Why printer specifications don't match your experience

Yesterday I read a detailed, helpful, well-written review of a new printer. The author covered all the things I'd want to know about when making a purchasing decision. A lot of the article was about how the performance numbers seemed exaggerated - testing showed that the printer didn't go nearly as fast as the published specifications would indicate. The author seemed surprised at this.

Excuse me while I change hats. There's a bit of dust on my printer R&D technician hat - ah, here we go. I will neither defend nor condemn the way that manufacturers arrive at performance specifications; but at least I may be able to explain them.

You will NEVER get the advertised speed from your printer. It does not matter who manufactures it; it does not matter whether it is a home-office desktop printer or a big high-throughput monster that the installers set on steel plates to keep it from wearing holes through the carpet. It does not matter whether it is a laser printer, inkjet, or some one of those obsolete technologies such as daisy-wheel. When a printer manufacturer publishes the technical specification for a printer's speed in pages per minute, that number represents the best speed obtainable in the test lab. If the printer is listed as a 20 page per minute printer, that means that it's able to drop 20 pages per minute into the output tray if you send it a huge print job consisting of nothing but page break characters.

How does that 20 pages per minute figure relate to actual experience?

Getting a print job started takes some time; you'll see lower average print speeds on short jobs than long ones, because the set-up time isn't all that much different for different sizes of print jobs.

If you ask your printer to actually make marks on the paper before ejecting it into the output tray - as most of us do, most of the time - that will slow things down. If your representative page in a long print job is about 30% covered with text (as for a letter or pages in a manual), your printer may do 15 to 17 pages per minute. This will be lower for short print jobs - maybe under 10 pages per minute.

If you want the printer to make marks on both sides of the paper, that will slow things down more. Even counting each sheet of paper as two pages, the speed will be lower than one-sided printing because of the time it takes the paper handler to flip each sheet.

So why do printer manufacturers not give comparison numbers from real-world testing? Would it be so hard to devise standardized testing - say, files consisting of varying numbers of pages of "Lorem ipsum"?

The other number that's likely to cause people to become unhappy with their printers is the "duty cycle" - how much of the time the printer is in use. The assumptions about how often a printer is serviced and how often it is likely to break are based on this purely hypothetical number, which is usually 25%. It's natural to assume that "25% duty cycle" means it's running 25% of the time, or 42 hours per week - but duty cycle is calculated on a 40-hour work week, not a full 168-hour week. So if your printer is rated for a 25% duty cycle, that means you are overworking it if it spends more than 10 hours a week printing. You'll hear about that from your printer service representative, who will probably urge you to get a high-output printer. It's probably good advice. You may feel as though you're being asked to trade an economy car for a luxury car, but it's more a case of not riding your moped on the freeway.

Vendors and leasing companies could save themselves a lot of explaining - and save their customers a certain amount of aggravation - by explaining how the speed and duty cycle numbers are derived; but printer manufacturers could nip these problems in the bud by adding text-printing test results to their specifications, and changing the way they calculate duty cycle.

Full disclosure: As mentioned earlier, I used to be a printer R&D technician. It has been many years since I had any connection to the printer business. Buy whatever printer you want, or hire a calligrapher - it's all one to me. :-)