Sponsor Message:
Aviation Technical / Operations Forum
My Starred Topics | Profile | New Topic | Forum Index | Help | Search 
How Is Simulator Software Programed?  
User currently offlineMastropiero From Spain, joined Dec 2005, 125 posts, RR: 0
Posted (5 years 4 months 2 weeks 3 days 14 hours ago) and read 4126 times:

Ok, I hope the title of the thread is not misleading, but it is the best I could do..... my question is basically this: how are simulators programmed to recreate to such level of precision the behavior of the aircraft? I would imagine that while designing the aircraft simulations could be run on how such and such feature would affect the handling characteristics but I would imagine there has got to be more?

During the test flights I guess the aircrafts are loaded with sensors and diverse equipment that gather plenty of data, is that used later in the process? Also, do test pilots have a say? I mean, do they get to test the sim and offer any input regarding the behavior of the plane? Can the software be adjusted according to this feedback?

25 replies: All unread, showing first 25:
 
User currently offlineDingDong From United States of America, joined Jan 2007, 661 posts, RR: 0
Reply 1, posted (5 years 4 months 2 weeks 3 days 13 hours ago) and read 4118 times:

I'm not familiar with full motion sim programming, but I do recall an interview with Austin Meyer, developer of the popular PC-based sim, X-Plane. He mentioned that he basically developed the flight physics model in X-Plane originally by taking out each aircraft he could get access to and, with an observer, doing timing for various performance-related aspects.

Then once he was back home, he developed a physics model fit to the datapoints as well as underlying aeronautical flight equations. X-Plane also breaks down the lift/drag to individual 'blocks' that's computed on the fly so it can get a more accurate overall picture of the aerodynamics in response to changing conditions. I can't remember right now what this approach is called, but it's rather effective.

Full motion sims would also have extra logic to more accurately model the interaction and response of systems. Most PC-based flight sims are only flight simulators and not really 'systems simulators'. I've seen a systems sim that faithfully reproduced the actual bugs seen with a particular revision of an actual aircraft -- impressive. Wink



DingDong, honey, please answer the doorbell!
User currently offlineAndz From South Africa, joined Feb 2004, 8453 posts, RR: 10
Reply 2, posted (5 years 4 months 2 weeks 3 days 13 hours ago) and read 4116 times:
Support Airliners.net - become a First Class Member!

I wonder how they get a simulator to fly like an aircraft that has never flown yet.


After Monday and Tuesday even the calendar says WTF...
User currently offlineMastropiero From Spain, joined Dec 2005, 125 posts, RR: 0
Reply 3, posted (5 years 4 months 2 weeks 3 days 13 hours ago) and read 4112 times:



Quoting Andz (Reply 2):
I wonder how they get a simulator to fly like an aircraft that has never flown yet.

That I would assume it just doesn´t happen, but that´s just my guess. I suppose they start working on the full sim once they´ve gathered enough data from test flights?


User currently offlineDavid L From United Kingdom, joined May 1999, 9524 posts, RR: 42
Reply 4, posted (5 years 4 months 2 weeks 3 days 12 hours ago) and read 4100 times:



Quoting Mastropiero (Reply 3):
That I would assume it just doesn´t happen

Hopefully Jetlagged will be along shortly to answer that but I'm pretty sure the models generated during design are usually good enough to get the bulk of it done before the first real flight.


User currently offlineDescendVia From , joined Dec 1969, posts, RR:
Reply 5, posted (5 years 4 months 2 weeks 3 days 10 hours ago) and read 4075 times:



Quoting Mastropiero (Reply 3):
I suppose they start working on the full sim once they´ve gathered enough data from test flights?

No.......
There are at least 3 787 level-D sims already out being used. NH and NW both have them at their training centers as well as another company ( I think its somehow a Boeing owned sim).


User currently offlineTdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 80
Reply 6, posted (5 years 4 months 2 weeks 3 days 2 hours ago) and read 3981 times:



Quoting Andz (Reply 2):
I wonder how they get a simulator to fly like an aircraft that has never flown yet.

All airplanes are governed by the same set of differential equations that relate control surface deflections to accelerations of the aircraft (for three linear axes and three rotational ones). The only thing that changes is what the coefficients are for the various equations. So you can write a generic numerical integrator (which is a pretty robust field of mathematics), plug in the right coefficients for your particular aircraft, and get very accurate outputs.

