Fullscreen HTML5Med Res (??MB)HTML4

FEATURE PROJECT: EPROM PROGRAMMER; PT.3 In this final article, we’re presenting the Windows-based software for controlling our new EPROM programmer. We also detail a few modifications which have been made to the circuitry to improve its performance. Pt.3: By JIM ROWE 72  Silicon Chip www.siliconchip.com.au D URING THE SOFTWARE development, we encountered a number of problems which delayed its release until now. Some of these problems were caused by timing errors that were encoun­ tered when using different settings for the “en­ hanced” printer port. They were fixed by adding a couple of RC filters. The rest were fixed by adding a few pullup resistors on the data input lines and by eliminating one capacitor. This means that we are now able to present the software and it’s the Windows-based version – not just a clunky interim ver­sion written in VBDOS, which looked a strong possibility when the second article was published. Before I start talking about the software though, let’s take a look at the modifications which have been made to the programmer hardware. Incidentally, the revised PC board pattern has been sent to board manufacturers, so the boards you buy or receive in programmer kits should have all the modifications included. Hardware modifications The first modification involves adding pullup resistors to the input lines from the printer port connector (CON1) to data buffer IC1 and mode decoder IC3. This prevents the chip inputs from “floating” at indeterminate voltage levels when the program­mer isn’t connected to the printer port. Leaving the inputs floating makes them prone to induced cable crosstalk and noise (and could even result in damage). The extra parts involved a SIL (single in-line) array of eight 10kΩ resistors for the data inputs of IC1 and discrete 2.2kΩ pullup resistors on the three inputs to IC3. These are shown at upper and lower left in Fig.11, which shows the modified input and decoding circuitry. There is one small complication with these changes: if the programmer is now powered up when it isn’t connected to a PC, the fact that IC3’s inputs are all pulled high sets the programmer into programming mode. This means that if an EPROM is also plugged into the ZIF socket, it will be subjected to the program­ming voltages continuously. So it’s now important that an EPROM should not be plugged into the ZIF socket until the programmer is connected to a PC’s printer port, www.siliconchip.com.au Fig.11: the modified EPROM Programmer circuitry. The modifications involved adding extra pull-up resistors on the data lines to IC1 & IC3, plus RC filters to the READ & PROG* lines. A capacitor is also deleted from another section of the circuit (see text). powered up and set to read mode by the software (which does this as soon as it starts). Another modification involved adding a couple of RC filter networks. The original circuit showed low-pass RC filters on five of the outputs from mode decoder IC3 (LAL*, LAM*, LAH*, LCF* and LPD*). They were necessary to suppress narrow glitches from IC3, for reliable address and configuration downloading. Similar RC filters have now also been added to the remain­ing two mode February 2003  73 This revised version of Table 1 shows the corrected entries for the CF1 configuration bit. lines (READ and PROG*). They provide small time delays which match those caused by the existing RC filters. Both these extra filters are shown in Fig.11 and have exactly the same component values as the originals. The remaining hardware modification involves removing a component. The original circuit showed a 100nF bypass capacitor from pin 24 of the ZIF socket to ground. However, this line cannot be fitted with a bypass capacitor because it’s the EPROM OE* line (output enable/disable) and must be toggled rapidly during programming and data verification. A bypass capacitor slows down the voltage changes quite seriously, so it can’t be used here. The solution to this one was easy – remove the capacitor. It was originally mounted at the very front of the PC board, just below the ZIF socket – see Supplementary Parts List Capacitors 2 1nF metallised polyester Resistors (0.25W 1%) 3 2.2kΩ 2 100Ω 1 8 x 10kW SIL resistor array, pin 1 common (SIL1) Delete 1 100nF multilayer monolithic capacitor (near ZIF socket) 74  Silicon Chip the PC board overlay (Fig.8) on page 34 of Pt.2. It was also shown on the original circuit diagram (Fig.3) on page 27 of Pt.1 and should be deleted. The revised PC board overlay is shown here as Fig.12 and includes all the corrections. Most of the added components are just above IC1, IC2 & IC3, with the remaining components just to the left of IC1. Note that to fit in SIL1 (the 8 x 10kΩ pullup resistor array), I had to reduce the mounting centres for nine of the 100Ω input suppressor resistors. This means that these no longer mount horizontally but are instead tilted at about 45°, so their leads mate with the board holes. Note too that SIL1 is