RE: More overshooting of boost - long/nerdy

[QSHIPQ@aol.com]

In a message dated 12/8/99 1:01:25 PM Central Standard Time, 
scottmo@teleport.com writes:

>  >I'm intrigued by the claims of audi, Scott, since a second or two of 
>  >"overboost" readings really isn't affecting a measureable time. 
>  
>  Well, these are not "claims" by Audi, as the overboost is a "reality" that
>  is a result of the feedback control system and it is something that can
>  easily be measured or observed.

I'm still intrigued by the use of 'overboost', when in reality we are 
speaking of an input/output loop delay resulting in a boost 'overshoot'.  A 
faster processor and/or a better WGFV could control this better, even in the 
early WGFV years.  If my lowly dodge GLH could do a timed "overboost" mode 
without the boost "overshoot" in 1985, I certainly think a 30kUSD 5ktq could 
do it.

>  
>  Having a boost profile that has more area under the curve, "IS" going to
>  make a difference if you imagine that boost translating into a Torque graph
>  with more area underneath it. Whether you can measure the difference
>  "accurately" and repeatedly, is another matter.

A really tough variable to isolate.  I used a vericom and a G-analyst.  The 
"overshoot" didn't make a difference to either.   Given the "overshoot" 
doesn't "have" to happen (per your audi quote), and a bunch of other 
variables start entering into multiple runs, really hard to make any claims 
of it, IME.  Comparing 'overshoot' to an actual turbo map (see below) makes 
the above a tough argument to make

>  I have not tested the HKS boost controller, so I can't comment on what the
>  boost output graph looks like, versus the boost output with the stock MAC11
>  boost controller solenoid. If the HKS provides a longer time in overboost,
>  and gets the boost ramped up quicker, then it surely could result in a
>  quicker 1/4 mile time. 

The faster feedback loop and "faster" FV application of the HKS really shows 
the difference between the 80's and the 90's technology.  The new HKS offer 
fuzzy logic (learns boost profile of turbo), overboost mode, and a disabled 
mode.  I will be adding one of these to my toyota 4Runner turbo, and will 
report more on the results (but it is an automatic, so not a good comparo).  
I still don't recommend them for the audi turbos however, I'd sure like to 
see better WGFV hardware (yea right:)

>  
>  Getting consistent 1/4 mile times with the quattros can be tough to do and
>  require making a lot of runs with the same setup. The cheaper G-analyst
>  (G-Tech-Pro) didn't give me consistent results for the 1/4 mile times when
>  used back to back and compared to the actual time slips at the race track.
>  I don't know how the comparisons were made, but the ability to make
>  repeatable, accurate measurements, should always be kept in mind.

Agreed.  To baseline a given car, I use the Gtech pro.  To evaluate and test 
a car comprehensively, IMO, the track is best, the Ganalyst is the best off 
track, the vericom second, the Gtech third.  IME, the G-analyst software 
gives some really accurate (to track slip) results

>  Well, to say that "it doesn't really make a difference" is pretty silly. I
>  can't imagine finding too many people who would look at the boost graph of
>  the two different scenarios (with and without the ECU controlled boost) and
>  think that the profile which goes higher and stays higher for 2 seconds, is
>  going to have NO difference? 

The problem really is testing a 2 second anomoly on a turbo car, IMO.  I 
couldn't isolate a difference between the 2 second overshoot, and one 
without.  I certainly can attest, that a stiffer spring, and/or a Boost 
Controller makes a bigger and expressly measured difference.  

>  Translate that boost graph into a torque output graph for the same
>  scenario, can't you imagine that the torque is going to be higher when the
>  boost is higher? 

Not necessarily Scott, that's somewhat oversimplifying a turbo boost map.  If 
you look at a turbo map of the 'standard' audi k26, increasing the PR from 
1.8 to 2.0bar before 3000rpm decreases the turbo efficiency from 70% to 50%, 
that surge line is really tight, and the Turbo Efficiency (TE) is right at 
the 70% line at 1.8bar.  At 50% you are on the surge line, a really bad thing 
in terms of turbo car performance, and ouput (air volume) from the turbo 
doesn't increase in the above scenario (rule of thumb: increasing boost 
increases PR, not necessarily flow).  That means the 'overshoot' boost 
profile should yield a measured loss, not gain.  A k24 map is even more 
sensitive to these overshoot ramp profiles, I had many a complaint of surging 
in k24 with boost controllers.

>  I did not do the modifications to these 944 Turbos, as I said, the mods
>  were from a "particular tuner" and I have told the owners, that I don't
>  recommend doing this. They have been told by others that headgasket failure
>  is common with this "Overboost" mode.

Most likely from a lean condition.  Bleed systems are no better than resistor 
mods, sometimes worse.  A properly configured turbo car, shouldn't have to be 
"fooled" into any OOP (out of program) parameter.
 
>  Well, the term "Overboost" certainly can have different meanings for
>  different people depending on the context of the conversation. The brief
>  "Overboost" found in the 10V Turbo is due to the Feedback control system
>  particulars, the 15 seconds "Overboost" mode found in the 1992> S4/S6 is
>  due to the programmed values inside the ECU and is much different in its
>  execution. This doesn't mean that my definition when applied to the 10V
>  turbo is somehow incorrect, as my boost graphs surely show that it exists
>  as defined by me.
>  
>  Over and out
>  Scott Mockry

I agree that you properly used the term as you defined it.  I might look to 
define the rogue "overboost" as an "overshoot"  (tm - DLawson), and leave 
"overboost cutoff" and "overboost mode" as separate and accepted definitions. 
 The side effect of a slow feedback loop, isn't really relevent to the 
accepted definition of overboost, nor necessarily a bragging right on a slow 
controller, or to performance gains.

My .02

Scott Justusson