ATARI 1040STFM
Additional pictures. Note that loading the page could take a while...
(pictures copyright by M.A.Grundke)
 |
ATARI®
1040STFM
ATARI CORPORATION
MADE IN TAIWANATARI CORPORATION, CA 94086
SERIAL NUMBER: A106L 4071908 |
 |
board overview (with PSU, shielding and diskdrive removed):
on the board's left side there's the expansion port and the two MIDI ports. Moving right from there, you can see the RAM bank with 32 pcs. 256Kx1 RAMs. Note that normally the RAM bank's covered by the machine's integrated power supply (thus the openings in the case behind). Below the RAMs are silkscreenings for up to six 256Kx1 TOS ROMs (occupied with 2 pcs. 1024Kx1 'BLiTTER TOS in this machine, U63 and U67). Right from there, held in its PLCC-socket by metal stripes, the machine's MMU (U56) and the MC68000 CPU (U64) below.
Right above MMU, by the way, the connectors for power supply and disk drive power. Let's get back to the CPU. Moving right from there, you can see GLUE (again fastened in its PLCC socket at U65), and the spare BLiTTER pads (U66), which are unoccupied in this machine. Above MMU there's Shifter (U49), with its video logic circuitry and completely shielded. Still above, the RF-modulator with its MC1377P below. Right from there the diskdrive data cable, and the machine's I/O area. Normally covered by the EPSON SMD-380 diskdrive, it contains (again from left to right) a WDC 1772-PH floppy controller (U15) with the DMA-controller (U31) below it, the Yamaha YM2149F (U16) and the MC68901P (U43) below. Above the Yamaha and to its right are the tiny serial line driver & -receiver (U8, U17).
Below the MC68901P, finally, the two ACIAs (U52, U54) and the case opening for mouse & joystick ports (which can be found on the keyboard's lower side, which additionally holds a Hitachi HD6301V1) |
 |
board silkscreening:
ATARI®
ACTION F
C070789-001 REV.F |
 |
keyboard PCB overview:
well, it seems there hasn't been enough space on the 1040STFM's mainboard, so the designers decided to put joystick and mouse ports onto the keyboard PCB (left). Right from there is the keyboard controller, Hitachi HD6301V1 (see below). It's placed on the keyboard PCB in all ST-series machines. On the lower right, by the way, the keyboard PCB silkscreening |
 |
keyboard PCB silkscreening:
C070777-002 REV.A |
 |
| The Motorola 68000P8 (8 MHz, DIP-package version) is the CPU used in most of the ST-series - exceptions are the MegaSTE (which came with PLCC-packaged, 16 MHz 68000), and of course TT and FALCON 030 (which are not exactly ST-architecture). Like in this particular machine, the MC68000 was soldered directly to the board, making replacement and expansion (e.g. with CPU-cards) difficult. By the way, the 'SC'-prefix is used by Motorola to indicate that a part is produced for a certain customer (in this case ATARI, of course). Note also that the MEGA-series machines had a socket for an
optional floating point unit (MC68881 or MC68882) |
 |
| C070714-001 'GLUE' is the so called 'control logic' in the ST-architecture. It provides basic logic operations to, roughly spoken, make the 'main system' work with 'graphics subsystem', 'music subsystem' and 'device subsystem' (from 'ST computer Engineering Hardware Specifications', ATARI 1986). On the picture's lower right you can see part of a jumper, W3. Together with W4, it is used to determine whether the machine comes with a BLiTTER equipped or not (W3 & W4 open = no BLiTTER, both soldered = BLiTTER fitted) |
 |
| Although the MC68000 could access 16 MBytes RAM, the ST's MMU (C100109-001 in this case) can only handle a maximum of 4 MBytes, limiting the ST-series expandability (an exception to this, again, is the FALCON 030, which can be expanded to 14 MBytes). Right from there, and not related to RAM, jumper CE. It's one of three jumpers to choose between 256Kbit or 1024Kbit TOS-ROMs (the other two being A16 & A17) |
 |
