MENU

NXP focuses Cortex-M0 32-bit microcontrollers on the 8-bit space

Interviews |
By eeNews Europe


Last month NXP revealed the company is aiming to open the door for a broader range of applications previously closed to typical 32-bit MCUs due to package footprint or manufacturing constraints.  The new LPC111x devices, which feature SO20, TSSOP20, TSSOP28 and DIP28 options, claim to be the world’s first 32-bit ARM microcontrollers in low-pin-count packages.     

Starting at $0.49, NXP’s low-pin-count devices claim to deliver 50 MIPS of performance compared to the 1 to 5 MIPS performance typical of 8/16-bit MCUs, at a highly competitive price point enabled by NXP’s exceptional capacity in manufacturing high-volume commodity packages.     

Target applications include human interface devices (HID), consumer electronics, alarm systems, small appliances and simple motor control, among many others.     

eeNews Europe:  What is the significance of producing the ARM Cortex-M0 LPC1100 family of microcontrollers in low-pin-count package options?  

Bezemer: With this introduction we are showing a commitment to drive Cortex-M0 32-bit microcontrollers straight into the 8-bit space. We have been looking to take away the boundaries one by one of traditional 8-bit users to enable them to adopt and embrace 32-bit Cortex solutions. 

These boundaries really lie in three areas.  One is cost, one is power and one is simplicity.  With this announcement we have made tremendous progress on the cost side.  We already had the world’s lowest cost 32-bit microcontroller at 65 cents.  But now by introducing SO,TSSOP and DIP packages we can tap into the company’s biggest capacity bracket for high volume, low-pin-count commodity packaging.  That has helped us to further reduce our price points of our 32-bit Cortex-M0 microcontrollers in the market.

Power consumption is very important.  We have been able to achieve over the past few years tremendous improvements in power consumption.  Our current Cortex-M0 microcontrollers are best-in-class when it comes to dynamic power.    Current products run at 130 μA/MHz which really makes them suitable for low-power applications and that is typically what 8- to 16-bit users expect.

The third barrier, which is the one of simplicity, is the one that is most difficult to bring down.  The 8/16-bit users have been familiar with those architectures for long time and have become very loyal customers.  Cortex-M0 is the simplest ARM core there is and, by design, simplicity was put as the top priority on the list when it was developed.  It only has 56 instructions which makes it very comparable to the number of instructions of traditional 8- and 16-bit architectures and makes it much simpler than some of the 32-bit architectures out there. 

Providing design tools has also been a major point that drives simplicity.  It helps developers to work with these microcontrollers and we have introduced LPCXpresso, which is a new development toolchain to help support this family.  We have made that tool chain very low cost and easy to use.

Making another step towards simplicity are the form factor of the SO,TSSOP and DIP packages.  They allow for much easier handling, single layer low-cost boards and hand soldering.  They are also easier for assembly, prototyping and manufacturing.

eeNews Europe:  What are the key benefits of the new low-pin-count package options?

Bezemer: With the world’s smallest 32-bit MCU, the LPC1102, available in a 2-mm x 2-mm Chip-Scale Package (CSP), NXP now claims to offer the widest selection of package options for Cortex-M0 MCUs.

The introduction of the new low-pin-count package options provide reduced footprint and system-cost benefits to customers throughout the product development cycle.  SO and DIP packages provide ease of customer prototyping with the ability to hand-solder, simplifying hardware requirements for programming and debugging. 

The TSSOP packages eliminate potential reflow process in high-volume production. These easy-to-use and highly reliable packages are popular among 8/16-bit customers and help minimize the number of manufacturing processes while improving yield to further reduce overall system costs. Existing LPC1100 customers can easily convert their designs to the LPC111x low-pin-count devices and reuse their software due to the identical Cortex-M0 instruction set.  The low-pin-count packages are designed for easy PCB layout and scalability by sharing the same pin-out for VDD, VSS, GND, and XTAL.

The LPC1100 series are capable of executing sophisticated algorithms at low power, meeting the ever-increasing demands of cost-sensitive applications that 8-bit microcontrollers struggle to achieve, such as interfacing with sensors and performing complex control tasks. For example, a 16-bit multiply operation performed by an 8-bit microcontroller requires 48 clock cycles at over 770 µA/MHz, while an LPC1100 device can complete the same task in 1 cycle at 130 µA/MHz.

eeNews Europe:  Are there different design challenges facing European design engineers?  

