Author Topic: What is the smallest antenna that still performs well?  (Read 8793 times)

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
What is the smallest antenna that still performs well?
« on: October 16, 2016, 06:57:15 PM »
As there has been a lot of discussion about antennas lately, I'm wondering if anyone has played around with an inverted F antenna?  I notice in the wikipedia article that there's one used in DECT phones that looks pretty simple and relatively small:

Since DECT phones run around 900Mhz, it would seem to be extremely relevant to those of us running at 915Mhz.  DECT 6.0 is a different frequency altogether (1.9Ghz)

There are also inverted F antenna's that can be traced on PCB, with the advantage being that they're a bit more compact (as in Figure B below) than just straight traces:


[Edit #1: here's a simplified guide I found on how to do a trace antenna:  http://colinkarpfinger.com/blog/2010/the-dropouts-guide-to-antenna-design/
Unfortunately, his example is for 2.4Ghz, so it would have to be adapted somewhat.  It does seem to indicate, though, that I would need a network analyzer, which I can't really cost justify: I could buy a lifetime supply of splatch antennas for the cost of one of one network analyzer.   ::)]

[Edit #2: On the other hand, TI does have a Design Nuide for how to do a 915Mhz Inverted F-antenna:  http://www.ti.com/lit/an/swra228c/swra228c.pdf
and so perhaps that can be more readily adapted.  In fact, it gives the exact dimensions of the antenna trace that would be needed.  Further, it states in the conclusion, "Measurements of reflection show that the center frequency is easily adjusted by trimming the antenna length."  The part that I'm puzzled over though is exactly what all those little discrete components need to be, and their values, between the radio and the antenna. Is there a similar way to wing it when it comes to those?]

[Edit #3: If all else fails, I'm guessing that some kind of folded monopole or meandering monopole antenna should be possible (as briefly outlined in http://www.atmel.com/Images/Atmel-42332-ISM-Band-Antenna-Reference-Design_Application-Note_AT09567.pdf) by making the trace longer than it needs to be and then shortening it a bit at a time until the sweet spot is reached.  It may not be perfect, but perhaps it would work good enough (?). ]

[Edit #4: I just found a Design Note for a chip antenna, and on its face it sounds quite good: PCB footprint is just 8mm x 6mm in sizehttp://www.ti.com/lit/an/swra160b/swra160b.pdf
915Mhz Johanson chip atennas cost $0.92 each on Digikey in quantity 1.  On the other hand, a uSplatch antenna is 9.14mm x 12.7mm but apparently can be mounted vertically if so desired to help keep a small footprint.  The regular splatch antenna is 13.7mm x 27.94mm.  Is a chip antenna the smallest antenna that still performs reasonably well?]
« Last Edit: October 17, 2016, 08:32:24 AM by WhiteHare »

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: What is the smallest antenna that still performs well?
« Reply #1 on: October 18, 2016, 05:26:58 PM »
Okay, this may be a bit silly, but this 'antenna' worked for me.



It's two 51ohm resistors in parallel.

It closely matches the radio's 50 ohm load impedance and has a short enough range for me to test signals on the edge of the reception horizon. This way I don't have to separate tranmiter and receiver by hundreds of meters in order to simulate intermittent loss of signal.

Why 2x 51 ohm resistrors in parallel? Isn't that half the value that's required?

Well.. it depends. These particular resistors are metal-film and if you look at how they are made you'll see that they are basically a spiral cut out of a film-cylinder by a laser etch. The end result is that you end up with a resistor that looks like a coil (and behaves like one!). At DC you'll measure the expected normal resistance, but at rf things look very different.

I don't have the plot at hand anymore, but when I put a single 50 ohm metal-film resistor on the VNA I was seeing an impedance of around 100ohms at 433 MHz. This is purely due to the inductor effect which is not negligible with these types of resistors:



So if they were 100ohms at 433MHz, simply adding another one in parallel would half it back to around 50 ohms.

There you go, dummy load with minimal range (about 3 meters).

If you're planning to make your own dummy load, try to find carbon resistors, as they are not inductive as these metal-film resistors. That way you'll only need 1x 50 ohm resistor.  I think most SMD resistors are still carbon based so they would be a better candidate.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #2 on: October 18, 2016, 06:13:08 PM »
Great post! 