| That's the 1040STFM's RAM bank, consisting of 32 pieces 256Kx1 bit RAMs (=1024 KBytes). The rarer 2080ST and 4160ST used 1024Kx1 RAMs. Some 1040STs, by the way, came with 8 pieces of 256Kx4 RAM (and thus 24 spare sockets) making memory expansion easier. What I've not yet figured out is, however, how the machine can handle these kinds of differently organized RAMs (256Kx1, 256Kx4, 1024Kx1) - there must be some way to determine which RAMs are used. But there doesn't seem to be any jumpers for that purpose - perhaps ATARI did some manual modifications directly to the PCBs? |
 |
| These are the ATARI 1040STFM's TOS-ROMs, C300689-002A and C300690-002A, containing PAL TOS 1.2 (the so-called 'BLiTTER-TOS', because it supports a BLiTTER). Before them are jumpers A16 and A17. Together with jumper CE (near MMU), you can select operation of six 256Kbit TOS-ROMs (6x32 KB) or two 1024Kbit TOS-ROMs (2x128KB, with 192K used). Although most of the later machines seem to have come with two pieces ROM-sets, the 6 pieces sets of newer TOSes (like TOS 1.4) were also available |
 |
| The C070713-002 Shifter is part one of the ST's graphics subsystem (the second part is the 'video display memory', which is 32 KB RAM taken from main memory upon operation). Shifter's located in the board's middle, and, together with its video circuitry, completely shielded. It was, however, the ST-series' major weakness compared to their archrival, the AMIGA. Although it could create HighRes-graphics (640x400 monochrome) with 72 Hz (AMIGA: 60 Hz NTSC/ 50 Hz PAL), its color graphic capabilities were rather limited: there were only two color graphic modes, 320x200 @ 16 colors and 640x200 @ 4 colors (all from a 512 colors palette),
which additionally required another monitor (RGB) than HighRes - the AMIGA wasn't that fastidious. It could output all its resolutions to a RGB-monitor (although some interlaced), and featured up to 64 colors @ 320x200, even all 4096 colors of its palette simultaneously in HAM-mode (additionally, the AMIGA had Blitter & Copper for fast filling operations, soft scrolling et cetera, all perfectly suited for graphics and gaming). Although making the ST display all its 512 colors at once was possible with 'tricks', ATARI realized that Shifter had to be improved: the later STE Shifter was far more powerful, and came (what a surprise) with a 4096 color palette (although coming too late) |
 |
| The machine's modulator circuitry, which makes use of a Motorola MC1377P for encoding the RGB signals created by Shifter into composite PAL/ NTSC-signals (yes, it can create both). These signals can then be outputted by the RF-modulator. Note by the way that the 'P' indicates plastic DIP package - the MC1377 was also available in plastic SO20-L package as MC1377DW. Note furthermore that PAL/ NTSC operation must be selected via pin 20 - if grounded, the MC1377 is set to NTSC, otherwise PAL - this couldn't be done by software, so ATARI offered different PAL- and NTSC machines with appropriate modifications done to their video hardware |
 |
| These spare pads at U66 take the BLiTTER in the later 1040STs (to speed up filling-operations and thus graphic applications, including GEM). Since the mainboard design is already prepared for it, a BLiTTER could easily be added here (I don't know exact figures, but I think all 1040ST made past a certain production date came with a BLiTTER fitted). In order to enable a BLiTTER soldered here, you also have to set jumpers W3 and W4 (as mentioned earlier). If both are open, the BLiTTER's disabled, so both need to be soldered in order to activate it |
 |
| The Yamaha YM2149F or the compatible AY-3-8910 by General Instruments were used for producing sound in the whole ST-series, even in the later enhanced STs (520STE, 1040STE, 4160STE and Mega STE), although these had an additional PCM stereo soundchip. It features three independant voices and several sound effects. With its 2 MHz clock it can produce output frequencies between 30 Hz and 125 KHz. Compared to the AMIGA's Paula soundchip
(with 4 independant channels), the ST-series' Yamaha was less powerful and flexible. However, the Yamaha has I/O-ports for some additional tasks in the ST-architecture, such as receiving 'Request to send' and 'Data terminal ready' signals (RS232), or 'strobe' (Centronics) signal for the parallel interface |
 |
