RE: More on overboost

[QSHIPQ@aol.com]

Scott M writes:

>Yes, it is boost overshoot, that is done "purposely" by the use of a
>feedback control system that adds pressure to the top of the waste gate
>diaphragm, and results in a boost level that is slightly higher (initially)
>than you would get by the use of only a mechanical Waste Gate with a
>certain spring tension. 

I think we can agree that the reason this is done purposely, is that the WGFV 
as an exact boost controlling device is small and not designed for adding 
large amounts of pressure.  It also operates better when there is more boost 
in the manifold.  This "purposely" is because of the device limitations, not 
much else, IMO.  If a faster and better controller gets the same results or 
better, you are golden.

>Boost Overshoot, More Boost, Extra Boost, Extra strength Boost, Big Mama
>Boost, Holy Sh*t Boost, or OverBoost, call it what you want, or whatever
>makes you feel comfortable.
ha, I like big mamma boost.  Boost overshoot is more appropriate to what's 
happening.  When the "overboost" concept was presented by audi to SAE, the 
accepted definition of overboost changed already, not accomodating wg 
cracking or a slow feedback loop, but to accomodate an overboost profile that 
is consistent with a turbo map.  

>The whole point of my previous discussion was to point out that this
>feedback control system allows a boost profile that is "better" IMHO than
>would you get with only a mechanical Waste Gate with a given spring
>tension. 
TOTALLY agree Scott.  I'm just not really enamored with bad hardware 
requiring a rogue boost level being called overboost.  A WG controller that 
can tailor the more exact turbo boost profile (map) wins in performance.

>The system allows the Waste Gate to act "like" it has a stiffer
>spring "initially" to prevent waste gate cracking, but then will reduce
>pressure on top of the waste gate diaphragm to reduce boost after the boost
>is ramped up. 
Test:  Take manifold feed, hook it directly to the top of the WG.  How much 
cracking is there?  I'm thinking that the controller is the problem here, not 
wg cracking per sae.

>The use of some initial "overshoot" when designing a feedback control
>system is often standard practice and the designers of the system can have
>the system have a step response that is "overdamped", with a rolled off
>front edge, or with a "underdamped" response with 5% of overshoot, 10% or
>perhaps 30% of overshoot if they so desire. The designers of the system are
>well aware of the intended response.

Adapting hardware to the task requires some compromises to the task.  The 
intended response "works", but I'm not convinced at all it's better.

>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).  

>>Very good information, and it is good that you brought up this important
>>piece of the puzzle. It is important to be aware of the Turbo
>>characteristics before raising the boost on any turbo charged car. 

>>I don't agree with all of your assumptions, as the dynamic response of the
>>engine and turbo make it difficult to assume the engine/turbo will allow
>>the boost to overshoot to 2.0 when the engine is under 3000 RPM for the K26
>>Turbo. 

It *can* happen, bttt.  FYI, any overshoot to 2.0 bar  between idle and 
4500rpm decreases turbo efficiency according to the audi k26 turbo map. 

>>The other assumption is whether the "transient" boost condition will
>>match what is found on the boost map under steady state conditions. I guess
>>we can agree to "dis-agree" on this one until other evidence is available.

A turbo that spins 100,000 rpm, then spins 10,000 more rpm based on 
overshoot, will effect TE, whether that is measure AT the time it's happening 
or at a given time afterward.  

>That means the 'overshoot' boost 
>profile should yield a measured loss, not gain.  

>>Well, again, I don't agree with the above assumption. The acceleration
>>tests I did using the data acquisition system on the 1989 200TQ with the
>>K24 Turbo shows the recorded acceleration values (0.## g's) to max out, or
>>peak, the exact same moment when the boost was also at its max value during
>>this brief "Overboost" or "Boost Overshoot" period. This correlation was
>>true in 1st,  2nd, and 3rd gears when accelerating the car from 5-90MPH. 

To baseline this test, a boost controller would need to be tested back to 
back. You will see the controller win just based on it's faster response time 
throughout the rpm range.  I'm not a big fan of the stock WGFV loop, can you 
tell? :)  Why can't the boost map profile of the turbo be matched in first or 
second gear?  The WGFV circuit is too slow.  A boost controller will get you 
the boost map profile in any gear, bttt.

>>This "data" tells me the effect of the Overboost period isn't reducing
>>performance, at least during this test. Other tests should be done using
>>the data acquisition on the K26 equipped 87 5000TQ to see what occurs
>>during this "Overboost", or "Boost Overshoot" period.

Is it really helping?  A back to back with a faster controller might prove 
more insightful.  Quick shots of overboost aren't really going to hurt much, 
but the plotted curves on the boost profile within a map, are hard to argue 
with.  Anytime you stray from the plotted rpm v PR v output on a map, you are 
usually below maximum turbo efficiency for that set of parameters.  That 
doesn't mean you necessarily get a performance loss, it's still a turbo 
spinning applying pressure.  I argue, sticking on the proper rpm curve gives 
you a better performance gain.  WG cracking is the issue and the reasoning 
for this overshoot phenomenon in the Mac>11 cars.  That can be/ could have 
been solved with better hardware. 

I'll be getting a couple more controllers in for customers in the next couple 
months.  A ride in Chad Clark's ex 5000tq (k24 with profec boost control) or 
Gary Kalikians 4ktq (k26 with profec) sure has me convinced that the audi 
feedback on the wgfv circuit could benefit from some hardware/feedback 
improvement without "overshoots".  Watching a proper boost profile in first 
and second gear alone makes me a believer.  

Scott Justusson