fitted with its “common” pin to the left, just above pin 19 of IC1. If you fit SIL1 the other way around, you’ll almost certainly get rather strange problems. So that’s the story on the hardware modifications. Just follow the revised layout shown in Fig.12 and you won’t have any problems. Configuration table error There was also an error in Table 1 (Pt.1, p23), which showed the functions of the CF1 bit transposed. As a result, a corrected version of Table 1 has been reproduced in this article, to avoid any possible confusion. Any port won’t do! Before moving on to the software, let’s first take a look at the various printer port configurations and see how the differences between them can affect the operation of the EPROM programmer and its software. When the IBM PC first appeared in the early 1980s, it was provided with a fairly basic parallel printer port conforming to the standards originally specified by printer maker Centronics. This “Centronics” type of printer port became the standard for quite a few years but as processors and printers both became faster, various “enhanced” printer port configurations came into use. These provided both faster operation and fully bidirectional parallel data transfer. When the American IEEE came out with its IEEE-1284 stan­dards (see side panel), PC makers began incorporating multi-mode ports directly into their motherboard chipsets. That’s why the CMOS setup program in modern PCs gives you a number of options for configuring the onboard printer port(s) – usually as a “legacy” Centron­ ics port or SPP (standard parallel port), an enhanced EPP port or an enhanced ECP port with DMA options. We wanted the new EPROM Programmer to be backward compat­ible with older PCs having only an SPP, so that it could be used with just about any PC. As a result, I originally tested the design using an old 66MHz 486 machine, running a quick-and-dirty test program written in VBDOS (and running under DOS). This machine has only a bog-standard SPP/Cen­ tronics port and every­thing worked fine. However, when it came to developing the final software using Visual Basic 6, I used a Pentium 3 866MHz machine running Win98SE. This machine allows the onboard parallel port to be set to one of several different modes, so I automatically went into the CMOS setup and configured it as an SPP which seemed like the logical choice. Unfortunately, I was wrong! It turned out that on both this machine and a Pentium 2 266MHz machine which I tried as well, setting the port for SPP/Centronics mode operation gave all kinds of weird timing problems. In this mode, the ports couldn’t toggle the various bit lines in synchronism in response to soft­ ware instructions. It was only when I tried setting the port to ECP mode (or www.siliconchip.com.au Fig.12: this is the parts layout for the modified PC board. Most of the added components are just above IC1, IC2 & IC3, with the remaining components just to the left of IC1. The deleted 100nF capacitor was originally installed just below the ZIF socket (see text). “EPP+ECP” mode for the P2 machine), that these timing problems suddenly disappeared. This means that if you’re planning to use the EPROM pro­grammer with a PC that has a multimode parallel port, you’ll almost certainly have to set it for ECP or ECP+EPP operation. By contrast, you probably won’t be able to change the port mode on an older PC (eg, a 486) but this shouldn’t be a problem – it will almost certainly have an original SPP/Centronics type port which should work OK. Windows complication There’s another trap for the unwary in using devices like the EPROM Programmer which communicate with the PC via the print­er port – especially when the PC is running one of the many versions of Windows. This is because Windows and its printer drivers are very possessive about the port and simply won’t let another program www.siliconchip.com.au talk to it after they have previously used it for printing (during the current session, that is). This means that if you use the PC’s parallel port to print out from a Windows program during your current session, you can’t then simply connect the EPROM Programmer to the port and fire up the software to read or program an EPROM. The software simply won’t be able to communicate with the programmer if you try! Instead, you have to reboot the PC to unlock Window’s steely grasp on the port and allow the software and hardware to talk to each other. The software OK, let’s now look at the actual software. As mentioned earlier, the software is Windows-based and was written using Visual Basic 6. It makes use of a freeware DLL (dynamic link library) called IO.DLL to directly access the parallel port hardware. This DLL was written by Fred Bulback and works with Win95, Win98 and Windows NT/2000/XP. The DLL file can be downloaded from Fred Bulback’s website located at www.geekhideout.com and has also been made available in the software package on the SILICON CHIP website. Thanks to IO.DLL, the VB6 software can send data