Quoting Mastropiero (Reply 3):
That I would assume it just doesn´t happen, but that´s just my guess. I suppose they start working on the full sim once they´ve gathered enough data from test flights?

They have full sims long before they start flight testing. Flight testing will validate and, if necessary, tweak the sims but nowadays they're really really close. As a side effect of designing the aircraft, you have to determine the differential equation coefficients anyway (they're a key part of stability and control calculations).

Tom.


User currently offlineSpudsmac From United States of America, joined Feb 2008, 301 posts, RR: 0
Reply 7, posted (5 years 4 months 2 weeks 2 days 23 hours ago) and read 3949 times:

http://www.eagleworks.org/Projects/Aerosim/Aerosim.htm

Check it out. ERAU is doing the same thing for the DA-42 with the IO-360 engines right now. From what I understand, they put the instrumentation on the a/c, fly it around doing maneuvers, download the data, sell it to Frasca and they program the FTD with the data.


User currently offlineKELPkid From United States of America, joined Nov 2005, 6392 posts, RR: 3
Reply 8, posted (5 years 4 months 2 weeks 2 days 20 hours ago) and read 3897 times:



Quoting Mastropiero (Thread starter):
my question is basically this: how are simulators programmed to recreate to such level of precision the behavior of the aircraft?

At its base, a level D simulator is a highly accurate flight model of the real aircraft based directly on fluid dynamics computations of the actual aircraft.

Quoting Mastropiero (Thread starter):
I would imagine that while designing the aircraft simulations could be run on how such and such feature would affect the handling characteristics but I would imagine there has got to be more?

Well, you cannot predict 100% how the real thing will behave, but you can come awfully close, using fluid dynamics computations or even wind tunnel test data.

Quoting Mastropiero (Thread starter):
During the test flights I guess the aircrafts are loaded with sensors and diverse equipment that gather plenty of data, is that used later in the process?

Mostly for validation and verification of predicted flight behavior.

Quoting Mastropiero (Thread starter):
Can the software be adjusted according to this feedback?

I'm sure the flight model can be reprogrammed as flight test data is collected. In the case of unverified data, it is very likely that worst case assumptions are used, which often leads to the conclusion by test pilots that the real article flies better than the simulator would lead one to believe  Smile

Quoting Spudsmac (Reply 7):
Check it out. ERAU is doing the same thing for the DA-42 with the IO-360 engines right now. From what I understand, they put the instrumentation on the a/c, fly it around doing maneuvers, download the data, sell it to Frasca and they program the FTD with the data.

My experience with the Frasca has been that they are great for procedures training, but really cannot faithfully duplicate the actual IFR experience. I wish it bumped you around as much as a real 172 does in the clouds...not to mention the general squirliness of the 172 (she's not too stable about the yaw axis, and this becomes aparent when you fly the 172 around IFR, unless you happen to fly in one of the rare ones built with rudder trim...).



Celebrating the birth of KELPkidJR on August 5, 2009 :-)
User currently offlineFlypig687 From United States of America, joined Jul 2007, 56 posts, RR: 0
Reply 9, posted (5 years 4 months 2 weeks 2 days 11 hours ago) and read 3801 times:

From my experience motion sims are avaliable before the real aircraft is off paper. As stated before all aircraft are governed by the same general equations which as long as a general size of the plane/heli is known can be computed. As the aircraft comes off paper and more is known of its flight dynamics the sim can be changed to what the actual system dynamics are. I know at the company I have worked at we have a full motion sim that is simulating a future helicopter and it is very good at it, we are able to pilot test flight controls with it.

On a side note, where as X-Plane computes many of the equations on the fly MSFT FSX basically uses a look-up table for how the aircraft will react. Meaning that you can build your own planes in X-Plane and get a decent idea of how they will fly and not so much in FSX.


User currently offlineMax777geek From Italy, joined Mar 2007, 538 posts, RR: 0
Reply 10, posted (5 years 4 months 2 weeks 2 days 11 hours ago) and read 3796 times:



Quoting Andz (Reply 2):
I wonder how they get a simulator to fly like an aircraft that has never flown yet.

