Ye Olde Rocket Forum

Go Back   Ye Olde Rocket Forum > Work Bench > RockSim Asylum
User Name
Password
Auctions Register FAQ Members List Calendar Today's Posts Search Mark Forums Read


Reply
 
Thread Tools Search this Thread Display Modes
  #1  
Old 06-10-2009, 06:19 PM
Pyro Pro's Avatar
Pyro Pro Pyro Pro is offline
Rocket Scientist
 
Join Date: Apr 2006
Location: University of Michigan
Posts: 224
Default Buggy design?

Thought I'd just post this to see if anyone can get it working. For some reason, when I sim this, the rocket will appear to fly normally, but then I'll suddenly get "nan" for all simulation values, and the simulation will continue executing indefinitely. Because of this, I can't load a 2D flight profile, but I can look at the graphs of the simulation, which appear fairly normal (though, altitude is a bit lower than I'd expect, acceleration seems to be jumping around wildly towards the end of the burn, as well as the Cd, which starts off in a normal range, but starts rapidly pinging between ~0.5 and 2.5 towards the end of the first second of flight).

Ideas?

I simulated the tank tubes with non-ejectable pods for authenticity. I could redo the design with them as inside tubes, which would probably be more stable, program-wise, but... why shouldn't I be able to use pods

I also have a 3-stage version of this design (I've reduced it to a single stage due to time constraints on building the real thing), and it did virtually the same thing.
Attached Files
File Type: rkt 1-227 Saturn 1B SA-205.rkt (126.4 KB, 107 views)
__________________
David Hash
NAR#77967
http://www.Semroc.com
Reply With Quote
  #2  
Old 06-10-2009, 09:13 PM
Pyro Pro's Avatar
Pyro Pro Pyro Pro is offline
Rocket Scientist
 
Join Date: Apr 2006
Location: University of Michigan
Posts: 224
Default

Nope, inside tubes don't help either...

Attached Files
File Type: rkt 1-227 Saturn 1B SA-205 IT.rkt (116.9 KB, 95 views)
__________________
David Hash
NAR#77967
http://www.Semroc.com
Reply With Quote
  #3  
Old 06-10-2009, 10:03 PM
CPMcGraw's Avatar
CPMcGraw CPMcGraw is offline
BARCLONE Rocketry
 
Join Date: Mar 2005
Location: Mobile, Alabama
Posts: 5,357
Default

David,

I've been getting a bunch of designs that create a nan (not a number) condition recently. I sent Tim a message awhile back, but the suggestion was to check for any "Custom" material entries...

This is a bug in RS9, and it is probably related to the pods feature, because I never recall having nan on any RS8 design. Look for material entries with "Custom" material.
__________________
Craig McGraw

BARCLONE Rocketry -- http://barclone.rocketshoppe.com
BARCLONE Blogsite -- http://barclone.wordpress.com
BARCLONE Forum -- BARCLONE Forum

BARs helping BARs

SAM 0044
AMA 352635
Reply With Quote
  #4  
Old 06-10-2009, 10:44 PM
Gus's Avatar
Gus Gus is offline
7/21/61
 
Join Date: Sep 2005
Location: North of Detroit
Posts: 2,223
Default

Quote:
Originally Posted by CPMcGraw
David,

I've been getting a bunch of designs that create a nan (not a number) condition recently. I sent Tim a message awhile back, but the suggestion was to check for any "Custom" material entries...

This is a bug in RS9, and it is probably related to the pods feature, because I never recall having nan on any RS8 design. Look for material entries with "Custom" material.

Craig,

I've been getting the same thing and I'm not using the pods feature. I do have some "custom" parts. I downloaded 9.0.5f80 but no improvement.

I've also been getting a weird condition where if I try to hit a button in construction windows I get an error beep, and nothing happens. If I put the cursor off the window, click, and then come back to Rocksim and click the button it works fine. I've been using Rocksim since version 4, when I used it on a Mac using Virtual PC, and I've never had anything like this.

I'm also getting a freeze if I run an optimal mass sim and try to look at the graph.

I'm running Rocksim on the latest version iMac running system 10.5.7.
Reply With Quote
  #5  
Old 06-10-2009, 11:55 PM
blindeye2 blindeye2 is offline
Junior Rocketeer
 
Join Date: Jun 2009
Posts: 28
Default Suspect AT10-3 motor data

I was able to run simulations for all the available motor configurations except for the AT10.
I would speculate that something in the data is causing a numerical instablility in the solution. I do not know the detailS of the soulution method used within ROCKSIM.
This is not an uncommon problem in transient numerical solutions.
Hope I was of some help.
Bob
Reply With Quote
  #6  
Old 06-11-2009, 12:31 AM
CPMcGraw's Avatar
CPMcGraw CPMcGraw is offline
BARCLONE Rocketry
 
Join Date: Mar 2005
Location: Mobile, Alabama
Posts: 5,357
Default

Quote:
Originally Posted by Gus
Craig,

I've been getting the same thing and I'm not using the pods feature. I do have some "custom" parts. I downloaded 9.0.5f80 but no improvement.

I've also been getting a weird condition where if I try to hit a button in construction windows I get an error beep, and nothing happens. If I put the cursor off the window, click, and then come back to Rocksim and click the button it works fine. I've been using Rocksim since version 4, when I used it on a Mac using Virtual PC, and I've never had anything like this.

I'm also getting a freeze if I run an optimal mass sim and try to look at the graph.

I'm running Rocksim on the latest version iMac running system 10.5.7.


I've noticed one additional problem, when running RS9 on my laptop (4-yr-old Toshiba Satellite, XP Home). The program freezes when I try to edit the plot points of any fin using the mouse pad. If I edit the points in the X-Y table grid, it works OK. If I try to drag a plot point with the pad, the program quits responding, requiring a 'three finger salute' to break the ice.

No, RockSim is not perfect and has a long way to go in a couple of major areas. On the 'nan' issue, everyone here who has encountered this should send a note to Tim. I think he already has an idea where the problem might be, but obviously he has not found it for sure. He mentioned something about having seen the problem come up before, but that it was intermittent. I've had it happen on about a dozen designs since starting with RS9, which made me think it had to be something associated with a 9-centric feature.

Bob's observation about having it happen with one motor out of several is a new twist on what I've seen. Usually, once I start seeing the 'nan' runs, it keeps doing 'nan' until I make serious changes to the file, or try running the file on another computer. Using another computer sometimes works, sometimes doesn't. No immediate pattern.

All of my 'nan' runs to this point have been designs that used pods. If designs with no pods are also seeing 'nan', then the problem is much deeper in the core code.
__________________
Craig McGraw

BARCLONE Rocketry -- http://barclone.rocketshoppe.com
BARCLONE Blogsite -- http://barclone.wordpress.com
BARCLONE Forum -- BARCLONE Forum

BARs helping BARs

SAM 0044
AMA 352635
Reply With Quote
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump



All times are GMT -5. The time now is 11:47 AM.


Powered by: vBulletin Version 3.0.7
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Ye Olde Rocket Shoppe © 1998-2024