Sothius' Home

-

Museum

-

Commodore

-

CBM Model 3016

-

additional pictures

CBM Model 3016
Additional pictures. Note that loading the page could take a while...
(pictures copyright by M.A.Grundke)

serial plate: C= commodore 2001-16 N VDE 220V 0.5A 50Hz SERIAL NO. 1210442 MADE IN U.S.A. Below, the fixed-mounted powercord, the T 0.8A fuse and the power switch. As you can see, the serial plates says '2001-16 N', meaning that they used the same cases for all these early PETs (and the same boards, to a certain extent). The 'N' in the serial number, by the way, indicates a machine with graphics keyboard - machines with a 'B' in their serial number had the business keyboard

CBM 3016 opened up: the machine's circuit board fills the right part of the case, which has (besides the I/O ports on its back) an opening for the machine's internal expansion connector in its right. The left part is dominated by the integrated power supply, which apparently could get very warm - the plastic parts inside this particular 3016 are slightly yellowed, and most of the chips labels are hardly readable (see close-ups below). An interesting detail here is the metal plate above the PSU (below the keyboard) - its purpose was not only to shield the PSU, but also to carry a cassette drive, if required (i.e. if the case was used for a PET 2001 - as far as I know the only PET coming with integrated datasette)

ASSY 320351 'Dynamic PET P.C.B.' overview: this board is an improved variant of the original PET 2001/ 3001 boards (with 2K ROMs). It makes use of 4K ROMs (except for the screen editor ROM, which would otherwise collide with the I/O area), and has faster video RAM (see further below). However, it still doesn't have sound or a CRTC (first used in the next board revision), so video output had to be done using discrete logic. Now let's have a closer look at the components. The board's rear contains male edge connectors for (left to right) IEEE488, user-port and external cassette drive. Right below the latter you can see the internal expansion connector (2x 50 pins). The internal cassette connector (unused in the 3016) is located on the board's lower left, below the three large voltage regulators. Moving right from there, you can see the machine's RAM bank (two rows, UI2-UI9 and UJ2-UJ9), consisting of 16 pcs. 16Kx1 RAMs (=32 KBytes). When this 3016 was assembled, the lower row of the RAM bank (UJ2-UJ9) was unfitted, the additional 16 KBytes having been soldered there later. The middle of the board is almost entirely populated by the already mentioned discrete logic (gates, hex-inverters, flip-flops etc. of 74xx-type), responsible for doing the video timing and -output. Moving right from the internal cassette port again, and a little up, there's the character generator ROM (UF10), and still right, the two 2114-type SRAMs (used as video-RAM) at UF7 & UF8. Moving up from there, the ROM bank completely spreads from left to right: PET KERNAL (UD9), screen editor (UD8), BASIC 3 D000-DFFF (UD7), BASIC 3 C000-CFFF (UD6) and the three option ROMs (UD5, UD4 & UD3). Still up a row, from left to right: keyboard connector, 2x PIA (UC7, UC6), VIA (UC5), and finally, the CPU (UC4). According to the 320351 schematics, by the way, this particular board is an 320351-01 or -07 (since the machine came with 16K RAM and graphic keyboard), see table below. Why did they use different numbers for the same configuration, by the way?

PART NO.	DESCRIPTION
320351-01	DYNAMIC PET P.C.B. ASSY, 16K GRAPHIC
320351-02	DYNAMIC PET P.C.B. ASSY, 32K GRAPHIC
320351-03	DYNAMIC PET P.C.B. ASSY, 16K BUSINESS
320351-04	DYNAMIC PET P.C.B. ASSY, 32K BUSINESS
320351-05	DYNAMIC PET P.C.B. ASSY, 8K GRAPHIC
320351-06	DYNAMIC PET P.C.B. ASSY, WITH INTERNAL CASSETTE 8K GRAPHIC
320351-07	DYNAMIC PET P.C.B. ASSY, 16K GRAPHIC
320351-08	DYNAMIC PET P.C.B. ASSY, 32K GRAPHIC
320351-09	DYNAMIC PET P.C.B. ASSY, 16K BUSINESS
320351-10	DYNAMIC PET P.C.B. ASSY, 32K BUSINESS
320351-11	DYNAMIC PET P.C.B. ASSY, 8K GRAPHIC
320351-12	DYNAMIC PET P.C.B. ASSY, WITH INTERNAL CASSETTE 8K GRAPHIC

board silkscreening, part 1 (upper side of the board): COPYRIGHT 1978 COMMODORE INTERNATIONAL

board silkscreening, part 2 (upper side of the board): ASSY NO 320351 Above it, one of the DIP shunts (SH1) used for memory-size selection. Here are the DIP settings (from the 320351 board schematics). My machine, by the way, seems to be jumpered completely incorrect - these settings appear nowhere in the table! Nevertheless, it works, the gods know why... Note futhermore that the DIPs at SH2 also have to be set (see below)