I tried something similar once, but just for receiving:

https://lowpowerlab.com/forum/rf-range-antennas-rfm69-library/real-time-rssi-measurement-broken-rfm69cw/msg10960/#msg10960
In retrospect, we probably overlooked the inductive qualities of that resistor that your post so eloquently explains.

Ironically, our goal at the time wasn't to make an antenna but rather to not receive anything at all.  Consequently, Emjay later had me solder tack antenna directly to ground, but even then it still received packets!  In both cases, I was operating in Rx only.

How many ohms should be used at 915Mhz if I want to Tx using a carbon resistor antenna?  Does the length of the wire leads on either side of the resistor significantly affect the range?
« Last Edit: October 18, 2016, 06:30:27 PM by WhiteHare »

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: What is the smallest antenna that still performs well?
« Reply #3 on: October 18, 2016, 06:46:15 PM »
Hah, great to see you did the same! :-)

It's amazing that rf is leaked (and picked up) so easily. At home I'm experimenting with RFM12Bs (output power 1mW / 0dBm). If I put a quarter-wave vertical at one end and use the dummy load as pictured, I get a range of easily 30 meters. Whoa!... Heheh..

I would say carbon resistors can just be chosen with the intended value. I don't think they will do much at rf. However, you mention lead length. Yes, that would contribute some, but without a VNA it's hard to say. I would still just use one 50R carbon resistor and go with that. The radio is pretty forgiving. If you want to go smaller, 1206 smd is very easily hand-solderable and will minimise lead length even further.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #4 on: October 18, 2016, 08:09:50 PM »
As luck would have it, I just today received from China a giant sample book of 0805 SMD resistors, all nicely sorted by ohms.  It doesn't have a 50 ohm resistor, but it does have a 47 or a 51 ohm resistor.  It claims both are +-1% accurate.  I could give it a try if you think transmitting would work with one or the other soldered between Tx and GND on the RFM69HW module.  If it worked satisfactorily, then you're right: it probably would be the smallest antenna possible.
« Last Edit: October 18, 2016, 08:12:36 PM by WhiteHare »

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #5 on: October 18, 2016, 08:30:57 PM »
As luck would have it, I just today received from China a giant sample book of 0805 SMD resistors, all nicely sorted by ohms.  It doesn't have a 50 ohm resistor, but it does have a 47 or a 51 ohm resistor.  It claims both are +-1% accurate.  I could give it a try if you think transmitting would work with one or the other soldered between Tx and GND on the RFM69HW module.  If it worked satisfactorily, then you're right: it probably would be the smallest antenna possible.

BTW, which range is it that you got?  Your first post said 3 meters, but your second post said 30 meters.

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: What is the smallest antenna that still performs well?
« Reply #6 on: October 18, 2016, 09:54:50 PM »
If I use the dummy loads on both the transmitter and receiver I get about 3 meters.

If I change one of them to be a proper antenna the range jumps to about 30 meters (and go through a couple wooden/fibro walls).

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #7 on: October 18, 2016, 10:53:03 PM »
If I use the dummy loads on both the transmitter and receiver I get about 3 meters.

If I change one of them to be a proper antenna the range jumps to about 30 meters (and go through a couple wooden/fibro walls).

Thanks for the clarification.  In that case, this sounds very promising!  Hopefully, upgrading from quarter wave antenna to a dipole antenna on the one transceiver and keeping the resistor antenna on the other transceiver will prove more than adequate as a wireless link in a home environment.  I'll give it a try with an SMD resistor antenna and afterward post how it went.  I'll use a 51 ohms SMD, since that's closer to 50 ohms than 47 ohms is.

captcha

  • Full Member
  • ***
  • Posts: 108
  • Country: au
Re: What is the smallest antenna that still performs well?
« Reply #8 on: October 19, 2016, 06:56:37 AM »
Quote
I'll use a 51 ohms SMD

Yeah, that should work just fine. You may get a lot better performance as a dummy load that way (read: almost no range) if you can minimise the lead length.

On the topic of metal film resistors.. If you still want to make a dummy load with metal film resistors you can certainly do that but you'll have to parallel as many as you can. Reason being that when you put these metal film resistors in parallel their total inductance goes down. A common recipe for a home-brew 5W dummy load is to parallel 20x 1k/0.25W resistors.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #9 on: October 19, 2016, 07:54:23 AM »
Quote
I'll use a 51 ohms SMD

Yeah, that should work just fine. You may get a lot better performance as a dummy load that way (read: almost no range) if you can minimise the lead length.