out via one of the parallel port registers by calling the added subrou­tine: PortOut (ByVal PortAddress As Integer, ByVal Data As Byte) and can retrieve input data nibbles from the port’s status reg­ister by calling the added function procedure: ReadVal = PortIn (ByVal PortAddress As Integer) In each case, PortAddress has a value corresponding to the address of the port register in I/O space. So that’s how the software actually passes data out to the EPROM programmer and fetches it back again. Software interface A screen grab of the software’s main February 2003  75 Fig.13: the main interface for the software that’s used to drive the EPROM Programmer. It shows all the critical settings and includes a large “list box” that’s used to view a data file loaded from disk or read back from an EPROM. interface is shown in Fig.13. It has the standard Windows appearance, with a menu bar at upper left and a bunch of display windows and option button controls. There’s also a fairly large list box at lower left which can be used to view a data file loaded from disk or read back from an EPROM. The two small display boxes at top right are used to show the I/O base address of the selected parallel port (ie, the one that the programmer is connected to) and the current setting for the programming pulse width in microseconds or milliseconds. Beneath these boxes are eight pairs INSTALLING THE SOFTWARE On the SILICON CHIP website, in the software folder, you’ll find the following files: EPROMP1.CAB (1,283KB), EPROMP2.CAB (373KB), SETUP.EXE (137KB), SETUP.LST (4KB) and DEVCFGS.ZIP (1KB). Download these files and save them in your PC’s C:\Temp folder. Then from the Start menu, select Run, Browse to C:\Temp\Setup.exe and click OK to run this file. It’s then just a matter of following the simple instructions. Setup will unpack and install the executable EPROMProg.exe in the C:\ Program Files\EPROMProg folder, together with IO.DLL and all other files needed to run the program. It will also add a shortcut called “Silicon Chip EPROM Programmer” to your Start->Program Files menu. After the setup program has finished, open the file DEVCFGS.ZIP with Winzip, and extract the sample device config files in it to the C:\Program Files\EPROMProg folder, so that they’re ready for accessing by the program. The programming software is now installed, and ready to run. The compressed installation files can be deleted from the C:\Temp folder if you wish, or saved on floppy disks or a CD-ROM if you want to keep a backup. Incidentally, along with the EPROMProg.exe program and its associated files, the setup program also installs an uninstall log file, ST6UNST.LOG. This file is used by the Windows uninstall utility to uninstall all files, if you later choose to remove the software from your computer. 76  Silicon Chip of option buttons, which are used to set up the programmer’s configuration to suit each different type of EPROM. In the screen grab shown here, these have been set up automatically by opening the configuration file for a Microchip 27C256 device. Incidentally, I have produced a few of these configuration files for different EPROMs and these will be made available along with the software on the SILICON CHIP website. However, you’ll find that it’s also easy to produce other device configuration files yourself, using the software and information from the EPROM manufacturer’s datasheet. Basically, you only have to set the pin connections, the supply voltage levels for reading and programming, and the pro­ gramming pulse width. That done, you simply select File -> Save Device Config (from the menu) and the software guides you through saving the configuration as a file. Once the device configuration data has been saved to disk, you can set up the programmer for working with that type of EPROM at any time in the future simply by opening that configuration file again. This is done by selecting File -> Open Device Config from the menu. Two other options in the File menu allow you to open and save an EPROM data file (as you might expect). The software will not only load and save EPROM data in raw binary form but can open files in Intel and Motorola S-record hex formats as well. These latter formats are often used for saving EPROM data. The final option in the File menu is used to exit the pro­gram. Incidentally, when you do exit, the program automatically saves two handy pieces of information to a set-up file: the parallel port address currently in use and the device configura­tion file currently in use (if any). This information is then loaded by default the next time you fire up the program. The next main menu option (along from File) is Set. This gives you two options: the ability to set the parallel port base address and the desired programming pulse width in microseconds. There are three standard base addresses (278h, 378h or 3BCh) to choose from and a dialog which allows you to set the programming pulse width to any value from 1µs to 50ms. www.siliconchip.com.au