NPGS  : Solutions to Problems

The search function (Edit - Find) of the browser can be used to quickly find keywords in this list.

If you have a problem that is not discussed here or elsewhere in the NPGS manual, please call or send e-mail with a detailed description of the problem so that a solution can be found. Note that problems discussed below cover all aspects of SEM lithography, including the SEM and processing.

Please select topics from the following list or scroll down to see all of the information:

Pattern Exposures Astigmatism Missing Pattern Elements: Curves, Ellipses, Text, Etc. Pattern Aspect Ratio Pattern Distortion from Charging Pattern Noise from Environment Pattern Noise from SEM Pattern Position Offset Pattern Size Pattern Writing Time NPGS Software (Windows Version 8.0 and higher) Copying Custom Commands Between Projects Missing Project Directories Display of NPGS Help File Optimizing the Pattern Writing Speed Scanning Stops in Pg_Image NPGS Software (DOS Versions to 7.6) Executable Files Not Found Pattern Design with DesignCAD LT for Windows Design Units Filled Polygon Display Errors with Filled Polygons Off-Site Pattern Design Setting a Small Snap Grid Size Missing Zoom Menu Running DesignCAD LT from Desktop icon or from Start menu Pattern Design with DesignCAD 6.1 for DOS DC6: Memory DC6: Duplicated Patterns DC6: Filled Polygon Display DC6: Graphics Board Support DC6: Missing Zoom Menu DC6: Off-Site Pattern Design

Beam Blanker Exposed Dot at Origin Pattern Writing Time with a High Beam Current No Dynamic and/or Steady State Blanking Exposures Short by ~1 micro sec Alignment Invalid Matrix Calculation No Alignment Image Pattern Alignment when using the Global Rotation Mode Slow Movement of Overlay External Control (Stage/SEM/Etc) SEM/Stage Interface not Working Sharing Connections with EDS/WDS Systems SEM Input Relay Adapter not Working Delay Before Each Pattern is Written Computer and/or DT2823 Crash/Hang/Reboot ATLAB Error Corrupt ATL directory or ATL.dll Miscellaneous HASP Error Limited PC Memory DOS Prompt

Pattern Exposures

Astigmatism (go to Top)

· When I write fine lines, the exposure varies for lines at different angles.

· When I expose single dots, the dots are elongated rather than round.

· When I write fine circles, the line width varies between wide and thin every 90 degrees.

All of these problems are almost always caused by astigmatism in the beam. The astigmatism either is present when the samples are being viewed or was present when they were being written. If the pattern contains circles or radial spokes (see Sample0.dc2) the exposure is especially revealing. When the line width changes between wide and thin every 90 degrees it is a sure sign that there is/was astigmatism in the beam. In fact, a circle exposed by a single pass of the beam should be incorporated into every test pattern in order to detect astigmatism. Unfortunately, the results are not known until the sample is processed, but it still provides an excellent gauge of how well the microscope was set up for the pattern writing. With a little practice, most people can consistently set up the microscope to minimize the astigmatism well enough for very fine lithography.

For images showing the results of astigmatism during the exposure, please visit the "Guide to Common SEM Lithography Exposure Problems" on the "Sample Pictures" page.

Missing Pattern Elements: Curves, Ellipses, Text, Etc. (go to Top)

· When I use Curves, Ellipses, Text, or Dimensions in my pattern they are not written.

Curves, Ellipses, and Text must be designed as Vector entities in order to be written as part of a pattern. The DesignCAD LT command "Vector Convert" (or the DesignCAD 6.1 command "VECT CONVERT") may be used to convert these entities to vectors that will be written. Dimensions may only be used as comments in a pattern. For more information, see "Ignored Pattern Elements" in the NPGS Manual.

Pattern Aspect Ratio (go to Top)

· When I write circles they do no turn out round and squares turn out as rectangles.

The aspect ratio for pattern writing must be adjusted. For more information, see "PCI516 DAC Outputs" (for NPGS v9) or "DT2823 DAC Outputs" (for NPGS v8 and earlier) in the NPGS Manual and the section in the NPGS Installation Guide on "System Calibration".

If patterns only occasionally have a distorted aspect ratio, the cause may be that a "tilt correction" mode of the SEM was enabled during the pattern writing. For lithography, any tilt correction mode of the SEM should be disabled. For distortions on the order of a few percent, it is quite possible that the microscope's aspect ratio for the XY inputs changes as the magnification has changed, but errors on this level are typically within spec for an SEM. Another possibility is that the "scan rotation" option of many microscopes may introduce aspect ratio errors on the order of a few percent.

Pattern Distortion from Charging (go to Top)

· When I write a large pattern on a low conductivity substrate some parts of the pattern turn out distorted while other parts are fine.

The sample is probably charging during the pattern writing enough to deflect the beam in some areas. Please see the "Charging (during writing)" diagnostic image on the "Sample Pictures" page of the NPGS web site for more information.

One solution is to evaporate ~100 Angstroms of aluminum on top of the resist before writing and etch it off in NaOH or KOH after the writing and before developing. One recipe is to use several pellets of NaOH in ~100 ml of DI water and etch for about 30 seconds. (You can coat some test samples and try various concentrations and times to optimize the process, if necessary.) Note that NaOH may permeate the PMMA and chemically interact with some substrates. Other metals such as Au, AuPd, or Cr can also be used with the appropriate etch after exposure.

Pattern Noise from Environment (go to Top)

· My patterns have a significant noise problem (either line frequency or even a square wave at some other frequency) even though the PC chassis is grounded directly to the microscope chassis.

The most common source of noise is magnetic pickup at line frequency. Please see the "Line Frequency Noise" diagnostic image on the "Sample Pictures" page of the NPGS web site for more information.

Three possible solutions are:

1. Identify the source of the noise and prevent the interference from reaching the microscope by shielding the source or turning it off.
2. Shielding the microscope sample chamber with mu-metal.
3. Installing a magnetic field canceling system in the microscope room. These systems typically have large loops of cable around the room and a detector placed near the microscope column. They can cancel magnetic fields from DC up to a kilohertz or higher.

The images below show severe cases of 60 Hz pickup from a nearby power bus. Mu-metal shielding around the SEM chamber significantly reduced the noise. The lines were written with a single pass of the beam and the line widths in both images are about 150 nm.

In general, the appearance of line frequency noise pickup may range from the "choppy" interference shown here to almost perfectly smooth sinusoidal oscillations. The identifying characteristic is that the period will correspond to the local line frequency. To determine the frequency of the interference from a micrograph:

1. Measure the period of the noise in nanometers.
2. Divide the period by the "Center-to-Center" distance (nanometers) used when the pattern was written. This will yield the number of exposure points in one cycle.
3. Multiply the result from (2) by the exposure time per point (usec), as listed in the run file.  [For NPGS v8.0 and earlier: Multiply by the sum of the exposure time per point plus any settling times as listed in "Pg.sys", if appropriate. For example, if the pattern was written in the continuous mode (or without any blanker), just multiply by the exposure time (usec) per point. If the beam is blanked between every exposure point, first add (clock*delay + wait + normal_extra) to the exposure time.]
4. The result will be the period in usec. To get the frequency in Hz, multiply by 10-6 and take the reciprocal.

