[USRP-users] Amplifiers for USRP Transmission

Derek Kozel derek.kozel at ettus.com
Mon Jul 6 18:50:38 EDT 2015

Hello again Evan,

First, sorry for the flood of emails. I saw the website of Ghost Talk,
sorry we've been on the wrong track here. I've just read the paper and saw
the reference to the previous setup requiring a 10 Watt output to excite
the device sufficiently. This, as has been explained a number of times
above, is of course because the device under test isn't designed to be a
receiver in this fashion. Unsurprisingly pacemakers, ECGs, and other
medical and consumer devices have some amount of RF rejection built in in

Now that we've caught up and are on the page here's a few more thoughts
about increasing your output power. The SBX, at maximum, will output
approximately 20 dBm of power. In order to output 10 Watts of power you
would have to add 20 dB of gain, as you alluded to in your email. Yes, gain
does just add together, within the limits of the amplifier. For instance,
the ZX60-V63+ that you linked to maxes out at about +18 dBm, less than the
USRP itself! One better sized for the task is the ZHL-20W-13+ which has a
39 dBm minimum output power before it starts compressing the signal
significantly. It is a wideband amplifier, 20-1000 MHz, with very high
gain, about 50 dB, so more than you "need" to reach 40 dB at 400 MHz from a
20 dB signal. I put it forward only as an example.

Putting that much power out, especially after you count in the gain from a
directional antenna (which will need to be sized appropriately for this
much transmit power), does throw you into the realm where having a licence
to transmit is needed for most frequencies. If you are unable to find an RF
shielded room or box to use then you'll have to look at getting an
experimental licence from the FCC or restricting yourself to operating only
at frequencies and powers allowed as an unlicensed operator.

Filtering definitely comes into play as you add amplification. The USRPs,
and any radio, is non ideal and will produce undesired frequencies.
Especially if you have a wide band amplifier it is best, pretty much
necessary, to have filtering between the USRP and amplifier, and preferably
a high power capable filter on the output. This helps keep the spurious
signals from being transmitted at high power.

Sylvain's point about supplying (sending or transmitting being potentially
confusing words) the USRP with a higher than necessary sample rate is a
good general rule. The Nyquist rule states that you cannot reliably
represent a signal with maximum frequency N hertz without at least 2*N
samples per second. This (and some trivia about CD formats) is why
"standard" audio is 44.1 kHz, just over twice our nominal max hearing of 20
kHz. So if you are sending a 440 Hz signal to the USRP, make sure you're
using over 440 * 4 (or more) samples per second. Now that's actually a
silly low rate so use something higher like 1 Mega Sample Per Second (MSPS).

The gain setting on the USRPs is unitless dB. So an output at gain 33 will
be twice the power of one at 30. How that maps to dBm is approximate, but
at maximum gain it will be right around 20 dBm (100 mW). USRPs are not
precisely calibrated as there are a number of very complicated components
on board which make output and received power dependent on frequency, gain,
and sample rates to name a few parameters. What gain setting you use will
not affect the gain of an external amplifier, again unless you hit a
limitation of the amp. They will have maximum input ratings and also the
amplifier gain won't exceed it's maximum output power. Nothing
too surprising there.

The ARRL Handbook is a good general reference. (Disclaimer, I'm an
assistant section manager in the ARRL. Not that that makes any difference
to my opinion of the book.) It comes out every year, feel free to get an
older copy, these core concepts haven't changed in decades.

At 100 mW you aren't in danger of doing much to anything, however it is
worth noting that many RF devices don't appreciate having even that little
of power presented at their inputs. The USRPs for instance are best kept
with their inputs below -15 dBm (0.031623 mW!). Many other instruments can
tolerate up to 0 dBm (1 mW). Not putting enough attenuation in front of
input signals has been the end to too many pieces of wonderful equipment.

