THIS IS AN OLD UPDATE FILE CONTAINING NOTICES FROM 8/94-4/95. THE FILE CONTAINING NOTICES FROM 5/95 TO 2/96 HAS BEEN LOST. THE LOST FILE CONTAINED A NOTICE REGARDING A POTENTIALLY IMPORTANT BUG FOR CALCULATIONS OF HIGH PRESSSURE PHASE EQUILIBRIA, THAT WAS CORRECTED EARLY IN 1996. IF YOU ARE USING VERSIONS OF PERPLEX COPIED IN 1995 YOU SHOULD REPLACE THEM WITH THE NEWER VERSION. jamie@erdw.ethz.ch fax:1-632-1088 phone:1-632-3955 Dear Perplexed User: A list of the most recent changes to perplex follows. =========================================================== Mac PostScript Converter, April 17, 1995. epsConverter is a Mac program to convert PSVDRAW PostScript to a format that can be read by Adobe Illustrator 5+ or CorelDraw. It is similar to the Ghostview filter ai2ps.ps, but easier to use. You can copy it from the macapps subdirectory. NOTE: epsConverter output is not entirely correct for pre 5.0 versions of Adobe. =========================================================== IBM-PC and WORKSTATION/PERPLEX PROGRAMS. Stefano Poli (UMIPET@IMICILEA.CILEA.IT, Milan) has agreed to maintain current versions of PERPLEX programs compiled for pc-windows and risc6000 machines. Copy these programs by anonymous ftp from biko.terra.unimi.it (IP no 159.149.13.10). i) login as anonymous ii) type e-mail address as passwd iii) cd pub/perplex/risc6000 or cd pub/perplex/pcwin iv) set transfer mode binary v) copy files NOTE: Stefano only maintains the binaries, you must copy the documentation and data files as described above for non-IBM users. Also be careful to compare the date of the binaries to the list of corrections in the PERPLEX.UPDATE file to make sure that the binaries are current. =========================================================== VERTEX/TERTEX/BUILD revision. Mar 16, 1995. Changes were made to files vertex.f, build.f, tertex.f, rlib.f, and vlib.f. The primary effect of these should be to decrease the likelihood that VERTEX or TERTEX does not completely trace univariant curves. The "improvements" cause a 50% increase in the time required for a calculation. Additionally, an error that caused VERTEX/TERTEX to misstate the reaction labeling convention for chemical potential diagrams has been corrected. I have also experimented with, and changed, the numeric tolerences specified in the datafiles, so if you update your sources you should also update the data files. I still don't really understand how these work and they can have a dramatic effect on performance. In particular the last three tolerences in the data files (line 6) are rtol, utol, and ptol. rtol must always be less than utol which is the tolerance for testing delta(G) = 0 for a univariant equilibrium, and ptol must be about 5 times larger than rtol. Other than this there are no rules, increasing utol means there will be more slop in the location of equilibria, but also means the program will run much faster. =========================================================== MacIntosh applications updated. Feb 13, 1995. With current dimensioning vertex requires ca 4.8 Mbytes RAM. =========================================================== FRENDLY BUG. FEB 13, 1995. The parameter k1 was = 1000 and should have been set = 1200. This cause various confusing errors. To repair the source change all occurences of the string: k1=1000 to: k1=1200 =========================================================== BUILD BUG. FEB 13, 1995. A counter in file build.f was incorrectly set = 1, causing build to prompt for more components than necessary for constrained bulk composition calculations. To repair the source change line 95 from: isct=0 to: isct=1 =========================================================== VERTEX BUG. Jan 26, 1995. A routine ASSPTX in file rlib.f had an error that stopped, in certain circumstances, VERTEX from outputting points along univariant curves with small or zero slopes. As a consequence of the error VERTEX may have computed incomplete diagrams. To repair the source change the line (1270, file rlib.f) in ASSPTX: if (dabs(v(iv2)-ptx(ipt2))/delv(iv1).gt.0.01.or. * dabs(v(iv1)-ptx(ipt2-1))/delv(iv2).gt.0.01) goto 10 to: goto 10 =========================================================== =========================================================== =========================================================== =========================================================== OLD