Also, some SEMs will have internal pickup on the xy scan lines depending on the SEM scan mode and magnification that is used for the pattern writing. In theory, the SEM scan mode should have no effect on the pattern writing because the microscope is under external control, but in some cases it does matter. For example, on some JEOL SEMs, interference from the internal scan generator may be superimposed on the pattern if the microscope is in the SR (Super Rapid) or TV mode during writing, while a Slow Scan mode will appear fine.

The magnification used for pattern writing can also affect the interference that is seen in the patterns. SEMs have different circuits that are used for different ranges of magnification and the low end of each range will have the best signal to noise, and consequently, the least sensitivity to internal interference from the microscope. Typically, when the SEM changes from one magnification range to another, an audible click may be heard and/or a glitch may be seen on the SEM image. For more information, see "Pattern Noise from SEM" below and "System Installation: Check List" in the NPGS User's Manual.

Pattern Noise from SEM (go to Top)

· My patterns were writing perfectly at 1000x, but now have a noise problem at 800x.

· My patterns were writing perfectly at 100x, but now have a noise problem at 80x.

The typical design of the scan coil amplifier circuits in SEMs is such that the signal to noise for pattern writing will vary as the SEM magnification is changed. For example, in older JEOL microscopes (models 6400 and earlier), patterns over a field of view of 900 micrometers square may be conveniently written at a magnification of 100x, however, slightly larger patterns should not be written at 90x or the noise level will be about ten times worse. Typically, for older JEOL microscopes, the best signal to noise will be found from 10x to 30x, 100x to 300x, and 1000x to 3000x. Other brands may have the same effect, but at different values in magnification. For example, on an FEI XL30 SEM, the signal to noise changes significantly around 750x, with the exact value depending on the direction of the last change in the magnification and on the working distance and accelerating voltage being used. A clue to where the range is divided is that the SEM image will often shift and/or an audible click will be heard as the SEM is changed across the range boundary. The best signal to noise will be found just above the magnification value where the image shift and/or audible click is observed. For more information, see the section "System Installation: Summary - Check List" in the NPGS Manual.

· My patterns exhibit periodic noise on fine lines and/or small gaps in filled polygons, while other people in the group do not have these problems.

A typical cause for this is that some microscopes (JEOL 6000 series, Leica/LEO S440, and probably others) will have internal noise that distorts the external scan signals when the microscope scan setting is left on TV or Super Rapid, etc. Ideally, when the microscope is in external scan mode, the internal scanning should not have any affect on the beam, however, that is not the case. The solution is to put the microscope in a Slow scan mode when doing lithography. Some new models will completely turn off the internal scan generator when the SEM is in external mode, which eliminates this problem.

Pattern Position Offset (go to Top)

· When I write a pattern, there is an unexpected offset between different parts of the pattern.

The following causes for this type of problem have been reported:

1. A small particle in the column would collect a charge during dynamic point to point blanking of the beam during the pattern writing and cause a consistent offset. Since the beam is not dynamically blanked during the alignment procedure, this offset would appear even after a careful alignment to the registration marks. The offset was so consistent that initially, the user simply designed a compensating offset into the pattern. This approach worked until the column was cleaned and the particle was removed, which eliminated the original offset.
2. A related problem was also seen where different layers of the pattern had up to a 1 um offset. The offset was again caused by charging related to the dynamic blanking of the beam. In this case, the user had designed a pattern with two very different exposure times on the two layers of the pattern. This effectively caused a different blanking frequency for each layer, which resulted in a different steady state charging and corresponding beam offset. The quick solution was to keep the exposure times in the different layers more similar and the long term solution was to clean the blanker/column to remove the contamination that was charging.
3. A nearby freight elevator would cause a distortion in the Earth's magnetic field and consequently affect the position of the beam in the microscope. Other than disabling the elevator or working at night when the elevator is not in use, the only approach to solving this problem is to use an active magnetic field canceling system. These systems have coils around the room and a sensor near the SEM chamber. The coils are driven to null the magnetic field at the sensor. Such systems can cancel fields from DC to a kHz or so.

Pattern Size (go to Top)

· My patterns are not exposed at the size I designed them to be.

The "mag scale" parameter in "Pg.sys" tells the NPGS software what the microscope field of view is at each magnification. See the section in the "NPGS Installation Guide" on "System Calibration" to adjust the "mag scale" parameter. If the calibration has been successfully completed at one magnification, but the patterns are not correct when written at a different magnification, then the microscope may need to be serviced. Each type of scope will have a different specification on how accurate the magnification is at different settings. A high performance SEM should be accurate on the order of 1 to 3%.

An alternate cause may be that the microscope magnification was not set at the value entered in the run file. Other possibilities are that a microscope setting has been changed that affects the scaling of the XY inputs when in external control. Examples of such settings may include the scan speed (the TV mode on older Hitachi SEMs has been observed to do this), tilt correction (this SEM feature can be used to intentionally distort the XY aspect ratio and should not be used during lithography) , or a magnification calibration mode (on PC based SEMs, this adjusts the displayed magnification value so that it may match the output when viewed on the monitor or when the image is printed on a video printer, or reproduced on a Polaroid - in each case, the "magnification" value is adjusted so that it corresponds to the viewing size of the image).

Pattern Writing Time (go to Top)

· The pattern takes much longer to expose than I expected.

<This issue only applies to NPGS v8.0 and earlier.>

To achieve the fastest overall writing time for a given beam current, the exposure time should be comparable to or greater than the typical calculation time per point (this varies depending on the computer being used). Note that any newer PC will have calculation times that are typically much less than 1 microsecond, so the calculation speed is seldom an issue. However, if a slow PC is being used, if the exposure time per point is 5 microsecond and the calculation time per point is 10 microsecond, then the overall speed will be limited by the calculation time and it will take approximately twice as long to expose as it ideally should. In such a case, the overall speed can be improved by increasing the exposure point spacing while keeping the same dose. A general rule for lines made by a single pass of the beam is to have the center-to-center distance about 1/2 to 1/4 of the final line width. This reduces the number of points to calculate and increases the exposure point time per point, without sacrificing the quality of the line edge. Use the NPGS Menu program command "Commands - Process Run File - Error Check Mode (or for older versions, the 'c' option when running PG.exe) to obtain calculation times for your patterns. One of the common mistakes of people new to SEM lithography is to think that setting the point spacing to the smallest possible value will improve the pattern quality. When writing in a 100 micron field, the smallest spacing is about 15 Angstroms, which certainly cannot be seen in any typical pattern writing.

· The pattern should take about 3 minutes to write, but when written, the beam moves much too fast and finishes in about 20 seconds. The runfile has been checked and the NPGS Menu program command "Commands - Process Run File - Time Test Mode (or for older versions "pg runfile t") confirms the expected 3 minute writing time.

