This is a software tool to load memory images onto the memSIM2 EPROM emulator from http://momik.pl .

MemSIM2 is an EPROM emulator. It emulates EPROMs from 2764 (8K/28 pin) to 27040 (4M/32 pin). This is accomplished by downloading your code with this tool.

WARNING: Do not use the memSIM2 in an EPROM burner. It will be destroyed by the programming voltages. To 'program' the EPROM just use the command line to update the emulators contents via the USB. Also memSIM2 may work fine in place of EEPROM but it can't be programmed in system like an EEPROM or Flash chip can.

WARNING: Take great care when using an adapter to simulate 2716 (2K) or 2732 (4K) ERPOMs (24 pin) some of the really odd chips put the voltages in different places or, worse yet, use voltages (+5, -5 and +12v) that will destroy the memSIM2. See the TMS2716 below as an example of an EPROM with triple voltages.

make
sudo make install

The Makefile is written for GNU make. On some BSD systems, you'll have to call gmake instead of make to get a GNU make. Tested on Linux and GhostBSD, which is based on FreeBSD. Should build on most POSIX systems out of the box.

The binary installs to the PREFIX path defined in the Makefile, usually '/usr/local/bin'.

sudo cp 62-memsim2.rules /etc/udev/rules.d/

This installs an udev rule. Once you attach the EPROM emulator, a symbolic link '/dev/memsim2' will appear which points to the usual '/dev/ttyUSBn', with n being any digit starting at zero.

If you try to access the simulator and get an error message 'permission denied', you probably don't have the required rights to access the EPROM simulator. You may gain these rights for example by adding your user to the appropriate group.

Ensure, the simulator is not plugged in. Plug it in then, then issue

	sudo dmesg

You should get a number of messages, the last ones should tell you the device name used by the EPROM simulator, '/dev/ttyUSB0' in this case:

[ 4789.932478] usbserial: USB Serial support registered for FTDI USB Serial Device
[ 4789.932504] ftdi_sio 1-1:1.0: FTDI USB Serial Device converter detected
[ 4789.932520] usb 1-1: Detected FT-X
[ 4789.932659] usb 1-1: FTDI USB Serial Device converter now attached to ttyUSB0

Show the group of that device:

	ls -l /dev/ttyUSB0

which gives here:

	crw-rw---- 1 root uucp 188, 0  9. Jun 07:20 /dev/ttyUSB0

so your user must be in the 'uucp' group to gain the required access rights. Run 'groups' to show the current set of groups you're part of. If your user name is not a member of the group used by the EPROM simulator, you may add this group to your user like so:

	sudo usermod -aG uucp <your-user-name>

Run 'su ' to apply the new group immediately. As an alternative, you may log out and in again.

Ensure, the simulator is not plugged in. Plug it in then, then issue

	sudo dmesg

You should get a number of messages, the last ones should tell you the device name used by the EPROM simulator, '/dev/ttyACM0' in this case:

[ 4789.932478] usbserial: USB Serial support registered for FTDI USB Serial Device
[ 4789.932504] ftdi_sio 1-1:1.0: FTDI USB Serial Device converter detected
[ 4789.932520] usb 1-1: Detected FT-X
[ 4789.932659] usb 1-1: FTDI USB Serial Device converter now attached to ttyUSB0

Show the group of that device:

	ls -l /dev/ttyACM0

which gives here:

	crw-rw---- 1 root dialout 166,  0 Mar 19 07:14 /dev/ttyACM0

so your user must be in the 'dialout' group to gain the required access rights. Run 'groups' to show the current set of groups you're part of. If your user name is not a member of the group used by the EPROM simulator, you may add this group to your user like so:

	sudo usermod -aG dialout <your-user-name>

Run 'su ' to apply the new group immediately. As an alternative, you may log out and in again.

My notes and ramblings. The code is a work in progress.

The memSIM2 can emulate up to 512Kbyte EPROMs (32 Pin), the smallest is an 8K 2764 (28 Pin)

The memSIM2 comes with a 32 pin cable also says that a 28 pin cable is available.

I've stacked the 32 pin end of the cable to a high quality machine socket to the socket. I learned that trick from using CPU emulators. It gives a sturdier set of pins to insert into the EPROM sockets.

