Compare ARM MCU Vendors

l selling the

Atmel has more

Marvel has the fastest ARMs (previously xscale). But getting basic datasheet needs NDA (it was true sometimes ago) and they don't seems to be interested in supports of small users.

Actel also supports hard and soft ARM cores with FPGA.

Apple, Samsung and Qualcomm makes propriety ARMs.

There are many others as well.

ally shocking.

Because it's changing every day.

y.

As needed.

The same as for any other processor: study data sheets, look at manufacturer's sites, recall previous experience with that manufacturer, read EDN "processor roundups", etc.

selling the

There are a lot of different ARM cores, there are ways and ways or realizing cores, and there are a lot of different ways to build a microprocessor around a core. So yes, there are going to be big real differences.

the same?

Not at all. In fact, price, support, and longevity are often what matters the most, and what you're going to get the least information on (except for today's pricing) from the vendors. Hence, it helps a lot to know the vendor's history, as that's really the only window you have onto their future.

shocking.

Actually it's really typical. There's no "Consumer's Reports" for microprocessors -- the industry press makes its money on advertisement, so you're never going to see an article about XYZ Microprocessor Company having terrible delivery, or promoting a chip only to discontinue it, or whatever. That would cut the revenue stream right off.

I generally make decisions on a per-product basis. I put _weight_ on a company that I've had success with in the past (and a different kind of weight on companies that I've had bad experiences with), but it's rarely a make-or-break thing.

Usually when I'm putting a processor on a board I look at everything that I need the board to do, then I look at how much external circuitry I can scrape off the board if I use any given processor with its peripherals. That gives me a price for the _whole board_; then I look at how much I trust the processor vendor to (a) do what they say they do and (b) still do it next year, then I make my decision.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

One indirect method is to check out the variety of dev boards at places like

and Certainly not the only qualifier but finding lots of dev boards, or finding none, could give some indication of both what's popular and what's obtainable.

--
Rich Webb     Norfolk, VA

No two vendors make a pin-for-pin compatible ARM processor.

So, how would you compare one chip to another ?

Pick what you need for a given project and evaluate the company for service and price. (ie availability and overall cost )

What else are you going to use ??

h

selling the

Atmel has more

shocking.

Have you tried talking to the distributors? My first thought on a question like this would be to ring an FAE at Arrow, EBV, Future, Silica, etc. (pick your own favourite). They should be able to tell you a bit about the pros and cons of the manufacturers, including technical issues, non-technical issues (prices, lead times), and non-public issues (future plans, reliability questions, etc.).

selling

Atmel has

like this

favourite).

manufacturers,

non-public

Yes, but I'm looking for other perspectives as well. Rich Webb's criteria of looking at the variety of development boards looks brilliant and a distributor won't tell you that. It seems to me that the more 3rd party development boards there are for a part, the more popular that part is, right?

I've had people shoo me away from NXP because they are not reliable. Linnix is right, if you're not high volume Marvel wants nothing to do with you. Freescale focuses on their own proprietary 32-bit. ST and Atmel have the lowest power. TI has Luminary's Cortex which I've been told is lower quality to everyone else.

Energy Micro has the lowest power ARM devices, as far as I know. I haven't used them myself for comparison.

Freescale are coming out with a whole new Cortex M4 family that looks very interesting.

A really great part is the Cypress PSOC5 which gives a great deal of flexibilty because of its configurabilty.

Unfortunately it appears to be made of pure unobtanium.

--------------------------------------- Posted through

Not true. I've heard it's being designed by the engineering firm of Tuttle and Dunsel. (Capt, Retired)

e and Dunsel.

Dave,

you heard strange things such as Luminary (TI) being low quality. They manufacture on one of the highest quality production lines in the world TSMC. Marvell does not design and manufacture MCUs, they do high end application processors, no flash but lots of MHz. Atmel started strong with ARM7 and ARM9 but is weak in Cortex-M3, their focus shifted very much towards AVR32. NXP offers the fastest Cortex-M3 with Flash, btw. did you know that Toshiba has the fastest M3 running from internal SDRAM? Did you know that Energy Micro achieve better power numbers using the Cortex-M3 then any other vendor even those using Cortex-M0?

PSoC5 is a great product and if your volume production does not start before 2011, you might want to order a FirstTouch for PSoC 5, just $49 free tools, several sensors for acceleration, temperature, capacitive touch and readily available. Got one on my desk, like it.

is a good place to start.

