FastFieldSolvers Forum

FastFieldSolvers Forum

All Forums

FastFieldSolvers

FastHenry2

Possible errors in writing files

Note: You must be registered in order to post a reply.
To register, click here. Registration is FREE!

Screensize:

UserName:

Password:

Antispam question:

What do MOONwalk and MOONdance have in common?

Answer:

Format Mode:

Format:

Message:

* HTML is OFF
* Forum Code is ON

Check here to include your profile signature.

T O P I C R E V I E W
dnl	Posted - Sep 17 2007 : 15:59:47 My Zc.Mat file contains different data to that displayed in the program window Program window: Frequency = 1e+007 Forming sparse matrix preconditioner.. conductor 0 from node ng1c_pseudo_11 Calling gmres... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 conductor 1 from node ng1c_pseudo_10 Calling gmres... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 Computed matrices (R+jL) Row 0: n1c_c1 to n1e_c1 Row 1: n1c_c2 to n1e_c2 Freq = 0.1 Row 0: 0.000896206+8.2479e-009j 0.000516544+4.89758e-009j Row 1: 0.000516544+4.89936e-009j 0.00136594+1.18344e-008j Freq = 1000 Row 0: 0.000898476+8.19934e-009j 0.000517682+4.87085e-009j Row 1: 0.000517685+4.87259e-009j 0.00136915+1.17781e-008j Freq = 1e+007 Row 0: 0.00177828+3.72265e-009j 0.00115711+2.32861e-009j Row 1: 0.00115721+2.32849e-009j 0.00268639+4.63719e-009j All impedance matrices dumped to file Zc.mat Times: Read geometry 0.25 Multipole setup 8.485 Scanning graph 0.046 Form A M and Z 0.0469999 form M'ZM 0 Form precond 4.954 GMRES time 15.03 Total: 28.812 and the file, as read into the Matlab editor Row 2: n1c_c2 to n1e_c2 Row 1: n1c_c1 to n1e_c1 Impedance matrix for frequency = 0.1 2 x 2 0.000896206 +5.18231e-009j 0.000516544 +3.07724e-009j 0.000516544 +3.07836e-009j 0.00136594 +7.43577e-009j Impedance matrix for frequency = 1000 2 x 2 0.000898476 +5.1518e-005j 0.000517682 +3.06044e-005j 0.000517685 +3.06154e-005j 0.00136915 +7.4004e-005j Impedance matrix for frequency = 1e+007 2 x 2 0.00177828 +0.233901j 0.00115711 +0.146311j 0.00115721 +0.146303j 0.00268639 +0.291363j So the floating point exponent appears completely wrong! Writing files appears a little troublesome at present: Some .mat are ascii text, others are Matlab binary form Some .mat binary files are not readable by Matlab Use of the "-k text" option still appeared to produce binary files Hopefully, they are all linked to the common bug causing the first fault listed above. David Lee Power Electronics, UK
2 L A T E S T R E P L I E S (Newest First)
dnl	Posted - Sep 27 2007 : 09:46:15 Thanks for your response. I should have picked up that impedance is not the same as R+jL. I had skiped over the passage on ReadOutput in the manual because it talked about the source code, "ReadOutput.c" which I assumed was there for integrating into other programs - and I do not have a c compiler on my computer. As a stand-alone program, does it do anything more than reproduce the R+jL values displayed in by FastHenry? There are two other utility programs supplied: MakeLcircuit and MakeLnegative. MakeLcircuit is in 2.2.2 of the FastHenry manual, but I can not find MakeLnegative details anywhere. Can you assist? David Lee Power Electronics, UK
Enrico	Posted - Sep 25 2007 : 18:05:32 Wait, wait! Please don't mix different problems together. Let's first make some clarity about types of files. The default output file from FastHenry is called Zc.mat. This file contains 'raw' data, in the sense that results are in the format R+jwL, where R is resistance, L is inductance, w is 2fpi (where f is frequency and pi = 3.1415...) To have data in more readable and useful format, you can use the utility program 'ReadOutput', which will output the results in the more common R+jL format. FastHenry visualization just uses ReadOutput internally to provide you with the results directly, avoiding you one passage. As a matter of fact, if you take your results for 1e+7: Impedance matrix for frequency = 1e+007 2 x 2 0.00177828 +0.233901j 0.00115711 +0.146311j 0.00115721 +0.146303j 0.00268639 +0.291363j and divide the imaginary part by 21e73.1415 you get exactly 0.00177828 +3.72265e-009j 0.00115711 +2.32861e-009j 0.00115721 +2.32849e-009j 0.00268639 +4.63719e-009j that is the result in FastHenry. The format of Zc.mat and the usage of ReadOutput are detailed in the FastHenry User's guide, paragraph 2.2, 'Processing the Output' Please be always sure to have read the User's guide thoroughly, it will save you from committing errors. On other points, see the 'get_new_x_y: Vector never hits edge of box?' topic. Hope this clarifies Enrico

FastFieldSolvers Forum