I've got the 32 pin cable version of the memSIM2 and I also need to work with 28 pin EPROMS. I see in the flyer that it supports a Vcc of 2.7v to 5.0v. It is possible to connect the 32 pin cable to the 28 pin socket by using an additional 28 pin machine socket using the lower 28 pins of the 32 pin cable (pins 1, 2, 31 & 32, the top 4 pins are exposed). (Yes)

Similarly it is possible to also connect the 32 or 28 pin cable to a 24 pin socket by again using the lower 24 pins (pins 1, 2, 3, 4, 29, 30, 31 & 32, the top 8 pins are exposed). Be careful not to short the exposed pins to anything else. Also I'm pretty sure that the Vcc pin 24 needs to be connected to pin 28 of the cable for voltage sensing. (Untested)

• When loading code into the simulated EPROM where does it get loaded?
• I can have multiple address sections that may not be contiguous. Will it load correctly?
• How does it know where the EPROM starts?
• I'm not sure I have this coded with the s-record.
• Need to provide pictures of the stacked machine sockets adapters.

MemSIM2 supports the Intel hex format and straight binary. Motorola S19 format is a work in progress.

$ memsim2 -h
Usage: [OPTION].. FILE
Upload image file to memSIM2 EPROM emulator.  Where file can be .ihx or .bin.

Options:
        -d DEVICE     Serial device
        -m MEMTYPE    Memory type (2716 -2K, 2732 -4K, 2764 -8K, 27128 - 16K, 27256 - 32K,
                      27512 - 64K, 27010 - 128K, 27020 - 256K, 27040 - 512K)
                      2716-2732 are 24 Pin, 2764-27512 are 28 Pin, 27010-27040 are 32 pin.
        -r RESETTIME  Time of reset pulse in milliseconds.
                      > 0 for positive pulse, < 0 for negative pulse
        -e            Enable emulation
        -o BYTES      Specify an offset value with different meaning for:
                      binary files: skip first n bytes of file
                      Hex files: start address in memory map of simulated memory chip
        -h            This help
$

If the provided defaults work for you, the usage is as easy as plugging everything in while your devices are powered off, switching them on and running

        memsim2 imagefile.ext

to upload and reset the device. The kind of imagefile is solely detected by its file extension. These formats are currently supported:

There's not too much to say about these. Usually they're a binary dump of a memory chip. The type of memory chip (e.g. 2764) will be auto-detected if the data size matches the size of a supported chip. If not, a warning will be written to the screen and the next larger chip will be simulated.

It can't be easier than just simulating one chip of the very same size of the provided binary data, for example:

        memsim2 myrom.bin

If you don't provide a specific chip to simulate, it will be auto-detected by the amount of binary data but you may of course specify one:

        memsim2 -m 27128 os16k.bin

If your build system emits a larger binary than you actually would like to simulate, you may specify an offset to the beginning of the file indicating where the simulated data starts and the type of the memory chip to simulate. As an example, let us assume, your operating system occupies 16 KB and is spread across two 2764 chips of 8 KB each. You're currently working on code in the upper half, so you'd like to simulate the upper memory while your build system provides the full 16 KB:

        memsim2 -m 2764 -o 8192 os16k.bin

If you prefer, you may specify the offset in hex if you prefix the value with '0x' or '0X' such as used in the C programming language:

        memsim2 -m 2764 -o 0x2000 os16k.bin

This example will simulate a 2764 chip (8 KB size). The -o option specifies an offset of 8192 to make the simulator skip the first 8192 bytes of the provided file os16k.bin, so that only the upper half of the binary file is transferred to the simulated memory.

For historical reasons, it's even possible to provide the value in octal but I won't go into details here. You only need to know that a leading 0 triggers octal interpretation so if you mean 10 decimal you need to write it as 10 and not as 0010 because that would be octal 0010 = decimal 8 which would obviously break whatever you intend to do.

If the file extension is .hex, the file will be interpreted as Intel hex file:

        memsim2 mydata.hex

These may come in three variants:

• I8HEX (16 bit addresses)
• I16HEX (20 bit addresses)
• I32HEX (32 bit addresses)

Intel hex files usually consist of several records (text lines) with data, each secured with a checksum and provided with an start address indicating where the following data bytes should get stored.

