Hi Bob: A coworker in Denmark informs me that Bang & Olufsen forbids transistors to drive capacitive loads without some form of series impedance. ("A foolish consistency is the Hobgoblin of small minds." /rap) This came about due to my use of an LM7805 with the usual filter capacitors and no series resistance in between either the load or on the input to the LM7805.
• Ken Harstine (via e-mail)
• Pease: Just about every IC linear circuit can be made to oscillate under some worst-case circumstances. So even an IC can be misapplied, and you have to make sure your circuit—whether IC or transistor—is stable and has good safety margins for being stable under all working conditions. But a heavier capacitive load usually does more good and no harm at all for the output of a voltage regulator, such as an LM7805 or LM317. This is because these ICs have the series resistance built in. That's what integration provides, as an advantage. And you can tell that to your colleague in Denmark!
Dear Bob: This is a more technical question than the administrative foolishness that you recently suffered. I use a debit card for shopping, and recently the swipe readers have increasingly failed to read the magnetic stripe. In most cases, this forces the sales assistant to type in the numbers manually, but it is a slow process and prone to errors. A week or two ago, I was surprised when the girl on the checkout, on failing to get the swipe reader to read it, just popped the card into a plastic bag and reswiped. Amazingly, it worked!
Since then, several others have done the same thing. It is evidently a widely known trick. One variation is to stick some Scotch tape over the stripe. Again it magically makes the card work! So what's going on here? At first, I thought it might be simply wear on the read head, causing it to "see" more than one symbol. Moving the card away would attenuate the contributions from the adjacent symbols more than those from the symbol opposite the reader and improve the S/N ratio. On further reflection, however, I realized that the fault was with my card, not the reader. The stripe is looking a little worn, but not excessively so. Any ideas?
• Martin Williams (via e-mail)
• Pease: I would have guessed the same theory you did. Maybe that's a bad theory, in which case I can't guess.
Dear Bob: Re: "What's All This Fuzzy Logic Stuff, Anyhow? (Part VI)" (Electronic Dsign, Dec. 4, 2003, p. 20), I wanted to make some comments about your racecar driver example. It is doubtful that any driver can consistently hit the start-braking mark better than 0.1 s, which would imply that the cars would be all over the place going through a turn. Because most of the top performers can hit the apex of the curve within an inch or two, something else has to be working here.
That ability to hit the mark happens because the drivers don't work right at the limit of adhesion, but somewhere below that. This gives them "room" to modulate the braking to allow them to hit that mark. The better drivers use smaller margins most of the time, but other factors determine the minimum time around a course (given the same car).
The best drivers ever, Jim Clark and Jackie Stewart, would seem slow through the corners because they were so smooth. Actually, they were slower than other drivers in certain parts, braking a little earlier and less abruptly. This seemed to keep the vehicle in better balance and allowed them to get more acceleration coming out of the turn than the others could manage. This was surprisingly effective, and I can remember one GP of Canada in the 1970s when Stewart pulled out a 10-s lead on the first lap, with the next two in line having a clean shot too.
• Don Rumrill (via e-mail)
• Pease: It's true that braking at the last possible millisecond might cause the car to run too slow in the corner, bounce around as the brakes are cut off, and—as you point out—to be less balanced and slow coming out.
[email protected] —or:
Mail Stop D2597A, National Semiconductor
P.O. Box 58090, Santa Clara, CA 95052-8090