Indeed, a lot of the value created by a microcontroller company is in
these reference designs and the choice of peripherals alongside the CPU
core. Of course Microchip has made its reputation with the PIC range of
At the 8- and 16-bit level these have a proprietary
architecture derived from Microchip's origins in the microelectronics
division of General Instruments. The 32-bit version of PIC is based on
the M4K core from MIPS Technologies Inc. MIPS was recently acquired by
Imagination Technologies Group plc, but the message has been that it is
business as usual.
Sanghi commented: "The core has become
irrelevant. The CPU core is a small part of the microcontroller. There
are the peripherals, the I/O, USB, Wi-Fi, A-to-D converters and on and
on. Most customers use compilers." The last point indicates that for
their code most software developers no longer need an intimate knowledge
of CPU core and generate their final version of their code using a
"You want ARM? We could put ARM in any time. We
already have it in the company through acquisition of Roving Networks
and SMSC [Standard Microsystems Corp.]. But why change? PIC32 is built
on MIPS and has a large ecosystem."
Sanghi is adamant that the
reputations certain CPU architectures have for low power consumption
have become moot for MCUs because of the minor part the active
consumption of the core plays in overall power consumption of the MCU
and in a system.
"For a start, the MCU current consumption is
dominated by the consumption when it is not active, when systems are in
sleep mode," Shanghi said. Referring to what is most important in low power
implementation Sanghi adds: "It's not the core, it's not the memory,
it's the overall MCU design and the manufacturing process and the
interaction of the two." And that interaction is exemplified by the
extension of Microchip's XLP range--extreme low power--of
microcontrollers to the 32-bit CPU range.
XLP is a range of PIC
MCUs that has been developed with particular thought to power
consumption efficiency and long life in battery-operated applications.
This is achieved by having specially crafted low power monitor circuits
and real-time clocks so that current is reduced to the minimum while the
MCU is asleep, where most embedded systems spend 90 to 99 percent of
Right now, XLP microcontroller variants are offered
with 8- and 16-bit PIC MCUs but 32-bit is coming, said Sanghi. "We're
working on that at 32-bit," he said, but declined to give any indication
of when products would be available.
Thank you Peter. A least one voice of reason here.
Not everything Silicon Valley (where I leave and work) makes or thinks of is all we need, and not all we need comes from Silicon Valley.
Steve Sanghi is one of the smartest people in this industry. Just the fact that he knew how to turn around a failing NVM company and make good dough for over 20 years as its head, navigating it safely around troubled waters and keeping it financially strong, is deserving a lot of credit.
The Tesla example is perhaps misused by Peter, or Maybe overly emphasized. Don't just use one quote to dismiss the whole person. Read carefully, there's some wisdom in there.
You are entitled to have an opinion, although clearly "junk" is a generally considered a derogatory term.
It would be helpful to other readers to elaborate on any issues you see.
People do buy a considerable numbers of Microchip ICs and one person's "cheap" is another person's value for money.
I wish the exec described a more broader and appealing use case than the one with Tesla. To that end, to make some of the vision at Microchip happen, the question on "MEMS technology and multi-die packaging is on the horizon from Microchip?" is a relevant one. For one, I have not come across initiatives or efforts from Microchip to take part in industry standards bodies and events to make its efforts known.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.