rec.autos.simulators

I've been trying to get to the bottom of why I'm always
blowing this Cosworth engine. It happened again on lap
12 at Imola this week.

I decided to use the Telemetry option in GPLRA to save
my first 12 laps, load them up into a spreadsheet and
basically check for time spent at more than 8,000 revs,
max revs etc. etc.

GPLRA saves telemetry info every 0.0278 seconds so theres
lots of data.

I was shocked at how much time I'd spent at > 8,000 revs
and at having touched the 8528 (max revs) at least ever
so briefly on each lap (0.0834 in fact)

I decided to compare my lap against one of Paul H.'s.
I chose one from about lap 8 when Paul and I did a mid 1:39
lap. We were both in Lotus's so thought it would make a good
comparison.

Loading it into my spreadsheet it was clear that Paul also
had spent quite a bit of time > 8,000 revs but not quite
as much as me, and never hitting 8528.

I then started looking at these 2 laps in much greater detail.
Without going into great detail Paul spent about 3.5 secs > 8k
and I spent a whopping 5.0 secs > 8k. I think the telling
thing though is that Paul only spent just over a tenth > 8.4k
whereas I spent over a second.

AFAIK GPL uses some "counting" method to determine abuse
related engine failures. I've no idea what that count entails,
whether it is for time at revs > 8k, maybe each 10th is 1
point, each 10th over 8,100 is 2 points, etc. etc. and once
you accumulate a certain number of points, then pop. I'm sure
it is different for the different cars. I believe that, although
the F2 cars all have the same engine  that they inherit this
"count" <reliability factor?> from their F1 counterpart. I
generally drive the slower cars, BRM & Honda and rarely if ever
have an engine failure.

OK, so that's probably how it works. I then needed to figure
out what I was doing wrong. I do lift between gear changes,
so I first suspected too early downchanges.

It turned out that, although a number of downshift did result in
the engine revving to greater than 8k, this rarely went above
8,200 and Paul and I were about equal on those.

My problems were with the upshifts. More specifically I deduced
from the data I was getting on the power about 0.0834 (3 RA units)
quicker than Paul which was ALWAYS resulting in an engine rev rate
of around 8,500 and this is where the 3 x 8528s were.

This does not show on the tach via the redline at all, my gauge
looked the same as Paul s, to all visual intents and purposes I
wasn't revving above 8,200.

I also deduced from the data that as the gear became engaged I
would occasionally still be revving at 8,000+ and in effect GPL
was modelling a slipping clutch. Furthermore my early application
of the throttle resulted in me dropping MORE revs than Paul H, in
effect I WAS SLOWER.

In every instance I lost 100revs+ to Paul on upchanges.

I'd always maintained that a slick upchange with a well timed
lift and re-apply was as quick as speed-shifting. I've never
tested this out but may well do some testing now. Yes, I think
a speedshift is perhaps going to minimise to a small degree your
speed loss from lifting before upchanging but it looks to me as
if this can easily be negated by the "slipping clutch" effect
that appears to take place.

Anyone ever tried analysing this?

To summarise, I obviously have a timing problem with my throttle
re-application, only slight but a) it means I'm generally going
to blow an engine in an hour long race (although I didn't at Kya)
and b) If I correct it, I can go that little bit quicker :)

Just thought I'd share that knowledge and see if there are any
other opinions out there.

Maxx

PS. Will also post to vrocF2 list, GPLEA forum and r.a.s.

[...]

[...]

I don't know that this is necessarily evidence of a slipping clutch.
Couldn't it just be wheelspin?  If you're getting on the throttle earlier,
your tires are probably carrying more cornering load at the time you do
it, and so perhaps have less grip available for accelerating.  I know I
was popping Ferrari engines, usually very reliable, at Rouen for a while.
I discovered I also had unusually high rear tire temps.  A close look
at the replay, and a bit more consciousness of the issue while driving,
led me to realize that I was overrevving the engine with wheelspin while
trying to accelerate hard out of fast corners where I was already sliding
(tires had already exceeded their ability to grip).

That doesn't explain why it doesn't show on your tach, but then neither
does the clutch slip hypothesis.  Perhaps the tach just isn't that
accurate.

Just some thoughts...

        Marc

I've discounted wheelspin as we are talking F2s here, sorry I didn't
mention that specifically.