<This issue only applies to NPGS v8.0 and earlier.>

This strange problem has been reported only one time (when using the DT2823 of NPGS v8.0 and earlier) and the exact cause was never determined. However, rebooting the PC cleared the problem and it worked fine after that. For any unusual problems with a computer, rebooting is always a good thing to try.

· The pattern seems to alternate between being properly exposed and underexposed, with the underexposed parts of the pattern being written too fast. The change occurs consistently from pattern to pattern and even within a single pattern element.

<This issue only applies to NPGS v8.0 and earlier.>

This problem was seen many years ago when the Expanded Memory Manager EMM386.sys was added to the config.sys file. This driver should not be used with NPGS, because it can affect the DMA (Direct Memory Access) used to communicate with the DT2823 board in NPGS v8.0 and earlier. (Few programs require an Expanded Memory Manager, and NPGS is compatible with the Extended Memory driver HIMEM.sys.) In this case, the pattern elements were filled polygons that were automatically broken into two groups of exposure points. The first group had about 25,000 exposure points and the second about 15,000. Apparently, the expanded memory manager affected the writing of the second group for each pattern element, such that it was exposed much too quickly. The problem disappeared once EMM386.sys was removed from the config.sys file.

NPGS Software (Windows Version 8.0 and higher)

Copying Custom Commands Between Projects (go to Top)

· I have created a new project and I would like to copy all of the custom commands from my other project.

To copy all of the NPGS custom command definitions from one project to another, do the following steps. Be aware that all user defined settings for the project will be copied.

1. Set the "Current Project Directory" to the project with the configuration to be copied.
2. Select "File - Copy File(s) From Project".
3. For the "Destination Directory" select the project where the configuration will be copied.
4. For the "Display File Type" select "All Files".
5. Highlight "PG_MENU.sys" and hit "Copy Files", then "Yes".

To copy individual Custom Commands between projects, use the Cut/Copy/Paste buttons on the Custom Command dialog box.

Missing Project Directories (go to Top)

· I previously had several NPGS Projects defined, but now one (or more) do not show up in the pull down project list in the NPGS Menu program.

NPGS maintains the file "\NPGS\Program\Pg_PrjDir.sys" that contains the list of active projects. In some cases, this file has become corrupted, thus resulting in missing projects within the NPGS Menu program. If only one or two projects are missing, the easiest approach is to run "Projects - Create New Project" and type in the name of a missing project, and hit "Create". The NPGS software will then see that the project already exists and will give you the option of using it without any changes (as long as the full list of system files exists in the project). If many projects have disappeared from the list, the easiest approach is to replace the corrupted "Pg_PrjDir.sys" file with a good version from the most recent backup of the system. Note that the most recent versions of the NPGS Menu program automatically create a backup of the project list in "\NPGS\Program\Pg_PrjSav.sys", which will automatically be used if NPGS detects that the original list has been corrupted.

Display of NPGS Help File (go to Top)

· When I use the HTML Version of the NPGS Online Manual, the information is not formatted nicely in Windows 95.

<This issue only applies to NPGS v8.0 and earlier.>

The HTML version of the NPGS User's Manual uses the Microsoft compiled HTML help format (Version 1.2). While this format has several more advanced options than the traditional Windows "WinHelp" format, unfortunately, it may not be fully supported under Windows 95 (it should work fine on Win98 and newer).

Consequently, the NPGS User's Manual is supplied in the PDF file format as well as the HTML help format, so that everyone should have full access to the information.

Optimizing the Pattern Writing Speed (go to Top)

· I am writing patterns with many structures that are each defined with only a few hundred exposure points (or less) and I would like to minimize the total writing time.

· I am writing arrays of patterns with hundreds or thousands of stage moves and pattern exposures and would like to minimize the total writing time.

<This issue only applies to NPGS v8.0 and earlier.>

When patterns are exposed while the Windows 95/98 operating system is in use, Windows will slow down calculations and add milliseconds to each structure written by NPGS v8.0 (this is not an issue for NPGS v9.0 and higher). For most patterns the extra time is negligible, however, when patterns have very large numbers of small structures or when many repeats of the same pattern are exposed, the small delays caused by Windows can become quite significant.

When running NPGS v8.0 or earlier, a solution is to reboot the computer and run directly from DOS by hitting "F8" during the initial boot sequence and selecting "Command Prompt Only" from the boot menu. After changing to the appropriate project directory, run the desired run file using the following format:

         \NPGS\Program\NPGS runfile

Note that long names for run files and patterns should not be used when using this method.

Scanning Stops in Pg_Image (go to Top)

· Occasionally, when Pg_Image is scanning, the scan will terminate before it is complete.

<This issue only applies to NPGS v8.0 and earlier.>

In NPGS v8.0, the user interface for Pg_Image is a 32 bit Windows program, while the actual image acquisition program is a separate 16 bit program. This arrangement was necessary to preserve the compatibility with the v8.0 pattern writing software, while adding an advanced Windows user interface to Pg_Image. The consequence of this arrangement is that the communication between the two programs will occasionally become corrupted, apparently when the operating system does some background operation during the scan. The good news is that on most PCs, such problems will be very rare, but on some PCs, the problem happens more frequently. In any case, no damage is done, however, the current scan must be repeated. In general, Windows 95 is the least realizable.

In one case, a user with Windows 95 reported that the problem could be avoided by disabling Norton AntiVirus, however, Norton and other virus checkers on other PCs have not caused problems. This example shows just how subtle the cause of problems can be in the Windows environment.

NPGS Software (DOS Versions to 7.6)

Executable Files Not Found (go to Top)

· When I try to run any of the programs (NPGS, , MENU, MRF, PG, AL, etc.) from within a pattern directory, the executable programs are not found.

< This issue only applies to NPGS version 7.6 and earlier. >

All of the executable programs for NPGS (except those supplied with DesignCAD) are in the directory y:\pg. The NPGS Installation Guide describes how to add or modify the path statement in the c:\autoexec.bat file to include y:\pg so that the programs can be used from directory. Do not copy the pattern files to y:\pg! Do not copy the executable files to the pattern directory! Instead, simply modify the c:\autoexec.bat file so that the y:\pg directory is included in the path. (The reference to "y:" should be replaced with the correct drive letter of where NPGS is installed, typically "c:".)

Pattern Design with DesignCAD LT for Windows

Design Units (go to Top)

· What are the units displayed in DesignCAD?

Within DesignCAD, the units should be set to "inches", however, they will be interpreted as "microns" by NPGS. Consequently, the user can simply consider all dimensions displayed in DesignCAD as microns.

· When I write the same pattern at two different magnifications, why are the structures written at the same size? I had expected the sizes to scale with the magnification when I change it on the SEM and in the NPGS run file.

In some simplistic SEM lithography systems, the pattern size (and doses) vary as the SEM magnification is changed. However, NPGS has always been designed to write the patterns at an absolute size which is defined in the pattern itself. This approach makes it unnecessary to make calculations for the pattern size, while instead the user knows that a pattern element designed to be 10 microns in DesignCAD will be 10 microns when exposed (as long as the magnification listed in the NPGS run file matches the magnification of the SEM when the pattern is written). Of course, if NPGS is not properly calibrated or if the SEM magnification does not match the value in the run file, the patterns will be written at an incorrect size. See the printed NPGS Installation Guide that was provided with the system for information on calibrating the pattern size for NPGS.

