Date:         Wed, 13 Jun 2012 04:57:45 -0400
Reply-To:     tinkerman007@GMAIL.COM
Sender:       Vanagon Mailing List <vanagon@gerry.vanagon.com>
From:         tinkerman007@GMAIL.COM
Subject:      Re: Any way to trick the Digijet ECU into generating injection
              pulses without the RPM/ignition signal, for test bench purposes?

On Sat, 2 Jun 2012 15:37:50 -0400, David Beierl <dbeierl@ATTGLOBAL.NET>
wrote:

>At 05:08 AM 6/2/2012, tinker man wrote:
>>I'll get down to testing that myself soon, but was wondering if anyone
>>has already tried this or has other ideas.

OK, I tested it myself and found that only a simulated external signal let
me cause the ECU to generate injection pulses.

>The output driver is FET equivalent of an open-collector,

actually, in my ECU it indeed is an open collector device, not a FET.

> i.e. it
>supplies a ground when it's active and open circuit when it's
>inactive.  Connecting and disconnecting things while it's powered is
>a no-no.

In general you're right, but the high power output transistor doesn't seem
to mind. The internal logic circuits and CPU would be much more vulnerable
due to the well know LATCH-UP scenario known in silicon devices, but they
are protected by the interface circuitry.

> Damage is cumulative and actual failure may occur some time
>later.

I'm not sure it's severe as you describe because I'm doing it all the time.
And don't forget that the power relay is connecting and disconnecting the
ECU from power thousands of times during the vans life.

> (General rule on the ECU is not safe unless they specifically
>say it is.)

Yep.

>Suggest trying a logic pulser set on TTL and powered from 5 volts.
>Check Bentley for specs but I believe it's a 5-volt pulse, so TTL
>logic margins may be enough.  Or in CMOS mode, adjusting supply
>voltage to get desired output voltage.  Otherwise build a triggerable
>555 circuit that will output a pulse stream for you at the desired
>voltage.  That may be easiest if you don't have a scope to measure
>the output of the pulser.  If the datasheets at National Semi etc
>don't have a suitable circuit there will be one somewhere on the web.

I did a much simpler solution - just connecting +12V and ground a few times,
it worked...:D
BTW, the signal is probably 12V and not +5V, because it's generated by the
hall amplifier that also drives the coil in my setup (a modified digijet).
cheers!