Architecture for Real Time operation
For all electronic systems, and especially for Real Time Systems, the architecture of the hardware can make it simple, difficult or even impossible to achieve the operations that you need.
For many years I have been trying to get specifiers to consider the architecture of possible solutions at the beginning of a project. More often hardware is chosen at the beginning of a project and it is left for the programmers to fight with that architecture throughout the project development.
Starting with the right architecture for the job can bring you many benefits :-
- It can be the difference between the project being possible or impossible
- It will affect the software/firmware development and debug time
- It can affect the performance and reliability of your end project
As you can see from my background I was involved in defining the TIM-40 specification with Texas Instruments. Sadly the group involved were not prepared to go beyond the specification of a module, and as a result different companies adopted different architectures, and ultimately fell into the same pitfalls, and came up with similar solutions. Luckily for us, I was already working on the architecture of our products as the module specification was being completed, so we had a head start on the competition.
Emboldened by the TIM-40 successes, and equally frustrated by being unable to offer some of the features that my customers were asking for, I started to define and work on a new architecture. When Texas Instruments announced the C6000 DSP family, they didn't have any on-chip interfaces that were defined for real time connection between the processors and between processors and other parts of the system. I saw this as an opportunity to offer our own architecture that had been designed specifically for real time systems. I wrote a paper about the reasons for the new architecture and jointly with our US distributor at the time presented it at several DSP conferences. The resulting HEART architecture was the basis for our HERON products where we successfully combined FPGAs, DSPs and high speed I/O devices into Real Time systems that have been programmed by HUNT ENGINEERING customers all around the world to implement a huge variety of systems.
But HERON and HEART reached their technical limits and I worked on a completely new architecture that better used the features of modern FPGA and DSP devices. The HAWK system architecture never saw daylight outside our lab, but the results we got in the lab were really astounding.
With the amount of time and experience I have in studying the architecture of Real Time systems, I am perfectly placed to
- Help you define what your system needs from a hardware architecture
- Help you to choose between FPGA and DSP or how to split your system across both
- Research, study and assess different off the shelf architectures from the perspective of your project
- If necessary define and develop an architecture specifically to meet your needs
As with everything I can offer, I can work on this alone, generating and/or presenting a study report, or I can work with your team in an advisory or mentoring capacity.