UPDATES AND BUGS: =========================================================== TUTORIAL UPDATE. Jan 17, 1995. The PERPLEX tutorial has now been expanded to include chapters on: (6) Phase diagrams for graphitic rocks (7) P-T projections for a system including a fluid of variable composition (8) Thermobarometry (9) Calculations with fixed bulk compositions =========================================================== NEW PROGRAMS COHSRK, ISO, SPECIES. Jan 17, 1995. PERPLEX includes three programs for manipulating COHS fluid equations of state. COHSRK computes fluid properties and speciation as a function of a number of different variables; ISO computes bulk isochores for variable speciation fluid inclusions; and SPECIES produces a plot of species and atomic fractions of fluids as a function of f(O2), f(S2), X(O), X(C), or X(S). see file README.COHSRK, README.ISO, and README.SPECIES =========================================================== UPDATE. Jan 3, 1995. Fluid routines 7 & 8 for COHS fluid speciation now permit the user to enter his/her own function for the P-T dependence of f(O2) or f(CO2). =========================================================== BUG. Jan 3, 1995. Fluid routine 11 did not return the correct value for f(H2) if the user projected through carbon. =========================================================== UPDATE. Jan 3, 1995. The parameter XMAX(I,1) for pseudocompound generation by symmetric transform subdivision is now used if -1 < XMAX(I,1) < 1. If you use this method check the revised program documentation (file vdoc.ps), Doc Sect 4, p 17, heading "IMD(I) = 1, ISP(I) = 2" for the significance of XMAX(I,1). There was also a typo in the 4th line of the preceding paragraph (heading "IMD(I) = 0, ISP(I) = 2 or 3"), XMX(I,2) should be XMX(I,1). =========================================================== UPDATE GRAPHICS. Dec 23, 1994. For those of you who want to edit PostScript generated by PSVDRAW with something other than IDRAW, here is a cheaper (free), though more complex and less general, solution than TransverterPro (see PERPLEX README). You can convert generic (encapsulated) PostScript generated by PSVDRAW to Adobe Illustrator format, the converted file can be read and edited by Adobe Illustrator 5.0 (Macs, but not by the 3.0 version), or Coreldraw (PC's). To do this you must copy the file ps2ai.ps by ftp from toby.princeton.edu directory /pub/olszewsk. This file is a filter that can be used (as described in the comments in ps2ai.ps) in conjunction with Ghostscript. Even without the ps2ai filter, Ghostscript permits you to view PSVDRAW output and to write PICT and PPM formats. You can copy Aladdin Ghostscript for UNIX and IBM-PC's (MS-DOS, OS-2, Etc) from ftp.cs.wisc.edu by anonymous ftp from directory /ghost/aladdin; or you can copy GNU Ghostscript for UNIX, MS-DOS, and Mac's from directory /ghost/gnu. It's possible that recent versions of Ghostscript already incorporate the filter ps2ai.ps, in which case copying ps2ai.ps is unnecessary (but saves you from having to figure out the documentation). Ghostscript includes a generic PostScript viewer that also permits you to view PSVDRAW output without conversion. Although I'm not entirely sure of the legal aspects, I've put a copy of Ghostscript and ps2ai.ps for Mac's in the archive MacGS.sit in my directory: /home/imp/jamie/thermo/export/macapps There are so many versions of Ghostscript for IBM-PC users that it's not practical for me to maintain copies, however there is now a copy of ps2ai.ps file in /home/imp/jamie/thermo/export/graphics. Note that you can also run ps2ai on a UNIX workstation and then transfer the file to your Mac or PC for editing. =========================================================== SUPCRT DATA BASE, Dec 23, 1994. SUPCRT data base revision for K-Na-Al-Si-O-H phases by Vitalii Pokrovskii. The new data is in the file sup93ver.dat. =========================================================== PSVDRAW, Dec 15, 1994. The old version of PSVDRAW may draw the first tick mark on the x-axes of diagrams incorrectly (parallel to the axis). =========================================================== =========================================================== =========================================================== =========================================================== OLD