Tips and Tricks for RTL-SDR - Written by: Jay Moore
DXing With The RTL2832U
So, you’re new to RTL-SDR? Wondering just how to make the best of your experience?
I have written down some tips that have helped me that I often share with people who are seeking advice. I won’t say any of this is true from a technical standpoint; but it seems to hold up from an experience point of view.
This one is important only because I’ve seen people get it so wrong. A lot of people want to think that there’s been some magical advancement that allows you to pick everything up with a tiny antenna. That’s just not true; there are still basic laws of physics that apply. Too often I hear of people trying to pick up shortwave or AM with a whip antenna; that’s just too electrically short. There are ways it can be done, as my car exhibits fantastic AM reception, but it requires some clever engineering.
The same applies for all frequencies; it’s just the higher up you go, the smaller the wavelength. The telescoping whip that came with my dongle measures 12 inches fully extended. This is fine for very strong FM stations and will work relatively well over 300mhz since you can shorten the whip as you go higher; but it’s not going to perform very well. I, for example; have a 39 inch whip antenna mounted to the base that came with my dongle and I’ve received FM stations from 50+ miles away with it. It’s closer to quarter-wavelength than the included antenna.
Half the fun of the reception side of things is you’ll either have good reception, or you won’t; it’s much different on the transmit side of things. Various antennas can be built cheaply and perform well, and you’ll learn a lot in the process.
- Antennas and electricity DO NOT mix!
This should be an absolute given; but I have read many horror stories about people ignoring this. Antennas and electricity do not mix; at all. So should you decide to mount an antenna outside; you have to be absolutely mindful to not mount it *anywhere* near electrical lines. There’s a general rule to follow when doing this:
“If your antenna runs over a power line, assume it will fall. If it runs under power lines, assume they will fall.“
There have been too many people lost in this hobby because they ignored the warning or played the chance and mounted an antenna where it’d come in to contact with power lines. Not only will power lines raise your noise floor; but you don’t want supply voltage running in to your shack blowing everything up; and this hobby isn’t worth dying for.
Please, don’t be another name added to list of those who lost their life over something as stupid as letting an antenna come in to contact with electricity. At the kind of voltages you’re dealing with in the grid supply; anything and everything will become conductive. The insulation on your wire will not protect you; I’ve seen demonstrations where something like a shoestring will become conductive enough to kill. It’s better if you just avoid it entirely and place the antenna out of harms way. As I said earlier, assume that either the power line, or the antenna, will fall. If your antenna is mounted in such a way that neither of those would be a problem; you’re ok.
- The “RF Gain” slider is NOT your friend.
Depending on which SDR software you use; you will have likely seen the “RF Gain” in your SDR radio options, and many of you have probably thought “oh, i’ll just crank this all the way up and pick up everything”. WRONG! There are many problems with this thinking. First of all, an RF amplifier doesn’t know the difference between noise and actual RF signals; so you get a “garbage in/garbage out” ordeal; it will amplify RF by also amplifying *everything* that comes in to it. This can and will raise your noise floor. The problem is compounded by how the RTL-SDR RF gain slider works. There are three gain settings that slider controls; an LNA (Low Noise Amplfier), a Mixer amp, and a “VGA amp” (which I’m not sure what it means). With standard RTL-SDR drivers; the RF gain adjusts all three of these using pre-set variables; however, all three don’t get along well. While the LNA gain is pretty useful even if it does raise the noise floor; cranking it up too much causes “images” of other signals to bleed through to the other bands. The Mixer gain is also handy when used in moderation; but in combination with a high LNA gain, it just makes the imaging and bleeding even worse. The VGA gain seems to be the most pointless; I’ve never seen it do anything but act like it multiplies the overall signal; while the other two do raise the noise floor, the rise in the NF isn’t the same ratio as the gain in RF levels for actual signals. So, for example, raising the gain to bring the NF 5db might raise the RF levels of a signal an additional 10db. There is a modded driver for SDR# that allows you to access these three gain settings independently; allowing you a bit more control over how much each amp applies it’s gain.
The trick to setting the gain is to find a weak(er) signal and monitor it while adjusting the sliders, I highly recommend you do this with the volume raised up. Sometimes increasing the RF gain will make the signal seem “louder” at one volume simply because you’re pumping more signal in to it; even if you’re not actually raising the signal above the noise floor. So you’ll hear a signal getting louder without there actually being any other improvement. Keeping the volume all the way up on a weak signal will allow you to hear how much the noise is increasing as well as the signal. Do this with the AGC in your software turned off; sometimes that will throw things off. It’s also a good idea to watch the waterfall while you’re doing this; paying attention to the noise floor and the received levels. You will likely find that raising the noise floor a few dB might have the advantage of making even some weaker signals come in a bit better.
