Andy Ross

> 
> I wrote:
> > We're about to go in circles again, and my blood pressure is rising.
> > So try this: DON'T reply to my message paragraph by paragraph.
> > Start from scratch, post a configuration file that you want to use
> > that does not work.  Explain why.  Use numbers.  Ask for
> > suggestions.  Don't touch the C++ code until you have convinced me
> > of what you want to do.
> 
> OK, I gave this some thought on the train on the way to work, and I
> think I understand now what you are trying to say: the cutout system
> is Nasal-based, and won't run during solution.  It also engages during
> the specified cruise parameters, something that I was expecting it
> would not*.  So you want to use the wastegate setting as a proxy for
> the boost control, but you are stymied because if you use the
> wastegate, it then becomes active at runtime.

Yes, good old train. Nearly right, except we now have a Boost Control
(nasal) which replaces the function of the wastegate. The Boost Control
Cutout is part of this implementation. This is just terminology - your
analysis is correct. We no longer need a wastegate for the supercharger (and
only the supercharger - not the turbo), but have to have an accurate
'wastegate-mp'(perhaps we should call it max-mp) setting so that YASim
solves correctly.  

> * Are you sure it does?  The boost control will be actively modifying
>   the throttle at low altitudes and high throttle settings.  Cruise is
>   generally at high altitude and middle throttle.  My guess is that it
>   would *not* be active, honestly.  High performance cruise is what
>   the engine was tuned for -- the boost control is there to prevent it
>   from damaging itself in non-optimized situations.

Yup. Absolutely sure - we have curves of boost against altitude at max power
rpm which show exactly at which altitude the Boost Control stops operating
for most of the Merlin variants. This altitude is called the "full throttle
altitude". So we set the cruise values at the "full throttle altitude". We
know exactly the relevant data at this point. With a bit of experimentation
we can derive an accurate turbo-mul parameter and everything falls into
place for the full speed supercharger gear ratio, including non-full
throttle settings. We also have full throttle altitudes for the medium speed
supercharger gear, so we can experiment further to get the right number for
the lower boost setting. The combat power also falls out along the way when
the Boost Control Cutout is operated. We have a very good simulation at the
cost of very little effort or code.
 
> Does that sound like what you want?  If so, then I argue that the way
> to do this is to map a property input to the wastegate pressure.  You
> can then use an otherwise-unused property to set it to whatever value
> you want for solution, and leave it at a very high value in the
> property system for runtime usage.

I'm not sure about this on 3 counts: 

1. Does Nasal initialise before YASim runs the solver or after?

2. I'm not clear what you mean by "an otherwise-unused property"

3. We have a perfectly good solution in C++ already. As you have said we
will need the supercharger Boolean when we need to differentiate between
gear-driven superchargers and turbos. This is nigh, because Josh has reached
the point where he would like to do the B29 turbo-charged engines. We (I
that is) will shortly be testing a more appropriate curve for turbo output
pressure v rpm. 

It would also be highly desirable to fix the ambient pressure at zero rpm
and the windmilling outputs. The diff does this, albeit I realize that you
consider it imperfect, it is better than that in cvs right now, and would
surely do as a temporary fix until you can find the time to come up with
something better. The diff certainly doesn't break anything.

Vivian 





 



_______________________________________________
Flightgear-devel mailing list
[email protected]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d

Reply via email to