Author Topic: Connecting SMA Pigtail to Moteino  (Read 2679 times)

Humancell

  • Newbie
  • *
  • Posts: 25
  • Country: us
  • The future is awesome!
Connecting SMA Pigtail to Moteino
« on: October 11, 2016, 11:51:45 AM »
Hello,

I'm curious if there are any specific instructions/details that I ought to consider if I want to connect an external dipole antenna to the Moteino via a SMA pigtail.  I'm not an RF engineer, and so I'm not sure of some of the implications of doing this.  My primary intention is to have the Moteino in an enclosure, and have the SMA bulkhead connector exposed to connect the SMA dipole.

The pigtail is like this:  https://www.aliexpress.com/item/New-SMA-Female-Jack-Connector-Switch-SMA-Female-Jack-Cable-RG316-Wholesale-Fast-Ship-15CM-6/32229453413.html

The antenna would be something like this: https://www.aliexpress.com/item/3G-GSM-Antenna-5dBi-850-900-1800-1900-2100-MHZ-SMA-Male-Floding-Omni-aerial-For/32405312060.html

I'm mostly curious about where both conductors of the pigtail coax would connect ... the antenna connection (center) would be to the ant connection, and the gnd/shield of the coax to any ground?

Are their issues with length, etc. of the coax?

Anything else I ought to consider (besides radiated power limit issues?)

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: Connecting SMA Pigtail to Moteino
« Reply #1 on: October 11, 2016, 05:38:21 PM »
The antenna in the link does not look like a dipole but seems to be a (quarter-wave) vertical. If it isn't a dipole but a vertical it needs a ground plane. Important question: do you want to use a dipole or do you want to use the antenna pictured in the link?

The connectors can just be soldered on to the moteino board. If you can't mount the connector directly onto the board you can have a short length of of feedline from the board to the bulkhead connector; Coax shield to GND closest to the antenna pad on the radio module, centre conductor to the ANT pad. Keep connections neat and short.

Length of coax doesn't matter if the antenna has a feedpoint impedance that matches the radio and the feedline (in our case: 50 ohms), but it's rare to have an antenna with exactly that. Therefore, to be on the safe side (to minimise mismatch and reflections) we recommend to use a feedline of a multiple of an electrical half wavelength of the frequency of interest. The length is measured from the point where the signal leaves the radio until it enters the driven element of the antenna.

Google feedline velocity factor and you'll get an idea what length coax you need.

For a node that only receives signals the above is far less important but in general, an antenna that transmits well will also receive well.

Humancell

  • Newbie
  • *
  • Posts: 25
  • Country: us
  • The future is awesome!
Re: Connecting SMA Pigtail to Moteino
« Reply #2 on: October 12, 2016, 12:38:29 AM »
Thank you for the reply ... and yes, you are testing the limits of my RF knowledge.

As I read the link about the antenna, I actually *assumed* (most likely incorrectly) on the terminology of the antenna type.  I'm actually open to antenna recommendations, but want something that can attach to a bulkhead connector on the enclosure, and provide decent gain/signal.

When you say this antenna in the link needs a ground plane, how are these used for other applications?  For example when I see other products that have these types of antenna is the ground plane a part of the PCB inside of the enclosure?  Or?

The coax in the link I posted is RG316, which I looked up and has a vF of 0.79.

Using the calculator here:  http://n1gnn.com/webData/CoaxVelocityFactorLength.html

... it says "For coax with Vf = .79, 1/2 wavelength = 12.95 cm, 0.425 ft & 5.1 in" so then I need to ensure that I have that much coax from the Moteino PCB to the SMA?  Or do I count the internal lengths of the SMA and Antenna SMA in this number?

Am I on the right track here?

Thank you again for your assistance in learning this!

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: Connecting SMA Pigtail to Moteino
« Reply #3 on: October 12, 2016, 03:35:05 AM »
Yes, you're definitely on the right track there. And thanks for trying out some calculations, it shows us you're keen to learn!  ;). RF isn't rocket science but it does have a number of variables that usually don't play a significant role in many other hobbyist projects.

Quote
I'm actually open to antenna recommendations, but want something that can attach to a bulkhead connector on the enclosure, and provide decent gain/signal.

The gain and signal properties can work just as well (maybe even better, because you're not putting an antenna right on top of the Moteino) when using bulkhead connectors. In one of my projects (not Moteino based, but it does use a HopeRF radio module), I specifically *want* to be able to easily interchange antennas and I went with a custom pcb design that has pads for putting an sma connector over the edge over the circuit board and the connector just pokes out the die-cast aluminium box.

An important difference between all the various antennas is radiation pattern. Just to get the dialog going, you have three main choices:

- If you want to transmit and/or listen all around you (from above and below, but not straight up or down): go dipole in a vertically polarised mounting position.
- If you want to transmit and/or listen in two opposite directions (e.g. North-South): go dipole in a horizontally polarised mounting position.
- If you want to transmit and/or listen all around you, bu within a limited angle of 'elevation' (and not straight up or down): go quarter-wave vertical
- If you want to concentrate the signal into a narrow beam (hmm.. more fat sigar-like) for listen or transmit: go yagi

If you look closely at a yagi, you'll see that the driven element is just a dipole.

