The first question to settle when choosing a microcontroller is usually
Answer: A
A part that lacks a CAN controller or a twelve-bit converter is ruled out whatever its price or speed, so the peripheral mix narrows the field first. Speed and cost then choose among what remains, and the board is a convenience rather than a criterion.
Please log in or register to report this question.
Which of the following is NOT normally a microcontroller selection criterion?
Answer: B
Die size affects the vendor's cost but is neither published nor relevant to the buyer; the package outline is what constrains the board. Temperature grade, second sourcing and sleep current are all live considerations in a real selection.
Please log in or register to report this question.
The criterion that most often eliminates candidates before any other is
Answer: C
Peripherals are a pass or fail test, whereas price, longevity and tools are matters of degree on which a compromise is possible. Adding an external device to make up a missing peripheral is always an option but usually a costly one.
Please log in or register to report this question.
Consider the following statements about selecting a microcontroller.
1. A faster clock always gives more work done per second 2. Flash and RAM should be sized with headroom for later changes 3. Package choice affects how the board can be assembled
Answer: D
Headroom and package are both correct and both routinely underestimated. Clock rate alone says nothing about work done: cycles per instruction, memory wait states and the pipeline all intervene, which is why benchmarks exist.
Please log in or register to report this question.
DMIPS and CoreMark are quoted because clock frequency alone does not indicate
Answer: A
Two parts at the same megahertz may differ threefold in throughput according to their cycles per instruction, their memory wait states and their pipeline depth. A benchmark measures the product of frequency and efficiency, which is what an application feels.
Please log in or register to report this question.
A benchmark figure should be treated with caution because it
Answer: B
Optimisation level, library choice and whether the code ran from flash or from RAM all move the number substantially, and vendors quote their best case. A benchmark is a screen rather than a decision, and the honest test is the application's own hot loop.
Please log in or register to report this question.
Moving from an 8-bit to a 32-bit core often reduces code size for arithmetic-heavy work because
Answer: C
A 16-bit add on an 8-bit machine becomes several instructions with carry handling, and a 32-bit multiply becomes a library call. Doing it in one instruction saves both space and time, which often outweighs the wider instruction encoding.
Please log in or register to report this question.
For a design whose only computation is reading a switch and driving a relay every 100 ms, the appropriate core is
Answer: D
The computation is trivial and the deadline enormous, so nothing beyond the smallest part is justified and the pennies matter. Specifying performance nobody needs is one of the commonest and least defensible cost overruns in embedded design.
Please log in or register to report this question.
Flash size should be chosen with headroom because
Answer: A
Field updates, new features and bug fixes all add code, and running out of flash late in a programme forces either a redesign or a painful optimisation. Twenty to thirty per cent spare at first release is a common target.
Please log in or register to report this question.
The RAM requirement of a system running a real time kernel is dominated by
Answer: B
Each task needs its own stack sized for its deepest call chain plus an interrupt frame, and a dozen tasks at half a kilobyte each is six kilobytes before any buffers. The program image lives in flash and constants need no RAM at all.
Please log in or register to report this question.
