Do RF modules eliminate the need for RF Engineers?

The truth is that an RF module cannot replace the know how, experience, and instinct of a seasoned RF engineer. I'm not an RF engineer myself but I have a lot of respect for them. RF engineering is one of the most complex specializations in engineering, at high frequencies every detail is important. Simple Capacitors and Inductors become complex impedance networks, heck the ordinary transformer begins to behave as a third or fourth order bandpass filter whose influence on the design must be taken into account.

I like modules because they are easy to experiment with and are good for setting up a basic link, but I would never be so audacious as to say that thanks to little Zigbee modules we no longer need RF engineers.

I don't feel I have moral authority to post anything on RF(not my specialization, and frankly a weakness in my knowledge base) but I can appreciate how hard it must be to design an RF device. Any manager (or other higher up) who fires his RF engineer, because he has discovered X brand RF module, is ignorant and will pay a very high price for it.

Those are my 2cents.

Regards,

Jorge Garcia

Cadsoft Computer

P.S. If anyone can recommend some reading material on basic RF, I appreciate any suggestions, just so I can learn a little more on this subject and hopefully have something to contribute.

This covers a lot!

Yes the current batch of modules save a lot of work, but only in the sense that previous generations of RF modules did the same. They still, on the whole, need to be integrated into a product and have antennas attached etc and they will still interact with the rest of the circuitry in a wanted or unwanted manner. EMC can be a factor in the box as well as out of it. This is going back to the days of Wood & Douglas and even Microwave Modules.

You can look at the design problems with such as the iPhone 4G and see that the rest of the product can go wrong. There is still a lot of work needs doing in matching the module to the real world, so to speak, as most aren't completely self contained (ie no antenna on board).

Hmm ... it's an interesting question. Speaking as an RF engineer who works for a manufacturer or modules, I should say 'no: our modules solve all the issues'.  But that's rubbish. A module is a good choice for small to medium volumes and short design cycles, but it isn't going to compete with an in-house design (for cost, or functions) if you have the resources. Also, ignorance of basic RF techniques will make successful implementation of even a module based design questionable, and a good knowledge of RF design is essential to understand what is wrong (in specification terms) with so many of the  'single integrated circuit' radios that companies are trying to use.  I'd like to see a lot more "RF trained" engineers out there, for all our sakes.

I am an RF engineer and developer of modules.  I worked first at Linx Technologies and then founded Radiotronix.  Since I no longer do RF consulting work and have no financial motivation, I think I can give an honest answer to the question.

Firstly, in many cases a module can replace the need to have an RF guy on the team for the purposes of getting the product to function.  Keep the antenna trace short and follow a few basic design principles with the power supply and grounding and you will be fine.  This is true for modules sold by Radiotronix, Radiometrix, Aerocomm, Maxstream, etc.

That said, unless the module has a modular license (it is already approved by the FCC), your product will need intentional radio certification (for any product with a transmitter) and unintentional radiator Class B certification (for any product without a transmitter.  That is not always true, but it is most of the time.  For certification, you should definitely bring in an RF guy.  He needs to evaluate your product through a process we call pre-certification to determine if there are any radiation problems that need to be fixed.

Lastly, if you use some cheap SAW transmitter the size of a thumbnail, I can tell you that its tolerance to an antenna with a VSWR >1.1 is not good.  And, it needs an external LPF to mask harmonics.  These things require a good RF guy.

The point is that is probably good to have a RF consultant in your pocket regardless of your approach, but using a pre-approved module will reduce or eliminate how much you need him.

In the end, it is a trade.  You can use a pre-approved module and have quicker and cheaper development, but at a significant per unit cost increase for your product.  Or you can roll-your-own or use a cheap module, which will make development take longer and will increase the cost because of the RF guy, but your per unit cost will be a lot cheaper in production.

I can tell you this, as a matter of observation.  The RF chip designers are getting so good at building stable RF chips, it is now possible for a lot of sharp non-RF guys to do a roll-your-own design without a module or the help of an RF guy.  I have seen that a lot.  It sucked because it took business away from my module company and my private consulting business, but it is the truth non-the-less.  The ultimate goal of the RFIC companies is to eliminate the need for either of us.  In fact, look at how many of them are now offering their RFICs as modules.  They want to remove the RF engineer and the third part module OEM from the equation and totally control that design socket. I have been a preferred partner of Xemics, then Semtech, and Renesas in the past.  But I can tell you I have given more than I gained in those situations. This is the primary reason I got out of both businesses.

What we have to do now is use our knowledge to add value in other ways if we want to prosper.  That is my goal for the time anyways.

Hope that helps.

Regards,

Steve The RF Guy

God Bless