15

I know that the Sega Master system had two primary game formats: cartridges just like the NES and other game consoles of the time and the "Sega Card" which was a 2mm thick plastic card with contact pads on the back making it look sort of like a credit card sized SD card.

These were supposedly cheaper for publishers to distribute games on (though limited to games small enough that they wouldn't require banking/mapper chips) but I'm wondering why that was the case.

From all that I can tell it seems like the cards were just a single 4KB to 32KB rom chip epoxied into a plastic shell. That would remove the extra cost of a circuit board and larger plastic body but compared to the cost of the custom ROM chip it doesn't seem like it would make much of a difference if the game was already small enough to not need a mapper or other special chips.

Were there other differences that made the price difference big enough to justify having a second game slot and special format just for small games?

David
  • 1,949
  • 1
  • 9
  • 14
  • 1
    At first I thought it was just for backwards compatibility with their computer systems, but those took cartridges too. That said, from images it looks as though the cartridge format (or just design) changed, so maybe it was a way to provide a path forward for at least some titles? – Matt Lacey Nov 05 '19 at 22:59
  • Re: why, I'm sure I've read that it was Sega's ambition that the card based games be cheap enough that children would collect and swap them like other kinds of card. That doesn't help answer as to how the cost savings were achieved though. – Tommy Nov 06 '19 at 02:47

1 Answers1

11

The price difference was indeed entirely in the packaging. The important thing to remember about mask ROMs is that, though the setup costs and lead times were high, they had the lowest marginal cost of any form of memory. The reasons behind this are fairly clear if you look through this ROM, EPROM, & EEPROM Technology document from Integrated Circuit Engineering: mask ROMs not only used the minimum number of transistors per cell (one) but also had simpler processes, higher densitities, and higher yields.

Unfortunately, historical pricing information for mask ROM seems very hard to come by. This has been asked here before, where one answer gives the price of 2 KB (16 Kbit) mask ROMs as $15/unit in 1976.

At that time a 2708 1 KB PROM retailed for $68. By 1983 2708s and 2 KB 2716s were both retailing for $3.95. A year and a half later in mid-1984, 8 KB 2764s could be had for $5.95.

While PROMs and retail pricing are not directly comparable to mask ROMs and manufacturer pricing, this gives us a sense that there was probably something like a hundred-fold decrease in ROM pricing, so we can guess that 16 KB mask ROM would certainly cost well under a dollar, perhaps even under fifty cents, to produce.

Given that PC boards at the time were not as cheap to manufacture as they are now, and the costs of dealing with the assembly of a PC board, cartridge halves and their associated fasteners (which is mainly what was removed in the Sega Card), it doesn't surprise me at all that there would be considerable manufacturing savings with the Sega Card and similar formats.

cjs
  • 25,592
  • 2
  • 79
  • 179
  • 1
    Interestingly, DRAM density is often greater than that of mask ROM, since every transistor in a mask ROM needs a DC ground path, but DRAM can get by with just a capacitive path to ground. In 1990, I asked my VLSI design professor about the possibility of "DROM", and he said that it's easier to tell the difference between a cap that's charged to 5V vs ground, than to tell if a cap connection is simply absent. On the other hand, some multi-level flash storage media can now hold multiple bits per transistor. – supercat Apr 02 '20 at 14:59