Filled Polygon Display (go to Top)

· When I use the BasicCAD program "PolyFill" to make a filled polygon, the dashes (or dots) of the lines defining the polygon are smaller than I would like them to be.

· I have imported patterns from other CAD formats and the dashes of the filled polygons are too small or too large.

Filled polygons are denoted by dashed (linetype=1) or dotted (linetype=5) lines. The "Line Scale" parameter determines the size of the dashes or dots. If the display of the line appears slow and solid, then the Line Scale parameter is too small. Note that this parameter only alters the display in DesignCAD and has no effect on the pattern writing.

Within DesignCAD LT, you can click on a drawing element and hit "Ctrl I" to bring up the Info Box. From there, you can select the dashed (1st choice below solid) or dotted (5th choice after solid) linetypes to change between the serpentine sweep mode and the one-sided sweep mode, respectively. You can also change the Line Scale parameter to any desired value. Note that for filled polygons, the Line Width should be zero.

To change the Line Type, Line Scale, and/or Line Width for all lines in a drawing layer, use the layer options window. This window is accessed by hitting "L" or using the menu command "Options - Layer". Once there, the "Assign Line Style" prompt will allow any desired changes to be made for all lines in the selected layer.

Errors with Filled Polygons (go to Top)

· I have made Filled Polygons in DesignCAD LT, however, the patterns generate errors when I try to write them.

When problems are encountered with Filled Polygons, the reason is almost always that the "NPGS - PolyFill" command in DesignCAD was not used to create the structures. When Filled Polygons are designed manually or are converted from other pattern formats, it is possible that invalid structures can result. In DesignCAD LT, the command "NPGS - CheckAll" can be used to check the entire pattern for elements that will be ignored when the pattern is written, as well as for invalid or questionable Filled Polygons.

Off-Site Pattern Design (go to Top)

· I would like to be able to design patterns in my office or at home, rather than in the microscope room.

· I have a collaborator who would like to design patterns at another location for us to write on the microscope here.

By purchasing additional copies of DesignCAD LT 2000 from Upperspace, formerly ViaGrafix, (800) 233-3223, patterns can be designed on any PC. If patterns are to be designed on a PC other than the pattern writing system, the "Office Installation" mode can be selected when running the NPGS Installation program which will then install the appropriate NPGS files so that patterns and run files can be created just as on the original PC that is connected to the microscope.

After installing the "Office Installation" of NPGS, DesignCAD LT 2000 must be installed in the directory "\NPGS\DC2000LT". The NPGS files that are installed may be used on any number of secondary PCs without violating the license agreement, as long as they are only used in conjunction with the original NPGS system. Each secondary PC must have a separate licensed version of DesignCAD LT 2000 installed! (DesignCAD LT 2000 is currently available through mail order for under $200US, or universities can purchase direct from Upperspace, formerly ViaGrafix, at a discounted price of about $60US.) Contact information can be found at www.UpperSpace.com. Note that DesignCAD LT 2000 and the NPGS files for secondary use may be installed under Windows 95/98, as well as Windows NT/2000/XP.

Setting a Small Snap Grid Size (go to Top)

· When I set the Snap Grid Size to a small value, it instead snaps to a grid of a larger spacing than what I have specified.

Unfortunately, in DesignCAD LT 2000, the Small Step Size interacts with the Snap Grid Size, however, there is an easy fix. If you go to the menu command "Options - Options..." in DesignCAD, then select the "Cursor" tab and set the "Small Step Size" to be equal to or smaller than your desired snap size, the Snap Grid Size you have defined will be used.

Note that the "Large Step Size" defines the displacement when the arrow keys are used and the "Small Step Size" defines the displacement when the "Shift" key is held down while the arrow keys are used. If you never use the Shift-Arrow combination, then you can simply set the "Small Step Size" to a very small value, and the "Snap Grid Size" will always work as expected.

Missing Zoom Menu (go to Top)

· I was using DesignCAD LT 2000 and the icons in the lower right part of the window disappeared, however, the commands actually function when I click where the icons are normally displayed.

The zoom icons will be displayed properly if you minimize and then restore the DesignCAD window. Note that this sort of bug looks suspiciously like other bugs that come from using Microsoft compilers for Visual Basic or Visual C++.

Running DesignCAD LT from Desktop icon or from Start menu (go to Top)

· When I start DesignCAD LT 2000, it sometimes uses my project directory, but other times it uses the projects of other users.

This will happen when either the DesignCAD icon on the Windows desktop or the DesignCAD entry in the "Start - Program..." menu is used to run DesignCAD. It is recommended that any desktop icon and/or the entry in the Start menu for DesignCAD LT be deleted to prevent this from happening.

DesignCAD LT should only be run from within the NPGS Menu program! The NPGS Menu program makes sure that DesignCAD LT uses the correct project and that any changes in DesignCAD are uniquely saved for each project.

Pattern Design with DesignCAD 6.1 for DOS

DC6: Memory (go to Top)

· When I try to run DesignCAD 6.1 an error message flashes that there is not enough DOS memory, even though I have 16 Mb of RAM.

· DesignCAD 6.1 will run when I type "dcpat" at the DOS command line, but an error message about DOS memory appears when I try to run it from within MENU.exe (which is the same as NPGS.exe without any options). <Applies only to NPGS v7.6 and earlier.>

The amount of RAM in a PC is divided between conventional DOS memory and Extended and/or Expanded memory. Conventional DOS memory is the lowest 640 kb of RAM. DesignCAD 6.1 requires about 500 kb of conventional memory to be available when it loads. (DesignCAD 2D v6.1, Release 7-17-97, requires at least 510 kb of available DOS memory.) Device drivers and TSR's will use up the conventional memory. Also, in NPGS v7.6 and earlier, when MENU.exe is calling DesignCAD, it still occupies about 8 kb of memory, so less is available for DesignCAD. Once DesignCAD is loaded, it will make use of all available Extended memory for pattern processing. Newer versions of DOS provide the "MEM" command (use "mem /c | more" or "mem/c/p"), which will display a detailed listing of the programs currently occupying memory. In order to increase the available memory for DesignCAD 6.1, a program must be removed. Typically, to remove a program from memory, the "config.sys" or "autoexec.bat" file must be modified and the PC must be rebooted.

· I run out of memory in DesignCAD when I try to make a large array of individual dots.

Typically, the best way to make an array of single dots is to define a filled rectangle that covers the desired area and then set the "Center-to-Center" and "Line Spacing" parameters in the Run File Editor to the desired spacing for the dots. For this approach, use the "Point Dose" option in the Run File Editor. If the microscope in use has a hysteresis problem, i.e., dots written in adjacent passes are out of sync, use the "One-side only" option when defining the filled rectangle. If the dots to be written are random, then this approach cannot help and the solution is to provide more memory for DesignCAD to use

· I run out of memory in DesignCAD when I try to make very large patterns.