Oh that'e easy, they build the simulator first


User currently offlineAirbuske From United States of America, joined Jun 2007, 466 posts, RR: 0
Reply 11, posted (5 years 4 months 2 weeks 2 days 9 hours ago) and read 3770 times:



Quoting Spudsmac (Reply 7):
http://www.eagleworks.org/Projects/Aerosim/Aerosim.htm

Check it out. ERAU is doing the same thing for the DA-42 with the IO-360 engines right now. From what I understand, they put the instrumentation on the a/c, fly it around doing maneuvers, download the data, sell it to Frasca and they program the FTD with the data.

That is correct. Eagleworks already did it once for the DA-42 with Thielert engines, but Thielert went bust so that's why they're redoing it with the Lycoming engines.


User currently offlineAirbuske From United States of America, joined Jun 2007, 466 posts, RR: 0
Reply 12, posted (5 years 4 months 2 weeks 2 days 9 hours ago) and read 3764 times:

Quoting Tdscanuck (Reply 6):
All airplanes are governed by the same set of differential equations that relate control surface deflections to accelerations of the aircraft (for three linear axes and three rotational ones). The only thing that changes is what the coefficients are for the various equations. So you can write a generic numerical integrator (which is a pretty robust field of mathematics), plug in the right coefficients for your particular aircraft, and get very accurate outputs.

Tom, I don't believe that this is an entirely accurate explanation. It isn't a matter of just "plugging in the right coefficients for your particular aircraft thereby getting very accurate results".

The equations of motion do not make the flight model, however; they merely set the limit on what is and is not possible. In order to support these equations, you must also have good models for finding the lift-curve slope, drag coefficients, stability derivatives, and other parameters.

Today what is used is a 6 DOF model - A 6 degree of freedom flight model provides for a fairly accurate modeling of the motion and flying characteristics of an airplane.
It is generally used when the airplane is to be modeled as a "rigid body." It considers both rotational (yaw, pitch, and roll) and translational motion, both centered around the center of gravity. Since there are three axes to consider in each case, this is referred to as a six- degree-of-freedom model. This model actually considers twelve variables, since both the instantaneous rate of change and position have to be considered. These are referred to as the state variables, which are applied to varying matrices of coefficients to get the desired fidelity.

[Edited 2009-05-11 09:06:02]

User currently offlineDavid L From United Kingdom, joined May 1999, 9524 posts, RR: 42
Reply 13, posted (5 years 4 months 2 weeks 2 days 7 hours ago) and read 3728 times:



Quoting Max777geek (Reply 10):
Quoting Andz (Reply 2):
I wonder how they get a simulator to fly like an aircraft that has never flown yet.


Oh that'e easy, they build the simulator first

In other words, they get a simulator to fly like an aircraft that's never flown by building a simulator that flies like an aircraft that's never flown.  Smile


User currently offlineJetlagged From United Kingdom, joined Jan 2005, 2556 posts, RR: 24
Reply 14, posted (5 years 4 months 2 weeks 2 days 5 hours ago) and read 3668 times:



Quoting David L (Reply 4):
Hopefully Jetlagged will be along shortly to answer that but I'm pretty sure the models generated during design are usually good enough to get the bulk of it done before the first real flight.

Correct. The design data for the aircraft makes a good basis for the "predicted" simulator data package. The first simulators of each type are built using this data and can be used for training until flight test validated data becomes available. Predicted data packages often used to bear very little resemblance to the real aircraft performance, but these days they are very good.

Quoting Flypig687 (Reply 9):
On a side note, where as X-Plane computes many of the equations on the fly MSFT FSX basically uses a look-up table for how the aircraft will react. Meaning that you can build your own planes in X-Plane and get a decent idea of how they will fly and not so much in FSX.

X-Plane uses simplified aerofoil data to compute lift, drag, pitching moment, etc for each wing element, and these are integrated to form the overall coefficients. However, the aerofoil data comes from look up tables of course! MSFS does use look up tables to calculate coefficients but uses equations of motion to simulate dynamics, just like X-Plane. MSFS uses simplified EOM compared to X-Plane.

Austin Meyer would probably like you to think full flight sims are built his way, but in practice they use the same techniques as MSFS (though with much more depth and sophistication). Some people deride MSFS look-up tables, but this is much more efficient than attempting to solve complex aerodynamic equations in real time.