The idea is you want to stop before the noise starts getting louder than the signal you’re monitoring; or before you start hearing other signals bleed through. With a standard RTL-SDR driver; this can occur with as little as 1.4dB of gain; where as with the modified “gains” driver for SDR#; you don’t get this until you crank your LNA all the way to max and/or you begin fudging with the mixer setting.
- That 75-ohm 50-ohm difference isn’t a big deal.
If you’ve started researching antennas, you may have figured out your RTL-SDR stuff is using 50-ohm SMA connectors and cable, rather than the 75-ohm stuff you’re probably used to. Yes, there is a difference between the two; how big of one depends on the application. Typically speaking the impedance of everything should match in an RF environment; a mismatch will cause a loss of signal or reflected power, which is a big deal in transmitting.
The reception side of things is much different, all you deal with is signal loss; the amount depends on how large the difference is. 75-ohm and 50-ohm aren’t a very large difference, so the amount of loss isn’t anything to worry about. RG-6 is getting cheap these days; and adapters to connect the typical F-type connector to SMA stuff are available.
- Your modern way of life is ruining your reception.
As much as I hate to say it; our modern way of life is actually not helping our RF reception at all. Most power supplies these days are known as a “switched mode power supply”; rather than having a transformer do the bulk of the voltage reduction with some filtering and voltage regulation on the output; these devices use a high-frequency signal that “flips” the output of the power supply on and off very rapidly. The advantage to this is it’s much more efficient; both in how much power you draw in vs how much you get out, but also in the operation of the power supply since transistors work most efficiently when “all on” or “all off”.
So, these devices produce a wide range of radio frequency interference; most of which interferes with your dongle. Likewise, the PC power supply is usually pretty noisy and is the product of an additional switching supply in the computer. So before you go thinking your device is garbage or your antenna’s suck because you can’t pick up a signal you want to; take some time to find out if your local environment is causing it. I have one laptop I can’t attempt to DX with because it’s power is so dirty I get nothing but interference below 3mhz.
It’s also worth noting that this interference doesn’t have to be “local”; with a big enough antenna and most houses being spaced close together; you might sometimes be picking up interference from the neighbors through the air. That’s not to mention the kind of interference you get from the numerous power lines running all over the place.
There’s not a “one-size-fits-all” solution for this; you may have to try several things..and in some cases, you may have to live with it.
I have not tried all the SDR programs out there; I’ve tried a couple and stuck with SDR# as it works for me. But not all SDR programs are created equal; and they all won’t perform the same. Sometimes it’s amazing how much your performance improves when using a different program or a different configuration. So, you may need to try several things before you find one that works best for you.
- These are not $1200 radios, nor perform like one.
If you’re expecting your RTL-SDR to outdo a more expensive radio or SDR unit; you’re in for some disappointment. The mere fact is, there are actual hardware-based radios out there that outperform these dongles; and there are SDR solutions out there that WAY outperform the RTL-SDR. These dongles are being used in a manner they were not designed to; and the limitations of the dynamic range of the ADC means you will have a higher noise floor and more imaging of signals than you would with a unit with a higher-resolution ADC. They will outperform most cheap radios, and they make excellent SDR units for the price you pay. That being said; the low-cost is allowing them to be used in place of even more expensive SDR hardware; as well as a cheap way to get in to the world of what SDR can do in the first place. Quite often, online streams of scanner traffic or weather radio is done using a couple of these cheap dongles with specalized software; at a much lower cost than even a basic police scanner with modern digital decoding.
Mine performs about as well as a 70’s era receiver; much better than the cheaper digital radios made in the last 15 years. But it does not perform to the same quality as a very expensive Yaseu set; or even a more expensive Sangean portable radio. Close, but not quite. For me, it was between RTL-SDR and a Tecsun PL-880; the RTL performs almost as good as the Tecsun; but by the virtue of it’s much wider frequency range and the fact demodulation is done in software, it’s a much more capable radio that will “pick up more”.
To put it in perspective; you’ll pay around $100 or so for an RTL-SDR setup to do HF; much less if you go the direct sampling mod or get a cheap upconverter (both of which can seriously hinder your performance); a more proper SDR solution starts at around $300 and goes up to thousands of dollars. A similar performing radio might start around $200 on the portable end and get up there in to the thousands of dollars.
If you’re wanting to get in to SDR; these dongles are a great way to start without worrying about investing a large chunk of change. But if you’ve got the money to invest in a better SDR (which I don’t), then go for it.
- Beware the edges of your tuning rage.