Hopefully this is useful to you. Do let us know how things work as you
progress along. I read that you were finding GNU Radio tricky. I recommend
asking questions on the GNU Radio mailing list. They're a good resource for
answers about all sorts of signal processing and the gottcha's of the
software. And of course this list for any USRP or UHD questions you


On Mon, Jul 6, 2015 at 10:31 PM, Marcus Müller <usrp-users at lists.ettus.com>

>  Hi Evan,
> I'm a bit confused over your use of a microphone, though. I'm making a
> sketch, and would ask you to correct me if I got anything wrong:
> [image: Drawing of the System]
> I'm presuming you're actually tring to use the microphone to emit sound,
> thus making it a speaker?
> There's a bit of a problem, here:
> The energy the antenna, even under perfect circumstances, will pick up is
> very little. That's a good thing, in principle -- your antenna is not very
> selective, and if the power coming from your wire would be sufficient, all
> speakers that unconnectedly lie around would instantly receive a lot of
> signals from the spectrum around you. If you look at the classical AM
> radio, you will notice that people (for much lower frequencies) used large
> antennas, which they were able to *tune*, very much like you tune the
> string of a violin, to emit and receive a specific frequency. This made the
> antenna work good as a receiver for a small bandwidth, the channel they
> wanted to receive, and bad for other channels. This way, you could listen
> to a station from Paris in London, even though the BBC of course had an
> overwhelmingly large signal power -- but at a different frequency, so by
> means of frequency selectivity, people were able to pick up the "right"
> channel only.
> The fact that your receiver works well if you put the transmitting and the
> receiving antenna very close to each other doesn't rely on this principle
> of reception: Whilst all the radio waves picked up by your wire are still
> very weak electrical signals in the wire, if you put both antennas
> sufficiently close to each other, they form a capacitor. Capacitors are
> very good conductors at high frequencies, so almost all energy produced by
> the SBX transmitter amplifier (it already has one; these 100mW don't come
> from nowhere :) !) is directly fed into microphone-used-as-a-speaker.
> That's not really radio -- almost all audio amplifier you typically see
> have capacitors between the actual amplifiers output and the speaker
> connectors (mainly to protect speaker and amplifier).
> What then happens is that a small amount of the energy passed through the
> speaker to ground actually moves the coils in your microphone (or the
> crystal, if you're using a piezo mic/speaker), because these mechanical
> things have high mass and thus large inert mass and low-pass character. In
> principle, if your microphone was a theoretically perfect thing, you would
> hear nothing -- for example, let's assume your radio wave has a frequency
> of 500 MHz, and you want to hear a 500 Hz tone -- in the time the membrane
> should have made a full oscillation, one million oscillations of the
> electric current induced in the antenna would have happened. If you take an
> average about whole oscillations, it's zero, no matter what you do. What
> you need to do is actually build what we call an envelope detector:
> [image: Envelope detector]
> <https://commons.wikimedia.org/wiki/File:Amplitude_modulation_detection.png#/media/File:Amplitude_modulation_detection.png>
> As you can see in C, the audio signal is the envelope of your RF signal.
> So to make the best of your experiment, we have to do three things:
> a) we need to build something that selects a specific bandwidth equal to
> the bandwidth of your audio signal, so that you don't receive a mixture of
> all the signals that your antenna picks up, and
> b) we need an envelope detector that "removes" the radio frequency from
> your signal, and
> c) build an amplifier, so that the small electrical signal from the
> antenna can actually move the membrane.
> Point c) is the only point where you could "cheat" with more output power
> (I really don't recommend doing that; legal, health and neighborhood issues
> would be the consequence); in fact, some early radios with small earphones
> could do completely without an amplifier, but since this is 2015 and not
> 1905, the amplifier is actually the easiest to get part.
> a) is called a filter. You can build a filter by tuning your antenna
> (which really doesn't make much sense, since you're not doing the low
> frequencies usually associated with AM), or by actually using a normal
> antenna like yours and having a filter further down the signal processing
> chain; either directly after the antenna, or after a only slightly
> filtering radio frequency amplifier has already increased the voltage of
> the signal. Easy filters can be built by using an inductance/capacitor