Several options are available: Remove any TSR programs (these are typically installed in the c:\autoexec.bat file); Use Extended memory by installing "HIMEM.sys" in the "config.sys" file; Add more RAM to the PC. Typically, DesignCAD 6.1 will draw about 5,000 line segments for each 1 Mb of Extended memory. When run from within Windows 95/98, DesignCAD 6.1 typically will not use the full amount of available extended memory, however, when run from DOS, the full amount will be used.

DC6: Duplicated Patterns (go to Top)

· When I reload my pattern with "F9" (or "Files - Retrieve"), everything gets duplicated.

In DesignCAD 6.1, the retrieve command will load the new file on top of any existing pattern on the screen. This makes it easy to merge files, but can be confusing if you just want to reload a pattern or change files. To prevent the duplication, you must clear the screen before retrieving the file. Use "yy" or "Display Clear Screen" to clear the screen.

DC6: Filled Polygon Display (go to Top)

· When I use the BasicCAD program "PolyFill" (or Alt F5) to make a filled polygon, the polygon is displayed as a solid line and it is relatively slow when it is drawn in DesignCAD 6.1.

Filled polygons are denoted by dashed (linetype=1) or dotted (linetype=5) lines. The "Line Type Scale" (LTS) parameter determines the size of the dashes or dots. If the line appears slow and solid, then the LTS parameter is too small. Note that this parameter only alters the display in DesignCAD and has no effect on the pattern writing. Either redraw the polygon and set the LTS parameter to a larger value (for an absolute zoom of 1, use LTS ~1; for zoom of 0.1, use LTS ~10; etc.) or set a point on the polygon and use the CHANGE command to edit the LTS parameter.

DC6: Graphics Board Support (go to Top)

· The graphics board in my PC is not listed as one of the choices within the "setup2d" configuration program for DesignCAD 6.1.

Graphics boards are continually being improved and it is unlikely that newer boards will be directly supported by the DOS version of DesignCAD. However, almost all boards will support one or more of the VESA modes, which are supported by DesignCAD 6.1. Typically, the 800x600 or 1024x768, 16 color VESA modes will work with DesignCAD 6.1, even on newer boards. Change to the "\dc6" directory and run "setup2d.exe" to change the graphics mode. If an incorrect mode is selected, the DesignCAD program will either not load, will have a distorted display, or may lock up. In any case, run "setup2d" again and try a different mode.

Note that some boards will require a special driver to be loaded in order for the VESA modes to be used. For example, the ATI graphics boards (Fall '99) require a driver called M64VBE.exe to be loaded before a 16 color VESA mode is used, while for the 256 color modes, the command "M64VBE S VGA" is required. Such commands can be listed in the "autoexec.bat" file so that they will be loaded when the PC is booted. This driver is available for download at the ATI web site. Other brands may also have similar drivers.

Note that the resolution and color depth set under Windows 95/98 does not affect the operation of DesignCAD 6.1.

DC6: Missing Zoom Menu (go to Top)

· When I select 256 color mode for the graphics board, the custom NPGS "Zoom Menu" which is normally displayed in the drawing area of DesignCAD is no longer visible. Alternately, if the "Zoom Menu" displays in the 256 color mode, it does not display in the 16 color mode.

· When I click the mouse to set a point in the upper left part of the screen, the point is not set and then DesignCAD tells me to set points for a zoom command.

The custom command menus that can be displayed in the drawing area of DesignCAD 6.0/6.1 must be designed for the color depth that is in use. Consequently, the "zoom_3.sym" menu is designed for use with the 16 color modes and the "zoom_3hi.sym" menu is designed for 256 color modes. If the incorrect menu is loaded, it will not be visible on the screen, however, it will still be active.

The following steps show how to change from one menu to the other:

1. Start DesignCAD through NPGS so that you are working in the "\pg\pat" directory.
2. Go to "Menu - Delete", then set a point in the upper left part of the black drawing area and hit Enter. This step will remove a menu that doesn't show up in the current color mode.
3. Again, go to the "Menu - Delete" command, but this time it should just beep, which will indicate that all custom menus have been deleted. If it does not beep, repeat the previous step.
4. Now, go to "Menu - Load" and hit Enter to get a list of available files. You should see "zoom_3.sym" and "zoom_3hi.sym" as choices. Highlight "zoom_3hi" when using a 256 color mode or "zoom_3.sym" when using a 16 color mode, and hit Enter or F2. You should then see a list of zoom choices in the upper left part of the drawing area.
5. Finally, go to "Options - Config Save". This will save the changes so that each time DesignCAD is loaded, you will see the new menu.

If you have made other pattern directories, repeat these steps for each directory.

DC6: Off-Site Pattern Design (go to Top)

· I would like to be able to design patterns in my office or at home, rather than in the microscope room.

· I have a collaborator who would like to design patterns at another location for us to write on the microscope here.

DesignCAD 6.1 is no longer available, however, DesignCAD LT 2000 can be used with the free upgrade to NPGS v8.0.

Beam Blanker

Exposed Dot at Origin (go to Top)

· The SEM has a fast blanker, however, after writing a pattern or just doing an alignment, there is an exposed spot in the middle of the field of view.

This has been seen when the blanker has been "leaky". Instead of completely turning off the beam, enough current hit the sample to cause the extra exposure. The extra exposure appears as a dot at center of the field, because that is where the scan coils are normally set after writing when the system has a fast blanker. The best solution is to fix the blanker. A less desirable solution is to set a dump point in all patterns so that the beam is never "parked" in an important area of the writing field.

Pattern Writing Time with a High Beam Current (go to Top)

· When I use a very high beam current, the required exposure times are longer than they should be (i.e., it appears to take more dose at higher beam currents!).

In some beam blankers, the rise time for turning on the beam will depend on the beam current, if the blanker is not installed properly. To check for this problem, connect a two channel oscilloscope to the Secondary Electron Detector Photomultiplier Tube (PMT) output and to the Blanking Control Circuit (BCC) output. Then write a pattern consisting of a single large filled rectangle with the exposure time per point set to about 50 usec. If the blanker is working properly, the PMT signal should lag the BCC signal by about the same amount independent of the beam current. Magnetic blankers typically have about a 1 usec rise time, while electrostatic blankers typically have rise times of 100 ns or less.

No Dynamic and/or Steady State Blanking (go to Top)

· I have a fast beam blanker installed but it does not respond when a pattern is written unless the "continuous" mode is used in the run file. The blanker works fine when run manually and from within the NPGS Menu program, i.e., "Commands - Set Blanker", or from DOS using "BCC 0" and "BCC 1".

