Date: Wed, 12 May 1999 12:37:00 -0400
Reply-To: David Beierl <synergx@IBM.NET>
Sender: Vanagon Mailing List <vanagon@gerry.vanagon.com>
From: David Beierl <synergx@IBM.NET>
Subject: Re: Hey all you shop mechanics with Oscillosopes? VW GREMLIN?
In-Reply-To: <000701b7e29c$64ba8960$9e1e9bd1@hpcustomer>
Content-Type: text/plain; charset="us-ascii"; format=flowed
At 01:05 5/12/93 -0700, Debi wrote:
> I am reading that Oscilloscopes let us see problems in circuits-
>problems that can't be detected by other means, Right?
Well...oscilloscopes show a graph of voltage against time. Essentially
they are extremely fast voltmeters which do not have either the pointer
mass of an analog meter or the integrating time of a digital meter. They
are very good for showing regularly recurring behavior/events. Special
types called storage 'scopes are very good for showing isolated single
events. However...
> One good example is electronic noise, an elusive and deadly enemy,
>especially when it invades the engine computer through input sensor wires.
...actually they're not very good at showing random events mixed in with
regularly recurring ones, particularly if the random events happen rarely
compared to the regular ones. Also they are very poor at displaying
typical digital signals on a bus, where there are many patterns arising
from many sources. The 'scope is only good at looking at one pattern at at
time, so the digital signal just looks like mush. However, the DigiJet
takes primarily analog inputs which ordinarily vary smoothly and
comparatively slowly, so perhaps all is not lost. Looking at the signal
from an AFM, for example, should show a regularly varying signal, perhaps
with some oscillation as the AFM vane oscillates when idling. If there is
poor contact at some part of the AFM travel, the scope should show
irregular "hash" when the vane is in that area. If this is happens
reliably the scope will soon show it. Likewise it should easily show
switching noise on the idle/WOT microswitch(es). OTOH if a noise signal
happens only occasionally or is very short, there is a very high chance
that you won't be looking at the scope when it *does* happen (or indeed
that the noise trace will be too dim to see, or that it will happen during
the time that the 'scope beam is inactive (retracing or waiting to
trigger). For digital systems there are two easy answers: a logic probe
(cheap) or a logic analyzer ($$$$). And nowadays there are little handheld
digital scopes with memory for a few hundred dollars which may be very
useful. My experience is with lab oscilloscopes for fairly big bucks which
have very high-quality analog circuitry but don't remember anything beyond
the persistence of the screen phosphor.
>Electronic noise on the input sensor wire make the computers heart skip a
>beat here and there (called computer resets).
Usually the noise from erratic contacts will simply give a false value to
the input, with attendant bad results; and DigiJet inputs [excepting the O2
sensor] are mostly pulled high inside the ECU, with the various sensors
acting as voltage dividers and providing a varying-resistance path to
ground. This method is fairly resistant to static noise which might force
the signal outside its normal limits (causing erratic ECU operation and
possible damage) . Resets are much more likely to be caused by noise on
the power inputs.
> This causes an erratic operation of the computer
[or actually of the entire system -- the ECU is probably just doing its job
based on the input it receives]
> and often leads to the
>replacement of the ECU [[and several other parts]]...in a misguided effort
>to fix the problem by replacing hardware...
> Is this our VW GREMLIN?? Could it finally be, this the dreaded
>intermittent problem that has driven us insane on more than one occasion??
>And a simple visit to our ECU with an Oscilloscope could help locate and
>Identify him??...Naw...too easy!
Well, sometimes it's easy and sometimes not. If a 'scope shows a problem,
it is real. If it doesn't show a problem, that doesn't prove it's not
real. For learning about what's generally going on, though, they can't be
beat. They're like a set of eyes that can see electricity (or in fact
anything that can be turned into electricity). If you're generally curious
about time-related effects you may decide that you just have to have a scope.
This is of course way oversimplifed and ymmv <g>.
david
David Beierl - dbeierl@ibm.net