Testing 2.4GHz Bluetooth RF?

[bronson]

Unfortunately, it looks like I have a Bluetooth design that could be prepared for production in the next few months...  Never done anything with antennas or intentionally radiated anything.  I have some learning and preparation to do.

For the prototype, I just banged together the patch antenna based on recommendations from a TI app note and taking coarse measurements from other bluetooth devices.  It seems to work.   

I suppose I should actually tune the antenna on its production PCB, so I should find a 2.4 GHz VNA?  Or maybe a 2.4 GHz spectrum analyzer and signal generator would be enough?

If my employer pays, I figure the Rigol DSA832-TG would be the low-end of what I would need.  If I'm paying for it, I might try to make old equipment work, maybe a 8594E and 8616A/8340/8341.  I avoid renting because I always need a device a week after returning it but, if the alternative is having $25k of RF equipment lying around, then no problem!

I can ensure the rest of the board is quiet.  Anything else I need to worry about?  Random thoughts from anyone who's been here before?  Thank you!

[Dago]

I think you might have jumped off the deep end...

If you're doing this so that you need to get the design certified (which you have to do to be able to legally transmit with it without being a certified ham radio operator) you basically need to do the measurements way higher than the fundamental. In the minimum you need to see at LEAST the second harmonic (4.8 GHz), preferably up to the fifth or something to have any idea what the thing is transmitting. Measuring the radiated emissions quantifiably without calibrated equipment is practically impossible. These measurements are usually done at the certification place with the same equipment the actual certification measurements are done as well.

It is quite difficult and requires careful planning to get your design certified, even with a team of people of whom part are RF engineers and someone specialized in the certification process. A single day of certification measurements will cost you several thousand. Without an understanding of the requirements and the certification process you WILL fail it.

For an FCC+CE+bluetooth certification set you're looking at around 40000€. If you use a premade bluetooth stack then maybe 20k.

If you just want to get your antenna tuned then that Rigol with a VSWR-bridge should do the job.

Source: my friend designs bluetooth modules for a living.

[bronson]

Thanks Dago.  Yes, I'm definitely in over my head.  I'm using the CSR bluetooth stack so that should help a little.

Pre-cert does seem to reduce lab time.  Even if I don't get any good readings at all, scary blips encourage me to notice things like loopy clock traces and potential issues that should probably be fixed, whether or not they are actually causing a problem.

So, I guess I'll tune the antenna, attempt to silence the rest of the board like usual, and hope for the best.  Sounds like there's not much else that I can do on my own.

[Teledog]

Whatever SA you get, make sure it has a tracking generator - for tuning your antenna with a directional coupler (AKA: low buck VNA)
Try and scrounge a small TEM cell for doing "in house" tests prior to any certification, if you so choose.
Find an existing well designed board/antenna & clone it  (like any company doesn't do that now!)
Be careful with casing materials, some of them - even benign looking plastics can really screw up your SWR!
G' luck!

[bronson]

Thanks Teledog, that sounds like excellent advice.

[Dago]

Also design for an RFI shield. You will fail certification for the first and maybe second harmonics without an RFI shield.