From granite.sentex.net!sentex.net!hyper Fri Nov 11 05:05:19 1994 From: granite.sentex.net!hyper.hyper.com!granite.sentex.net!IASI.RM.CNR.IT!CONTINIG Date: Fri, 11 Nov 1994 09:50:20 +0100 Subject: information To: granite.sentex.net!hyper.com!hyperchem Dear, Please, I'd like to receive a demo disk, general information and a quotation of the software "hyperchem". My address is: Dr. Giorgio Contini C.N.R. I.T.M. via Bolognola, 7 00138 Roma Italy Tel. +39 6 8804361 Fax. +39 6 8804463 Syncerely, Giorgio Contini ________________________________________________________________________ From granite.sentex.net!wscgate.wsc.edu!jdunning Thu Nov 17 14:46:17 1994 From: granite.sentex.net!wscgate.wsc.edu!jdunning Date: Thu, 17 Nov 94 13:19:27 cst To: granite.sentex.net!hyper.com!hyperchem Subject: Hyperchem key Our institution has just purchased Hyperchem and have installed it on our Netware 3.11 Network. I have installed the license key on a workstation (which also has a printer attached to the same parallel port). When Hyperchem attempts to aquire a license, my printer spews about 1 page of garbage and Hyperchem announces that it can't aquire a license for the software. HELP. Any suggestions would be greatly appreciated. |\______/| John B Dunning - Network Systems Specialist / /\ \ Wayne State College | o_| |_o | Home of the Wildcats | ( \/ ) | \__\|==|/__/ voice: (402) 375-7447 fax: (402) 375-7411 \|==|/ e-mail: jdunning@wscgate.wsc.edu ________________________________________________________________________ From granite.sentex.net!watson.Princeton.EDU!llavelle Mon Nov 21 12:06:48 1994 From: granite.sentex.net!watson.Princeton.EDU!llavelle Date: Mon, 21 Nov 1994 11:19:11 EST To: granite.sentex.net!hyper.com!hyperchem Subject: aqueous baths larger than 56.1A Greetings. Has anyone made an aqueous and/or nonaqueous 'box' that has at least one side larger than 56.1A. Any advice on how this was done would be much appreciated. Hope to hear from you soon. Thank You. ________________________________________________________________________ From DBellnie@hs1.buffalo.edu Tue Nov 22 11:50:24 1994 From: "Bellnier, David" To: hyperchem@hyper.com Date: Tue, 22 Nov 94 10:49:00 PST Is there a program that will actually dock a ligand to a binding site (automatically; ie., will rotate bonds + rotate/translate ligand/water molecules) that uses IBM-PC(Pentium)/Windows as a platform? We've tried Chem-X but its current incarnation is FUBAR. Thanks. David Bellnier, Ph.D. Cancer Research Scientist ________________________________________________________________________ From 100046.123@compuserve.com Tue Nov 22 12:21:24 1994 Date: 22 Nov 94 11:17:17 EST From: Michael Taylor <100046.123@compuserve.com> To: "Hyperchem User's Group" Subject: Windows NT Dongle Problems Hi there. We have some problems using the single-user hardware key (dongle) under Windows NT. Graham Hurst at Hyperchem has provided the appropriate Rainbow Port NT driver for the lock and it seems to have installed OK, but every menu command (which I understand from Hyperchem causes the dongle to be checked for a valid licence) gives a delay of 16 seconds (for example, when calling up File | Open) before anything visible happens. When the program is first started, the delay is even longer (a couple of minutes) which led me to believe at first that the program had crashed. Does anyone else use the single-user version under Windows NT? Any similar experiences or tips on what I might have wrong in Windows NT configuration to cause this? Any ideas would be appreciated. Cheers, -=Mike Taylor=- Paint Research Association, Teddington, UK ________________________________________________________________________ From llavelle@watson.Princeton.EDU Tue Nov 22 12:42:58 1994 From: llavelle@watson.Princeton.EDU Date: Tue, 22 Nov 1994 11:43:08 EST To: hyperchem@hyper.com, support@hyper.com Subject: Aqueous boxes larger than 56A. Amber Parameters for the Protonated Nucleic Acid Bases: dA+ and dC+. Greetings. (A) Has anyone made an aqueous or nonaqueous box that has at least one side larger than 56A. Any advice on how this was done would be much appreciated. Even better would be the implementation into HyperChem that allows the user to vary the box size with no limitation. For those of us looking at fairly linear polymers (DNA) long rectangles are what is needed, boxes are suitable for globular molecules (proteins). (B) If anyone knows where to find Amber parameters for protonated nucleic acid bases, in particular dA+ and dC+, please let me know. Here again HyperChem has many functions for peptides and proteins but is lacking when it comes to Nucleic Acids. The implementation of looking at the DNA backbone using a Ribbon type feature would also be Very Helpful. Hope to hear from you soon. Thank You. Laurence Lavelle Princeton University Depts. of Chemistry and Molecular Biology Princeton, NJ 08540 Tel:(609)258-3935 Fax:(609)258-3980 P.S. Enjoy Thanksgiving. ________________________________________________________________________ From hurst Tue Nov 22 15:29:03 1994 Date: Tue, 22 Nov 94 15:29:03 -0500 From: hurst (Graham Hurst) To: llavelle@watson.Princeton.EDU Subject: Re: Aqueous boxes larger than 56A. Amber Parameters for the Protonated Nucleic Acid Bases: dA+ and dC+. Cc: hyperchem On Tue, 22 Nov 1994 Laurence Lavelle asked (in an expanded version of a message sent the previous day): > Greetings. > (A) Has anyone made an aqueous or nonaqueous box that has at least > one side larger than 56A. Any advice on how this was done would be much > appreciated. > Even better would be the implementation into HyperChem that > allows the user to vary the box size with no limitation. For those of us > looking at fairly linear polymers (DNA) long rectangles are what is needed, > boxes are suitable for globular molecules (proteins). This request was entered into our wishlist the first time that you asked! HyperChem uses two water boxes: one that is 18.70136 A per side with 216 water molecules, and the other which is 56.10408 A per side. When we first wrote the solvation code, we didn't realize that the large box is 27 periodic images of the small one so HyperChem ended up with a finite maximum box size! You could build a larger box by using the ChemPlus crystal builder. Start by doing File/New and then Setup/Periodic Box to make the 18.70136 A box with 216 water molecules. Then start Crystal Builder and use the HyperChem/Get menu. You can then change the number of unit cells in each dimension and use HyperChem/Put to create the larger box in HyperChem (which may take a very long time!). You can then save this and merge it with the solute, but you will have to find some way of deleting water molecules that are too close to the solute. > (B) If anyone knows where to find Amber parameters for protonated > nucleic acid bases, in particular dA+ and dC+, please let me know. > Here again HyperChem has many functions for peptides and proteins > but is lacking when it comes to Nucleic Acids. The implementation of > looking at the DNA backbone using a Ribbon type feature would also be Very > Helpful. Your request for DNA ribbons is also in our wishlist. > Hope to hear from you soon. Thank You. > > > Laurence Lavelle > Princeton University > Depts. of Chemistry and Molecular Biology > Princeton, NJ 08540 > Tel:(609)258-3935 > Fax:(609)258-3980 > > P.S. Enjoy Thanksgiving. I did - it was on October 