Leibson's Law

HP introduced its first inkjet printer, the ThinkJet, in 1984. No one mistook it for a high-quality printer. It competed with 9-wire impact dot-matrix printers (primarily from Epson) at the time. However, HP’s marketing and engineering teams recognized the vacant market niche for a “low-cost” printer (defined as under $1000 at the time) that could deliver laser-quality printing. So HP’s DeskJet engineering team went about re-engineering the existing ThinkJet technology to boost the print resolution from 96 to 300 dpi, which required several mechanical, electrical, and electronic innovations.

You can read about these innovations, plus the straightforward approach to “real” market research (as opposed to HP’s traditional “next-bench” approach to marketing) in this PDF of the October 1988 issue of HP Journal, which provides several detailed articles coinciding with the introduction of HP’s first DeskJet printer. (Note: it’s a long download. Be patient!) The first of the several stories starts on page 51. You’ll find some really interesting articles on the mechanicals, the electronics, and on the highly profitable printhead technology that still supplies a big chunk of HP’s profit two decades later.

For the purposes of this blog, the really interesting historical documentation resides in a sidebar on page 77. Here’s a quote from the sidebar written by HP’s Tom Pritchard:

“Even a very fast microprocessor would not have the time to perform all of the required character enhancements and other dot manipulations at the 300-dot-per-inch resolution of the DeskJet printer. So instead, a relatively slow, inexpensive, 4-MHz [Zilog] Z80 processor is used to control a large custom IC.

Approximately 85% of the logic in the custom IC is there to handle the dot data, as described in the accompanying article. The IC also handles serial and parallel data communications controls many logic functions required by the paper and carriage motors, provides timer functions to the Z80, and performs several external chip selects.

Contained within an 84-pin plastic leaded chip carrier package is logic laid out as two standard cell blocks, as shown in Fig. 1, and a large custom 50-bit wide path corresponding to the 50 nozzles of the printhead. There are approximately 80,000 field effect transistors in a die measuring 6.8 millimeters by 7.6 millimeters. A high-density CMOS process is used to fabricate the chip. Gate widths are 1.2 micrometers.”

I get several really interesting system-design tidbits from this sidebar. First, note the use of hardware to enhance and accelerate the processor’s ability to generate dots. This hardware is configured to be 50 bits wide, because that’s the “width” of the printhead and therefore, that’s the natural width of the data in this printer. Chopping the native 50-bit word into seven 8-bit chunks for manipulation by the Z80 processor would have made things much slower. Intolerably so.

The 32-bit processors of the day, such as the Motorola 68000 and Zilog Z80000 weren’t much faster, from a clock-rate perspective; they were a lot more expensive; and they still couldn’t handle 50-bit native data without chopping it in two. So the thoughtful design approach taken by the DeskJet engineering team made a lot of sense.

It’s also really helpful to get the historical perspective of the definition of “large custom IC” for the time period. Today, we’d hardly call 80,000 CMOS transistors (about 20,000 gates) “large” nor would we be impressed by 1.2-micron gate lengths, but that’s what system designers had to work with back then. How things have changed in two short decades!

HP’s DeskJet ushered in the era of low-cost, ink-jet printing for HP and for several other competing vendors (including Canon, Epson, Lexmark, Brother, and Alps, to name but a few). This 1988 issue of HP Journal does a nice job of giving some perspective on the early days of innovation in ink-jet printing.