Quoting Mastropiero (Thread starter):
During the test flights I guess the aircrafts are loaded with sensors and diverse equipment that gather plenty of data, is that used later in the process? Also, do test pilots have a say? I mean, do they get to test the sim and offer any input regarding the behavior of the plane? Can the software be adjusted according to this feedback?

Flight tests instrumentation provides actual performance data which corrects the predicted data produced at the design stage. This objective data should be sufficient, though pilot feedback could be used to "fill in the gaps" between test points or at the edges of the envelope such as stalling.

Flight models are fairly standard in construction. Each aero coefficient and stability derivatve is built up from a series of functions which define it in terms of various independent variables including: alpha, beta, pitch/roll/yaw rates, Mach No, normal accel, alpha dot, beta dot, flaps, gear, control surface position, thrust coefficient, ground effect, etc. So, as you can imagine, there's an awful lot of data involved.

The flight model is normally provided by the aircraft manufacturer to the simulator manufacturer. Occasionally the sim manufacturer has to generate its own model (not ideal as flight testing is expensive and time consuming).

Quoting David L (Reply 13):
In other words, they get a simulator to fly like an aircraft that's never flown by building a simulator that flies like an aircraft that's never flown.

True, though the simulator now plays an active role in the design of aircraft, especially in refining fly-by-wire control laws, so inevitably the aircraft ends up flying like the simulator!



The glass isn't half empty, or half full, it's twice as big as it needs to be.
User currently offlineTdscanuck From Canada, joined Jan 2006, 12709 posts, RR: 80
Reply 15, posted (5 years 4 months 2 weeks 1 day 21 hours ago) and read 3581 times:



Quoting Airbuske (Reply 12):
Tom, I don't believe that this is an entirely accurate explanation. It isn't a matter of just "plugging in the right coefficients for your particular aircraft thereby getting very accurate results".

The equations of motion do not make the flight model, however; they merely set the limit on what is and is not possible. In order to support these equations, you must also have good models for finding the lift-curve slope, drag coefficients, stability derivatives, and other parameters.

We're agreeing, I think. The models for lift-curve slope, drag coefficients, stability derivative, etc. *are* the coefficients that go into the differential equations of motion. These coefficients are not constants (I should have been clearer about that), they're functions of other parameters like alpha, Reynolds number, etc.

Tom.


User currently offlineMax777geek From Italy, joined Mar 2007, 538 posts, RR: 0
Reply 16, posted (5 years 4 months 2 weeks 17 hours ago) and read 3436 times:



Quoting David L (Reply 13):

In other words, they get a simulator to fly like an aircraft that's never flown by building a simulator that flies like an aircraft that's never flown. Smile

Uh... I guess they build an aircraft that flies like its simulator.... when all bugs are hopefully fixed. I think a380, every airbus after the a300/310, Boeings 777 and 787 have been built like this.


User currently offlineDavid L From United Kingdom, joined May 1999, 9524 posts, RR: 42
Reply 17, posted (5 years 4 months 2 weeks 15 hours ago) and read 3419 times:



Quoting Max777geek (Reply 16):

Uh... I guess they build an aircraft that flies like its simulator....

Yes but I'm only saying that they need to have a reasonable idea about the design of the aircraft before they can build the simulator. Simply "building the simulator first" wouldn't really seem to solve the issue. As Jetlagged said...

Quoting Jetlagged (Reply 14):
The design data for the aircraft makes a good basis for the "predicted" simulator data package.



User currently offlineBravo1six From Canada, joined Dec 2007, 397 posts, RR: 2
Reply 18, posted (5 years 4 months 1 week 6 days 20 hours ago) and read 3345 times:



Quoting Jetlagged (Reply 14):
The flight model is normally provided by the aircraft manufacturer to the simulator manufacturer. Occasionally the sim manufacturer has to generate its own model (not ideal as flight testing is expensive and time consuming).

Bingo. The OEM licences a data package to the simulator manufacturer.

The same is true for at least some of the pre-packaged aircraft for MSFS, although I believe the data package in that case is less sophisticated given the nature of the ultimate end use, and some variables in that data package are purposely modified by the OEM prior to providing it to Microsoft.


User currently offlineMax777geek From Italy, joined Mar 2007, 538 posts, RR: 0
Reply 19, posted (5 years 4 months 1 week 6 days 14 hours ago) and read 3316 times:



Quoting David L (Reply 17):
Yes but I'm only saying that they need to have a reasonable idea about the design of the aircraft before they can build the simulator. Simply "building the simulator first" wouldn't really seem to solve the issue. As Jetlagged said...