Size	RAM to be used	A	B	C	D	E	F
32KB	16x 4116 (16Kx1)	OPEN	OPEN	CLOSE	CLOSE	OPEN	CLOSE
16KB	8x 4116 (16Kx1)	OPEN	OPEN	CLOSE	CLOSE	OPEN	CLOSE
16KB	16x 4108-30 (8Kx1)	OPEN	CLOSE	OPEN	CLOSE	CLOSE	OPEN
16KB	16x 4108-31 (8Kx1)	CLOSE	OPEN	OPEN	CLOSE	CLOSE	OPEN
8KB	8x 4108-30 (8Kx1)	OPEN	CLOSE	OPEN	CLOSE	CLOSE	OPEN
8KB	8x 4108-31 (8Kx1)	CLOSE	OPEN	OPEN	CLOSE	CLOSE	OPEN

board silkscreening, part 3 (upper side of the board): BS838054 written manually (and most probably meaning nothing...)

board silkscreening, part 4 (lower side of the board): PART NO 320350 REV E

board silkscreening, part 5 (lower side of the board): 49-79 Most probably the production date of the PCB (week 49, year 79)

The CBM 3016's CPU, MOS MPS 6502 (made with MOS' NMOS-process), with its label hardly readable. The 6502 is probably the most popular 8 bit CPU ever, not only because Chuck Peddle (its creator) convinced Jack Tramiel to build the PET-series 'around it'. In comparison to its strongest opponent, the Zilog Z80, the 6502 works memory-oriented, thus executing most of its operations directly in system RAM, instead of fetching commands and data into CPU-registers first (the Z80 and its ancestor, the Intel 8080, are so-called 'register-oriented' CPUs). That's why the 6502 and its derivates (6510, 8500, 7501, 8501...), although being clocked at only about 1 or 2 MHz in CBM's 8 bit designs, are not significantly slower than a Z80 is at 4 MHz (e.g. in AMSTRAD's CPC464) - an efficient design advantage. In addition to that, members of the MOS 6500 family have up to 16 bit address range, allowing 64 KBytes to be addressed at once - not bad for the time!

The machine's RAM bank, consisting of 16 pcs. 16Kx1 DRAMs. The eight NEC µPD416C-2 of the lower bank (UJ2-UJ9) have been soldered there later, effectively upgrading the machine to a 3032 (you can probably spot some of the brown soldering traces left there between them).The eight Texas Instruments 4116-25NL (upper row, UI2-UI9) have been there 'from the fab'. This ASSY 320351 board can be configured to 8, 16 or 32 KBytes (Remember? 'Dynamic PET P.C.B.'!), so it could also be used as a replacement board for the older 2001/ 3001 machines (I don't know for sure if the board was already available in the 2001/ 3001 lifecycle, but latest machines could have been based on it)

This is the 1 KByte dedicated video RAM (2 pcs. National Semiconductor MM2114N 1Kx4 SRAMs) used on this PET-board (of course, 80 column machines required double the amount of video RAM, i.e. 2 KBytes). All early boards (before 1981) had dedicated video-RAM, and came without a MOS 6545 cathode ray tube controller (using discrete-logic for output instead). Latest boards (in the 128K-machines, for example) didn't require that dedicated RAM anymore, since their CRTC was able to draw up to 8 KBytes video RAM from main memory - apparently without any drawbacks. The problem with video RAM in the early PET boards, however, was speed. In the very first PET-design, video memory was as fast as main memory. The problem was that while accessing video memory, there were no CPU cycles left for main memory access, causing what the german PET expert André Fachat calls 'video snow'. However, with the next revision PCBs (like this one), only one CPU clock cycle was required to access both video- and main memory, and the problem was solved

The CBM 3016's ROM area, consisting of: MOS 901465-01 (4KB, contains BASIC 3.0, C000-CFFF range), MOS MPS901465 (4KB, contains BASIC 3.0, D000-DFFF range), MOS 901447-24 (2KB, contains screen editor, E000-E7FF range) and MOS 901465-03 (4KB, contains KERNAL, F000-FFFF range). Note that the screen editor ROM at UD8 is different for each type of keyboard and character generator ROM used - the schematics mention 'business-' and 'graphic ROMs' for the US-ASCII version (business or graphic keyboard)

The BASIC 2/4 character generator ROM, MOS 901447-10 at location UF10, containing an extended ASCII character set (also called 'PETSCII'). There were also versions with country-specific characters (e.g. german, hungarian, swedish, norwegian) - Commodore knew it was essential to allow the PET-series the entrance into professional markets everywhere in the world (e.g. business and governmental use). However, as mentioned earlier, different character sets required different screen editor ROMs as well

The machine's function ROM sockets (also called option ROMs), which could contain additional programs (often used for third-party software, like word-processors or for special IEEE488 peripherals). It was Commodore's plan to make the computers more versatile, more individual by providing these additional sockets. However, when in 1984 the Plus/4 was launched (which was planned to come in different flavours with different option ROMs), Commodore dealers refused to stockpile all these different versions, because nobody knew if they would sell equally good (or bad). After that, some machines still came with option ROM sockets (like the C128), but there were no further plans to fit them from the fab. Don't know, by the way, what these particular ROMs contain - perhaps someday I'll read them out, for curiosity's sake...