Bezemer: I don’t rate the level of skill differences between the European and US designers as super high.  But I think when you consider the type of applications they work on then there is definitely a difference.  The European market is much more dominated by industrial type of applications.  I believe the way European and US engineers think and the needs of the engineers together with the way we communicate with those engineers is very similar. 

The differences are less relevant than if I compare North America and Europe with the Asian engineer working in say, China, Taiwan or the South Asian Pacific Rim.   There I do see a difference.  I think it is also related to applications they play in.  I see the Asian customers being more opportunistic, faster in adopting new ideas.  Some of the North American or European applications have a tighter regulatory focus.  The regulations that the North American and European governments set are more focused on making sure the qualifications are done.  On the other hand in Asia people are now building more and more for the European and North American markets and need to address a lot of these specifications.

eeNews Europe:  Is there some aspect of the technology that is attractive to European design engineers in particular?   

Bezemer: I think there are three things in our technology that are especially attractive to European engineers.  One of these is low power which has major relevance.  We see it all over.  We see it in applications that really mandate it such as battery driven applications like glucose meters or sensors for security systems or gas meters.  In these we see low power is critical and this links very well to our ARM Cortex-M0 low power technology.    

Secondly, packaging is absolutely of interest to European engineers.  We have seen with the wafer level chip scale technology that we introduced 18 months or so ago which was very new and coming from the mobile space we could make super small 32-bit parts.   They were 2 x 2 mm parts and we have seen more and more interest in these because people take their dimensions down not only for mobile related devices but also for smart sensors or modules.  We see size being more and more an interesting parameter.      

Then thirdly, in general we see a trend for a need for more performance which is being driven for two reasons.      

One is communication and the other is user-machine interfacing.  In our parts you will see the need for more performance to do more and to communicate better whether that is linking with USB, Ethernet or wireless technologies.  We are seeing more and more customers who want to make interesting user-machine interfaces.  We have the LPCXpresso technology in our microcontrollers but work is needed to make those screens, to make menus, to make buttons to make fade-in, fade-out features.  Everybody wants their application to look like and iPhone. Having that best-in-class graphics library now available for customers has been a big thing and has been very well received in Europe and we have developed a very good partnership with SEGA.

eeNews Europe:  Are there any power management benefits?   

Bezemer: Looking at Europe specifically and focusing on applications relating to motor control and applications relating to solar, which is rapidly growing I think that both are related in a way to power management. Having very flexible timers or PWMs and having route analog sensing is extremely important so we are really focusing on that. The new products offer timers with PWM generation – For each timer, up to four match registers can be configured as PWM, each timer supports up to three match outputs as single edge controlled PWM outputs.   

In an earlier announcement we talked about a new NXP innovation called State Configurable Timers where we recognized that if you look across all the applications that manage and control power.   

Whether it is a motor control or whether it is in a solar device, the needs for those timers are a bit different every time and engineers can be frustrated that in whatever product they are working on it solves one problem but it is not the right timer or PWM for the other one.  So what we decided to do is to develop a very flexible timer or PWM and we called it a State Configurable Timer which combines a normal timer function with a state machine and that allows engineers to make it flexible and program it to their needs.  This introduced the concept of events, of states and inputs and outputs to make their timer event dependent.  That allows them to fit the timer to meet the needs they have for that particular application.  That approach has resonated very well.  The beauty of that approach is that it means there is no CPU involvement.  The timer acts as a stand-alone unit and does not impact the performance of the controller.  It is very deterministic and very fast and that is what you want to have in these critical applications that require a motor controller.  Solar has a high importance in Europe.     

In the industrial side I think there is a lot of innovation going on in Europe.

eeNews Europe:  How important is code density?  

Bezemer: Is code density important?  Hell yeah, very important.  But why is that?  In a way it is cost because the smaller your code is the smaller microcontroller memory size you can buy and generally that also means a lower price.  So code density is very relevant.  We have had a lot of success with that because we have been able to prove to our customers that coming from traditional 8-bit and 16-bit architectures you can have a major code size advantage in moving to Cortex M0 that can be as much as a 50 percent code size saving.  Smaller code does improve the internal cost structure.  Smaller code means a possibility to adopt cheaper components.

eeNews Europe:  How important is a clear migration path?  

Bezemer: A beautiful attribute of the total ARM continuum that we offer and the total ARM architecture in general is that we have lined up Cortex M0, Cortex M3 and Cortex M4 products as one continuum. 

