Linear Technology’s Chief
Technology Officer Robert
Dobkin has been contributing
Ideas for Design (IFDs) since
his youth. He’s still enthusiastic
about them. “I remember
one I wrote when I was a kid.
It was a current source that you
could pulse on or off. I don’t know how many letters
I got from that,” he said.
“It had just two or three transistors in it, but it was
really effective in terms of people coming back to
me after reading it, which means there were a lot of
people out there who couldn’t do that,” he added.
“Also, of course, it was useful. There weren’t many
couple-transistor circuits that provided a current
source you could turn on and off.”
Dobkin breaks down IFD articles functionally into
those that provide a basic framework for inexperienced
designers to follow and those that go beyond
some component’s intended function to provide a
novel solution.
With respect to the former, he says, “I think those
IFDs are one of the best parts of the magazine—parts
that get torn out and stuffed in peoples’ drawers.
They’re important because if you look at the population
of engineers, most of them, when they first came
out of college, they can look at a circuit, they can
understand it, they can modify it, but they can’t really
design it from scratch.
“So all these guys ought to want to collect ideas for
circuits that do something, and now they’ve got some
seeds to start from. So sometimes the IFD gives a
starting point for a lot of different kinds of solutions...
and hopefully it was done by somebody who really
knew what he was doing so it really works when it
gets put together.
“Or they’re digital engineers,” he mused. “They
can do this complex digital thing, but if they try to
put three or four transistors together, that’s not what
they’re trained to do. So IFDs provide a nice way of
getting solutions to these people, too.”
The second kind of IFD, Dobkin says, “is one that
provides kind of a neat solution. Somebody came
up with a way to use an IC ‘off-application.’ They
found out that the ‘XYZ’ pin will also do something
undocumented if you feed a signal into it. So they
wrote a little IFD about that.”
Building Blocks
Along those lines, I asked Dobkin about IFDs that
take advantage of some non-ideal characteristic of
the part, for example, when a transfer characteristic
exhibits negative resistance. He said those IFDs
generally originate with the applications engineering
department at the manufacturer.
“That’s because most customers can’t get inside
the part. There isn’t enough information in datasheets
nowadays to allow customers to get inside the part
and use it off-application,” he said. “At one time, all
datasheets included the actual circuit diagrams. That
doesn’t happen now. So, you can’t tell exactly what’s
going to happen on the pin without that.”
Since Linear is best known for basic “building
block” parts, I asked whether they lent themselves
more to IFDs than more application-focused parts.
“I hope so. We try to come out with generalpurpose
parts that people will use for years. Even
the new ‘3080’ is one of those. (The LT3080 LDO
is a descendant of the classic LM317. Dobkin was
directly involved in designing both. See the figure.)
That’ll still be in use years from now,” Dobkin said.
“As a building block, it’s not an end in itself all
the time. A power-supply circuit can be used as an
amplifier. It can be used as an isolator. It can be used
as many things besides just a power supply. It can be
used as a power control. It can be used as a battery
charger. You don’t have to put too much around it to
change its overall function,” he said.
Getting back to IFDs and educating recent engineering
grads, I mentioned that ED receives many
IFD contributions from universities around the world.
As someone involved in hiring designers worldwide,
did he perceive any differences in the way universities
teach young engineers these days that affect
their approach to basic components? For the sake of
talking about that basic design instinct, he chose to
address the engineers who come from countries that
were formerly behind the Iron Curtain.
“Those engineers, those students, they didn’t have
access to a lot of common ICs, so they used to build
things out of transistors, which they could get,” he
said. “So you’ll see more transistor-level designs
coming out of there because the engineers have
more hands-on experience with bipolar discretes that
we get here. I think all engineers should have some
experience designing at that level.”
Finally, I had to ask him what he didn’t like
about IFDs. “The thing I dislike most are the IFDs
that don’t work. The contributor hasn’t thought it
through, hasn’t built it, doesn’t understand or doesn’t
talk about the potential problems. It’s not hard to
come up with an IFD and put it on a piece of paper,
but if you haven’t built it, you don’t know about all
its problems. You need a capacitor here to make it
work, the supply bypass leads have to be close by,
details like that,” he said.
“That’s a difference between a magazine’s IFDs
and a company’s app notes. If somebody takes a
reference design out of one of our app notes and it
doesn’t work, he knows who to call. If he gets an idea
out of a magazine and it doesn’t work, he should at
least have the e-mail address of the author.”