10th in Canada! ;-) Cheers, Graham ------------ Graham Hurst (hurst@hyper.com) Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" to hyperchem-request@hyper.com ________________________________________________________________________ From DBellnie@hs1.buffalo.edu Tue Nov 22 16:05:13 1994 From: "Bellnier, David" To: hyperchem@hyper.com Subject: returned e-mail Date: Tue, 22 Nov 94 15:29:00 PST I see that my e-mail query regarding software other than hyperchem was promptly returned to me. Does that mean that it was *not* sent out to members of the users group? If so, I'm unclear of the user group policy regarding such matters. Could you please fill me in? Thank you. David A. Bellnier, Ph.D. ________________________________________________________________________ From hurst Wed Nov 23 10:54:13 1994 Date: Wed, 23 Nov 94 10:54:13 -0500 From: hurst (Graham Hurst) To: "Bellnier, David" , hyperchem@hyper.com Subject: HyperChem FAQ (long) David Bellnier wrote: > Is there a program that will actually dock a ligand to a binding site > (automatically; ie., will rotate bonds + rotate/translate ligand/water > molecules) that uses IBM-PC(Pentium)/Windows as a platform? We've > tried Chem-X but its current incarnation is FUBAR. Thanks. > David Bellnier, Ph.D. > Cancer Research Scientist I'm not aware of any Windows based programs that can do automatic docking but you can approach the problem with HyperChem. At the risk of wasting bandwidth, I'll append a prototype FAQ that I've cobbled together. Docking is discussed in question 6. Cheers, Graham ============================================================================== Frequently Asked Questions for HyperChem, ChemPlus and HyperNMR ============================================================================== This is a list of frequently asked questions compiled by Hypercube's technical support staff. Most of the answers have appeared before on the HyperChem Email Users' Group (which you can join by sending the message "subscribe hyperchem" or "subscribe hyperchem your_email_addr" to hyperchem-request@hyper.com. The questions are: 1. Can I get Hyperchem for MacIntosh? 2. Will HyperChem for Windows run on Windows NT for Intel? 3. What does the message about trying to obtain a license mean? 4. I've got trouble with the hardware lock? 5. I read MOL2 files from HyperChem into Spartan or Sybyl and get wrong types? 6. How can I sustitute inhibitors or do docking? 7. How can I simulate a reaction? 8. I get a warning mesage about missing AMBER, BIO+ or OPLS parameters? 9. How much memory do HyperChem calculations need? 10. How can I read in a file of atomic Cartesian coordinates? 11. Why don't the semi-empirical methods allow elements beyond Xenon? 12. I don't get the plot that I'm expecting? 13. What is the format of the force matrix in the log file? 14. I get a bad structure from ZINDO/1? 15. Why won't ChemPlus work in Europe? 16. What are the HIN extensions that Molecule Prensentations uses for orbitals? 17. Why do I get out of memory errors with HyperNMR? 18. How accurate is HyperNMR? ============================================================================== 1. Can I get Hyperchem for MacIntosh? ------------------------------------------------------------------------------ There is no MacIntosh version of HyperChem at this time and the Windows version will not run under SoftWindows on PowerMacs because SoftWindows does not (yet) support Windows in 386 Enhanced mode. We've been considering a Mac version for over 6 years. So far our estimated porting costs outweigh our estimates of increased revenue, but there are some automated porting tools on the horizon which may change the equation... ============================================================================== 2. Will HyperChem for Windows run on Windows NT for Intel? ------------------------------------------------------------------------------ HyperChem for Windows will run under Windows NT for Intel provided that you install an NT device driver for the parallel port lock. Hypercube will supply this driver free of charge via uuencoded email or floppy disk upon request. ============================================================================== 3. What does the message about trying to obtain a license mean? ------------------------------------------------------------------------------ The message should changed to something like: "No local hardware lock found and no network license is available. Please install a hardware lock on your parallel port or wait for a free license if you are using network licensing." It is in our bug list to improve this message. ============================================================================== 4. I've got trouble with the hardware lock? ------------------------------------------------------------------------------ Two users have reported that they needed to move the dongle from LPT2 to LPT1 to get it to work with Release 4. There are also known problems with IBM microchannel bus, where the CMOS setup has to be changed to disable Parallel Port Arbitration Level for the dongle to function. Rainbow Technologies (the manufacturers of the lock and its driver software) has also advised us that there may be dongle trouble with fast Pentium machines using a fast, local-bus parallel port which can usually be solved by using an ISA bus parallel port instead (no HyperChem users have reported this problem so far). Please check your parallel port configuration, keeping the above in mind, as the dongle should work. ============================================================================== 5. I read MOL2 files from HyperChem into Spartan or Sybyl and get wrong types? ------------------------------------------------------------------------------ When HyperChem writes MOL2 files, it writes the element correctly, but uses the default atom type (Du) for all atoms. This was a deliberate design compromise, to avoid having to map all of the four HyperChem forcefield type sets (each with completely user definable types!) to the set of types in the current Tripos force field (which could change, though we don't know if a Tripos user can change their definitions or not). At one point we considered making the MOL2 filter read-only because of this, but opted for writing with default types instead. Too bad neither Tripos or Wavefunction offer the ability to assign types based on user-editable rules like HyperChem can. ;-) ============================================================================== 6. How can I sustitute inhibitors or do docking? ------------------------------------------------------------------------------ Let's assume that for your substitution you want to place the new inhibitor in the same place as the old one. You can do this by using the draw tool to make the minimum changes needed to change the old inhibitor to the new inhibitor. You can then select only the changed parts and model build (the unselected atoms will keep the same coordinates). If you are interested in semi-empirical calculations you could then start modelling. For molecular mechanics simulations you would first need to set the atom types and partial charges of the new inhibitor (you don't need charges for MM+ if you are using bond dipoles for non-bonded electrostatics). The types can be set by selecting the inhibitor and using Build/Calculate Types (checking that none are ** unless you are using MM+). The charges are trickier - you could enter "standard" values manually if have them, select the inhibitor and use ChemPlus QSAR Properties to calculate partial charges, or isolate the inhibitor and use a semi-empirical calculation. A feature that you might wish to use, particularly to "isolate the inhibitor" for charge calculations, is File/Merge once you have set it to not translate the merged-in molecule. By default, File/Merge and Edit/Paste translate the added part (in the viewer's X direction) so that it does not overlap with the current system. You can turn off this behaviour via a HyperChem script command. Perhaps the simplest way to do this is to use Script/Open Script to read in a text file that consists of the script command: translate-merged-systems no If you always want this behaviour, put this command in a CHEM.SCR file (in your ChemIniPath) so that the command is executed whenever you start HyperChem. Once you have turned off translate-merged-systems, you can save, delete and merge to work with parts of the system independently without losing relative position information. For instance, to get charges after model building the changed portions of the inhibitor, you could: 1. Save the total system, eg. as both.hin. 2. Select and cut (via Edit/Cut) the inhibitor, then save as protein.hin. 3. Use Edit/Paste, Select/Complement Selection, Edit/Delete and then save as inhibit.hin. 4. Now use a semi-empirical calculation (single point or optimization) to calculate charges (and perhaps a better geometry). 5. It's probably a good idea to save, perhaps overwriting inhibit.hin. 6. Use File/Merge to read protein.hin and save, perhaps overwriting both.hin. Now you should have a reasonable starting point for molecular mechanics calculations of the new inhibitor and the protein. When you don't already have an inhibitor or substrate to replace, HyperChem allows two methods of "docking" a small molecule (substrate) onto a larger molecule (receptor). These are manual docking, and restrained energy minimization. Manual docking is carried out purely through moving molecules around on the screen: no actual simulation is carried out. Any "docked" structure obtained in this way should then be optimized to remove bad contacts and other unrealistic aspects of the structure. Restrained minimization requires a docking hypothesis (atom A on the substrate is assumed to be close to atom B on the receptor) which is incorporated into a simulation using restraints. A combination of dynamics and optimization subject to these restraints allows you to force proximity of substrate and receptor atoms. After restrained optimizition you can turn off the restraints and use further optimization and dynamics to test the docking hypothesis. If you select only the substrate to freeze the receptor, you may prefer to use one-atom restraints to tether substrate atoms to points in space that you define to be in your proposed docking site. ============================================================================== 7. How can I simulate a reaction? ------------------------------------------------------------------------------ The way to set up reactions is by using selections, Setup/Set Velocity and checking Restart in the MD dialog box. Of course you also have to be using a semi-empirical QM method that can handle the system you want to simulate! Thus to set up an SN2 reaction: 1. Set up the initial geometry. Create the methyl chloride molecule and a chlorine atom. Use Edit/Translate, Edit/Rotate and/or Edit/Align Molecules to position them. For instance draw and build methyl chloride, us Align Molecules to align the primary axis with the X axis. Select the carbon and translate it to the origin (assuming whole molecule translation is on in the Tool Preferences dialog box). Then draw the chlorine atom, select it and translate it to -8,0,0. Clear the selection. 2. Choose the calculation options. Set up the semi-empirical options (eg. PM3, charge -1, singlet) and do a single point calculation to make sure all is okay. 3. Define the initial velocities to cause a collision. Select the carbon and use Select/Name Selection to name it as POINT. Clear the selection and then select the isolated chlorine atom. Use Setup/Set Velocity and choose At POINT for direction and 50 for the velocity. Clear the selection. 4. "Fix" the display. To have sphere intersections calculated correctly, draw bonds wherever you think spheres will intersect during the reaction. In this case connect both chlorine atoms to all hydrogens and the carbon. Change to CPK rendering (and perhaps adjust the view. 5. Run dynamics and save snapshots. Use Compute/Molecular Dynamics and choose: Heat time 0, Run time 0.5, Cool time 0, Step size 0.0005 ps, temperature 300, *no* Constant temperature, data collection period 1, screen refresh period 1, and check Restart. Click on Snapshots... and type sn2.hin for the file name (or equivalent but be sure to use hin as the extension!). Then run the simulation! You can Cancel it after the reaction has proceeded as far as you want to follow it. 6. Play back the simulation. Open sn2.hin and it should warn you that modifying the system will cause snapshot data to be lost. You can play back the simulation by checking Playback in the MD dialog box (use Snapshots to choose a subset of of the trajectory). Obviously there are lots of variables you can play with: initial velocity, "target" location (perhaps off-centre), collection periods, etc. ============================================================================== 8. I get a warning mesage about missing AMBER, BIO+ or OPLS parameters? ------------------------------------------------------------------------------ You have run into the most common problem in molecular mechanics: missing parameters. See Chapter 3 of the HyperChem Computational Chemistry manual (in Part I) for a description of what parameters are. We simply provide published parameters for the published force fields AMBER, CHARMM (which we call BIO+ for trademark reasons) and OPLS. For MM+ we also provide a default parameter scheme for generating parameters when published MM2 parameters are not available, so that you never get warnings about missing parameters. If you are using a selection so that the atoms with the missing parameters are not selected in the calculation then it is safe to ignore missing stretch, bend or torsion parameters. Otherwise proceeding without parameters can lead to funny optimized geometries, because the program ignores any interactions that it doesn't have parameters for. See the "Warning Messages" part of the Compute Menu section in Chapter 7 of the Reference manual. If you decide that you need parameters, you can search for them in the literature (we only include parameters from the force field authors and you might find some published by other groups using the force field) or develop your own. At its crudest level this means guessing based on parameters for similar types, but parameter development is an area of active research. Finding or developing parameters for you is beyond the scope of product support for HyperChem. The following are some basic instructions for adding parameters. You should also see the Parameter Sets and Compiling Parameters sections of the Molecular Mechanics chapter in Part 2 of the Computational Chemistry manual, the Setup/Select Parameter Set and Setup/Compile Parameter File menu item descriptions in Chapter 7 of the Reference manual and Appendices B and F of the Reference manual. To add new parameters to an existing parameter set you should: 1. Note which parameters are missing. You should note the kind of interactions (i.e. stretch, bend, non-bonded, etc.) and the atom types involved from the warning messages. You may find it convenient to save a log file so that messages are recorded there. (You can also set WarningType=Log in CHEM.INI or by a script so that warnings from the calculations only go to a log file and not to the screen, but remember to switch it back later or you won't get any warnings unless you save and read a log file!) The warning messages will probably have much duplication; you just need to identify the unique set of missing parameters. 