Hello all, apparently I've never bothered to introduce myself.

I'm an Embedded Field Application Engineer specializing in embedded systems and security.

I work for a little company who, among other things, sells and supports SiLabs parts.

Pleased to be here!

Hello all, apparently I've never bothered to introduce myself.

I'm an Embedded Field Application Engineer specializing in embedded systems and security.

I work for a little company who, among other things, sells and supports SiLabs parts.

Pleased to be here!

---

dasimon wrote:

Hi,

 

2. Theoretically you can calculate it by using the following formula: Z = 1 / (2*pi*f*C).

Actually, I used Murata SimSurfing application, looked for the exact component and its Z curve:

http://ds.murata.co.jp/software/simsurfing/en-us/index.html?intcid5=com_xxx_xxx_cmn_nv_xxx

 

3. We do not test external antennas generally. The 2.4GHz EFR32 Radio Boards contain an inverted-F antenna, beside that a lot of antenna solutions exist, but usually a whip monopole antenna is the best choice if the required ground reference is provided.

 

Daniel

---

Excellent information. Thank you.

---

dasimon wrote:

Hi,

 

1. So your test showed that using the inverted-F antenna, the range is ~200m, while using the whip monopole, you got ~500m range. Did you use the same settings on TX and RX side during the test, so the only difference was the TX antenna, right?

2. The impedance of a 3.9pF capacitor at 2.4 GHz is ~10ohm, so the actual loss is negligible.

3. Unfortunately, we do not have any material regarding this issue. The important point is that the monopole whip antenna needs a good ground reference close to the feeding point, this should result the best range. Any additional cable that connects the radio board with the antenna will cause range degradation on one hand due to the cable loss, on the other hand due to the antenna impedance detuning.

 

Daniel

---

Hi dasimon,

1. Yes, I got about 200 meters with the inverted-F antenna and about 500 meters with the monopole. I suspect that both would have gone farther had I sufficient unobstructed line of sight and resources. (It really takes two people to test this sort of thing well.) The same settings were used both times; the only difference being both the transmit and receive antennas. Again, my goal was not a comparison to any particular baseline, though I may do that at some point.
2. Good to know! How do you calculate that?
3. A solid ground reference makes good sense but is, of course, somewhat limited when using the dev kits. Obviously, if designing a custom board this could be taken into account. Does SiLabs have a set of best practices at least? By what means does SiLabs test external antennas (if at all) for this sort of thing?

Thank you for the comments and feedback!

Hi dasimon

1. I did perform the range test with the built in inverted-F antennas out to about 200 meters; I'm sure they would have gone farther. My goal was not to exhaustively test possible configurations; rather, to see what practical range limitations might be with excessively large antennas, to learn, and of course to have a little fun 
2. Fascinating! Is there any particular advantage to using a discreet 0 Ohm component at all or would a solder blob do just as well?
3. Do you have a knowledge-base article or reference material that might help explain this in simpler terms? I am not an RF expert by any means but it does sound interesting.

Hi andrasbiro, I went with 2FSK based upon my understanding of this knowledge base article:

http://community.silabs.com/t5/Wireless-Knowledge-Base/Modulation-choice/ta-p/144373

While I didn't have a spectrum analyzer handy, I do know that the 2.4 GHz band is very crowded in the middle of a big city.

Is there a particularly recommended modulation scheme for low-bandwidth applications in very noisy environments?

Thanks jose!

I think I'd like to see if I can get a kilometer at some point!

Below are the conditions and results of some range tests performed with a pair of BRD4151A +19.5 dBm EFR32MG WSTK boards.

Tests were performed in the most glorious of non-scientific manner befitting of real-world conditions.

The range test application was flashed to the boards and I verified that it was working properly:

Next, a bit of data from the BRD4151A Reference Manual is required:

3.3.4 Inverted-F Antenna

The BRD4151A Radio Board includes a printed Inverted-F antenna (IFA) tuned to have close to 50 Ohm impedance at the 2.4 GHz band.