UPDATES AND BUGS: =========================================================== GRAPHICAL EDITING: There is a commerical program (TransverterPro) that will convert psvdraw PostScript into images that can be edited with your favorite drawing program. For a Mac (also available for Windows and Unix) it costs 295 US$ from: TechPool Software 1463 Warrensville Center Road Cleveland, Ohio 44121-2676 Fax: 216-382-1915 Supposedly some of the newer drafting tools (e.g. Adobe Photoshop) also can read generic postscript, but I'm not sure if they permit pixel or object oriented editing. =========================================================== MAC USERS: The self-unstuffing archives for PERPLEX that were in place between Sept 9 - Oct 4 (1994) did not unstuff on some Mac's. This problem has been repaired and the archives have been reorganized into a smaller files in a new directory (see guide in /home/imp/jamie/thermo/export). =========================================================== IBM-PC USERS: Daniele Castelli & G. Franceschi have written an interactive graphics program VertexView that runs under Windows for displaying and manipulating the output of Vertex (at present only composition and Schreinemakers diagrams). VertexView can be obtained from Daniele Castelli at: Dipartimento di Scienze Mineralogiche e Petrologiche Universita di Torino Via Valperga Caluso 37 10125 Torino, Italy Daniele's (sometimes unreliable) E-mail address is FERRARIS@ITOCSIVM.CSI.IT =========================================================== UPDATE AND CORRECTION TO B92VER.DAT, Oct 15, 1994: The b92ver.dat file containing Berman's data (1988, revised for TWEEQU 6/92) has been updated (by a person who shall remain anonymous for security reasons) to include grunerite (gru), Fe-clinochlore (da), greenalite (gre), iron (iron), and minnesotaite (min). The anonymous person also noted that a heat capacity coefficient for glaucophane is in error in the TWEEQU files and, therefore, in b92ver.dat. =========================================================== CORRECTION TO HP94VER.DAT: Due to an error in my translation program, the version of Holland & Powell's data in file hp94ver.dat has errors for the phases fayalite (fa), ferrodolomite (fdol), ulvoespinel (usp), and methane (ch4). These errors were corrected Oct 14, 1994. the correct entries are: fa 1 0 0 0 H= -1476430. 0.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00 2.00 0.00 0.00 0.00 -1377215. 151.0000 4.630000 59.90000 0.7062000E-01 -5743700. 0. 2012.100 0. 0. 0.1520000E-03 0. -.3400000E-05 0. 0. fdol 1 010 0 H= -1968880. 0.00 0.00 0.00 0.00 0.00 1.00 0.00 0.00 1.00 0.00 0.00 2.00 -1816707. 185.6000 6.572000 442.1000 -.4743000E-01 2037000. 0. -4931.000 0. 0. 0.1900000E-03 0. -.6500000E-05 0. 0. 0. 1273.000 8.000000 0. 0. 0. usp 1 0 0 0 H= -1505830. 0.00 0.00 0.00 0.00 0.00 0.00 1.00 0.00 2.00 0.00 0.00 0.00 -1408427. 168.9000 4.682000 -102.6000 0.1425200 -9144500. 0. 5270.700 0. 0. 0.1920000E-03 0. -.2600000E-05 0. 0. 0. CH4 1 0 0 0 H= -74810.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 -2.00 2.00 1.00 -50695.63 186.2600 0. 150.1000 0.2062000E-02 3427700. 0. -2650.400 0. 0. 0. 0. 0. 0. 0. 0. =========================================================== CORK IMPLEMENTED, Oct 5, 1994. Earlier versions of PERPLEX were using an outdated version of Holland & Powell's fluid equation of state. Holland and Powell's (CMP, 1991) CORK equation is now implemented as fluid equation of state choice 5. CORK is probably the best high pressure H2O-CO2 EoS available in PERPLEX, and it should certainly improve the high pressure performance of (at least) Holland and Powell's data. =========================================================== BUG IN FIXED BULK COMPOSITION CALCULATIONS, Aug 20, 1994. A bug in VERTEX caused mineral modes and back-calculated compositions to be reported incorrectly (although the calculated equilibria were correct). This bug was fixed Aug 20, 1994. There is still no documentation for the fixed bulk composition option except that in the README file. It's actually a very useful option (especially for thermobarometry) and I'll add it to the tutorial soon. =============================================================