2. Figure which parameter file to add new parameters to. First you need to know which force field you are using (MM+, AMBER, BIO+ or OPLS) and which parameter set for that force field (by using the Setup/Select Parameter Set menu and the associated dialog box). You can see which files are used for that parameter set by looking in the CHEM.INI file. For instance, for the AMBER forcefield and amber3 parameter set, you would look in the [amber,amber3] section of the CHEM.INI file using a text editor. The parameters for each type of interaction (i.e. stretch, bend, non-bonded, etc.) is stored in a different file pointed to by a CHEM.INI keyword. For instance the [amber,amber3] section should have a line like: QuadraticBend=am89aben.txt This tells HyperChem to find the bend parameters for the amber3 parameter set in the file am89aben.txt. All the keywords are given in Appendix F of the Reference manual and the formats for the associated files are described in Appendix B. 3. Add new parameters to the appropriate file or files. It is a good precaution to back up the original before modifying a a parameter file. Use a text editor to edit the text parameter file or files. (You can also keep parameters in dBASE 3 format files, in which case you would need to use a database program to modify the files.) The file formats are described in Appendix B of the Reference manual and the equations that the parameters are used in are given in the Force Fields section of the Molecular Mechanics chapter in Part 2 of the Computational Chemistry manual. 4. Compile all the parameter sets that use the modified file or files. Use the Setup/Select Parameter Set and Setup/Compile Parameters menu items to choose and recompile all the parameter sets that use the files you have modified. You can see which sets use which files in the CHEM.INI file. The Parameter Sets and Compiling Parameters sections of the Molecular Mechanics chapter in Part 2 of the Computational Chemistry manual also describe adding and compiling new parameters or parameter sets. You should use a new set if you want to modify existing parameters or be able to easily switch back and forth between parameter sets. ============================================================================== 9. How much memory do HyperChem calculations need? ------------------------------------------------------------------------------ The details of this application note refer to HyperChem Release 4 for Windows, but the calculations work the same way for earlier Windows releases, so you should be able to transfer the results to your system. Types of memory =============== The memory available to HyperChem is the sum of random access memory (RAM) and swap space, or virtual memory. The latter is a section of a hard disk set aside for use by programs during their execution. Virtual memory is much slower than RAM, as disk access is required. However, the size of calculation that can be done depends only on the total memory available, RAM + swap space. In Windows, the size of your swap space is set in the Control Panel application, that is part of the "Main" application group. Choose the "386-Enhanced" item, and click on the "Virtual Memory" button to set this size. Memory Requirements: an Outline =============================== The total memory that needs to be available in order to carry out a calculation using HyperChem is as follows: 1. Memory required by Windows, DOS and items specified in your config.sys and autoexec.bat files, such as Smartdrive. 2. Memory required by the HyperChem executable (CHEM.EXE) 3. Memory required by the computational "backend" executable (NDO.EXE, EHT.EXE, MMPLUS.EXE or NEWTON.EXE) 4. Memory required to carry out the particular calculation. Items 1, 2, and 3 are independent of the molecular system under study. Item 4 depends on the particular calculation you are carrying out. You should note that any other applications running will also consume memory: when doing large calculations it may be necessary to ensure that you are not using other Windows applications. Finding the Amount of Memory Available on your Computer ======================================================= If you click on the "About" item under the Help menu of Program Manager, you will see the amount of memory currently available to you. In general, this does NOT add up to the total of your RAM + swap space, but is substantially less. On a machine with 8 Mb RAM and 16 Mb of swap space, (which "should" have 24 Mb free) the value may be about 18 Mb. The difference is due to drivers and other items loaded in your config.sys and autoexec.bat files, as well as the basic DOS and Windows operating system requirements. Fixed Memory Requirements ========================= While running any calculation in HyperChem, memory is needed to load both the "front end" program (CHEM.EXE) and the computational "back end" program into memory. Fixed Memory Requirements for HyperChem Programs =================================================== Program Calculation Type Memory Required / KB --------------------------------------------------- HyperChem All 640 (CHEM.EXE) EHT.EXE Extended Huckel 670 NDO.EXE Other semi- 2,500 empirical MMPLUS.EXE MM+ 350 NEWTON.EXE Other molecular 200 mechanics =================================================== To find the memory available for a particular calculation, use the following procedure. 1. With no applications running in Windows, look at the About box in Program Manager to find the available memory. 