If you look closely at a quarter-wave vertical, you'll see that it's just a dipole with one of the legs replaced by a ground plane.

Check out this awesome link from John K2OX explaining many aspects on the topic of antennas:
https://lowpowerlab.com/forum/rf-range-antennas-rfm69-library/antenna-tutorial-or-antennas-in-a-mote/

I guess the question is: What's your application?


Quote
For coax with Vf = .79, 1/2 wavelength = 12.95 cm

Yep, for 915MHz and that velocity factor that looks correct.

Quote
have that much coax from the Moteino PCB to the SMA?  Or do I count the internal lengths of the SMA and Antenna SMA in this number?

My suggestion would be to minimise the length of coax where possible. Especially on those higher frequencies the losses in that type of coax is quite substantial.

In your enclosure I would use maybe a few cm's of coax, just to connect the radio module to the sma connector. The sma connector should have PTFE (Teflon) dielectric (same as the RG-316) so it should have a similar velocity factor. Then, add all the lenghts all up and make sure the total electrical length is close to (a multiple of) a half-wavelength.



Humancell

  • Newbie
  • *
  • Posts: 25
  • Country: us
  • The future is awesome!
Re: Connecting SMA Pigtail to Moteino
« Reply #4 on: October 12, 2016, 12:10:17 PM »
My application is to make a gateway for now, with an omni-directional antenna.  I do have a shed on the mountain above my house, and can see where I might want to test with a yagi there, to test across the valley.  :-)

One question ... when you looked at the link for the antenna that I had found as an example, what led you to believe it is not a dipole?  For example, I found this one doing some more Google searches, and it appears to be a dipole ... how can I spot the differences?  https://www.lsr.com/downloads/products/330-0017.pdf

I had already read the link from John K2OX and found it a good resource ... lots of the common basics.  Now I'm trying to buy some antennas to test with.

One last question about the coax and lengths ... so the coax can be 1/2 wavelength - or a multiple - so it coul dbe a 1/4 wavelength, correct?

And when adding up the lengths of the various pieces ... am I measuring from the Moteino to the bottom of the antenna?

Thanks again!

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: Connecting SMA Pigtail to Moteino
« Reply #5 on: October 12, 2016, 07:17:16 PM »
Quote
what led you to believe it is not a dipole?

It could be a sleeve dipole, but it often is just a single bit of wire and that then is a 1/4wave vertical. My main reason for thinking it wasn't a dipole it because nowhere on the page does it say so. The link from your previous post explicitly says so, so no doubt there.

Gain-wise I don't always trust what is printed next to these cheap antennas. 5dBi sounds nice, but it is a substantial gain over a normal dipole (maximum gain of about 2dBi). If they were able to boost up the gain to that number they had to do some serious engineering. In that case I think they wouldn't be sparse with details.

There are two other antenna designs that I can think of that can be fed from the bottom, look like a monopole but aren't. One is a J-Pole, and the other is a collinear. Like the sleeve dipole, both the j-pole and the collinear also do not require a ground plane. The 1/4wave vertical definitely does.

J-pole gains are around the same as for a dipole (around 2dBi) but they need some additional work to prevent rf on the feedline from coming back to the radio. If you're on low powers (e.g. <10W) it's a really simple design with a reasonable omnidirectional radiation pattern, but you probably won't find these as commercial antennas..

Collinears are nice because you can stack them and increase gain quite substantially, but they are very tricky to construct and I would only look at buying one, rather than making one.

Quote
so the coax can be 1/2 wavelength - or a multiple - so it coul dbe a 1/4 wavelength, correct?

To be on the safe side, I would not do so.

I modeled this in NEC recently:



The second picture shows a quarter-wave section of feedline. What the curved lines show is the current distribution. For a dipole this is highest at the feedpoint (in the centre). Where current is maximum, voltage is minimum. So if you look at the feedpoint of the coax (bottom end, where the radio would connect) it has the lowest current and therefore the highest voltage. THAT can damage your radio.

Unless your antenna has the same feedpoint impedance as your feedline and your radio the length of coax doesn't matter (well, not too long, cuz you're gonna lose signal strength), but in all other cases I would stick to feedline lengths of (or multiples of) a half-wavelength.


Quote
And when adding up the lengths of the various pieces ... am I measuring from the Moteino to the bottom of the antenna?

Depends on the type of antenna. Usually you are looking for the point where the coax/braid/outer metal sleeve stops and the inner conductor can radiate outwards. To be honest I haven't looked into sleeve dipoles and how to determine where the inner conductor comes out. Referring to the pictures in your links: maybe at the point where the antenna become thinner, sort-of halfway.
« Last Edit: October 12, 2016, 08:37:11 PM by captcha »

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Connecting SMA Pigtail to Moteino
« Reply #6 on: October 13, 2016, 05:20:24 AM »
I was under the impression a 1/2 wave dipole (i.e. two arms each a 1/4 wavelength) can be purely resistive at the feedpoint at resonance, and can match the impedance of a transmission line and that transmission line lengths don't matter if they are terminated correctly at both ends (i.e. at the transmitter and antenna). So I'm confused where the dependence on coax length comes from.
Mark.
« Last Edit: October 13, 2016, 01:58:57 PM by perky »

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: Connecting SMA Pigtail to Moteino
« Reply #7 on: October 13, 2016, 06:55:27 AM »
From what I know, a dipole, resonant for a specific frequency of interest is 73ohms resistance + 43 ohms reactance. Such a dipole is not a good match for a feedline like RG-317 nor for the radio which expects a load impedance of 50ohms.