<This issue only applies to NPGS v8.0 and earlier.>

When a pattern is written with a fast blanker in the normal mode, the beam is flashed at each exposure point rather than being left on as the scan coils are moved. The signal that synchronizes the blanking and the DAC outputs is a 0.5 msec TTL pulse, carried by the small coax cable soldered to the DT2823 board and plugged into a socket on the BCC board. The symptoms above suggest that this cable is not properly connected, or the BCC clock chip is not working, or that the 74LS123 (U8) chip on the DT2823 or the 74LS04 (U8) chip on the BCC is not working properly. First, check the cable, then check that the BCC clock chip (in rectangular metal enclosure) is properly inserted in the socket, then try replacing the 74LS04, then check the 74LS123. If these all are OK, then another chip on the BCC may have gone bad. In that case, replace the chips on the BCC (they are all in sockets and cost less than ~$20US total).

· The blanking was working fine, but now even the manual from within the NPGS Menu program, i.e., "Commands - Set Blanker", or from DOS using "BCC 0" and "BCC 1" do not work.

<This issue only applies to NPGS v8.0 and earlier.>

First check the output of the BCC (center pin of the BNC relative to the PC chassis) to determine if it is the blanker or the BCC. The following only applies to NPGS v8.0 and earlier. If the BCC output is not responding properly using the manual control, the problem is probably with the output chips OP1 and/or the optional OP2. In the manual mode, the only other chips that are used are the 74LS00 (U12), 74LS04 (U8), and the 74LS32 (U6). Another possibility is the signal that comes into the BCC on the 25 pin connector. Either the connector may not be attached properly, or the DT2823 may not be producing the proper signals.

Exposures Short by ~1 micro sec (go to Top)

· When I connect the output of the BCC to an oscilloscope and measure the exposure time per point while writing a pattern, the measured time is about 1 microsecond shorter than it should be.

<This issue only applies to NPGS v8.0 and earlier.>

The "normal extra" parameter in the system file "Pg.sys" must be ~4 msec or more. Use the NPGS System File Editor (PG_EDIT.exe) to change the parameter. If this is not the problem, then one of the 4 bit counters (74LS163) on the BCC may need to be replaced. U4 controls the programmable delay, U2 controls the lowest 4 bits of the exposure time, U1 controls the middle 4 bits of the exposure time, and U3 controls the highest 4 bits of the exposure time.

Alignment

Invalid Matrix Calculation (go to Top)

· The alignment program responds with "Invalid Matrix Calculation" when I recalculate the alignment matrix.

This error message results when the calculated alignment matrix would cause the alignment windows to be scanned outside of the allowed field of view. For example, if the alignment windows are positioned only 3 um away from the edge of the field of view, then an overlay offset of >3 um would require the windows to move outside of the field of view, thus causing this error. This error message results when the calculated alignment matrix would cause the alignment windows to be scanned outside of the allowed field of view. For example, if the alignment windows are positioned only 3 um away from the edge of the field of view, then an overlay offset of >3 um would require the windows to move outside of the field of view, thus causing this error.

When the run file is being created, an easy way to check how much room is available for moving the windows during the alignment is to look at the limit that is displayed on the status bar at the bottom of the Run File Editor window when the "Origin Offset" entry is highlighted. For both alignment patterns and writing patterns, the value displayed on the status bar will be the maximum amount that the pattern can be moved in X and Y during the alignment that guarantees that the pattern will not "hit" the edge of the writing field.

Note that the magnification used for the final alignment must be the same as the magnification used for the subsequent pattern writing and that adequate space must be left for moving the alignment windows, as well as the pattern(s) to be written. The best approach is to ensure that the "Origin Offset" limit for the subsequent patterns is equal to or greater than the limit for the alignment windows, and that the alignment window limit is large enough to accommodate the uncertainty of the initial stage position when the alignment windows are first scanned.

Another cause for this message is when four alignment windows are open and the user recalculates the matrix after only one window has been aligned. In some cases, this partial alignment will result in an invalid matrix. For the best results, when four windows are open, all overlays should be aligned before recalculating the alignment matrix.

No Alignment Image (go to Top)

· When I use the AL program, I do not see an image on the PC screen and the "DATA" values just slowly drift or are constant with a 1 or 2 point difference.

Check that a cable is connected between the "Input" connector on the NPGS hardware at the back of the PC (for NPGS v9.0, this will be the SMA connector farthest from the 34 pin connector, and for NPGS v8.0 and earlier, there will be a BNC connector labeled "Input") and an image intensity signal on the microscope. The image intensity signal should change as the beam hits areas on the sample of different "brightness". The signal can be from a secondary or backscattered electron detector as long as the maximum signal is within +/-10 volts. Ideally, as the beam goes from bright to dark, the signal will change by 0.5 volts or more. In AL, the "DATA" values correspond to the 16 bits of the ADC. Consequently, a 1000 point range in the "DATA" indicates a 0.3 volt data range, when the input voltage range of the ADC is +/-10 volts. It can be quite informative to connect an oscilloscope to the signal to observe its response to the contrast and brightness settings, as well as to the structure on the sample.

Pattern Alignment when using the Global Rotation Mode (go to Top)

· After using the Global Rotation mode to compensate for a ~90 degree rotation of my sample, the alignment windows are also rotation by 90 degrees.

The alignment feature expects that the scan axes of the microscope will initially be at least nominally aligned with the axes of the sample. If the sample is rotated by some large angle, the "Scan Rotation" feature of the microscope should be used to make the scan axes nominally align with the sample rotation. If the microscope does not provide a "Scan Rotation" feature, or if the "Scan Rotation" feature distorts the pattern writing and cannot be used, you should use the stage rotation to nominally align the sample with the scan axes.

Slow Movement of Overlay (go to Top)

· The response in AL is very slow when the overlays are moved by hitting the arrow keys.

<This issue only applies to NPGS v8.0 and earlier.>

This problem was discovered after a power outage hit a user's Zeos 486 PC. After checking for hardware damage (and a possible virus infection, since a game program had also recently been put on the hard disk by other users), it was discovered that the "Shadow Video ROM" and "Shadow BIOS" options in the PC setup had been disabled. After enabling these options, the AL program worked quickly, as it had before the power outage. PCs have basic system configuration parameters (BIOS) stored in CMOS memory that is backup up by a battery. A "setup" program is used to modify these parameters, however, accessing the program varies from machine to machine. Often during bootup, a message will say to hit a specific key to enter setup. Others will have a special key sequence (such as Ctl-Alt-Esc) that will enter the setup program from DOS. Check your PC user manual for details. The "shadow" option tells the PC to use part of RAM to hold information that is otherwise held in slower ROM.

External Control (Stage/SEM/Etc)

SEM/Stage Interface not Working (go to Top)

· When I try to have NPGS control the SEM and/or the automated stage, the SEM/stage driver reports an error.

For many models of SEMs, NPGS communicates with the SEM by calling a batch file with specific commands that control the microscope and/or stage. When using NPGS version 8.0 or newer, i.e., with a Windows user interface, any batch files called by NPGS must have a full path specified when calling other programs. It is recommended that any non-standard programs used with NPGS be placed in the "\NPGS\Projects" directory.

Another issue that may arise is that if a DOS window is opened and it is used for serial communications, it must be closed before NPGS v8.0 can communicate with the stage. Apparently, within Windows, only the first DOS window that performs serial communication with an outside device can access the serial port, consequently, other attempts to do serial communications will fail.