Each checksum is verified, bad files are rejected.

The data records may appear in any order. For example data at lower addresses may follow data at higher addresses. Spaces between occupied memory areas are allowed also. memsim2 will determine the lowest and highest address at which data is stored and use this information to auto-detect the simulated chip if no -m option is given.

There are two possible ways to specify the addresses inside the hex file: seen relative to the storage position inside the memory chip or absolute inside the system's memory map. For example, think of a 2764 chip of 8 KB size storing code from 0xE000 to 0xFFFF within the system's memory map of a 6502 system. If seen relative to the storage position inside the chip, the addresses inside the hex file will start at 0x0000 and rise up to 0x1FFF (8191 decimal). If the addresses inside the hex file are seen absolute to the memory map, they will start at 0xE000 and rise up to 0xFFFF.

Either way, relative or absolute addresses, memsim2 will auto-detect the amount of data and the used addresses to determine the appropriate chip to simulate.

This detection may however fail if you want to simulate only a subset of a larger provided rom-set or if there is some unused, reserved space before the provided data starts.

Remember the 8 / 16 KB example from the binary files where we simulated only the upper 8 KB? If we had this data provided as hex files with addresses starting at 0x0000 (relative addresses), simulating only the upper half would look like this:

        memsim2 -m 2764 -o 8192 os16k.hex

As another example with absolute addresses inside the hex file, think of a boot ROM for a 6502 system located at the top of memory, stored in an 8 KB 2764 chip with addresses from 0xE000 to 0xFFFF.

You're working on a monitor code starting at address 0xF000, leaving 4 KB of free space reserved for further extensions in a range from 0xE000 to 0xEFFF.

memsim2 will notice that the provided data starts at 0xF000 so it will wrongly assume a start address for the memory chip of 0xF000 which will put your code in the wrong place: because the simulated chip is visible from 0xE000 to 0xFFFF, your data will now be placed at 0xE000 instead of 0xF000 and you won't have any useful reset and other system vectors at the top of memory, thus breaking your system. Just tell memsim2 where the simulated memory actually starts inside the system's memory map with the -o option:

        memsim2 -m 2764 -o 0xF000 monitor.hex

This may sound confusing. As a rule of thumb: if the auto-detection doesn't work, provide the kind of chip and the starting address within the system memory map:

        memsim2 -m 2764 -o 0xF000 monitor.hex

This file format is detected by a couple of file extensions: .s19, .s28, .s37, .srec and .mot, for example:

        memsim2 mydata.s19

It doesn't make any difference which actual file extension you provide, they're all treated in the same manner.

The S-Record file format is very similar to the Intel Hex file format, so don't be surprised to find most of the provided description here just being a copy of the former paragraph describing the Intel Hex file format. The most notably difference is that a S-Record file may contain a header with ASCII text that is written to the screen and can be used to provide a description, version number etc. Another difference is that there is a provision to detect data corruption if the actual number of records stored inside the file varies from the specified number of records. If this should happen, a warning is written to stderr.

SREC files usually consist of several records (text lines) with data, each secured with a checksum and provided with an start address indicating where the following data bytes should get stored.

Each checksum is verified, bad files are rejected.

The data records may appear in any order. For example data at lower addresses may follow data at higher addresses. Spaces between occupied memory areas are allowed also. memsim2 will determine the lowest and highest address at which data is stored and use this information to auto-detect the simulated chip if no -m option is given.

There are two possible ways to specify the addresses inside the hex file: seen relative to the storage position inside the memory chip or absolute inside the system's memory map. For example, think of a 2764 chip of 8 KB size storing code from 0xE000 to 0xFFFF within the system's memory map of a 6502 system. If seen relative to the storage position inside the chip, the addresses inside the hex file will start at 0x0000 and rise up to 0x1FFF (8191 decimal). If the addresses inside the hex file are seen absolute to the memory map, they will start at 0xE000 and rise up to 0xFFFF.

Either way, relative or absolute addresses, memsim2 will auto-detect the amount of data and the used addresses to determine the appropriate chip to simulate.