We offer customers pin and software compatible migration between those cores.  With that we have a very simple message to customers.  We have 100% ARM cores ready.  We do not care about boundaries between 8-bit, 16-bit, 32-bit and DSP.  We believe that a customer has a problem and wants a solution to that problem at a certain price point. And it does not matter whether the architecture that drives that is an 8-bit, a 16-bit or a 32-bit architecture.  We explain that it is possible to solve those problems with ARM-based microcontrollers that we develop. 

If you have a very simple problem such as a problem that does not have a lot of budget to be allocated to find a solution then a Cortex M0 entry level solution can be the right answer.  But from there on we often see that customers have an opportunity to develop more complex products and have a need for more features and a need for getting more ideas into their end products and they might want to migrate up into a Cortex M3 which provides a hardware design that provides more instructions and which can do more and usually operates at a higher processing frequency or even to Cortex M4 that adds a floating point unit and single cycle map instructions to the game. 

A single architecture may fit all but one core structure does not fit all.   With a line-up of M0, M3 and M4 solutions we offer the customer the possibility to re-use their code. To only have one toolchain and really migrate across that and really takes away the need of differentiating between 8-bit, 16-bit and 32-bit.  That concept has really picked up and been growing.  I think that idea explains part of the success that ARM and all its partners are having today.  Aside from the new announcement which is specifically focused at driving M0 further down into 8-bit space and effectively broadening the continuum further. 

The fact that the continuum is there and that people can easily migrate up is very important because many of our customers don’t just make one product.  They make a couple of them.  Sometimes they make a system where they have a sensor that requires a very simple microcontroller but in the same system they then may develop a user-machine interface or a communications hub that requires a bit more performance.  So it is very convenient for them if they can use two products that are offering different performance levels but originate from the same ARM architecture and where they can really re-use their software and re-use their tools.  It reduces their risk and reduces their time-to-market.

eeNews Europe:  What other obstacles will need to be overcome before the market switches over from 8-bit to 32-bit solutions?   

Bezemer: Of course every day I would like to believe that the old boundaries are gone now.  I am sure we will find new ones and we will work to break them down every day.     

I think with this release we have already set an expectation for costs to come down further.   We feel confident that within a year from now we can offer a 40 cent 32-bit microcontrollers which will help further on the price point side.      

The other thing which I think is very relevant and which will dictate the pace of the 8/16-bit migration relates to code conversion.  Engineers that come from school today are educated in working with C or C++ languages.  They will find it very easy to adopt these microcontrollers.  They will feel very comfortable working with compilers and the way our Cortex M0 or Cortex M3 and Cortex M4 products work.  Many companies we see are making the switch for flexibility reasons to C and C++ to write their code but there are still a lot of companies that have a lot of legacy code which is probably written in Assembly and often they are very scared to touch that because in many cases the people that originally wrote that code might no longer with the company or might even have retired.      

The barriers to the adoption rate of 8-/16-bit to 32-bit are not only the price, the power and the simplicity it is also the moment when people will be able to decide when I can move my code.  Taking that Assembly code which is proven and working but will not drive any further innovation is a risk.  For Europe I think it is important for us to drive innovation. The companies that I am working with recognize this and they realize they cannot just continue to just re-use the old code they have developed in Assembly.  They need to move on and everybody agrees with that but some companies say I will move today, some companies say I will move tomorrow while some companies say they will move the day after.  But everybody is in agreement they will move.  But this is a big trend with code conversion moving away from Assembly and going to C or C++.       

I think all the new projects that engineers are working on are in C or C++ but sometimes the company’s secret source is in Assembly and sometimes, although people might not like to admit it, they are very scared to touch that because it works now and you never know if once you start moving it around whether you may start breaking something.   But sometimes I think there is a need to start building something new.  I think the move from 8-/16-bit to 32-bit is going on and going very fast.  But on the other hand I think this will make it go not with a bang but with a whimper because it takes time.  People will need to make a decision and say: “Am I going to touch this code now to really innovate for our next platform?” Or are they going to copy the code just one more time just so it can to be done faster and reduce some risk?      

I think that is one of the situations we are seeing now and we support our customers as much as we can with the process.  Companies need to step away from their legacy code and step into another era that offers the opportunity to do more.  That is really the premise of the new Cortex M0 line that we are introducing. It is an opportunity to do more but still at 8-/16-bit cost levels.

World’s first low-pin-count 32-bit ARM MCUs in high-volume TSSOP and SO packages target 8/16-bit applications

More information about the LPC1100 family of microcontrollers at www.nxp.com/products/microcontrollers/cortex_m0/lpc1100l/

For further information on LPCXpresso and other third-party development tools, see www.nxp.com/lpcxpresso


Share:

Linked Articles
eeNews Europe
10s