[Diy_efi] boosted compression ratio

[Mike]

At 06:24 PM 23/12/2002 +0800, you wrote:
>> mmmm OK, I take it you mean the ECU reaches its own max advance
>> by virtue of a knock sensor - which we assume is operating correctly
>> (I say that as these are high impedance devices and dont need much
>> in way of connector contamination to go a little haywire), though
>> its weird an ECU would be arranged to keep advancing until knock
>> on a production car ! <phew>
>
>Well, advance at least until as far as the desired set-point.
>VW made a big thing about Digifant being able to adapt to a wide
>range of fuels. So it _might_ be that they do advance to the onset
>of knock and then back off for steady-state.

I take it you dont have access to this 'setpoint',

>> Eeeek - You get much better performance going for a hot wire type
>> AFM than the vane types ! If I recall the factory AFM on my VL
>
>It's not a standard Bosch AFM but a special version for VW's
>Digifant. Linearised output IIRC.

mmm Must be - because even the early vane types I used for EFI thesis
back in 82 were log output, I used one from a 2L Kombi for a ford
escort - the restriction was quite marked, never got anywhere near
as much power as a dual throat progressive carby :(

>> turbo is a pretty much generic Bosch type in that it was intended
>> to replace many vane types - I have a spare which you are welcome
>> to try (you are in PErth?)  - the pin configuration may well
>> be similar, you only need to put in a burn cycle ?
>
>One of the "projects" that I've had to shelve because of dire lack
>of free time and money, was to "map" the output of the present AFM
>against the now-cheap MAF by putting it in series (!!) with the aim
>of eventually replacing the AFM with the MAF and a small processor
>to handle the translation for what the ECU expects, provide an air
>temperature "reading"; and to do the special handling of the MAF of
>which the ECU is blissfully ignorant.

mmmm, I think I see, hot wire in front, vane in back and do a one-to-one,
there are heaps of processors with 10-12bit ADC and you'll get more
headroom from the hotwire ones, though the base point cal has a fixed
offset (which wanders a bit) and needs to be 'reset' in s/w 
plus that burn cycle.


>> >Engine management is mostly closed-loop so even changing the fuel
>> >pressure regulator didn't increase fuelling.
>> 
>> So why did you tweak the spring on the AFM, you did make the
>> spring a little weaker - right ?
>
>Backed off 3 teeth on the pre-tension cog. The tip-in "response" lag
>disappeared after that so I glued the cap back on... It's only for
>dealing with transient enrichment anyway.

Ah ok, so the cal range for the actual wasnt affected - I didnt think
they were that complex, I took one apart in 82 when I did my thesis,
and it was a real basic spring loaded pot - nothing special - can from
a 2L kombi.

>> Bear in mind most ECU's are designed to have only minimal shift
>> from initial settings so as to avoid wild fluctuations affecting
>> overall stability. So if you start out quite lean then the ECU
>
>I don't know about the "minimal" fluctuations. I was able to drive
>the car when the camshaft was installed retarded by 30 degrees...
>no mean feat for any engine management system. For some reason, the
>sub-contractor who pulled the head down to replace valve guides (the
>head was off the block anyway), took the cam sprocket off the
>camshaft. And then assembled it backwards... of course there were
>"timing" marks on the wrong side of the sprocket. But not in the
>right place.

What I was referring to was range when in closed loop, not overall.


>> may not have enough opportuntiy to make it richer, hence you
>> are hearing ping ! I've messed with mine to some degree and found
>> that it won't change fueling by more than about 8% from the
>> base settings - ie. It will try then hits an aysmptote. So
>> as long as I set it up to a median then it can go a little
>> lean or rich as needed but wont go far from the base settings.
>
>Well, it's operating closed loop and the control algorithm on my car
>tends to favour the lean side of stoich. That's steady-state
>operations; which is not where the (audible) ping occurs anyway.

Why not shift the o2 sensor output, add a buffer amp with a pot to
change the gain from say 0.9 to 1.1, but then I see you dont have
a problem with steady state,

I take it then you only have a ping problem with transient ?

>> There is an adjustment pot on my AFM (which goes only to the
>> ECU) which is a rich/lean trim and this has about a 15% affect
>> (or maybe a little mroe) on ECU injection DC - its a 22 turn pot.
>
>That'd be the base CO adjustment.

Well it affects everyting, CO is not independent of HC's that pot
handles the whole AFM cal setting - through the ECU.

>Space behing the engine of a Golf is very tight and the steering
>rack is quite close. All the LHD ones interfere with the steering
>rack. What's more, the engine moves around a lot so there has to be
>some flexible coupling; and it has to connect to the cat. converter
>because I'm not too chuffed at the prospect of a $20,000 fine.