| The popular Western Digital WD1772-PH is the floppy controller used in all ST's. It uses the machines' DMA-controller for memory access, and can control two daisy-chained drives. By the way, a harddisk controller (which would also use DMA) is not implemented in the ST's architecture - it had to come with external harddisks, and was called AHDI ('ATARI harddisk interface'). Communications between the DMA- and harddisk-controller were done by using 'SCSI-like' (ANSI X3T9.2-compliant) commands |
 |
| The ST's 'general purpose DMA controller', C100110-001. Controlling the system-bus together with the CPU, the DMA-controller is used for transfering data directly from the ST's 'device subsystem' to main memory. As described above, this is used for diskdrives as well as harddisks. Note that CPU and DMA controller are both 'busmasters', so that one of them has to wait until the other one has finished with its bus operation |
 |
| The ST EF6850P (aka MC6850P) are used for MIDI IN/ OUT, and to send commands to ATARI's 'intelligent keyboard' (see below). As the name implies, the ACIAs were originally designed for Motorola's MC6800-CPU (the P-suffix indicating plastic package, by the way). The MIDI ports, furthermore, were the reason why so much musicians used STs instead of the AMIGA: MIDI is a standard protocol for communication between musical devices (such as keyboards, mixers and other equipment). For the AMIGA, MIDI ports were also available, but were not part of its design and thus had to be purchased separately |
 |
The ST's serial line driver (right) and -receiver (left). Although manufactured by Goldstar with slightly different part numbers here, these are fully compatible to the common MC1488 and MC1489. These chips are used to provide hardware-level communication with RS232-compliant devices (i.e. providing the correct TTL signal levels for input and output). However, they need to be controlled by the machines MC68901P (which provides the 'higher level intelligence', such as baud rate selection, error monitoring, or detecting of the 'Ring indicator' signal). As mentioned earlier, the Yamaha is also responsible for some
of the RS232-compliant signals (such as 'Request to send' and 'Data terminal ready') |
 |
| Motorola MC68901P 'Multi-Function Peripheral' (aka Mostek MK68901) is probably the most versatile part of the ST-architecture. It's used for controlling the serial line driver and -receiver, providing serial RS232 communication, as well as data transfer on the machine's parallel port (both with YM2149F-support). Its 16 sources interrupt controller is used to generate interrupts for CPU and GLUE (i.e. to 'tell' them that a subsystem, like RS232, needs service). Additionally, it contains a single-channel USART which is used in the RS232 communication process, and provides (very important) the detection of the attached monitor (via 'monochrome detect', so that the monitor cannot be damaged
by driving it with the wrong sync signals). There's also a NTSC/ PAL detection pin |
 |
On keyboard PCB:
the Hitachi HD6301V1 'Microcomputer Unit' (aka ATARI C070122-002) is what ATARI called the 'intelligent keyboard' (ikbd). It scans for keypresses, mouse/ joystick-movements and other events, and provides a time of day clock. Additionally, it can also receive commands from the machine's ACIAs at 7.8125 KBits/s. Interesting to know is that one of its ports is 'combined mouse/ joystick', the other one 'joystick only'. The HD6301V1P is, by the way, not located on the machine's mainboard but on the keyboard PCB (keyboard's lower side) |
 |
On keyboard PCB:
the 1040STFM mouse/ joystick ports. As mentioned earlier, one of these can be used for both mice and joysticks, while the other one is for joysticks only. Interesting in the STFM is these ports' location: they have been moved to the keyboard's lower side, on the PCB were the HD6301V1 is. Reason for that were practical considerations, I assume: after all, these ports are controlled by HD6301V1, so why not put them together? Additionally, because of the integrated diskdrive, there's no room left on the machine's right side (were these ports were located in the older machines) |