This detection may however fail if you want to simulate only a subset of a larger provided rom-set or if there is some unused, reserved space before the provided data starts.

Remember the 8 / 16 KB example from the binary files where we simulated only the upper 8 KB? If we had this data provided as hex files with addresses starting at 0x0000 (relative addresses), simulating only the upper half would look like this:

        memsim2 -m 2764 -o 8192 os16k.s19

As another example with absolute addresses inside the hex file, think of a boot ROM for a 6502 system located at the top of memory, stored in an 8 KB 2764 chip with addresses from 0xE000 to 0xFFFF.

You're working on a monitor code starting at address 0xF000, leaving 4 KB of free space reserved for further extensions in a range from 0xE000 to 0xEFFF.

memsim2 will notice that the provided data starts at 0xF000 so it will wrongly assume a start address for the memory chip of 0xF000 which will put your code in the wrong place: because the simulated chip is visible from 0xE000 to 0xFFFF, your data will now be placed at 0xE000 instead of 0xF000 and you won't have any useful reset and other system vectors at the top of memory, thus breaking your system. Just tell memsim2 where the simulated memory actually starts inside the system's memory map with the -o option:

        memsim2 -m 2764 -o 0xF000 monitor.s19

This may sound confusing. As a rule of thumb: if the auto-detection doesn't work, provide the kind of chip and the starting address within the system memory map:

        memsim2 -m 2764 -o 0xF000 monitor.s19

The first device is assigned to /dev/ttyUSB0. The digit starts counting at zero and increments from that on, so that a second simulator would attach to /dev/ttyUSB1. Please note that this is a generic name used by the FTDI chip inside the simulator. If you use other equipment with FTDI chips, these might use the same naming scheme. For details about how to detect the actual device name refer to the installation section.

On Linux, the device name defaults to /dev/ttyUSB0 so you won't need to specify it in most cases. If necessary, it may given with the -d option:

        memsim2 -d /dev/ttyUSB1

The memSIM2 USB EPROM simulator comes with a pin connector that attaches to the reset line. Some systems use high active reset lines, whereas others use low active ones.

The default configuration for the reset line is low active, with a reset pulse duration of 200 ms, reset after receiving new data enabled.

The reset configuration is controlled with the -r option. Legal values are in a range of 1 to 255 milliseconds. Zero disables the reset line. A positive number issues a reset for high active reset lines, a negative number issues reset for low active lines.

For example, to configure the reset line for a low active reset line with a reset pulse duration of 100 ms:

        memsim2 -r -100

The -m option controls the memory type to simulate. If this option is not given, memsim2 tries to auto-detect the most likely chip. If the option is given but memsim2 thinks it doesn't match the provided data, it will issue a warning.

If you're in need of tweaking the internal defaults,

	memsim2 -h

provides this help text with a short description of all available options:

Usage: [OPTION].. FILE
Upload image file to memSIM2 EPROM emulator

Options:
	-d DEVICE     Serial device, defaults to /dev/ttyUSB0
	-m MEMTYPE    Memory type (2764,27128,27256,27512,27010,27020,27040)
	-r RESETTIME  Time of reset pulse in milliseconds.
	              > 0 for positive pulse, < 0 for negative pulse
	-e            Enable emulation
	-o BYTES      Specify an offset value with different meaning for:
	   binary files: skip first n bytes of file
	   Hex files: start address in memory map of simulated memory chip
	-h            This help

Numbers prefixed by '0x' are interpreted as hexadecimal numbers,
octal for numbers beginning with '0' and decimal for everything else.

Basic parameters of the memSIM2 simulator

When I ordered the memSIM2 device I didn't know that the device could come with a 28 adapter cable. Now I have to wonder how to hookup the memSIM2 to the device. At this time I'm assuming the only pin that has to be rerouted is the Vcc (power) pin.