Next along in the main menu options is Test. There’s just one option here and that’s to confirm that the software can communicate with the EPROM Programmer hardware – ie, that you have selected the right port base address and the programmer is powered up and ready to go (and that Windows hasn’t locked up the port)! It does this by making the programmer’s LED3 glow and then asking you to confirm that it is glowing. It then turns LED3 off again. The fourth of the main menu options is View, which again has only one option: to display the currently loaded data which has been either opened from a disk file or read from an EPROM in the programmer. The data is displayed in hex in the viewer wind­ ow, in lines of 16 bytes, with the address offset shown at the start of each line. The viewer window has a scroll bar so that you can examine every part of the data at your leisure. Read options The next main menu option is Read, which gives you two choices. One is to read the contents of any desired address range in the EPROM in the programmer, so you can view the data and/or save it to disk. This option asks you for the start and end ad­dresses for the range to be read but offers the EPROM’s own start and end addresses as defaults, in case you want to read all of its content (this depends on you having opened the right device configuration file, though). The second Read option is Check Erasure, which is provided so that you can quickly check if an EPROM has been properly erased. When this option is selected, the software quickly reads the contents of all addresses in the EPROM, checking to see if they all contain FFh (this is the sign of complete erasure). It’s always a good idea to use this option after you’ve wiped an EPROM in your UV eraser, to make sure it’s ready for reprogramming. The final main menu option is Program, which provides a list of three options. The first programs the complete EPROM with the data currently in memory – (ie, from opening a disk file or reading another EPROM). The second option programs just a range of addresses in the EPROM, again using the data currently in memory. You can specify just one EPROM address location if you wish, which is good for testing purposes. With both of these Program options, the software controls the programmer to perform a “fast” programming algorithm at each device address. This kind of programming algorithm does a “verify read” after each programming pulse, to check if the programming was successful. If it was, the address is regarded as properly programmed and the software sends the next address to the pro­grammer. On the other hand, if the verify read shows that pro­gramming wasn’t successful, the software directs the programmer to try again with another programming pulse. This “try again” routine continues until either that ad­dress has been successfully programmed, or a total of 10 program­ming pulses have been used without success. In the latter case, the software stops the programming operation and lets you know that programming failed at the EPROM address concerned. Fig.14: this oscillogram shows the key data and control line waveforms during a read operation – in this case, the read sequence for a single EPROM address at 7FFEh. www.siliconchip.com.au Don’t Skimp On The Programming Cable Be sure to use a good quality round DB25M-DB25M parallel data cable to connect the EPROM Programmer to your PC’s parallel port. Don’t use a 25-way ribbon cable with IDC connectors on each end if you can avoid it, because this type of cable has no shielding and degrades the noise and crosstalk performance of the programmer-PC link. A good-quality round cable, preferably one manufactured to conform to the IEEE 1284 specification, will give you virtually error-free EPROM reading and programming. I found this out the hard way, so be sure to follow this advice. In short, don’t skimp on the cable. This type of fast programming algorithm is now recommended by virtually all EPROM manufacturers, for most modern devices. Each manufacturer tends to call it by a different name, though – eg, “Turbo Programming”, “Programming Express”, “Flashrite”, “Presto II” and so on. The basic idea is that each address is only given just enough short programming pulses to ensure that the data is reliably stored. As you can see, this type of programming algorithm has built-in data verification while the programming is taking place. Despite this, the third option in the soft­ware’s Program menu Fig.15: the key data and control line waveforms during a programming operation. In this example, the address programmed is 1FE8h and the data is E8h. February 2003  77 About The Various PC Parallel Printer Ports The parallel port used in IBM’s original PC of 1981 – and in just about every clone made during the following few years – was very similar to the Centronics printer port. This in turn had been used on most of the earlier 8-bit computers like the Tandy TRS-80 and allowed 8-bit parallel data transfer from the PC to the