When using NPGS v9.0 under Windows 2000, a 32 bit interface program is required. Updated versions of the drivers for common SEMs and stages are available for use with NPGS v9.0 at no charge.

In general, there are five components to the serial or Ethernet connection between NPGS and the SEM. If a problem arises, you should check each of the following to ensure that the proper configuration is in place:

1. Software on NPGS PC. This will be a program that communicates with the SEM and/or stage and has detailed documentation that is included in the back of the NPGS Installation Guide.
2. Ethernet or serial hardware on NPGS PC. It is possible that serial and/or Ethernet ports can be damaged by voltage surges. Also, for Ethernet connections especially, the settings in Windows can affect the operation of the port.
3. Cable between the NPGS PC and the SEM. A surprisingly common cause for communication failures is that the cable has become unplugged! Also, for serial ports it is essential that the cable be plugged into the correct port on both ends, since both the NPGS and the SEM software will be set for a particular port.
4. Ethernet or serial hardware on the SEM or stage. (See #2.)
5. Software on the SEM or stage. A common cause for communication problems is that the SEM communication software is not enabled (see below). Unfortunately, most SEM vendors do not provide customers with detailed information on the operation of their interfaces. However, the NPGS documentation mentioned in #1 will often provide the necessary information for using the SEM software.

· The SEM/Stage interface had been working fine, however, now NPGS is not able to control the SEM/Stage.

The first thing to check is that the SEM and/or stage control has not been disabled within the Advanced Modes of the NPGS run file being used.

Alternately, newer PC based microscopes typically have separate programs that enable the external control of the SEM and/or stage. Ideally, the control program will automatically load when the microscope PC is rebooted, however, that is often not the case. Also, even when the correct program is automatically loaded on a reboot, a novice user may accidentally terminate or disable the external control interface while using the microscope.

For the LEO SEMs, be sure that the "RemCon.exe" or "RemCon32.exe" program is running and enabled.

For FEI (Philips) XL30 SEMs, be sure that the "SCS.exe" program is running and enabled.

For the JEOL 5900 Ethernet Interface, be sure that the "EDSI.exe" program is running and enabled.

For the JEOL 6500F Ethernet Interface, a JEOL technician must configure the NPGS PC in order for it to communicate with the JEOL SEM.

Note that often when you buy an SEM with an external control option (typically serial or Ethernet), the SEM vendors will NOT include the documentation that describes how to run it! You are encouraged to contact your SEM vendor and request that the complete documentation on the external control interface be provided. The documentation should describe how to enable the external control and it should also include a complete list of the available commands for external control of the SEM and/or stage.

Sharing Connections with EDS/WDS Systems (go to Top)

· NPGS has been working fine, but now we want to add an x-ray system to the microscope. What do we do?

The main issue is that other microscope accessories such as EDS/WDS systems will typically need access to the same connectors as NPGS. Some SEMs will have multiple sets of inputs for the analog voltages, however, most microscopes only have one XY input connector (or a pair of connectors when BNCs are used). In this case, either the cable(s) from NPGS and the other system must be physically changed before each system is used or a switch box can be installed that allows one system or the other to access the SEM XY input. The video signal output from the microscope can typically be connected to both NPGS and the other system simultaneously.

Some microscopes will have a serial input that allows external systems to control the microscope parameters such as magnification and focus. If the driver for NPGS is set to use the same serial communication settings as the other system, the simplest approach is to install a standard serial A/B switch box that will allow one system or the other to access the serial input to the microscope.

For all inputs to the microscope, such as XY, blanker, enable, and serial connections, you should never "tee" the cables from NPGS and the other system together.

SEM Input Relay Adapter not Working (go to Top)

· I have the optional SEM Input Relay Adapter that was supplied with NPGS and it no longer switches to the mode for external control of the SEM.

When the adapter is without power, the default condition is to stay in the internal SEM mode. Consequently, the most likely explanation for not switching the external mode is that the 5v power plug has come loose from the adapter or that the power supply itself has been unplugged from the wall socket. There should be a cable tie securing the power plug so that it is not accidentally pulled out. In one case, the power supply itself had failed and was promptly replaced at no charge.

Delay Before Each Pattern is Written (go to Top)

· When I start PG or AL there is about a 15 second delay when the program is loading, but the delay only occurs for some run files.

If a run file includes a command for an automated stage and a stage driver name is listed on line #1 of "Pg_Stg.sys" or a microscope driver name is listed on line #9 of "Pg_Cmnd.sys", an attempt to pass a command to the stage controller or microscope will be made when PG or AL begin. If a stage controller or microscope interface does not properly accept the command, the program will pause while the communication is attempted. In some cases, the driver will wait indefinitely and the computer will appear to have crashed. If the stage controller is already turned on, check that the cable between the NPGS PC and the controller is properly attached. Also, for newer microscopes where the stage is interfaced through the SEM PC, be sure that the interface program is running on the SEM PC. If no stage control is desired, either blank out line #1 of "Pg_Stg.sys" or disable the stage control using the "Advanced Modes" in the Run File Editor (for older versions, enter a "/" in the run file for the stage command).

Computer and/or DT2823

Crash/Hang/Reboot (go to Top)

· The computer hangs up every time I try to write a pattern or do alignment.

<This issue only applies to NPGS v8.0 and earlier.>

The only consistent cause for the NPGS software to hang is that whenever the diagnostics program DT2821.exe has been run, the PC must be rebooted before running PG or AL. Otherwise, the computer will hang every time when NPGS tries to access the DT2823.

If it is a new installation, other possible causes are:

1. If the DT2823 board passes the diagnostics, but after rebooting, the NPGS programs hang when they try to communicate with the board, there may be a hidden conflict that is not reported in the "System" option in the Windows Control Panel. Normally, any interrupt, address, or DMA conflicts will be reported, however, it is possible that Windows will report no errors, even when a conflict exists. The best approach is to reserve a different interrupt in Windows for the DT2823, change the interrupt jumper on the DT2823 to match, change the DT2823 driver in "c:\config.sys" to match, and reboot. In the only case where this problem has been seen, a sound driver that was using interrupt 9 conflicted with the DT2823 when it used interrupt 5. Changing the DT2823 to use interrupt 10 solved the problem. Typically, using interrupt 7 will work, even though Windows reserves 7 for use with a printer.
2. If the entire PC has been handled roughly (typically during shipment) the DT2823 board can become unseated. Open the case and check that the board is fully inserted into the ISA slot.

If the computer now consistently hangs up, but it had been working previously, possible causes are:

1. The DT2823 board may have been damaged and must be repaired or replaced. Damage can be caused by a power surge or static electricity, if the board has been handled improperly.
2. If the entire PC has been handled roughly the DT2823 board can become unseated.
3. The driver for the DT2823 board may have become corrupt or changed somehow. The driver is normally located in the root directory (C:\) for NPGS v7.6 and earlier and in "x:\NPGS" for NPGS v8.0 and usually has a name such as "pgdrv?.sys", where "?" is 15, 10, 7, 5, or 3. This driver must be installed in the "config.sys" file. From DOS or a DOS window, you can run "mem /c/p" to see if the driver is installed. For more information, see "Configuring the DT2823" in the NPGS Manual.
4. If another board in the PC has been added or changed, it may be conflicting with the DT2823 board. Conflicts are typically with the interrupt, but may also be with the address or DMA channels used.