PROM/ROM index and cross reference

(16k)  2k x 8 EPROM (2716)
                              TMS2516  TMS2716
A7   1 +-v-+ 24  Vcc          +5       +5
A6   2 |   | 23  A8           A8       A8
A5   3 |   | 22  A9           A9       A9
A4   4 |   | 21  Vpp          Vpp      -5
A3   5 |   | 20  /OE          /CE      A10
A2   6 |   | 19  A10          A10      +12
A1   7 |   | 18  /CE          PGM      /CE
A0   8 |   | 17  D7
D0   9 |   | 16  D6
D1  10 |   | 15  D5
D2  11 |   | 14  D4
GND 12 +---+ 13  D3

                                                                 GI9332
                                                                 TMM333
                                                                 TMS4732
(32k)  4k x 8 EPROM (2732)                                       VT2332
                              TMS2516  TMS2708  TMS2716  TMS2532 SY2332
A7   1 +-v-+ 24  Vcc          +5       +5       +5       +5      +5
A6   2 |   | 23  A8           A8       A8       A8       A8      A8
A5   3 |   | 22  A9           A9       A9       A9       A9      A9
A4   4 |   | 21  A11          Vpp      -5       -5       Vpp     CS2 *
A3   5 |   | 20  /OE Vpp      /CE      /CE      A10      PD/PGM  CS1 *
A2   6 |   | 19  A10          A10      +12      +12      A10     A10
A1   7 |   | 18  /CE          PGM      PGM      CE/      A11     A11
A0   8 |   | 17  D7                                              *polarity of
D0   9 |   | 16  D6                                               chip select
D1  10 |   | 15  D5                                               set at manuf
D2  11 |   | 14  D4
GND 12 +---+ 13  D3

(64k) 8k x 8 EPROM
     JEDEC 2764                    TMS2564
Vpp  1 +-v-+ 28  Vcc         Vpp  1 +-v-+ 28 Vcc
A12  2 |   | 27  /PGM        /CS1 2 |   | 27 /CS2
A7   3 |   | 26  -           A7   3 |   | 26 Vcc
A6   4 |   | 25  A8          A6   4 |   | 25 A8
A5   5 |   | 24  A9          A5   5 |   | 24 A9
A4   6 |   | 23  A11         A4   6 |   | 23 A12
A3   7 |   | 22  /OE         A3   7 |   | 22 PD /PGM
A2   8 |   | 21  A10         A2   8 |   | 21 A10
A1   9 |   | 20  /CE         A1   9 |   | 20 A11
A0  10 |   | 19  D7          A0  10 |   | 19 D7
D0  11 |   | 18  D6          D0  11 |   | 18 D6
D1  12 |   | 17  D5          D1  12 |   | 17 D5
D2  13 |   | 16  D4          D2  13 |   | 16 D4
GND 14 +---+ 15  D3          GND 14 +---+ 15 D3

(128k) 16k x 8 EPROM
       27128
Vpp  1 +-v-+ 28  Vcc
A12  2 |   | 27  /PGM
A7   3 |   | 26  A13
A6   4 |   | 25  A8
A5   5 |   | 24  A9
A4   6 |   | 23  A11
A3   7 |   | 22  /OE
A2   8 |   | 21  A10
A1   9 |   | 20  /CE
A0  10 |   | 19  D7
D0  11 |   | 18  D6
D1  12 |   | 17  D5
D2  13 |   | 16  D4
GND 14 |   | 15  D3
       +---+

(256k) 32k x 8 EPROM
       27256
Vpp  1 +-v-+ 28  Vcc
A12  2 |   | 27  A14
A7   3 |   | 26  A13
A6   4 |   | 25  A8
A5   5 |   | 24  A9
A4   6 |   | 23  A11
A3   7 |   | 22  /OE
A2   8 |   | 21  A10
A1   9 |   | 20  /CE /PGM
A0  10 |   | 19  D7
D0  11 |   | 18  D6
D1  12 |   | 17  D5
D2  13 |   | 16  D4
GND 14 +---+ 15  D3


(512k) 64k x 8 EPROM
       27512

A15  1 +-v-+ 28 Vcc
A12  2 |   | 27 A14
A7   3 |   | 26 A13
A6   4 |   | 25 A8
A5   5 |   | 24 A9
A4   6 |   | 23 A11
A3   7 |   | 22 /OE
A2   8 |   | 21 A10
A1   9 |   | 20 /CE
A0  10 |   | 19 D7
D0  11 |   | 18 D6
D1  12 |   | 17 D5
D2  13 |   | 16 D4
GND 14 +---+ 15 D3