2. Subtract 640 KB to account for loading HyperChem 3. Subtract the amount in the table above for loading the program appropriate for the method you are using. Calculation-Dependent Memory Requirements ========================================= NDO Calculations ---------------- The following figures are for AM1 or PM3 calculations. These methods are the most demanding of the semi- empirical methods: calculations on a similar size molecule using CNDO requires significantly less memory. SCF Calculation --------------- The major contribution to memory usage (not the only one) for a 'standard' RHF calculation is 104 * (Number of orbitals)2 bytes. If you are doing other kinds of calculation, additional memory is required: UHF: + 40 * (Number of orbitals)^2 bytes Convergence accelerator: + 160 * (Number of orbitals)^2 bytes Gradient Calculation -------------------- HyperChem computes the gradient for all kinds of calculation, so even for a single point calculation to complete, you will need space for a gradient evaluation. The gradient requirements are in addition to the SCF requirements. In the following expression, NHA = "non-hydrogen atoms", HA = "Hydrogen atoms". Memory required = 2400 * NHA * (NHA - 1) + 440 * NHA * HA + 24 * HA * (HA-1) bytes. For most systems, this is substantially less than the memory required for the SCF calculation. For systems with no hydrogens in, however (e.g., C60) the gradient requirements are comparable to that of the SCF. Examples --------- The "zero" in the following examples is the value given in the Program Manager Help with HyperChem running and NDO loaded. All calculations are AM1 RHF single points, with SCF convergence acceleration unchecked. NDO Memory Requirements ========================================== Molecular formula # of Extra memory orbitals used ------------------------------------------ CH4 (methane) 8 0 Mb C10H10 50 0.8 Mb (naphthalene) C16H10 (pyrene) 74 1.3 Mb C24H12 108 2.5 Mb (coronene) C30H18 138 4 Mb C40H22 182 8 Mb C48H24 216 9.5 Mb ========================================== Significant disk activity, and consequent slowing of the calculation, starts at around the amount of RAM in the computer. EHT Calculations ---------------- Extended Huckel calculations require significantly less memory than NDO calculations, because they do not include any three-electron integrals. The main data storage involves the orbital coefficients and various square matrices, such as the Hamiltonian and overlap matrices, each of which have N2 elements (N being the number of orbitals in the problem). The memory usage is proportional to the square of the number of orbitals involved in the calculation. Typical cases are shown in the figure. Extended Huckel Memory Requirements =============================== Number of Orbitals Memory / Mb ------------------------------- 0 0 90 0.4 180 1.5 270 3.4 360 6 =============================== MM+ Calculations ---------------- The memory requirements for the MMPlus program are proportional to the square of the number of atoms in the system. MMPlus keeps a list of all interactions in the system, and as each atom interacts with all others, this dictates a quadratic dependency. The graph below indicates requirements in addition to the static requirements described above. MM+ Memory Requirements ====================== Number of Memory atoms Used ---------------------- 10 0.58 60 0.58 258 0.8 652 2 904 3.2 1304 5.2 1956 9.7 ====================== Newton Calculations ------------------- The Newton program is more efficient than MMPlus in its memory requirements (as well as in speed) and maintains only information about each atom, not about each interaction. Consequently, the memory requirements for Newton are proportional to the number of atoms. The following graph indicates the requirements in addition to the static requirements outlined above. Newton Memory Requirements ====================== Number of Memory atoms Used ---------------------- 472 0.74 944 1.1 1888 1.8 3776 3 ====================== ============================================================================== 10. How can I read in a file of atomic Cartesian coordinates? ------------------------------------------------------------------------------ You can read in a cartesian coordinate file if you reformat it slightly, since a HIN file at its simplest is pretty much just that. Although there is a lot of other stuff in a HIN file, a minimal one can be built that looks like this (for ethylene, without the hydrogens): mol 1 atom 1 - C ** - 0.0 -0.67 0.0 0.0 0 atom 2 - C ** - 0.0 0.67 0.0 0.0 0 endmol 1 Here is a description of what you need: You need the mol 1 and endmol 1 lines at the beginning and end of the entire thing. Each line needs an atom number and element. You need to put in the dashes, but their column numbers are not important, just so long as there is white space (a space character or a TAB character) between the fields. The ** is the place for an atom type, and ** is just "unknown". The first 0.0 is the atomic charge (you can leave it zero). The next three fields are the x,y, z coordinates. (Here -0.67, 0, 0 and +0.67, 0, 0) The final column is the number of neighbours: zero here as we are not putting in any bonds yet. So after reading this in, one has a bunch of dots around the screen. The trick to making a guess at the bonding information is to save the system as a Z-matrix file, and then read it in again. The reason this less-than-obvious method works is that the Z-matrix file format has no room for explicit bonding information, and so we had to put a routine in to make a guess at the bonding whenever reading in such a file. Generally the guesses are quite good for organic molecules in regular geometries, but will be less good for inorganic systems or unusual geometries. Also, the user should check the valences after doing this: there is nothing explicit to stop 5-coordinate carbons, etc. If your coordinate file did not have the hydrogen coordinates in (as is commonly the case with X-ray data) then you can choose "Add hydrogens" from the Build menu to add them on without altering the coordinates of the heavier atoms. If your valences are correct this will work fine for organic systems. For other systems where valences are less well defined, you may need to add hydrogens yourself by drawing them in. If you select the hydrogen atoms and model build, then they alone will be fixed in position, again without altering the positions of the heavier atoms. ============================================================================== 11. Why don't the semi-empirical methods allow elements beyond Xenon? ------------------------------------------------------------------------------ HyperChem uses Slater orbitals and calculates integrals and their gradients analytically. Thus for each increase in principal quantum number we have to program new code for the integrals and their analytic gradients. With high quantum numbers the gradient code becomes very complex and invloves many more types of integrals (i.e. combinations of quantum numbers). So far nobody in the world has derived analytical gradient equations for Slater orbital integrals involving a principle quantum number of six. MOPAC does not use analytic gradients of Slater orbitals, but uses either Slater orbitals with derivatives evaluated by finite difference or analytic derivatives of Gaussian orbitals (where linear combinations of Gaussians are used to approximate a Slater orbital). Once one has taken the trouble to derive and code the integral and integral derivative expressions, as we have for HyperChem, the code is faster and/or more accurate than finite difference or using combinations of Gaussian orbitals to represent Slater orbitals. HyperChem represents the state of the art for Slater orbitals with analytic gradients. To go further in the periodic table using the same philosphy involves considerable effort and a large increase in executable file size. Adding the ability to use Gaussian orbitals would also involve considerable effort and increase in file size. Using finite difference gradients with Slater orbitals would involve less programming effort than the alternatives, but would decrease program execution speed. ============================================================================== 12. I don't get the plot that I'm expecting? ------------------------------------------------------------------------------ Without knowing the problem, the most common cause of weird-looking plots is the location of the plane of the plot. An often overlooked trick in HyperChem is to use selection to define the plane for the plot (after calculating the wavefunction without a selection!) The plane for the plot is parallel to your screen and relative to the center of mass of the selection or of the whole system if there is no selection. (You can see the centre of mass the selection via Display/Show Inertial Axes.) The offset of the plane from the centre of mass is set in the Grid Control dialog box. Thus to get "good" plots: 1. Get the structure that you want plots for in the workspace. 