The two MOS MPS 6520 'PIA' ('peripheral interface adaptor' or 'parallel interface adaptor'), some PETs came with MOS 6521 or Motorola MC6820/ 6821 chips, which are fully compatible. What they basically do is to provide I/O functions for controlling devices in a microcomputer. In the PET's case, one of them controls the IEEE488 bus (i.e. the three MC3446N transceivers), whereas the other one is used for keyboard scanning and controlling the cassette ports (together with the VIA)

This R6522P 'VIA' is an OEM-manufactured version of the MOS 6522. If the number printed on it is the production date (8412), then this must be a replacement part (I assume there was a defective MOS MPS6522 at this location in 1984). Apart from that, the 'versatile interface adaptor' provides a timer function and is used for user-port I/O and controlling the cassette interfaces, together with one of the PIAs. The VIAs were used in many different Commodore computers and devices, such as diskdrives and printers (that's what I call 'versatile'!). When they were later used in the VIC-20 and CBM 1540 designs, a serious bug in the so-called serial shift register was discovered, which was used for the serial data transfer. In order to solve this issue and retain VIC-20 compatibility, changes were done in the software (i.e. the ROMs). In the later MOS 6526, the problem was solved. However, this bug does not affect the PET-series, since they use IEEE488 peripherals. The later PETs, by the way, used the shift register for creating sound

These small MC3446N are the three IEEE488 bus transceivers used in the PET design. They create the required signals (=voltage levels) to drive the IEEE488 bus and the peripherals connected, and are controlled by one of the PIAs. IEEE488-compliance was an important feature these days, because this Hewlett Packard-developed serial bus was industry standard (although some claim that CBM's implementation is not entirely compatible with certain devices). It was also called 'HP-IB', and later (after the IEEE-standard had been created), 'GPIB' ('general purpose interface bus'). IEEE488 allows the connection of 15 devices to one bus, in the PET-series mostly printers and diskdrives. By the way, the later machines' CBM serial bus is in many ways similar to IEEE-488 (although slower)

A remnant of the original PET 2001 is the internal cassetteport (according to the schematics, 'cassette #1'), which was still present in the later PETs (like CBM 8296). It's absolutely identical to the second port on the machine's back, and was used for PETs with integrated cassette drive. Interesting to know is that CBM 3016 with integrated cassette drive and chiclet-style keyboard (PET 2001-style) did indeed exist, I recently spotted one in the internet (before that I believed that all machines, from the CBM 3001 onwards, would come with a 'real' keyboard). However, all machines after the 3000-series should have the full-stroke keyboard and thus no room for an integrated datasette

The machine's second DIP shunt (at location SH2), which is used for memory configuration (location SH2). Before it, you can see part of the internal expansion-port (2x 50 pin connectors, which can be accessed from the outside through an opening in the case). Note that, for memory configuration, DIP shunt SH1 also has to be set (see above). Note also that this machine seems to be jumpered incorrectly (but works, none the less). The settings are as follows (taking from the ASSY 320351 board schematics):

Size	RAM to be used	H	I	J	K	L	M	N	P	R	S
32KB	16x 4116 (16Kx1)	CLOSE	CLOSE	CLOSE	CLOSE	OPEN	OPEN	OPEN	CLOSE	OPEN	CLOSE
16KB	8x 4116 (16Kx1)	OPEN	CLOSE	CLOSE	CLOSE	OPEN	CLOSE	OPEN	CLOSE	OPEN	CLOSE
16KB	16x 4108-30 (8Kx1)	OPEN	CLOSE	CLOSE	CLOSE	OPEN	CLOSE	OPEN	CLOSE	OPEN	CLOSE
16KB	16x 4108-31 (8Kx1)	OPEN	CLOSE	CLOSE	CLOSE	OPEN	CLOSE	OPEN	CLOSE	OPEN	CLOSE
8KB	8x 4108-30 (8Kx1)	OPEN	OPEN	CLOSE	CLOSE	CLOSE	CLOSE	OPEN	CLOSE	OPEN	CLOSE
8KB	8x 4108-31 (8Kx1)	OPEN	OPEN	CLOSE	CLOSE	CLOSE	CLOSE	OPEN	CLOSE	OPEN	CLOSE

Also see the VC-20 CR page for a MOS 6502A picture!

Sothius' Home

-

Museum

-

Commodore

-

CBM Model 3016

-

additional pictures