Well, I hope they'll have much more than a reasonable idea... I think all the planning work, data models and every aspect of the physical project is been planned in first place.. I just meant the simulator will work before the actual building begins.


User currently offlineJetlagged From United Kingdom, joined Jan 2005, 2556 posts, RR: 24
Reply 20, posted (5 years 4 months 1 week 6 days 11 hours ago) and read 3298 times:



Quoting Max777geek (Reply 19):
I just meant the simulator will work before the actual building begins.

The manufacturer's engineering sim will be flying long before the aircraft, that's for certain. Launch customers for a new aircraft would normally have their training sims in service before aircraft were delivered too, so simulator design would have to start before the aircraft's first flight.

Years ago I was on the design team for Boeing's 757 and 767 training sims long before first flight, at a time when the 767 was still a three crew plane.



The glass isn't half empty, or half full, it's twice as big as it needs to be.
User currently offlineBorism From Estonia, joined Oct 2006, 431 posts, RR: 0
Reply 21, posted (5 years 4 months 6 days 3 hours ago) and read 3000 times:



Quoting Jetlagged (Reply 14):
Austin Meyer would probably like you to think full flight sims are built his way, but in practice they use the same techniques as MSFS (though with much more depth and sophistication). Some people deride MSFS look-up tables, but this is much more efficient than attempting to solve complex aerodynamic equations in real time.

Well, I don't think he cares much about what we think because he supplies basically the same version to FAA certified small scale full sim developers.

I'd guess you don't need to calculate same coefficients for the same model over and over in real-time, you should be able to just write them down when ready to RAM or HDD. Perhaps not as real as wind tunnel or test plane data, but good enough and much much cheaper  Wink

But didn't Scaled Composites use it a lot for testing of many of their own designs before they flew? There's a mentioning of this on Wikipedia, but no link to the original article.


User currently offlineDavid L From United Kingdom, joined May 1999, 9524 posts, RR: 42
Reply 22, posted (5 years 4 months 6 days 3 hours ago) and read 2986 times:



Quoting Max777geek (Reply 19):
Well, I hope they'll have much more than a reasonable idea...

Yes, I think I covered that in Reply 4. The point of my later reply was in the sentence immediately after the one you quoted.


User currently offlineJetlagged From United Kingdom, joined Jan 2005, 2556 posts, RR: 24
Reply 23, posted (5 years 4 months 4 days 5 hours ago) and read 2877 times:



Quoting Borism (Reply 21):
Well, I don't think he cares much about what we think because he supplies basically the same version to FAA certified small scale full sim developers.

Yes, as was MS ESP (the commercial training version of FSX now no longer in development). The FAA would not certify a full flight sim with X-Plane as the aero model unless it was validated to flight test data. What they currently approve are much more basic flight training devices (FTDs). AFAIK the devices X-Plane talk about are not approved for handling training, just navigation and systems procedures. Also the X-Plane website is very misleading in implying such sims are FAA approved as downloaded. Each such trainer is approved by the FAA on it's own merits as built.

Quoting Borism (Reply 21):
I'd guess you don't need to calculate same coefficients for the same model over and over in real-time, you should be able to just write them down when ready to RAM or HDD. Perhaps not as real as wind tunnel or test plane data, but good enough and much much cheaper

Putting coefficient data in memory or on HDD is exactly what MSFS and full flight sims do.

Quoting Borism (Reply 21):
But didn't Scaled Composites use it a lot for testing of many of their own designs before they flew? There's a mentioning of this on Wikipedia, but no link to the original article.

Very possibly. X-Plane is good for such experimental purposes, but not for high fidelity training.



The glass isn't half empty, or half full, it's twice as big as it needs to be.
User currently offlineZeke From Hong Kong, joined Dec 2006, 9109 posts, RR: 75
Reply 24, posted (5 years 4 months 4 days 4 hours ago) and read 2871 times:

Original A320 simulators used real aircraft computers (i.e. the FBW computers) and FMGC in a large computer room with a PDP-11 to drive it, this needed less programming as it was using the actual "aircraft" boxes to do the various calculations (and occasionally they would pull a box from a simulator to replace broken computer on an actual aircraft).

