Author Topic: Antenna's and Host Boards  (Read 2397 times)

DRCO

  • Newbie
  • *
  • Posts: 17
Antenna's and Host Boards
« on: January 21, 2016, 08:50:01 PM »
Hi Forum,
Iím looking to use the Moteino piggybacked onto a custom PCB and am looking at options for the antenna. Iíve found the dipole, standard cut monofilament wire supplied with the units probably gives me as good a range as anything, but I need to encase and waterproof the unit, so Iím going to need an antenna. If I use something like a rubber duck, screwed onto the case I need a tail that goes to the Moteino antenna solder through hole. I was thinking of using a uFL connector on the host PCB (something like this) and connecting it to the Moteino using a shielded RF wire at a through hole immediately next to the uFL connector. The alternative would be to have a throughole in the host PCB that aligns with the Moteino Antenna thoughole and have the Ufl connector as close to that as possible.
Please see the two attached images, thatíll probably better explain what Iím asking.
 If anyone can offer some advice on what solution will get better range, Iíd be grateful
Regards,
Chris

emjay

  • Full Member
  • ***
  • Posts: 119
  • Country: nl
Re: Antenna's and Host Boards
« Reply #1 on: January 22, 2016, 03:31:24 AM »
@DRCO,

Neither!  I know it sounds messy, but the pigtail center core should be soldered directly to the ANT pin on the radio module with the braid to the closest convenient ground plane pad.

Why?  It's all about the frequency you are operating at - the radio module has a cluster of L & C parts between the chip and ANT pad. A key function is to match the chip input/output to 50Ω at the output pad - just what you need for your coax cable pigtails.  Your layouts introduce an extra path before you reach the uFL connector (also carefully machined to have a 50Ω natural impedance). It becomes a partial antenna when the length is a significant part of the wavelength. Then the impedance presented to the ANT pad is no longer 50Ω.

This mismatch then reflects a portion of the signal backwards from both the ANT pad and the uFL connector, messing up what looks like a clean path to the physical antenna. This maps to a significant loss of signal (and extra stress on the chip since that reflected power warms up the output stage(s))

If you want a deeper analysis, you have to dive into SWR (standing wave ratio) etc. - hence some simplifications. The Tx as the example is easier to explain - the same loss of performance is true in the Rx case due to reciprocity.


Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6377
  • Country: us
    • LowPowerLab
Re: Antenna's and Host Boards
« Reply #2 on: January 22, 2016, 08:15:56 AM »
emjay, nicely put, great to see we have some RF masters around :)

TomWS

  • Hero Member
  • *****
  • Posts: 1925
Re: Antenna's and Host Boards
« Reply #3 on: January 22, 2016, 06:40:39 PM »
I agree with emjay here, but I believe Felix has shown a means to solder the u.fl connector directly to a moteino board with virtually no path length between the antenna pad and the u.fl connector (you do have to scrape away some solder mask to get a ground) and soldering u.fl connectors, especially where the pads aren't perfectly aligned with the connector, is a bear.

I like that Moteino Mega has u.fl connector pads and I use this on all my Gateways.  I go from this through a u.fl pigtail to an RP SMA with a long threaded shank.  That usually goes through a galvanized steel plate about 4" square to form my ground plane and, depending on where it's deployed, goop the heck out the bushing (on the outside of a PVC box) if it's an outdoor gateway and simply tighten it with gold plated lock washers if indoor.  RP SMA rubber ducky after that.  Works well enough and holds up well.  I do the same thing with my WiFi radios in the same gateway implementation but these all come with u.fl already.

Alternative plan would be put everything, including u.fl or SMA connector, on your custom board and not use Moteino at all.

Tom

DRCO

  • Newbie
  • *
  • Posts: 17
Re: Antenna's and Host Boards
« Reply #4 on: January 22, 2016, 07:01:33 PM »
Thanks emjay. OK, I think I get that Ė so in the second option, where a shielded RF cable jumpers from the Moteino ANT (and has itís shield grounded on the nearby GND pin) the very short (maybe 4 mm) track from the Host board though-hole to the Ufl connector has some impedance at 433 Mhz which causes a mismatch to the expected 50 Ohm input. I guess that then adding a SMA connector (as a case bulkhead through-put) adds another 50 Ohm to the equation and we are really upsetting the balance.
Is that right? And if so, is there an option to trim the antenna to match the impedance?
Or is the only option (for the R4 iteration) to solder a shielded RF tail to the Moteino that goes directly to a SMA bulkhead through the case (as shown in the attached diagram).

And Tom, I literally got your post as I was writing this, if I get what you're saying the uFL and SMA connectors do not (in combination) screw up the impedance (at least on the Mega - which is way overkill for my application).

As ever, I greatly appreciate the input of all forum members Ė and if I should be doing homework instead of posting here please let me know Ė and maybe point me in the right direction.
Thanks again

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Antenna's and Host Boards
« Reply #5 on: January 22, 2016, 08:57:03 PM »
@DRCO: It would be great if, when you're finished, you could post a photo of the connection you made so that others could benefit by seeing what you actually did.

DRCO

  • Newbie
  • *
  • Posts: 17
Re: Antenna's and Host Boards
« Reply #6 on: January 22, 2016, 10:32:26 PM »
Will do, if I go with the R4. The other suggestions of a fully custom build without the Moteino - or maybe even going with the Mega are things I'll consider.

emjay

  • Full Member
  • ***
  • Posts: 119
  • Country: nl
Re: Antenna's and Host Boards
« Reply #7 on: January 23, 2016, 01:32:32 AM »
@DRCO,

The revised diagram looks fine for a flying lead connect. Also think about what is happening on the motherboard immediately beneath the module. Avoid signal traces passing underneath - fast edges can easily couple into the radio section. Ideally an area of copper pour on the top side, with multiple, randomly-spaced vias to copper pour ground plane on the under side.