3.3.5 UFL Connector

To be able to perform conducted measurements Silicon Labs added an UFL connector (P/N: U.FL-R-SMT-1) to the Radio Board. The connector allows an external 50 Ohm cable or antenna to be connected during design verification or testing.

Note: By default the output of the matching network is connected to the printed Inverted-F antenna by a series component. It can be connected to the UFL connector as well through a series 0 Ohm resistor which is not mounted by default. For conducted measurements through the UFL connector the series component to the antenna should be removed and the 0 Ohm resistor should be mounted (see Chapter 4.2 Schematic of the RF Matching Network for further details).

The inverted-F antenna is actually quite great but I felt like going for something a bit more excessive.

For starters, we need to modify things a little.

Here is the 0 Ohm resistor mentioned above under a scope:

This resistor must be rotated 90 degrees anti-clockwise to the trace leading to the U.FL connector:

(It looks messy but it's just a couple square millimeters of flux reflecting light from the scope)

Two, 2.4GHz antennas were chosen (because they were what I had lying around and because both happen to have the aforementioned requisite 50 Ohms of impedance).

For the transmit side, an 8 dBi omni-directional antenna:

And for the receive side... a 24 dBi parabolic directional grid, which I'm sure everyone has in their junk drawer at home:

The pigtails and coax cable likely introduce a non-negligible amount of insertion loss but it was decided that I would just see how it worked out in lieu of a bunch of math.

A suitable location was required with significant distance of unobstructed line of sight which proved very difficult to find in my locale.

(The local airport was very confused about exactly why I wanted access to their runway and hung up on me.)

Fortunately, water is pretty flat and a nearby lake had a 535 meter clear path:

Cables were connected:

I managed to con my significant other into standing on the opposite side of the lake holding the omni-directional antenna and pressing the start button on the transmitter side.

Antennas were approximately 1.5 meters above ground level and were aligned optically (I pretty much just took a best guess).

Boards were configured to use the minimum data rate and maximum receive sensitivity.

FCC regulations were complied with to the best of my ability and understanding.

Results were quite positive and very consistent.

All transmitted packets were received!

A more thorough write-up may follow as I'm curious to find out what the practical range limitations are!

Hi mapleleafs,

I realize your post is a bit old but I figured I'd comment anyhow in case someone chances upon it at a later time.

Having done some range test stuff myself I can tell you that the height of the antenna can certainly make a difference, performance wise.

Some of it has to do with the wavelength of the band you're operating on, in this case the 2.4-2.5GHz range of the ISM band.

The formula for wavelength is:
λ = c / f

Where:

c = speed of light in vacuum, a constant ~ 300,000,000 meters/sec
f = frequency = 2.4 gHz
λ = wavelength in meters

So the wavelength at 2.4GHz is:
λ = 0.3 / 2.4GHz = 0.125 meter = 12.5 cm

It's usually a good idea to keep your antenna at least 1.5X to 2X times the wavelength above ground or your photons are going to do all sorts of silly things.

1.5 meters is a nice number because it gives a standard baseline of comparison that is easy for most humans to manage (it's about 5 feet) without an antenna mast.

Some of it also has to do with radiation patterns:

With sufficient height, you're less likely to be bouncing the strongest part of your "lobes" off of the ground in a way that would interfere with ideal operation.

For the most part, any obstruction will tend to reduce ideal performance, some more so than others.

Will your module explode if it's 5cm above the ground?

Probably not, but it's certainly not ideal and you won't see optimal performance.

As for your second question, I think I recall that the WSTK uses something like an LP3878 regulator, so I would suspect you should be fine with a single cell Li-ion.

Just be careful, those 18650s can often source a LOT of current (some up to 40A), so don't short it out!

I too am curious about this magical driver library.

I was working on porting the FastLED library over to Gecko for a while but if you've got something functional it would be great if you could share!