Creating your own dev board

Started by krk989 3 years ago8 replieslatest reply 3 years ago213 views
Hi everyone. I am an EE major working on my senior project and hope the experts here can lend some advice. My project entails the creation of a harmony effects pedal for guitar similar to the PS-6 Harmonist. We plan on using an ADAU 1446. The dev board is a whopping $700 so I wanted to see how practical it would be to have our own board made. How would one go about determining which components on the $700 board are necessary for our board and which are not? We are required to offload the ADC and DAC for complexity sake. We plan on having multiple user selections possibly handled by a FRDM-K64 that will need to be input into the 1446 as well. Any help would really be appreciated.
[ - ]
Reply by drmikeNovember 3, 2020

The first thing you do is get the data sheet and look at the schematics: https://www.analog.com/media/en/technical-document...

And then you cut out of the schematic everything you don't need.  And then you layout the board, send it to fab, get it back, and find out you forgot something and have to it over.  Probably 2 more times.  By the time you get it built and working, 2 semesters have gone by.  Do you have that much time?

One of the things you need to think about as an engineer is how much your time is actually worth.  What do you expect your salary to be when you graduate?  How many hours of work will it take you to build this custom board and how much money is that in terms of your salary?  That $700 isn't that big a deal in comparison.

One way to figure out how much time it takes to do something is to make a guess.  Then double the number and go to the next unit of measure!  If you think something takes 20 minutes - it's going to actually take 40 hours!  This is partly a joke, and partly experience.  A lot of things go wrong that you weren't expecting, and it takes time and effort to deal with it.

Maybe you can't afford that chip - that's an engineering decision too.  To quote one of the ancients from comp.dsp "Engineering is getting what you want from what you have". 

You can have a lot of fun with this project, but I think you want to eat and sleep too.

Patience, persistence, truth,

Dr. mike

[ - ]
Reply by ombzNovember 3, 2020

That's beautifully said. So true.

[ - ]
Reply by jbrowerNovember 3, 2020

Dr. Mike-

I'm one of those ancients, but not my quote :-)

I suggest you expand your reply into a blog. Really valuable words of advice.


[ - ]
Reply by ombzNovember 3, 2020

On what basis did you choose those chips and platforms?

The PS-6 guitar effect pedal seems to sell for roughly 150$. And it's produced for sure for a way much lower cost.

Why don't you try to find a way cheaper dev platform that has everything that you need? So you can first think of what main components you'd really need to build an effects pedal like that. Probably some kind of simple DSP or another cheap processor. But do you really need a graphical programming tool like what's provided with the ADAU? Why not try to find a low cost DSP that you truly need to learn to master (i.e. learn how to operate the development environment, writing code for it, making its peripherals and I/O's work, coding the effects which is truly the fun part etc.) When it's a low-cost processor it probably doesn't come with a floating-point capable ALU. So you might learn how simulated floating-point degrades latency a lot, it might be sufficient for your needs, or you would need to learn to write more and floating-point operations with integer arithmetics (fixed-point). This is a beautiful skill to master and very handy in the industry.

Then you need an ADC and DAC as you've figured out already. Just find and choose a platform that features all of it and that is (much) cheaper. I know these exist. And don't forget about the jack sockets. ;)

A part of engineering is to find and evaluate your options at hand. When you do it right, it can be very rewarding. So you should not rush this step.

Also ask yourself what kind of skills you wish to develop during this project: do you want to spend time making hardware (drawing schematics, layout the PCB, sending it to the assembly, debug it, do it over again)? With the downside of probably not finding enough time to actually program it or even worse: it's still not functional at the end of the semester. Or do you want to spend time to learn how to evalute and buy basically functional hardware and how to modify and program it the way you want it to be? With the downside that you don't get involved in electronics design too much.

You choose.


[ - ]
Reply by CrandlesNovember 3, 2020

I'm not sure of the cost but designing and fabricating your own development board would likely be enough of a project in and of itself. Have you considered something like the BeagleBoard/Bela combination? There seem to be a lot more open source designs for this type of stuff these days. This might also fit your bill: https://www.rebeltech.org/product/owl-pedal/

[ - ]
Reply by DaniloDaraNovember 3, 2020

Cool thing this Owl pedal!

[ - ]
Reply by DaniloDaraNovember 3, 2020

My humble opinion - the toolchain and libraries will make the difference.
If you opt for setting up your own board you are nowhere without a compiler, a smart debugger, and a rich library, which could change your day.
The toolchain that comes with a Dev board is perfectly tailored to the hardware you have.

[ - ]
Reply by lamabrewNovember 3, 2020

I'll second all of the excellent comments that have been made, and I'll throw out a couple of suggestions too.

I think for someone in your position is the journey is as important as the result. As a EE you will have many paths in front of you for software, hardware, and system design.  Exposing yourself to all of those early on is a great thing to do, and this project is like a Mini-me of things you'll do after graduation.

While I agree that dev boards cost are not a decision factor in most projects that are destined for production, it's a reasonable example of having to think about cost and look at alternatives and make tradeoffs.  

If you're mostly interested in the hardware design something from the SigmaDSP family is great because the SigmaStudio tools make it very easy to develop software.  There's other chips in the family and you should look at the possibilities versus your programming needs and cost targets.

If you are going to design and build/assemble your own board as part of what you need to do for the project then look for lower pin count chips to keep assembly simple. SMT assembly isn't hard once you know the tricks; keep it easy to start with (TQFP over LQFN, nothing smaller than 0805s or go through hole for as much as you can, etc).

If you want to learn more about embedded software and specifically embedded DSP software and have a well supported platform I encourage looking at the AMR M7 devices; there are ones out there that are about as fast for DSP as the SigmaDSP parts and have the range of I/O options you need.

If you want to focus more on software there are plenty of <$50 ARM boards you could use.  As a EE you become good at googling...

Good luck!