Core/Photon keepout zone

Hey,

I was just taking a look at the latest datasheet for the Photon. I was quite surprised when I saw there is now a specified keepout on my pcb for ground planes and signals.

Obviously I bring this up as I am concerned. My pcb is already designed and fabbed Has anyone at particle done anything to characterize performance with a ground plane and signals in this area. I need to understand the likely degradation in signal strength.

That’s a note to recommend keep-out and may not need to be the exact size of what’s shown.

However it is recommended not to have a plane in the area from the D0/A0 pin to the end of the board.

I’m not sure about the surprise this that was recommended since the Core :smile: If you are not soldering the Photon on directly it is fine but not sure how much performance degrading will be if that’s the scenario you are in…

@kennethlimcp

Well my pcb was designed for the Core and the Photon was expected to be a drop in replacement. So the next question is when was this keepout area specified for the Core? Do you have any idea?

@kennethlimcp,

I just looked at the Core datasheet I cannot find any specified keep out area?

I haven’t checked if the eagle footprint if it has the notes but it should be there. Ping @mohit

Keepout areas near antennas are standard. I think there may have been anecdotal design rules to implement a keep-out area for both the Core and the Photon, and this revision to the docs formalizes that for the photon.

If you are mounting the Photon to your PCB via the pins, especially via a socket on your board, then you probably do not need to force a re-spin, but if you have an opportunity, you should factor this into your next revision.

If you are mounting the Photon directly onto your PCB using the castellated pads, then I recommend you plan to revise your board, because the traces/backplane on your board is that much closer to the antenna.

The possible effects on performance when you violate the keepout area include: degraded wifi range, and unpredictable radiation patterns (e.g. unit may be more sensitive to orientation wrt wifi range)

1 Like

@AndyW,

Certainly I understand the logic. It really comes done to the specifics. I am using the header style photons, I had assumed the .1 or so spacing off the pcb that the headers provide was adequate as I had not previously seen any instructions to not place copper in that zone.

I am wondering if particle knows that copper in this area creates a problem when the photon is placed such that the header plastic strip touches the pcb or is it just a matter of particle erring on the side of precaution and there is no testing that has been done to prove disprove this part of the specification?

It is impossible to test every combination, but the particle engineers designed the keepout area on the modules according to the design rules for the chip antenna used, measured the results and iterated until the performance was optimal.

This guidance reflects extending the keepout area in order to maintain that performance.

Naturally, if you use an off-module antenna, then there is no need for a keep-out area of any sort.

@AndyW,

Well I was not hoping for anything like all the possibilities just the one highly likely possibility which is with a ground plane on the pcb when headers are used.

What I getting from your comments is that this has probably not been tested but someone thought “Hey let’s tell customers not to put copper there. It might not help them but it can’t hurt if we tell them that.”

@HardWater, that's sounds a little harsh. As @AndyW mentioned, there are design rules to follow and what the Particle team did is to bring this information to people using the footprint.

We can definitely do tests to under how much the signal degrades but the general rule for antenna is to have a keep out.

I certainly don't think that

this applies and i'm sure there's thoughts put into this during the creation of footprint.

If you look at the Photon closely, there's also a keepout area around the chip antenna. This is definitely be design and testing with the Antenna manufacturer/testing facility and not something random.

That’s not the process, and it was not my intention to make it appear that way.

The photon design has keepout areas in the 3 layers other than the top layer, this documentation simply expands that to the design guidelines when integrating the photon. The vertical spacing does matter, so a solid groundplane would behave differently depending on factors like the clearance between the photon and the board (was it socketed ? what kind of socket ?) and the impedance of the ground @ 2.4GHz.

I am sorry that the docs were not updated when you designed your board. Your board will probably work well since you are not directly surface mounting the photon to your board, even with copper in that area.

1 Like

Hey @HardWater I'm sorry to hear you did not see this piece of information sooner! It has actually been in the Photon Datasheet since the first release in May, and not something we recently changed. This info never made it into the older Core datasheets, but has been recommended to anyone that posts their PCBs for review here on the forums. These are recommended best practices and you can choose to error more or less on the conservative side. The chip antenna on the Photon has a recommended ground plane keep out specified by the antenna MFG, and if you look at how antennas radiate, the keep out of the PCB should be a bit larger to cope with the height of the headers. If the Photon was a headerless version, it could probably stand to have a smaller keep out than the recommended one in the red dashed lines.

Probably beyond late now, but I'll create a docs issue to add mention of the keep out area for the Core datasheets as well. We should simply be able to link to the new photon (with headers) land pattern with a note about the keep out.

As Andy says with a Photon up on headers, your ground plane won't affect the performance too much... unless you are trying to reach a WAP on the other side of that ground plane.

I'm going to rename this topic to Core/Photon keepout zone to help people searching for this thing in the forums as well.

EDIT: Github issue added: Core Datasheet & How to Build a Product are missing antenna keep out zone definition · Issue #103 · particle-iot/docs · GitHub

This is a pretty standard zone for a tiny-wifi device. I asked specifically about the core keepout about this time a year ago:

As mentioned, its basically impossible to test all the configs that people might put a RF device in, so keepouts are a recommendation vs. spec. It is pretty easy to come up with configurations where a design obeys the keepout completely and get little or no signal from the core/particle. These diagrams ought to be thought about as suggested boundaries on probable designs, your mileage my vary, especially with GHz signals.