One of the biggest issues with these units are the image signals that occur; both because of the design and also because the dynamic range for the receiver is a bit low. With that being said; the edges of your tuning window are usually filled with “wrap-around” signals from just beyond the sample rate limit; so looking at a signal right up against the edge of your tuning range isn’t very accurate. If you tune in an area with a couple of stronger signals, zoom all the way out, and drag your tuning range; you’ll likely find the signals move in a different direction than you’re tuning.
When you have your FFT waterfall zoomed in to get a better view; it can be quite easy to get to the edge of your 1 or 2 mhz (or whatever your sample rate is) window without realizing it; and nothing is more embarassing than trying to copy a signal you think is far away that’s actually just an image of a signal just beyond the tuned band that’s very strong.
This can also have an effect on your noise-floor. If you tune your range just beyond the tuning noise; due to the imaging you may have the same noise appearing. Dragging the tuning range in SDR# (the range, not the VFO) will sometimes allow you to see. In fact, the biggest tip is to zoom out and drag the tuning range and look at images and random “spikes”; somtimes they’ll move with you; sometimes they’ll move at a much different rate and can often see them begin to attenuate and disappear.
Of course, for that same reason; sometimes tuning in the upper or lower 1/3 of the range will give you a slightly clearer copy.
- Higher sampling rates aren’t necessarily better.
SDR puts a real strain on your system; it’s one of the most processor intensive things you can do. So it’s important to realize just because this device is capable of outputting 3.2MHZ of bandwidth (3.2msps); that doesn’t mean you should use it. Most computers have a real issue keeping up with that much data; plus some dongles can’t put that out without dropping samples. 2.8msps is great for some people; but still too much for others.
If you’re having performance issues; the first thing to do is lower your sample rate and see if reducing the load clears things up. It’s a double edged sword though; too high a rate works poorly; and too low a rate works poorly. I personally use 1 or 2mhz depending on what I do; those seem to produce the cleanest signals. There are various sample rates between them; but I’ve gotten pretty poor performance with usually extra errors/artifacts in the RF spectrum.
You’ll have to spend some time comparing a section of spectrum using various sampling rates to see which your computer can handle; as well as which ones give you less erroneous signals.
- Your PPM offset is only important if you decide it is.
I’m going to get some crap for this one; but I will say that determining your PPM offset of the oscillators in your dongle isn’t that important. Having the number right will not magically improve your reception. All that number does is allow the frequency displayed in the software to match what you’re actually tuned to; so if you’re tuned to say 1480khz; you’re tuned to exactly 1480000hz and not 1480244hz. If you’re attempting to use this as a highly accurate spectrum analyzer; or need to know the exact carrier of a signal; then you’ll want to calibrate your dongle (as well as reconsider your sanity). However, if you can live with “ballpark” estimations of frequency; then the worst you’ll do is be maybe a couple of kilohertz off.
Traditional analog radios also have this kind of calibration when setting up the RF front ends and the LO stages; and with old dial-radios, it’s never 100% exact. When I rebuilt my signal generator for example; the calibration instructions were really only syncing dial to what the signal actually was. Since I was using a digital frequency counter on it’s output; the dial was meaningless for fine signal adjustments; I could look at the frequency counter. Calibrating it didn’t improve it’s performance, it just moved the frequency output closer to what the dial said.
It’s the same thing here; adjusting the PPM offset will only make your displayed frequencies closer to the actual frequency you’re tuned to. I don’t think you’re going to get 100% accurate with these things to begin with, closest I’ve ever gotten is within 20hz of being accurate; but I usually run sometimes as much as 700hz off calibration and it does not affect the signal; just the displayed frequency.
- Radio stations are not a good source for calibration.
If you do decide to calibrate; many people seem to choose some kind of strong station and set it to that. Ok, close; but not quite. The frequency standard a lot of transmitters use isn’t 100% accurate; and not something you’d want to set an offset too. Sure, it’s *close*, but it’s not 100% accurate in laboratory terms. Of course, for all I know; modern transmitters may be using very stable frequency references and are in fact accurate and stable to within a couple of Hz.
There are however a few signals out there that use a highly accurate carrier signal; and often times there are signals that carry accurate time that can be used to fighre out the offset. GSM networks apparently have the ability to serve as an offset calibration with the right software. The only signals in the air I know of offhand that can be used to calibrate are some time signals. WWV/WWVH both use carrier frequencies that are derived from their atomic clocks; so they’re both highly stable and highly accurate to within a few Hz.
You could also pick yourself up a second-hand rubidium frequency source; which will output a very stable carrier. Most of these came out of old cell phone sites and are a few hundred bucks. But is it worth a few hundred bucks just to make sure 3455000hz is actual 3455000hz and not 3454857hz? Not in my book.