GM ALDL interface

[Terryk]

I don't know, I though a single $.01 transistor was about as elegant as you
could get. The opto is overkill. Powering the led might be a problem,
although some lpt lines could be used. But remember that CE data (D) is
inverted from serial data (E/M). That will need accounting for in the
interface or the code (preferred).

TK

I'll throw my C code out there if anyone wants it. It is processor speed
independent and handles the ALDL mode and CE/Serial data lines.

----------
> From: Jennifer Rose <javer96 at snowcrest.net>
> To: diy_efi at efi332.eng.ohio-state.edu
> Subject: Re: GM ALDL interface
> Date: Wednesday, December 10, 1997 7:44 PM
> 
> At 09:49 AM 12/10/97 -0800, you wrote:
> >Frederic Breitwieser <frederic.breitwieser at mcione.com> wrote:
> >>I created a single-transister circuit to make the proper voltage
> >>adjustments (12v to 5v), however over the next day or to I'll have an
> >>opto-isolator circuit that's even better.
> >
> >There are some ECM's where the ALDL 160 data swings between 0 and 5
volts.
> >Both of the ones I've looked at (2.5 L4 TBI and 2.8 V6 MPFI) do that. 
If
> >your's goes to 12 volts, could you trace the ECM schematic for that
output
> >back to a driver chip?  What series and pullup resistors does it use and
> >so on.  I'd like to add support for it in my ALDL interface.
> >
> >Both of the previously mentioned ECMs have an open collector driver with
a
> >1500 ohm pullup resistor to five volts.  They also have a series
resistor
> >between the driver and the ECM output - 300 ohms in one ECM, 120 ohms in
the
> >other.  This means that the output switches between 300 (or 120) ohms to
> >ground and 1800 (or 1620) ohms to five volts.
> >
> >If you are building an opto-isolated interface without a power source on
> >the ECM side, you have to deal with the very limited current available
> >during the 'high' state.  When driving the LED portion of an
opto-isolator,
> >there is a maximum of 1.6 mA available.  I found that most
opto-isolators
> >won't work or are too slow at such low current levels:  low gain photo-
> >transistors essentially don't work at all; high gain phototransistors
are
> >too slow (for the 160 Hz signal!); even the photodarlingtons are too
slow.
> >
> >Just about the only opto-isolator that worked is an H11L1 "Schmitt
trigger".
> >(They are $0.99 each from Jameco.)  It is meant to work on exactly 1.6
mA!
> >It is also good for up to 1 megahertz, so it even works for the 8192
baud
> >signal.  Hook the LED side directly across the ALDL pins A and E with no
> >extra series resistor.  To handle the 8192 ECMs, I've placed an 820 ohm
> >resistor between the M and E pins, although this hasn't been tested yet.
> >
> >The H11L1 output side has a unique pinout - it needs power and ground,
in
> >addition to an OC output.  It can run with a supply voltage of 3 to 16
volts.
> >I'm running mine on power stolen from a Mac's serial TxD+ and TxD-
signals.
> >This results in a 6 volt supply (actually +3 and -3.)  You can do
something
> >similar on a PC.
> >
> >>While there is nothing wrong with a
> >>complex circuit that converts to a serial port, this solution is just
more
> >>software based.  There's merit for both actually.
> >>
> >>I prefer the software method only because I can shove the teeny-tiny
> >>interface into the cable end rather than having a box hanging off the
> >>notebook.
> >
> >You can pipe the opto-isolator output to a parallel port and use a
> >software polling approach.  It is also just as easy to connect to an
> >RS232C serial port and use the UART hack I posted earlier.  An
opto-isolator
> >works great as an RS232 voltage level shifter.  It also provides the
correct
> >number of signal inversions for the 8192 mode (I think) or the UART
hack.
> >
> >PS: If you can't find an H11L1 locally, the NTE3090 is a direct
replacement.
> >
> >PPS: Has anyone found a source for the female side of the ALDL
connector?
> >The pin spacing is non-standard.  I doubt that even a silicon valley
junk
> >dealer will have anything that can be adapted.
> >
> >                   unsigned long BinToBCD(unsigned long i) {unsigned
long t;
> >Ludis Langens         return i ? (t = BinToBCD(i >> 1), (t << 1) + (i &
1) + 
> >ludis at netcom.com                 (t + 858993459 >> 2 & 572662306) * 3) :
0;}
> >
> >Hi
> 
> 	The 15 pin IP connector from under the dash can be hacked down to fit.
Noy
> very pretty.
> 
> Vance