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

[Bernd Felsche]

On Thu, Jan 02, 2003 at 05:36:48PM +0800, Mike wrote:
> At 03:26 PM 1/2/03 +0800, you wrote:
> >Also didn't want to risk installing a capacitor that may have caused
> >the ignition system to fail... adding capacitance may not be
> >"welcome" by the ignition amplifier and finding a replacement part
> >at this time of year is even less likely than me getting a day away
> >from work about now.

> I suppose that would be a concern if the +ve of the coil went into
> the 'bosch amp' etc. On my setup the cap goes from ignition +ve to
> ground somewhere near the ignition transistor. The ECU (by Bosch)
> just drives a 500v NPN transistor which drives the coil directly,
> but it is fully reliant on a 0.5uF cap near the coil +ve which is
> the same as the +ve ignition line. After changing to 8uF the
> difference is dramatic to say the least!

The coil driver is rather more complicated than an NPN transistor.
The amplifier is essentially the same as for direct Hall-sensor
distributor ignition; except that the ECU provides the timing
pulses. The "amplifier" contains a hybrid circuit comprising
dwell-angle control, coil peak-current cutoff, a driver and output
stage (pretty obvious), and a feedback loop to the driver and dwell
control sections. The ECU does none of that; no current sensing, etc.
There is no direct connection between ECU and coil.

The ignition amp is powered from the coil's +ve (battery)
connection.

I'm not sure what sort of car benefited from a larger capacitor on
the battery-side of the coil - no capacitor on mine or any similar
vehicles AFAICT. Even Bosch K and KE-Jetronic powered VW and Audi
don't have a capacitor there; and they use pretty much the same coil
driver - except it's triggered directly from the Hall sensor in the
dizzy. There's a fair number of them - at least 10 million cars.

The coil driver type was first used in the early 1980's IIRC. I
think that Bosch would have sorted out a problem of a missiing
capacitor when VW built my car almost 10 years later. :-)

[snip]

> In this respect the DC circuit is very different to the AC circuit,
> the assumption is I take it, that there is some suppression within
> the 'bosch amp', if its more than 10 years old and ove that time
> has been subject to our Aust. conditions then its probably reduced
> in capacity or suffers from resistance leakage - either way its
> lacking in suppression.

It's *always* pinged on ULP. Even when I had it "new" in 1992.
Build date is 08/90.

> It could be that were the car new and not modified then no extra cap
> may be needed to suppress ignition noise from posibly upsetting the
> timing start of ignition by the ECU. After a period of time, all
> sorts of things become marginal and who knows what tolerance was
> built into the initial setup - do you know what vintage is the
> ECU electronics and what year the 'bosch amp' was manufactured ?

1990. It was just as "broken" as it is now.

> I and other engineers can probably be safe to say they can guarantee
> that placing a good suppression cap from ignition +ve near your coil
> (if this is the positive ignition supply) to ground (also near your
> coil) wont do any damage but will improve dramatically the AC circuit
> your ECU and ignition amp is seeing !

The ECU doesn't see any of that.

> I take it you havent tried some other bosch amp as a swap ?

No. All the other Digifant Golfs do that when you don't give them
enough Octane. It's a feature.

In fact, Digifant deliberately drives the engine to knock
(at about 3000 rpm) periodically to test the knock sensor. I don't
know exactly how often it does it. It does because disconnecting the
knock sensor results in consistently-reduced engine power within a
short time (minutes). People who don't torque down the sensors lose
engine power because it can't "hear" the knock when induced.
Digifant reverts to a safe ignition map.

> >There is no capacitor fitted to the vehicle, shown in any schematic
> >or appearing anywhere in the parts catalogue for that type of engine
> >management system.
> 
> There would have to be something - probably in the 'bosch amp', on
> most schematics I've seen, its a separate cap due to issues like
> size and reliability, I recall for one car it recommended replacement
> each 100,000K ! Also over the last 25 years or so, I have
> seen many errors in automakers schematics, omissions being the most
> common.
> 
> How old is the car ?

Still "new"; only 11 years on the road this February.

> And what is the specfic indication you get that the ECU goes into
> retard mode, is there a light, beeper or just a dramatic drop in
> power - or that zippiness sensation ?

Lack of zip is the main clue. The next is hot starts. The knock
detection doesn't work until the engine has exceeded 2400 rpm three
of four times. The engine pings audibly for the first few gear
changes after the hot start.

Another clue is that putting a timing light on the engine through
the hot-start cycling clearly shows increased retard when the
knock-sensing is operating.

-- 
/"\ 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