 JEDEC 128k x 8 EPROM 27C010

Vpp  1 +-v-+ 32  Vcc
A16  2 |   | 31  /pgm
A15  3 |   | 30  nc
A12  4 |   | 29  A14
A7   5 |   | 28  A13
A6   6 |   | 27  A8
A5   7 |   | 26  A9
A4   8 |   | 25  A11
A3   9 |   | 24  /OE
A2  10 |   | 23  A10
A1  11 |   | 22  /CE
A0  12 |   | 21  D7
D0  13 |   | 20  D6
D1  14 |   | 19  D5
D2  15 |   | 18  D4
GND 16 +---+ 17  D3

JEDEC 27020 EPROM

VPP  1 +-v-+ 32  Vcc
A16  2 |   | 31  /PGM
A15  3 |   | 30  A17
A12  4 |   | 29  A14
A7   5 |   | 28  A13
A6   6 |   | 27  A8
A5   7 |   | 26  A9
A4   8 |   | 25  A11
A3   9 |   | 24  /OE
A2  10 |   | 23  A10
A1  11 |   | 22  /CE
A0  12 |   | 21  D7
D0  13 |   | 20  D6
D1  14 |   | 19  D5
D2  15 |   | 18  D4
GND 16 +---+ 17  D3

      27C040

VPP  1 +-v-+ 32  +5V
A16  2 |   | 31  A18
A15  3 |   | 30  A17
A12  4 |   | 29  A14
A7   5 |   | 28  A13
A6   6 |   | 27  A8
A5   7 |   | 26  A9
A4   8 |   | 25  A11
A3   9 |   | 24  /OE
A2  10 |   | 23  A10
A1  11 |   | 22  /CE /PGM
A0  12 |   | 21  D7
D0  13 |   | 20  D6
D1  14 |   | 19  D5
D2  15 |   | 18  D4
GND 16 +---+ 17  D3

I haven't tried this yet but I expect this will work.

      27C040                   64k x 8 EPROM               2k x 8 EPROM

Vpp  1 +-v-+ 32  Vcc               27512                       2716                   TMS2516  TMS2716
A16  2 |   | 31  A18                                                                           DON'T USE
A15  3 |   | 30  A17        A15  1 +-v-+ 28  Vcc  ........................  --+                THIS!
A12  4 |   | 29  A14        A12  2 |   | 27  A14                              | Jumper
A7   5 |   | 28  A13        A7   3 |   | 26  A13        A7   1 +-v-+ 24  Vcc--+       +5       +5
A6   6 |   | 27  A8         A6   4 |   | 25  A8         A6   2 |   | 23  A8           A8       A8
A5   7 |   | 26  A9         A5   5 |   | 24  A9         A5   3 |   | 22  A9           A9       A9
A4   8 |   | 25  A11        A4   6 |   | 23  A11        A4   4 |   | 21  Vpp          Vpp      -5  <- ***
A3   9 |   | 24  /OE        A3   7 |   | 22  /OE        A3   5 |   | 20  OE/          /CE      A10
A2  10 |   | 23  A10        A2   8 |   | 21  A10        A2   6 |   | 19  A10          A10      +12 <- ***
A1  11 |   | 22  /CE /PGM   A1   9 |   | 20  /CE /PGM   A1   7 |   | 18  CE/          PGM      /CE
A0  12 |   | 21  D7         A0  10 |   | 19  D7         A0   8 |   | 17  D7
D0  13 |   | 20  D6         D0  11 |   | 18  D6         D0   9 |   | 16  D6
D1  14 |   | 19  D5         D1  12 |   | 17  D5         D1  10 |   | 15  D5
D2  15 |   | 18  D4         D2  13 |   | 16  D4         D2  11 |   | 14  D4
GND 16 +---+ 17  D3         GND 14 +---+ 15  D3         GND 12 +---+ 13  D3

Written by Simon Berg [email protected], Nils Eilers [email protected] and Neil Cherry [email protected] memSIM2 is free software, licensed under GPL v2 or later