Dale has done quite a few odd ones with tight spaces, he'll look
at it for nothing...

>> >If I use normal ULP, the 7c/litre saving is offset by a 10 to 15%
>> >increase in fuel consumption. And then there's the crap in the fuel
>> >that gums up injectors to they need a "dose" every 15,000km as
>> >well... not been required since 99% of the fuel used has been 98.

>> consumption test setup ?
>
>They'd be inconclusive anyway; pretty much a waste because of day to
>day variations. The only proper way to do it would be to put the car
>in a controlled environment and run it in there.

Well thats what I meant on a dyno, just pick a day, note the temp/humidty and
run 500cc of fuel through a graduated line, or whatever amount is
selected whilst staying on say 40% power out in top gear - much more
defintive than say the psychological factors imposed by expectations ;)

Though I seem to be getting real world settings almost regardless of
how I drive, I used to get about 320 K's on a tankfull when ULP and
odd ping problem. Now on premium, no ping problem and get 450K per
tankfull on average - part city/highway cycle,

>> I 'seem' to get about a 5 - 10% improvement in economy when
>> going from ULP to Premium but havent noticed much at all when
>> going from Premium to BP 98 - havent tried optimax yet - partly
>> because here in Perth there arent many outlets.
>
>Not many? Are there _any_??

I think there's one on Alexander drive around truganini rd/st,

>Price difference between 98 and PULP is usually only one or two
>cents and the car seems happier (less audible ping and better
>throttle response) so it's worth it to me.

I usually fill up at Shell West Perth and get it about 6c on
average cheaper than 98 but havent been keep track of BP 98 as
its a bit out of my way,

>> i mean you've 'tweaked' the spring in your vane AFM and assume
>> the ECU has wide range over which to control fueling.
>
>On transients. It's been fooled into thinking that there's a greater
>inrush of air so it provides additional enrichment.

So I take it the main spring tension hasnt been changed,

>Steady-state and it reverts to (the lean side of) stoich.
>
>> What you may have here is a combination that results in the
>> ECU not being able to fuel because its base is out of its
>> median control range ?
>
>Only if it was made that way at the factory.

Yes they are all made that way so that in closed loop there arent
wild swings in fueling - imagine 80% change in injector duty due to
o2 sensor swings which arent detected as faults - most ECU's seem to
onyl 'trim' the fuel injector DC the main inputs being AFM and
engine temp - have you checked your coolant temp - mine went gagah
about a year after a bought the car - overfuelled badly...

>> Have you looked at the injector duty cycle change when
>> in stead state cruise and how this changes between fuels
>> for the same cruise conditions ?
>
>No... I don't have the diagnostic gear. Digifant is deaf and dumb,
>so it doesn't provide any cap[ability to read that information.

Thats not what I meant, get a cheap multimeter with Duty cycle
capacity or get a kit from Jaycar that does the same thing and
attach it to the injector (in parallel of course)

>Bosch Super 4. Brought in a set from the UK in 1997 to try and
>haven't had a reason to use anything else.

Yeah they are now quite common at Marlows, been using them for
a while and much happier at idle and the thing pulls much better
at higher revs - amazing difference between splitfires and 4 spot.
I'll prob go for a colder version when I get my next engine !

>> Just occured to me, have you changed/upgraded the capacitor
>> which is near the ignition coil, ie The one which goes from
>> coil +ve to ground. I had an odd problem for a few months
>
>There's no capacitor there on my car; nor on the factory
>current-flow diagrams for the car. Ignition is controlled by
>Digifant using standard Bosch ignition amplifier.

Ah ha !

If your engine was stock *and* the ignition circuit was in top
health then you probably wouldnt have a problem, now that you
have a modded cam and your car is older it might be worthwhile
taking a real close look at your ignition driver.

If the 'bosch amp' goes to both sides of the coil primary then
it might just have a cap in there, if not and the 'bosch amp'
only acts like a glorified NPN transistor then it might *need*
a cap to go from coil +ve (ignition) to ground - somewhat close
to the ground point of the 'bosch amp' (no more ascii art
for me today;) Because dont make the assumption the circuit is
perfectly accurate - I have *very often* seen errors in production
circuits over many years from HP, to IBM to Ford to Holden and
foreign (non english) etc etc

You might just jag your problem by trying this out, it wont damage
anything - you are quelling back emf from the coil going into the
ignition line and this nasty disturbance could well be enough to
jostle your ECU enough - also assuming its own noise suppression
is working fine..

I have a 4uF Bosch AC cap (Can type) you are welcome to try, and I'd
be interested in "hearing" your ping - it sounds similar to the
problem i used to have,  what year, no of cylinders, displacement
is your engine ?

rgds

mike




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