2. Align the molecule and your view. 3. Clear the selection or select the subset to be treated with QM. 4. Do a single point. 5. Select an atom that you want the plot to pass through. 6. Set Plane offset to 0 in the Grid Control dialog box (and bump up grid points to at least 60 x 60) and plot. Repeat 5 and 6 to build up a 3D understanding of the property. You can also change the view (NOT molecular coordinates) and replot. If you still think that the plots look like there is bug in the program, please send email to support@hyper.com or FAX +1 519 725 5193. Note that for problems with semi-empirical methods per se, we only implement established, published methods from the literature in HyperChem, and improving these methods is not part of our product support! Note that a bug has been found in electrostatic potential plotting of HyperChem Release 2, 3 and 4 for large systems. The electronic contribution is subject to an integral cutoff threshold, but the nuclear contribution is not. Thus for distances more than ~10 Angstrom from the center of a large molecule, the potential will only be shown as positive, and this will "swamp" the potential closer to the center. This will be fixed in the next HyperChem release. ============================================================================== 13. What is the format of the force matrix in the log file? ------------------------------------------------------------------------------ A 3N * 3N force constant matrix can be very confusing so HyperChem, like MOPAC, gives an N * N force constant matrix describing the atom-atom interaction forces. This N * N force constant matrix is formed from the 3N * 3N force constant matrix by calculating the norm of the tensor of each atom. You may reconstruct the 3N * 3N force constant matrix from HyperChem log file output using the matrix equation: F C = C E where F is the MASS-WEIGHTED force constant matrix (3N * 3N), C is the eigenvector (normal coordinate) matrix and E is the eigenvalue diagonal matrix. You can get C and E from the log file (with QuantumPrintLevel=9) or script output, you can calculate the matrix F by multiplying on the right by the inverse of C. The relationship between eigenvalue Ei in E and the vibrational frequency f is f = SQRT(Ei) / (2*PI*c) where c is the speed of light. ============================================================================== 14. I get a bad structure from ZINDO/1? ------------------------------------------------------------------------------ It turns out that ZINDO/1 often does a better job with 2nd row transition metals than 1st row transition metals. This seems to be a "feature" of Michael Zerner's INDO/1 method (called ZINDO/1 in HyperChem) that was originally implemented in the ZINDO program from Zerner's group. Some ZINDO experts use different sigma-sigma or pi-pi overlap weighting factors to get better results. One user has reported problems that are due to a poor initial guess for the wavefunction; he got a reasonable structure when he started from one with short bond lengths, but a subsequent single point did not converge. We are continuing to investigate this behaviour, but in general the ZINDO methods are not as robust for transition metals as AM1 or PM3 are for non-transition metal complexes. We have tried very hard (with lots of generous assitance from Tom Cundari and David Baker) to reproduce results of INDO/1 or INDO/S calculations that use Michael Zerner's ZINDO program (including the choice of the initial guess for the wavefunction). The optimization algorithms and wavefunction convergence acceleration options may differ from the ZINDO program, but we are confident that an INDO/1 (or INDO/S) potential surface will be same whether calculated from HyperChem, ZINDO or Argus (Mark Thompson's program). While not very satisfactory, we would advise caution when using ZINDO/1 for geometry optimization. In some cases the model builder, MM+, or even "hand-editing" will give better structures. ZINDO/S (INDO/S) was not developed for geometry optimization and should not be used for any geometry optimization. One should check, via a log file or by monitoring the status line, that all iterations in the optimization actually reached wavefunction convergence. Sometimes better results are observed by turning "Accelerate convergence" off (or on!). ============================================================================== 15. Why won't ChemPlus work in Europe? ------------------------------------------------------------------------------ If you change the decimal separator from "," to "." (and change the thousands separator from "." if necessary) in the International settings of the Windows control panel, all should be well. This problem occurs because of undocumented behaviour in Microsoft's programming products. We have informed Microsoft about this and are awaiting their fix. In the meantime one must use a period as the decimal separator for ChemPlus to work correctly with HyperChem. A problem has also been noted with Windows 3.1 when the language is set to Swedish or Danish - HyperChem's version command stops working which can cause problems with ChemPlus. Switching to a different language or using Windows 3.11 can solve the problem. There is also a problem when Spanish is chosen as the language. In this case the quotation characters required for Windows DDE are changed, breaking DDE communication with HyperChem. Changing from "Spanish" to "Spanish (Modern)" fixes this problem. ============================================================================== 16. What are the HIN extensions that Molecule Prensentations uses for orbitals? ------------------------------------------------------------------------------ The extensions for storing orbital information are: IsosurfaceValue: The value which the isosurface represents. The isosurface is built up from interpolating between the values on the grid points to find point in space with this value. OrbitalRendering: 0 for mesh, 1 for unshaded polygons, 2 for shaded polygons. GridStepCount: The three integers give the number of points in the x, y, and z directions (multiply the three to get the total number of grid points). GridStepSize: This real number gives the increment (in Angstrom) between grid points for each of the x, y, and z directions. GridOrigin: This gives the starting point for the grid (i.e. the point with the lowest x, y, and z coordinates). GridData: The values for the grid points, in the order of z varying fastest, then y and x varying slowest. Note that one set of data may span several lines but each line begins with "GridData" so that HyperChem can ignore it (it is just for line width that there are six numbers per