braindrizzle

[bcroe at juno.com]

Those are good thoughts, but here are some others
reflecting my opinions.  

A digital lookup table can modify a curve.  If the 
curve relating 2 different voltages ever changes
from positive slope to negative (or the reverse), you 
not only need the level, but also the slope, to convert 
to a new waveform.  The rising and falling parts of a 
sine wave have the same voltage levels, but the
2 corresponding cosine voltages are different.  A
lookup table doesn't cut it, a processor could do it.

A simple translation of voltage levels could be done 
much more simply with an OP amp and a couple of
resistors.  I already proposed (to BobR) a 
"universal" pattern where you chose the voltage
translation you want from a table, and it tells you
which resistors to solder into the pattern.  Inverting
or non, no problem.  There is already some space 
for that on the WB pc board, but not the pattern.  It
could also be included on a display board.

If you are interested in the LED bar graph, you already 
have a completely linear fuel added to display change
function.  There would really be nothing for a more 
complex circuit to do.  The operation of this is online, 
and although the theory is a little messy, the circuit is
straightforward.   By the way, this has been tested for
dozens of hours in actual vehicle engine tuning, and
the conclusion (so far) is that this is a far more useful
output than a digital reading of any kind.

If you need a digital display of A/F, you will need some
kind of conversion/look up table.  Instead of reinventing
the wheel, I suggest someone reprogram that JAYCAR
unit which could do the job.  

If you want to do really fancy frequency domain or level
clipping stuff, or RS232, you will need something more
fancy.  But I think these functions are not even defined
now, and would take lots of time to develop.  In the mean
time my opinion is that the simpler functions (and possibly
those most in demand) should be made available as soon 
as possible.  It would be ideal if these shaped up about the
time people are getting their basic WB boxes operational.
I built these things into my WB prototype, but putting it
all on the first board would have delayed an already too
much delayed project.  

Bruce Roe

On Fri, 28 Sep 2001 09:45:56 -0700 (PDT) Phil Hunter
<ilphayunterhay at yahoo.com> writes:
> I thought I'd offer a brain-drizzle (as opposed to
> a brain-storm) for the list's consideration, seems
> pretty obvious to me:
> 
> Imagine an Analog-to-Digital convertor, the output
> driving the address lines of an EPROM who's data lines
> drive a Digital-to-Analog convertor. In essence, any
> input voltage can be mapped to any output voltage
> depending what you program into the EPROM. A signal
> can be inverted, a sine wave converted to a cosine
> wave, and a wide-band O2 sensor output can be mapped
> into a narrow-band sensor output.
> 
> Now this is generally too simple for real life, things
> like settling times, access times and no sample-n-hold
> cause glitches in the output. But if we substitute a
> microcontroller for the EPROM, better yet a uC with
> built in A/D such as in the PIC family, things can
> stay fairly simple. I don't know of a PIC uC that has
> a D/A output other than Pulse Width Modulation, so we
> will still need an external D/A chip.
> 
> Assuming this is a good idea, I'd like to see the
> group focus on something like this before doing
> complete end-to-end projects such a WB-to-LED bargraph
> or WB-to-LCD. Get a core of hardware & software that
> has a lot of commonality and everyone more-or-less
> understands then build from there. I can invision
> software modules such as A/D, lookup, PWM output,
> serial D/A chip output, parallel D/A chip output,
> RS-232 output, etc. Hardware can be a 28 or 40 pin uC
> driving a 16 pin D/A for a lot of flexibility, or an
> 8pin uC driving an 8pin serial D/A for simplicity.
> Depending on what the user wants to do, they can
> pick-n-choose what they need, then build it. You want
> an analog display of A/F, you can probably do it w/
> just an 8pin uC and PWM output. You want a digital
> display of A/F, you could program the lookup to output
> in steps such that 1.125V is 11.25:1 and you mentally
> move the decimal point one position on your DVM. I
> want to eventually substitute the NB sensor w/
> a WB sensor, so I'd use the bigger uC driving a
> parallel in D/A chip so I don't have to worry about
> the latency of the serial input D/A chip confusing
> the ECM.
> 
> Perhaps the WB-LED & WB-LCD projects could be
> redesigned in a modular fashion, one using the large
> PIC and one using the small, then the user could
> "stuff" them as needed, only putting in the parts
> to accomplish what they want.
> 
> Get the picture...Kodak?
> 
> phil
----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org