I'm sure the tach is inaccurate in so far as it doesn't register more
than about 8,200 (on the F2s).

I found some more interesting stuff by delving a little further into
this which i'll post shortly

Maxx

Follow up post I did to GPLEA forum where there are other comments,
particularly in relation to if speed-shifting is slower than slickly
done lift'n'shift shifting.

---------------------------------------------------------------------------------

Not sure if the following table will come out OK but here goes.
This shows the same area on the track and the telemetry for mine
and Paul H's car.

It as quite hard to get a decent match, plus there are other
things at play here.

Firstly I have held onto 2nd a bit too long.

Secondly I have noted that Paul H's car is in neutral for 6 units
(0.1668) whereas mine is ALWAYS in neutral for 7 units (0.1946)
which is 0.0278 more. Not a huge deal at Imola with about 10
upshifts (although this is close to 3 x 10ths extra in neutral)
it would be perhaps seconds at the ring.

PH  Revs  ME  Revs
2 G 7996  2 G 8124
2 G 7996  2 G 8124
2 G 7996  2 G 8098
<0> 8124  <0> 8228
<0> 8255  <0> 8417
<0> 8362  <0> 8472
<0> 8335  <0> 8500
<0> 8281  <0> 8500
<0> 8202  <0> 8500
3 G 7846  <0> 8528
3 G 7469  3 G 7139
3 G 7097  3 G 6952
3 G 7014  3 G 6831
3 G 6932  3 G 6752
3 G 6831  3 G 6713
3 G 6851  3 G 6713
                3 G 6713
                3 G 6713
                3 G 6733

Now, there were in fact 2 or 3 occasions where Paul H's car was in
neutral for 9 units, not sure why unless it relates to finger speed
on the paddles/shifter.

Interestingly, it is those upshifts where Paul H gets most of his
greater than 8,300 over-revs.

OK, back the the table above. You can see that I in-fact lose a
small amount of revs just before going into neutral. having looked
at the whole lap however, Paul never loses any revs before going
into neutral, I only lose them on one other occasion

The above example is odd, and maybe skewd slightly by setup but
you can clearly see quite of bit of slip on Pauls chart.

There is also slip on mine but strangely the revs go from 8528 to
7139 in 0.278 of a second, very odd. I then get a similar amount
of slip ending up with less revs than Paul H.

I extracted the number based on 3 units before neutral and up to
the first unit that saw a rev increase.

This was not the best example of rev drop but was interesting for
the other issues mentioned.

Maybe GPL ALSO models a small power drop-off after a rev surge,
this is in fact fairly true to life on performance engines.

Regarding speed-shifting and hot-lapping. I said earlier that
speed-shifting probably helped in not losing revs on lift-off
before change-up. In fact, this doesn't seem to happen on a
slick lift'n'shift anyway.

It does appear that engaging a higher gear at max revs may gives
rise to some slip and maybe even a temporary performance hit ??

This could get very interesting, more testing to do I think.

Incidentally. On downshifts my car was generally in neutral
for 12 units whereas Pauls was in neutral for 13 units.

I'd be interested in knowoing what other figures people are
getting if they check a telemetry output, also what controller
they have/shifter etc.

Food for thought.

Maxx

I have looked at some of my replays and with horror
discovered that I am spending quite a bit of time
in 8000+ range. For example:
9000+ = 4.7sec
8500+ = 12.44sec (including previous values)
8000+ = 21.67sec (including previous values)
I seem to be very consistent in doing that, on other
laps results are very close (within 0.1sec)
Total lap time 81-82 seconds.

I mostly run Pro-Long races and I can't say I suffer much
from the blown engines. It happened to me once
or maybe twice.

Probably I would suffer it much more often in longer
races.

Alex
(alexti)

Unless I'm wrong, the actual RPM's and gear-ratios on *your* replay are
incorrect...
To get an accurate result, you should get hold of Paul's replay and use
*his* data for analyzing *his* laps.

If you look at N4 replays for instance - you will see that it completely
overlook clients gear-ratios and probably also RPM's..  (this is much easier
to discover in N4 due to the time spend at more or less constant rpm-range
(Daytona for instance)..

--
ed_
--
NOTE!
Return address is *not* my own - but the e-mail address to some spammers
that has bothered me one time to many!