line). Some (non-optimal!) psuedo code for reading in the grid would be: iPoint=0 for ix=1 to GridStepCount(1) x_coord = GridOrigin(1) + (ix-1)*GridStepSize for iy=1 to GridStepCount(2) y_coord = GridOrigin(2) + (iy-1)*GridStepSize for iz=1 to GridStepCount(3) z_coord = GridOrigin(2) + (iz-1)*GridStepSize iPoint = iPoint + 1 Value = GridData(iPoint) end for end for end for This question has come up from users wishing to use ChemPlus to plot isosurfaces of the density or other properties. ============================================================================== 17. Why do I get out of memory errors with HyperNMR? ------------------------------------------------------------------------------ The memory requirements for the calculation of a spectrum with HyperNMR come from three sources that depend on the number of nuclear spin states. The number of spin states that must be considered (let me call this N) can be up to 2^M (less if there are equivalent atoms defined) where M is the number of NMR atoms. Each spin state requires (in bytes): 4 * [(# of non-equivalent NMR atoms) + (# of possible total spin) + 2] Another large memory requirement is for three double precision square matrices (for a Hamiltonian, spin S eigenvectors and spin S+1 eigenvectors). The dimension of each is up to (depending on equivalancies) the largest binomial coefficient for N, namely N!/[i!(N-i)!] where i is the integer closest to N/2. The time required to calculate the spectrum is dominated by the diagonalizations of the N+1 Hamiltonian matrices (with dimensions given by the binomial coefficients for N). The third large memory requirement is 20 bytes per transition, for storing energy, intensity and degeneracy. The number of transitions may be as large as the sum of the pairwise products of neighbouring binomial coefficients for N, depending on equivalencies. For most PCs (i.e. 8-16 Mb), spectra for 12 non-equivalent NMR atoms are practical, with some use of virtual memory. You can go somewhat higher if you have equivalent atoms. You can also consider larger systems by reducing the set of NMR atoms. In comparison, similar programs from QCPE are limited to 8 nuclei. ============================================================================== 18. How accurate is HyperNMR? ------------------------------------------------------------------------------ HyperNMR, like almost any simulation method, does not perfectly reproduce reality. However it does represent a significant improvement in accuracy and speed over other wavefunction-based methods for predicting NMR spectra. We have tabulated some comparisons with experiment for proton and carbon spectra. For proton chemical shifts for 72 molecules (with 143 shifts) we find a standard deviation of 1.1 ppm, with maximum error of 4 ppm. For carbon chemical shifts for 55 molecules (with 118 shifts) we find a standard deviation of 19 ppm. The worst cases are for systems with 3 and 4 membered rings for protons and carbons and allene, furan or pyrrole functional groups for carbon. For systems other than these we expect proton shifts to be within 1 ppm of experiment and carbon shifts to be typically within 5-10 ppm of experiment. An advantage of HyperNMR is that accuracy can be improved by adding new atom types and associated parameters, and users can do this on their own if they wish. The above results have shown that we need some new carbon atom types and we expect that these will be added to the next release of the program. If you know of cases where HyperNMR gives poor results for chemical shifts or coupling constants, please send us the HyperNMR file so we can include those cases in the parameterization for future releases of the program. Please send electronic copies, either by email to support@hyper.com or on floppy disk to Hypercube Inc., 419 Phillip St., Waterloo ON N2L 3X2, Canada. Please include an explanation, including values from HyperNMR and experiment, with a reference for the experimental values that you are comparing to. The program is limited to the first two rows of the periodic table because of the substantial effort needed to derive, implement and test new Slater integral equations for chemical shielding with each higher principle and angular quantum number. ============================================================================== ------------ Graham Hurst (hurst@hyper.com) Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" to hyperchem-request@hyper.com ________________________________________________________________________ From hurst Wed Nov 23 11:17:42 1994 Date: Wed, 23 Nov 94 11:17:42 -0500 From: hurst (Graham Hurst) To: "Bellnier, David" , hyperchem@hyper.com Subject: Re: returned e-mail David Bellnier wrote: > I see that my e-mail query regarding software other than hyperchem was > promptly returned to me. Does that mean that it was *not* sent out to > members of the users group? If so, I'm unclear of the user group policy > regarding such matters. Could you please fill me in? Thank you. > David A. Bellnier, Ph.D. All mail is sent to the group, but when you post to the group you will frequently get bounced email back. This happens when email addresses in the list don't work, frequently because student accounts have expired or computers have been retired. We periodically flush out these rogue addresses, but can never seem to quite keep up, so PLEASE IGNORE ANY BOUNCED MAIL MESSAGES WHEN YOU POST TO THE HYPERCHEM GROUP! If you are worried that an email message didn't get distributed properly, send email to listadmin@hyper.com. TO UNSUBSCRIBE -------------- To unsubscribe from the HyperChem Email Users' Group, send a message with the body: unsubscribe hyperchem or: unsubscribe hyperchem subscribed_email_address (where "subscribed_email_address" is replaced by the address that you wish to unsubscribe) to hyperchem-request@hyper.com. Cheers, Graham ------------ Graham Hurst (hurst@hyper.com) Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" to hyperchem-request@hyper.com ________________________________________________________________________ From underhil@hp.rmc.ca Fri Nov 25 10:05:23 1994 Date: Fri, 25 Nov 1994 09:25:59 -0500 To: hyperchem@hyper.com From: underhil@hp.rmc.ca (Ross Underhill) Subject: energy surface generator If anyone is interested I have a visual basic module that runs from within Excel 5.0 that will generate a 3-D surface using Hyperchem by varying any two parameters while measuring a third (such as total energy, torsion energy, etc.). Free to a good home. Dr. Ross Underhill Royal Military College of Canada Kingston, Ontario (613) 541-6000 X6175 ________________________________________________________________________ From rconnors@WPI.EDU Fri Nov 25 11:04:40 1994 From: Robert E Connors Date: Fri, 25 Nov 1994 10:35:04 -0500 To: hp.rmc.ca!underhil, hyperchem@HYPER.COM Subject: Re: energy surface generator Sounds good. How does one get it? ________________________________________________________________________ From kellogg@uidaho.edu Tue Nov 29 12:32:35 1994 Date: Tue, 29 Nov 1994 09:08:13 -0700 (PDT) From: Scott Kellogg Subject: Pentium Problems To: Hyperchem Are there any known problems with Pentium calculations in Hyperchem or Hyperchem Plus? ========================================= Scott T. Kellogg, Dept. of Microbiol., Mol. Biol. & Biochem. Univ. of Idaho, Moscow, ID 83843 Internet: kellogg@uidaho.edu Phone: 