But wait, here I'm now trying to maximize range, not get almost no range.  So, for that purpose, are you saying I would be better off with the thru-hole carbon resistors, since they do have some lead length (unlike the SMD)?

[Edit: looking into it more deeply, if it's the thru-hole type, then apparently what I would need is the "carbon composition" type resistor.  The carbon film resistors would have the same helical spiral as the metal film resistors you highlighted.  It seems they also have lower tolerances (more like 10% maybe).]
« Last Edit: October 19, 2016, 08:31:08 AM by WhiteHare »

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #10 on: October 19, 2016, 10:20:16 AM »
Success!  I replaced the wire antenna on a Moteino with a 51 ohm SMD resistor (see attached photo), and it transmits fine.  A quick eyeballing and comparison of RSSI figures as compared to a typical monopole wire antenna suggests about a 22dB penalty, give or take.  Upgrading to a dipole antenna on a gateway would offset much of that, but not all of it.

Considering the minuscule footprint and minuscule cost, this seems like it could be a winner for use in low-cost ultra-compact wireless nodes.   :)
« Last Edit: October 19, 2016, 10:26:20 AM by WhiteHare »

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6404
  • Country: us
    • LowPowerLab
Re: What is the smallest antenna that still performs well?
« Reply #11 on: October 19, 2016, 11:27:22 AM »
@WhiteHare You just invented the world's smallest antenna, patent it before someone else does :D
Why didn't you just solder the resistor straight between the RFM module pads?

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #12 on: October 19, 2016, 11:55:43 AM »

Why didn't you just solder the resistor straight between the RFM module pads?

Would that work even better?  I'm wide open and welcome any and all suggestions. 

« Last Edit: October 19, 2016, 12:17:26 PM by WhiteHare »

Felix

  • Administrator
  • Hero Member
  • *****
  • Posts: 6404
  • Country: us
    • LowPowerLab
Re: What is the smallest antenna that still performs well?
« Reply #13 on: October 19, 2016, 12:08:49 PM »

Why didn't you just solder the resistor straight between the RFM module pads?

Would that work even better?  I'm wide open and welcome any and all suggestions.
Hmm, technically it should but perhaps the extra "brace" to the pad and ANT pin hole would act as a divider and produce some reflection? Maybe it's too short to really matter. But I was more hinting at soldering directly between the RFM pinholes to avoid the solder avalanche.

WhiteHare

  • Hero Member
  • *****
  • Posts: 1300
  • Country: us
Re: What is the smallest antenna that still performs well?
« Reply #14 on: October 19, 2016, 12:16:03 PM »
Hopefully others will now try this, and we can pool our datapoints to collectively figure out what works best.   :)

[Edit: I'm still confused from the earlier thread discussion as to whether longer leads are good or bad, so this was just a compromise for a first try it out.

I suppose this "SMD resistor antenna" could be especially practical on a LoRa module.  That thing has so much potential link budget that for a home environment you could easily workaround the antenna impairment and still have rock solid wireless links everywhere in the home. ]

In any event, I look forward to comparing a uSplatch antenna to it.  I should think a uSplatch would pretty much necessarily outperform it, because otherwise who would buy them?

In the spirit of KISS, my near-term goal is to fit everything, including 2x AAA batteries, into a 4x AAA battery holder case like this:

https://www.aliexpress.com/item/New-Plastic-4-AAA-Battery-Srorage-Box-Holder-4pcs-AAA-Batteries-Case-with-a-Lid-ON/32579574841.html?spm=2114.30010308.3.2.l7ZbuM&ws_ab_test=searchweb0_0,searchweb201602_5_10065_10068_10069_10017_10080_10082_10081_10060_10061_10062_10056_10055_10037_10054_10059_10032_10078_10079_10077_10073_10070_421_420_10052_10053_10050_10051,searchweb201603_4&btsid=53ae9427-799a-435b-9575-d6f3fba7ec42

I hope to fit a wireless node inside the case along with 2x AAA batteries positioned to produce 3V.  The case already has an exposed switch to turn it on/off. Total case size is relatively small, around: 6.2 x 4.8 x 1.5cm / 2.4" x 1.9" x 0.6"(LWH)  I think it will have enough space to fit a uSplatch in there, but if not this SMD resistor antenna is my fallback.  There's a good chance the new Moteino with the PCB antenna might fit nicely as well.
« Last Edit: October 19, 2016, 03:21:37 PM by Felix »