· The computer occasionally hangs up when running PG or AL, and in general patterns seem to take longer to write than they should.

<This issue only applies to NPGS v8.0 and earlier.>

The only known cause for occasional crashes is if a memory manager (such as EMM386) is being used. Memory managers should not be used with NPGS, however, it is recommended to install "HIMEM.sys. If a memory manager is not being used, disable any TSR (Terminate and Stay Resident) programs until the conflict is isolated. Another possibility is that a stage driver program is being called and the driver is not able to communicate to the automated stage. Normally the driver should time out and return control, however, it may just wait indefinitely.

ATLAB Error (go to Top)

· Sometimes when I write a pattern, I see an "ATLAB error #9" message either during the writing or after the pattern has finished, however, the patterns are properly exposed.

<This issue only applies to NPGS v8.0 and earlier.>

This problem has been seen in a very small percentage of computers when the exposure time per point is at or below the minimum dwell time (10 usec) for the DT2823 board. Basically, it seems that the error message will be generated in some PCs when the board is being run at its maximum rate. Increasing the point spacing slightly so that the exposure time per point also increases slightly should make the problem go away. For example, users have reported an error when running at 10 usec per point, but none at 12 usec per point. For most PCs, the error is never encountered, so it seems to be a timing problem with just some PC models.

The problem may also manifest itself by displaying the "PG(2) error in DT2823 operation" error message, which will terminate the pattern writing. Again, this is limited to only a few PCs and happens only at the fastest writing speed.

If this problem occurs with NPGS v8.0, the "cycle extra" parameter can be set to "-12", which enables a special compatibility mode. A value of "-12" will ensure that the DT2823 is never run faster than 12 usec, which should eliminate the problem. Also, NPGS v8.0 has been modified so that it can automatically correct for occasional communication errors with the DT2823 board, so the overall system is more robust. These changes have become necessary because the newest PC's do not always provide 100% compatibility with ISA boards.

Corrupt ATL directory or ATL.dll (go to Top)

· I've installed Microsoft Visual C/C++ and it has a directory named "ATL" which appears corrupt when I run ScanDisk.

· The file "ATL.dll" in "C:\Windows\System" appears to be corrupted and I cannot delete/copy it or even view it with QuickView.

· I cannot create a file or directory called "ATL".

<This issue only applies to NPGS v8.0 and earlier.>

These problems all stem from the Data Translation driver, which is used for communication with the DT2823 board. The default name within the Data Translation driver was originally selected as "ATL", which for many years did not cause any conflicts. However, Microsoft also picked that same name to use with their "Active Template Library" found in Visual C/C++.

Fortunately, other than the symptoms listed above, the NPGS software works fine and Visual C/C++ works fine, even with the apparent conflict. Note that the directory and/or file are not actually corrupt, since everything tests out fine after the conflict has been removed. In any case, for the newest version of NPGS, the DT2823 driver has been modified to use a driver name of "PG$" which should not have any conflicts.

Miscellaneous

HASP Error (go to Top)

· The "Menu(220) Security Check" message is displayed when I load the NPGS Menu program.

· A "HASP Not Found" or "This program only runs on an authorized NPGS PC" message is displayed when I try to run some of the NPGS programs from DOS (not a DOS window).

The HASP is a small, white, 25 pin security plug that must be attached to a parallel (i.e., printer) port on the PC. If the HASP plug is attached and the programs still do not run then:

1. Check if the parallel port is functioning (try controlling a printer or perhaps a portable Zip drive).
2. Note that it is possible to disable a parallel port in the BIOS of the PC, so if the port is not working, it may be a BIOS issue, rather than an actual hardware failure.
3. If the parallel port is on an expansion card, make sure the card is properly installed.
4. Check if the PC is infected by a virus.
5. Try moving the HASP to a second parallel port (the software will automatically scan all parallel ports).
6. Try running PG.exe on a different computer with the HASP plug attached. (The PC must be running Windows 95/98 for NPGS 8.0 and earlier or Win2000 for NPGS v9.0. The PG.exe program can be run from a floppy and no other files are required. If the HASP does not respond, a HASP error message will be displayed, otherwise one or more non-HASP error messages will be displayed.)

If the HASP plug proves to be defective it may be replaced by contacting JC Nabity Lithography Systems (or Aladdin Software Security 1-800-223-4277 or 1-516-424- 5100 USA). To obtain a replacement, the original HASP plug must be returned in good condition (plugs that have been tampered with will not be replaced). To date, only one HASP key used with NPGS has ever needed to be replaced.

· A "HASP Not Found" message is displayed when I try to run some of the NPGS programs when running in a DOS window under Windows 95/98 or when NPGS is started from a Windows 95/98 icon.

< This issue only applies to NPGS version 7.6 and earlier. >

In addition to the possible explanations listed above, one of the settings for the DOS window or NPGS icon may be causing the problem. While highlighting the icon in Windows 95/98, hit the right mouse button, then select "Properties". Next, select the "Program" tab and then the "Advanced..." button. The box in front of the "Prevent MS-DOS-based programs from detecting Windows" item should be marked. If not, the NPGS programs will normally not be able to locate the HASP plug. When the box is not marked, 32-bit programs will not be able to be run from the DOS window or be called by NPGS, however, this is typically not a problem. If it is necessary to run a 32-bit program, a Windows 95/98 driver for the HASP plug can be provided which will allow NPGS to function when the "Prevent..." box is not marked. If needed, please send e-mail and request the Win95/98 HASP driver.

Limited PC Memory (go to Top)

· The stage and/or microscope driver that I added to Pg_Cmnd.sys is not executed by PG or AL.

· When I write patterns using NPGS instead of PG, the tunes are not played.

· I added a TSR program to my autoexec.bat file and now the tunes do not play.

· The statement "Program too big to fit in memory" is displayed on the screen when PG or AL is running.

<This issue only applies to NPGS v8.0 and earlier.>

Commands that are executed while PG or AL is running must use the memory available while PG or AL is loaded. Before running an external command, PG and AL will minimize the memory they need, however, if there is not enough remaining memory, the tunes (or any program listed in "Pg_Cmnd.sys") will not be executed. Typically, about 180 kb or more will be available. To solve the problem, make more memory available for PG and AL by removing TSR programs. This problem was more significant with early versions of NPGS, and is seldom seen anymore.

DOS Prompt (go to Top)

· Sometimes when I abort a program with ^C, the cursor is no longer displayed at the DOS prompt.

The following only applies to NPGS v7.6 and earlier. The program "CLG" (CLear Graphics) may be used to clear the screen and restore the cursor.

<This issue only applies to NPGS v7.6 and earlier.>