[Diy_efi] Digifant (was:This turbo assist stuff...)

[Mike]

Bernd,

I wasnt looking to argue with you about the reletive merits of
VW's control systems strategy. I'm trying to understand your
problem and make an assesment as to what might make it rather
worse than what you'd expect from a 'drive out the showroom'
mode of operation. Incidentally the knock sensor on my Bosch/Hitachi
ECU is never tested by the ECU via pushing the advance. It seems
to be tested by virtue of noise detection within the rpm band its
designed to be effective if and only if the ignition is already
retarded after an initial knock detect and for my ECU its only 5deg.
It comes up with error code 34 (I think) to report some connection
problem or sensor problem or other problem in effect causing excessive
knock after a retard has alread occured.

On your system,
You are telling me the ECU detects knock frequently enough to cause
you some angst, even with hi octane fuel, to be a distinct irritation.
What I am focussing on is whether there is some aspect of your overall
system that gives the ECU an indication of excessive knock vs 'slight'
knock, because you are reporting the output falls dramatically in those
circumstances and you indicate this is relatively frequent.

What I am therefore trying to get a handle on is what you think
might be causing the ECU or sensor to (in effect) see an excessive
knock signal (which might be erroneous) vs a slight knock which
would (in theory) only require a minor retard - not the significant
retard you are reporting which you find so irritating.

And, the most probabilistic item I am focussed on at this point has
to do with suppression of ignition noise. I see a potential in
the probability that your 'bosch amp' has degraded to the point
that its inbuilt suppression cap is not functioning optimumly
causing possible coupling to the knock sensor - hence causing
the ECU to 'over-correct' for erroneous knock detection.

rgds

mike



At 08:18 AM 10/1/2003 +0800, you wrote:
>On Thu, Jan 09, 2003 at 10:43:23PM +0800, Mike wrote:
>> At 10:55 AM 1/3/03 +0800, you wrote:
>> >If you think that odd - schematics for earlier cars show a 10A fuse
>> >to power the ECU relay coil; and nothing else! Looks like the ECU
>> >was originally powered entirely via the ignition switch and somebody
>> >though better of it, putting a relay in to take the bulk of current.
>
>> >But the coil is still running via the ignition switch!
>
>> >Maybe that's deliberate - the extra cable length (inductance) may be
>> >being used to isolate the ECU from voltage fluctuations induced by
>> >the coil.
>
>> Nop, its the same on my car. The reason seems to be the ECU releases
>> the relay when its done, but the ignition can be 'killed' ad therefore
>
>Not on Digifant. At least the version in my car.
>The ECU relay is switched in parallel with the ignition power.
>
>> there isnt a 'hold' time. The other reason is the ECU is then
>> compatible with the hot wire AFM burn requirement. Inductance at
>
>Doesn't make sense. Not economically speaking and the design intent
>of Digifant - to be a low-cost digitally-controlled engine
>management system (i.e. Motronic's cousin from the wrong side of the
>track). Hot-wire was even more expensive than the flap-type meter.
>
>> those frequencies is not an issue IMHO. My schematic has a
>> capacitor, I am speculating yours is in your 'bosch amp' - which
>> if old may be marginal causing erroneous early and overly
>> sensitive determination of engine knock.  The ECU may also have an
>> eeprom which may need an update time after ignition off, mine
>> holds for about 12 seconds after ignition removed, but thats
>> mostly for AFM burn, though there is a small eeprom in the
>> bosch/hitachi ecu.
>
>No such EEPROM on mine. Later ones did have it (G60 engine), but
>AFAICT they use a power-fail circuit to initiate a save of any data
>that needs to be saved.
>
>> Onw thing which I dont understand with your setup given the point
>> you mentioned that your ECU occasionally does a knock sensor test
>> to drive the ignition excessively advanced to test the sensor etc.
>
>Yep. How else can you check it's working?
>
>VW and Bosch describe the process in their Digifant and
>Motronic-related SAE papers and literature dating back to about
>1986.
>
>> How does the ECU determine the difference between a slight knock
>> and a severe knock or are you implying it just doesnt care ?
>
>Of course it cares; otherwise the odometer wouldn't have gone this
>far without a major engine rebuild. :-)
>
>Slight knock conditions were at the time being recognized as not
>being damaging to engines and usually resulted in higher torque,
>especially at low to moderate engine speeds.  The way to detect
>slight knock is to back off the timing until no knock is detected
>and then incrementally advance until knock is just detected.
>
>i.e. suck it and see. :-)
>
>-- 
>/"\ Bernd Felsche - Innovative Reckoning, Perth, Western Australia
>\ /  ASCII ribbon campaign | I'm a .signature virus!
> X   against HTML mail     | Copy me into your ~/.signature
>/ \  and postings          | to help me spread!
>
>_______________________________________________
>Diy_efi mailing list
>Diy_efi at diy-efi.org
>http://www.diy-efi.org/mailman/listinfo/diy_efi
>
>

_______________________________________________
Diy_efi mailing list
Diy_efi at diy-efi.org
http://www.diy-efi.org/mailman/listinfo/diy_efi