It is very rare these days that I get permission to talk about what I'm doing, but today I did! Somos Internet has been designing and building their own MT79 based internet router and bridge for the past few months. They engaged me to take a look at the reference board, and we were really pleased to see how well all of OpenWrt, not just the make-wifi-fast fq_codel bits, was working nowadays.
That was last month. This month, they got back the new boards with a new antenna layout, and the 5ghz performance at range is terrible. I confess that I no longer have even the slightest grip on how to go about debugging and improving antenna performance - are there beamforming variables worth sweeping in the mt76 driver? A piece of test gear we could use to see what is going wrong? A design tool for antennas? A debugging strategy for antenna placement?
This is what the somos Orb looks like. The new on-board antennas were terrible at range - the additional antennas soldiered on for test not much better, neither approach is as good as the reference board.
To be honest, this is not the place you're likely to get the help you need, nor will you be able to optimize this yourself without considerable education and equipment expense.
Typically, you need to have a background in antenna theory and design. While a degree in the field isn't "required" (there are many very smart, self-taught people out there), this type of design work is usually the domain of people who have a masters degree or higher in Electrical Engineering (with a specialization in RF and antenna design) as this is an extremely complex field of engineering. Professional grade antenna designers will have test equipment and faraday-cage environments that may cost well north of $100,000 USD to deploy, and require extensive experience to operate said equipment.
It probably makes sense for you to pay a professional RF designer for their services to help you properly test and optimize the design, especially given that this is a commercial product development activity. This will allow you to get rapid and dedicated help from an expert rather than relying on volunteers who may or many not be able to dedicate their time and energy to your task (and that assumes that the RF experts who are in this forum would offer that advice in the first place). Further, when you hire an RF expert, you can specify in the contract the requirements and deliverables and you can hold them accountable.
I would higly recommend that you seek paid expert help for your product.
In general, I agree with you on finding far more (and paid) expert help. Can you make a recomendation of a firm or individual that has experience with the mt76 and mt79?
I don't have any recommendations to provide. However, it is possible that a user here may be willing to help you with your project as a paid expert. If that possibility arises, I would encourage you to have that conversation privately via PMs to agree on the details of such a professional engagement.
Otherwise, I would recommend that you begin leveraging professional employment networks (such as linkedin, talent acquisition/recruiter teams, other personal/professional connections you may have, and/or temp/contracting agencies that have appropriate talent pools to offer this expert area.
Here in FCC land, altering the original antennas is a no-no
You weren't provided this by the manufacturer?
I believe there's a thread discussion on white-label device manufacturers. I'll attempt to search, but the conversation diminished when the poster didn't want to consider the quantities required for such a purchase.
Re: FCC-land - the product is not manufactured in or intended for the USA market (presently). It is very early part in the design stage, I showed the very first boards back from the designer. There will be many more spins before it becomes a product that will need to go through the certification processes.