printer. However, there were only five lines for transfer of “status” information the other way. This made it capable of transferring outgoing data fast enough to keep up with a dot-matrix or daisywheel printer but it could only be used for transferring data back into the PC in 4-bit “nibbles” – ie, at less than half the speed. This type of parallel port is now known as the “legacy”, “Centron­ ics” or “standard parallel port” (SPP) and is probably the only type found on every model of PC. When IBM subsequently released its PS/2 models in 1987, it provided them with an improved version of the SPP which allowed fully bidirectional 8-bit data transfer on the eight main data lines. This type of paral- lel port was adopted by other makers as well and became known as the “PS/2 parallel port”, “simple bidi­rectional port” (SBP) or “byte mode parallel” (BMP) port. Not long after this, when Intel produced its 386 processor and its associated chipset, it collaborated with computer makers Zenith and Xircom to produce an enhanced parallel port which would allow much faster bidirectional data transfer than the SPP or SBP. This new port became known as the “enhanced parallel port” or EPP and was designed to speed up both printing and communication with peripherals like scanners, memory, tape drives and even external disk drives. The EPP specification allowed for fast byte-wide transfer of both data and address information in both directions and was widely adopted in computers based on the 486 and later proces­sors. Then in the early 1990s, yet another type of enhanced parallel port appeared: the “extended capability port” or ECP. Originally proposed by Hewlett-Packard is Verify, which allows you to do a complete re-read of an EPROM and have its contents compared with the data currently in memory. This option might seem unnecessary but it has two uses. The first is obvious: it allows you to make doubly sure that an EPROM has been correctly programmed with the data. The second use is for re-reading an EPROM you’ve just read, to make sure the Read operation did indeed retrieve a faithful copy of the EPROM cont­ents. If Verify gives the “all clear”, the re-read data compares exactly with the data you read the first time. OK, so that’s a quick rundown on the EPROMPrg.exe program, the features it provides and how it’s used. Signal Oscilloscope/Logic Analyser. They show key data and control line waveforms during a read operation (Fig.14) and a programming operation (Fig.15). In Fig.14, you can see that a read sequence for a single EPROM address (here address 7FFEh) takes about 40µs. The first 15µs are taken by the address downloading (high bits, middle byte and low byte in that order), while the high data nibble is read after the address has stabilised and then the low data nibble at the end of the sequence. Fig.15 shows a programming sequence which is successful after a single 100µs pulse, so it is complete after about 160µs. The address programmed is 1FE8h and the data programmed into it is E8h. You can see the address download at the start, the 100µs long programming pulse in the centre (PGM_N and CE_N lines) and Key waveforms Figs.14 & 15 are screen grabs from SILICON CHIP’s Agilent 54622D Mixed 78  Silicon Chip and Microsoft, this soon became incorporated into the new IEEE 1284 standard specification entitled “Standard Signalling Method for a Bidirectional Paral­ lel Peripheral Interface for Personal Computers”, which was released in March 1994. As well as establishing a full standard for very fast byte-wide transfer using an ECP, IEEE 1284 also formalised the speci­ fications for the existing EPP, SBP and SPP port configurations. At the same time, it also established standards for the cables and connectors to be used with each type of port, for optimum performance. In most PCs manufactured in recent years, the primary parallel port is incorporated inside one of the motherboard chipset ICs. Many of these onboard parallel ports can be config­ured by the BIOS firmware to emulate any of the types of port we’ve mentioned above – or at least some of them. This configura­tion is done using the CMOS setup utility, which can usually be entered during bootup by pressing the <DEL> or <F2> key. then the verify read operation at the end (far right). Note that during EPROM Read, Check Erasure, Program and Verify operations, the software displays a graphical “progress” bar at lower left (just below the viewer window). This lets you observe the progress of the selected operation. Finally, the program is available for download from the SILICON CHIP web-site (www.siliconchip.com.au), together with IO.DLL and some sample EPROM config files to get you going (see panel). We are also providing the VB6 source code for the program, for those who’d like to see for themselves just how it does what it does. Hotshot programmers might also want to improve it in vari­ous ways –feel free to modify it to meet your needs. Happy EPROM programming! SC www.siliconchip.com.au