I could write a lot more about ARM / Cortex-MCUs because that's what I have been dealing with since the first ARM7 MCUs hit the market. If you need professional help, with the selection write an email to microcontroller (skip this at gmail) -dod comm It would go a long way if you would list your requirements, you get better answers.

For a list with many articles about Cortex based MCUs check out this one:

Cheers, An Schwob

selling

Atmel has

really

like this

favourite).

manufacturers,

non-public

looking at

you that.

the more

What kind of things did they say were unreliable about NXP ??

Too many Device errata or something like that ? Or something else ??

boB

Linnix is right, if

on their

Cortex

and

I never doubted TSMC. When we de-capped the Luminary products we found that they had stripped some logic to reduce die size and get their costs down. I believe that TI has since corrected the issues, but I'm not sure.

Gotcha. But I gotta disagree with you on Atmel, they will be introducing a lot of new Cortex parts in 2011

NXP's errata is unacceptable, while Atmel's SAM3U is 96MHz, Atmel's ADC is better. The Toshiba parts are nice.

Nice parts. Too early to judge the company, though. I hope they do well.

Nice parts.

2010-09-02 00:37, Dave Graffio skrev:

selling the

Atmel has more

shocking.

You will find the difference in the peripherals. Most toolchains worth their salt, support all major ARM manufacturers.

If you can afford it, you want second source of everything you do.

As for speed, most current flash based ARMs/CM3s are limited by flash waitstates and there is very little performance increase once you reach those 60-70 MHz.

While a CM3 at zero waitstates is 1.25 Dhrystone MIPS/MHz, the performance at 100 MHz with 4 waitstates is more like

0,85 Dhrystone MIPS/MHz.

Using a Keil compiler, I measured the difference between running at 84 MHz and 96 Mhz to be less than 1%. This was more than one year ago, but I doubt that this will change until people start to put faster flash memories into the products.

--
Best Regards
Ulf Samuelsson
These are my own personal opinions, which may (or may not)
be shared by my employer Atmel Nordic AB

Dunsel.

I first heard about the PSOC5 three years ago when they said it would be out in a year. That year turned into 18 months, then 2 years,... until in early 2010 they announced the PSOC5 would be shipping in Q2. That date never changed that I read. Did they ship in Q2? I can't say, but right now, when Q3 is nearly over, if you *really* want a PSOC5 badly, you can get them for $64.84 each in qty 1, or qty 100 or qty whatever by buying a CY8CKIT-010 and unsoldering the CPU chip.

So they are shipping, just not very many.

I figure they will actually be ready to ship for production about the time they are thoroughly obsolete. I still have no idea if they will be affordable for any of my designs.

Rick

Or putting instruction cache there.

...

y.

You will find the selection often is based more on the Peripherals or other features, not the core.

For example, we like 12b ADC and 5V operation (for direct PowerMOSFET drive), and this new family, adds to the list :

Their NUC1xx family development tools, and devices are showing in stock at Nu Horizons. NUTINY-SDK-120 is ~$21 A 32KF M0 device shows in stock at $1.70

Even their NUC501, which targets toys, looks useful. ARM7, 32K RAM, with USB/ADC/DAC/UART/SPI/i2c, it could become a more flexible FTDI type device.

At $1.54, that's got potential as USB debug-dongle etc, but it currently lacks a low cost eval path. Needs a USB-Stick... (unclear if the N55UL501 is the same, or a new-spin?)

-jg

any.

For a FTDI type device I find the AT90USB162 just amazing: small, cheap, easy to use, USB firmware out of the box, widely available, and is a USABLE 8-bit MCU, not simply a peripheral.

JaaC

linnix skrev:

That might work but I am not aware of this solution beeing implemented anywhere.

It was not neccessarily a good idea with the ARM7. If you added an insruction cache you added 1 waitstate to all accesses.

Good for top performance on some apps, but certainly reduced the worst case performance, which sometimes is more important.

--
Best Regards
Ulf Samuelsson
These are my own personal opinions, which may
or may not be shared by my employer Atmel Nordic AB

A better solution for micros like that is a wider flash design with an sram buffer in the flash module - that is certainly how some manufacturers handle the problem. It is a simpler solution than a full instruction cache because you have only a single "tag" (or perhaps two, if you have two such buffers), and there are no issues with coherence or anything else. The buffer of perhaps 256 bytes gets filled whenever you access a new "page" in the flash, so that the processor then reads from the buffer rather than directly from the flash. And if space/economics allow, you have have a wider flash-to-buffer bus to keep up a high bandwidth even with slow flash and a fast processor.