> circuit (google "LC resonant circuit"), but it's a bit hard to make the
> antenna work well directly with these; but basically: if you play around
> long enough, you'll build a filter that works. You're lucky, because you
> can just build a filter, and then use the USRP to transmit at a lot of
> different frequencies, until you hit a frequency that your filter lets
> through well. If you then adjust the parameters of your components, you can
> tune your filter to the desired frequency. That's a variable filter, and
> these are painful to get very exact. Still, it's what is used in very
> simple radios.
> Modern Radios often use a different approach: They do what is called
> "mixing". That's basically taking the Radio signals (all radio signals
> picked up by the antenna) and moving the whole spectrum down (and up, but
> that component can be eliminated easily) by an adjustable frequency, so
> that fixed filters can select a channel.
> b) Hit Wikipedia about "Crystal radio". The trick here is that you simply
> need a rectifier to make the average-zero signal an average-varying one --
> a simple diode (given it's fast and sensitive enough) will do.
> c) now that's just a normal amplifier. Use something with a high input
> impedance -- a microphone preamplifier might do well, or a guitar amp or so.
> All in all, building a crystal radio is a fun project -- but the signal
> theory behind and the physical basics of why the individual components
> (electromagnet waves, antennas, filters, mixers, detectors, semiconductors
> like diodes, amplifiers ...) work are something that electrical engineers
> study for years, literally. I've only selected a very small subset of the
> physical and engineering problems that you might meet.
> Don't let yourself be discouraged by initial hardships -- Marcus Leech has
> already pointed you to the ARRL handbook, I think. Although I'm in Europe,
> I also happen to have one of these -- they really introduce what we learned
> "the hard way" at university [1] much more from a radio "technicians" point
> of view: What does make the radio work? How can I built a simple one myself?
> Best regards,
> Marcus
> [1] (e.g. from linear algebra and differential equations->linear
> electrical networks->field physics->wave physics->wave propagation->antenna
> theory and from stochastics&integrals->integral transforms->signal theory->
> communication theory->receiver mechanics in a 5 years degree)
> On 07/06/2015 09:32 PM, Evan Chavis via USRP-users wrote:
> Hello USRP users,
>  I've been working with a USRP1 to transmit an amplitude modulated signal
> to a very basic improvised receiver, basically just a wire of length
> corresponding to the frequency I'm transmitting at attached to a microphone
> such that the microphone is picking up the signals on the wire as it would
> audible noises.  I'm using an SBX daughterboard (which can produce 100 mW)
> and a log periodic antenna (with 5-6 dbi gain).  I can get a decent
> transmission at a range of up to about 25 cm, although it is best at a
> distance of just a few cm.  I've got the gain set with GnuRadio as high as
> possible without causing distortion, and I'm looking to improve the range
> of possible transmission.
>  My questions are, what do I need to consider when looking for an
> amplifier to use with a USRP, are there any special problems a beginner
> would be likely to miss when trying to use those 2 things together?  Also,
> if an amplifier's gain is listed at 20 dB, would that be 20 dB on top of
> the 31.5 the daughterboard is capable of outputting, or would the amplifier
> be useless in that it is capable of providing less power than the
> daughterboard itself?  And third, how is a device such as this one:
> http://www.minicircuits.com/pdfs/ZX60-V63+.pdf powered?  The data-sheet
> gives the numbers 5 V and 69 mA but I don't see an obvious way to power it
> aside from just soldering wires onto the marked bumps that say +5 and
> ground, is that the actual intended way to do it?
>  Many thanks for any input and have a great day
> _______________________________________________
> USRP-users mailing listUSRP-users at lists.ettus.comhttp://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
> _______________________________________________
> USRP-users mailing list
> USRP-users at lists.ettus.com
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20150706/ae13c0e0/attachment-0002.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: drawing_evan.svg
Type: image/svg+xml
Size: 17615 bytes
Desc: not available
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20150706/ae13c0e0/attachment.svg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 485px-Amplitude_modulation_detection.png
Type: image/png
Size: 99277 bytes
Desc: not available
URL: <http://lists.ettus.com/pipermail/usrp-users_lists.ettus.com/attachments/20150706/ae13c0e0/attachment.png>

More information about the USRP-users mailing list