208-885-6966 Fax: 208-885-6518 ________________________________________________________________________ From hurst Tue Nov 29 16:41:32 1994 Date: Tue, 29 Nov 94 16:41:32 -0500 From: hurst (Graham Hurst) To: hyperchem, chemistry@osc.edu Subject: Comp. Chem. and the Pentium FDIV bug Here are some observations on the pentium FDIV bug and computational chemistry calculations. The results are for HyperChem, though I would expect similar results for calculations with other modelling programs. A regression test that we use for the testing of HyperChem uses HyperChem's scripting to perform around 3,000 calculations, including geometry optimizations, CI, molecular dynamics, vibration calculations, and uv-vis spectra, logging all results to files that we can then compare to reference files. When we run this test on a Pentium 90 with the bug (verified by finding 4195835 - (4195835/3145727*3145727) = 256 with the Windows calculator) and on a 486 without the bug, we do find differences for 4 (out of roughly 3000) calculations. The four cases are all geometry optimizations, (two with MNDO and one each with AM1 and CNDO) where the final optimized electronic and core-core repulsion energies differ. For one the difference is one in the 12th significant digit and for the other three the differences are in the 13th siginificant digit. For comparison, we find larger differences (but still relatively insignificant) comparing results from 486, SGI, DEC Alpha or IBM PowerPC computers, which we have found is due to very small differences in math library routines (like sqrt and arccos). Based on the above, I'd suggest not panicking about the Pentium FDIV bug. The magnitude of the error as manifested in molecular modelling calculations seems almost immeasurable, and is certainly far smaller than the "errors" introduced by the numerous approximations that are used routinely (like integral cutoff thresholds or non-bonded distance cutoffs). This kind of thing is hardly new - the Paranoia program was written long ago to characterize floating-point weaknesses of chip/library combinations. (Has anyone run this on a buggy Pentium 90? I've misplaced my copy, but would guess that it passes with flying colours.) I remember Clemens Roothaan showing me that the ROM for the table lookup square root in FPS-164 and FPS-264 array processors had several errors in it, leading to small but measurable imprecision, but I wouldn't advocate tossing out the wealth of published results from the mid-80s that used those once-popular computers! For what it's worth, my advice (if you've got a buggy Pentium) is to call Intel and tell them that the bug is detectable in some computational chemistry calculations (including off-the-shelf, shrink-wrapped HyperChem) and see if they'll replace it, but in the meantime carry on with calculations! Cheers, Graham ------------ Graham Hurst (hurst@hyper.com) Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" to hyperchem-request@hyper.com ________________________________________________________________________ From DBellnie@hs1.buffalo.edu Tue Nov 29 17:09:51 1994 From: "Bellnier, David" To: hyperchem@hyper.com Date: Tue, 29 Nov 94 15:50:00 PST Where can I purchase glasses for viewing stereoimages? Thanks. ________________________________________________________________________ From kellogg@uidaho.edu Wed Nov 30 00:14:37 1994 Date: Tue, 29 Nov 1994 20:04:05 -0700 (PDT) From: Scott Kellogg Subject: Postscript Slides To: Hyperchem We use a driver for Hyperchem output to our various 35mm slide makers (Agfa, etc), which creates postscript output from Hyperchem or any other Windows program. One problem is that the stick form creates sticks or bonds that appear to be a single pixel wide, which makes for poor final slides under postscript. Is there a way to increase stick width? I have run windows in 800x600x256 and 6409x480x256 modes but that doesn't work and there is nothing in the manual. All other molecular forms come out beautiful. Thanks. ========================================= Scott T. Kellogg, Dept. of Microbiol., Mol. Biol. & Biochem. Univ. of Idaho, Moscow, ID 83843 Internet: kellogg@uidaho.edu Phone: 208-885-6966 Fax: 208-885-6518 ________________________________________________________________________ From hurst Wed Nov 30 10:08:44 1994 Date: Wed, 30 Nov 94 10:08:44 -0500 From: hurst (Graham Hurst) To: hyperchem Subject: Re: Where to get stereo glasses David Bellnier asked: > Where can I purchase glasses for viewing stereoimages? Thanks. Elizabeth Yuriev sent the following to hyper@sentex.net (but I think she meant to send it to the group at hyperchem@hyper.com - please check the To: address when replying!): > From: "Yuriev, Elizabeth" > To: hyper@sentex.net > Date: Tue, 29 Nov 94 11:18:00 PST > > >>Where can I purchase glasses for viewing stereoimages? Thanks. > > Company called "NU 3-D VU" sells them. > Adress : 71 East 28th Str. > Eugene, OR 97405 > > ************************************* > Elizabeth Yuriev > Department of Environmental Management > Victoria University of Technology > Melbourne Australia > ************************************** I'm not sure if "NU 3-D VU" sells the same thing, but stereo viewers for large images are made by VCH Verlagsgesellschaft, P.O. Box 1260/1280, D-6940 Weinheim, Germany. The product is "Reflecting Stereoscope" (in English) or "Spiegelstereoskop" (in German). We have purchased them from the Aldrich catalog (catalog # Z15,675-2) and the last time I looked I think they were around $40 (US). Hope this helps, Graham ________________________________________________________________________ From hurst Wed Nov 30 10:21:11 1994 Date: Wed, 30 Nov 94 10:21:11 -0500 From: hurst (Graham Hurst) To: Scott Kellogg Subject: Re: Postscript Slides Cc: hyperchem Scott Kellogg writes: > We use a driver for Hyperchem output to our various 35mm slide makers > (Agfa, etc), which creates postscript output from Hyperchem or any other > Windows program. One problem is that the stick form creates sticks or > bonds that appear to be a single pixel wide, which makes for poor final > slides under postscript. Is there a way to increase stick width? I have > run windows in 800x600x256 and 6409x480x256 modes but that doesn't work > and there is nothing in the manual. All other molecular forms come out > beautiful. Thanks. Hmm. My recollection was that if you did a File/Print the sticks were widened... You could try changing selections to "Thick Line" under File/Preferences and then select everything before printing. A side effect is that you will lose the display of multiple bonds. A better solution is to use the Molecule Presentations part of ChemPlus, where you can specify the line width in pixels. Hope this helps, Graham ------------ Graham Hurst (hurst@hyper.com) Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" to hyperchem-request@hyper.com ________________________________________________________________________ From yurel@stamail1.vut.EDU.AU Tue Nov 29 20:14:55 1994 From: "Yuriev, Elizabeth" To: hyper@sentex.net Date: Tue, 29 Nov 94 11:18:00 PST >>Where can I purchase glasses for viewing stereoimages? Thanks. Company called "NU 3-D VU" sells them. Adress : 71 East 28th Str. Eugene, OR 97405 ************************************* Elizabeth Yuriev Department of Environmental Management Victoria University of Technology Melbourne Australia **************************************