These days that is all replaced by a couple of "PC" style computers that have a lot more computing power, as Tom mentioned above these are built on 6 DOF mathematical models, engine models, aerodynamic models etc. I have been told that some people have used tools like Matlab to build the dynamics/contol models (using Simulink), and then get Matlab to automatically generate the code for the simulator. As this allows the engineer to adjust the gains in the control loops, aerodynamics etc at a high level.

Whilst engineers can predict to an extent how an aircraft will behave, real life flight test performance is fed back into the simulators once it obtained to improve the fidelity.

I read recently that current research thinks that full motion simulators are a waste of time, just as much training value is obtained from a fixed based simulator.



We are addicted to our thoughts. We cannot change anything if we cannot change our thinking – Santosh Kalwar
User currently offlineJetlagged From United Kingdom, joined Jan 2005, 2556 posts, RR: 24
Reply 25, posted (5 years 4 months 3 days 4 hours ago) and read 2761 times:



Quoting Zeke (Reply 24):
Original A320 simulators used real aircraft computers (i.e. the FBW computers) and FMGC in a large computer room with a PDP-11 to drive it, this needed less programming as it was using the actual "aircraft" boxes to do the various calculations (and occasionally they would pull a box from a simulator to replace broken computer on an actual aircraft).

Even though black boxes were used (and still are in many cases), the aerodynamic and systems simulation is the same as with any other non FBW aircraft. Black boxes alone can't drive a flight simulator.

Quoting Zeke (Reply 24):
These days that is all replaced by a couple of "PC" style computers that have a lot more computing power, as Tom mentioned above these are built on 6 DOF mathematical models, engine models, aerodynamic models etc. I have been told that some people have used tools like Matlab to build the dynamics/contol models (using Simulink), and then get Matlab to automatically generate the code for the simulator. As this allows the engineer to adjust the gains in the control loops, aerodynamics etc at a high level.

You're right that tools like Simulink can write the code, but of course the aerodynamic data still comes from the manufacturer, it can't be derived in some way by the tools used.

Quoting Zeke (Reply 24):
I read recently that current research thinks that full motion simulators are a waste of time, just as much training value is obtained from a fixed based simulator

I heard that too, but the theory is that motion cues can largely be replaced by visual cues (not for better training but to save money). Think about it, most of the time you are in the box the vis is set to zero. No vis, no motion cues. Not a good idea.

It's a controversial concept, and one which is not widely accepted. Certainly not by most regulatory bodies.



The glass isn't half empty, or half full, it's twice as big as it needs to be.
Top Of Page
Forum Index

Reply To This Topic How Is Simulator Software Programed?
Username:
No username? Sign up now!
Password: 


Forgot Password? Be reminded.
Remember me on this computer (uses cookies)
  • Tech/Ops related posts only!
  • Not Tech/Ops related? Use the other forums
  • No adverts of any kind. This includes web pages.
  • No hostile language or criticizing of others.
  • Do not post copyright protected material.
  • Use relevant and describing topics.
  • Check if your post already been discussed.
  • Check your spelling!
  • DETAILED RULES
Add Images Add SmiliesPosting Help

Please check your spelling (press "Check Spelling" above)


Similar topics:More similar topics...
How Is Aircraft Retirement Decided? posted Sat Mar 22 2008 11:35:33 by XJETFlyer
How Is An FMC Updated? posted Thu Feb 14 2008 10:09:43 by LIFFY1A
How Is Cargo Decided Which Hub It Will Go To? posted Fri Aug 24 2007 20:25:37 by RampGuy
How Is This Possible? posted Mon Jul 2 2007 03:27:16 by 777DadandJr
Q: How Is A Plane Weighed To Determine Fuel, Etc.? posted Sun Apr 8 2007 21:09:27 by JAAlbert
New Jets... How Is Warranty Covered? posted Sun Mar 18 2007 03:42:58 by 747hogg
How Is An IL96M Take Off? posted Mon Jan 22 2007 23:40:15 by 747400sp
Quality Controls Of Fuel..how Is This Done? posted Sun Jan 21 2007 03:38:41 by Scalebuilder
Dewpoint: How Is It Calculated? posted Wed Nov 22 2006 04:24:49 by LTU932
SIL - How Is It Measured. posted Fri Aug 18 2006 13:13:19 by HAWK21M

Sponsor Message:
Printer friendly format