If you could somehow cancel out the reactance in the dipole and find a radio and transmission line that were designed for 73 ohms then yes, the length of the transmission line would not matter.

There are a few methods to match the 73 ohms impedance of the dipole with 50 ohms transmission line. Baluns, quarter-wave stubs, gamma matches, etc..

When you say:
Quote
if they are terminated correctly at both ends

Are you referring to everything being 50ohms? Then, yes, if the antenna is indeed 50ohms and only reactive resistive the length of the transmission line would not matter. However any nearby objects, even ground, plays havoc with the antenna being symmetric and would upset impedance. Not even all transmission line is always exactly 50 ohms. Hence my main aim to protect the transmitter from rf coming back and stick to specific lengths of feedline.


« Last Edit: October 13, 2016, 05:01:43 PM by captcha »

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Connecting SMA Pigtail to Moteino
« Reply #8 on: October 13, 2016, 07:26:30 AM »
You can eliminate the reactance of a dipole by slightly shortening the arms by a 'k' factor from their 1/4 wavelengths, which is roughly 5%. The impedance then is roughly 75R resistive only. There'll be a mismatch between that and the coax which will cause some reflections back, but it's not very much at the low powers we're using. As soon as those reflections reach the transmitter they won't reflect back again anyway due to the transmitter having a 50R output impedance. This makes the antenna's feedpoint 'see' an infinitely long piece of coax as it is correctly terminated, so I question the dependence on the length of the coax for this.

There's a potential problem with common mode currents in the coax caused by slight mismatch of impedances between the two arms at the feedpoint, which are inevitable and result in the coax radiating and thus diverting power from the antenna. In that case a current balun (common mode choke effectively) at the feedpoint gets rid of it. You could also use a balun structure whose windings also form an impedance matching network, 50R on the transmitter side and 73R on the antenna side, that's your best solution. Note that a current balun (common mode choke) by itself is not impedance matching, you also need a transfomer structure in there as well to match impedances.

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: Connecting SMA Pigtail to Moteino
« Reply #9 on: October 13, 2016, 08:19:00 AM »
Very much enjoying this conversation and I'm learning more about dipoles as I go.. :-)

Quote
but it's not very much at the low powers we're using

Correct, a 75 to 50 ohm mismatch gives about a 1.5:1 swr which is about 4% reflected power coming back to the radio.

Quote
As soon as those reflections reach the transmitter they won't reflect back again anyway due to the transmitter having a 50R output impedance.

That is a very interesting point you raise there and I must say I have never heard it said before in that manner. I'm happy to state ignorance on my part, but I've always read that any swr will reflect back and forth in the feed line until all the energy has dissipated. As heat warming up the coax and going back to the antenna to radiate a bit more.

Quote
so I question the dependence on the length of the coax for this.

For swr the length of feed line doesn't matter, it will remain the same whatever the length. My observations have just seen that impedance mismatches can do weird things if the coax length is chosen arbitrarily. I think I'm explaining that with the images modeled with NEC. Unless I've completely missed the point?

Quote
50R on the transmitter side and 73R on the antenna side, that's your best solution.

From a matching point of view, yes. Interestingly I was just looking at small 1:1 baluns-in-smd-package and see insertion losses of 1.5dB and more with is taking away a lot of power from the already low-power output signal to begin with. You win some, you lose some..

JOHANSON TECHNOLOGY  0896BM15A0001E  Baluns Transformer, SMD, 863 MHz, 928 MHz, 1.5 dB, 0805
http://au.element14.com/johanson-technology/0896bm15a0001e/balun-928-863mhz-1-5db-180-deg/dp/1885513
$1.23 Cheap as chips! :-)

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: Connecting SMA Pigtail to Moteino
« Reply #10 on: October 13, 2016, 01:49:23 PM »
If you guys ever settle on an optimal solution that you confirm through testing, I'd be interested in exactly which components to order so I can hook one up to a Moteino.  I had generally assumed that the insertion losses at both end of the coax would negate the benefits, but I guess not.  How much improvement is left after insertion loss is accounted for?  I'm not sure how to evaluate that.

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Connecting SMA Pigtail to Moteino
« Reply #11 on: October 13, 2016, 01:56:54 PM »
I really don't think it's worth bothering with for these low power systems, like you say baluns introduce losses and it's quite sensitive to other things. If you're transmitting hundreds or thousands of Watts then that's a different story. I do think it's worth slightly shortening the arms by the 'k' factor to make the impedance resistive only, i.e. make the antenna resonant in the centre of the band you're using.
Mark.