RTL-SDR Blog V.3. Dongles User Guide

Version 3 of our customized RTL-SDR dongles brought out some new interesting features. In this guide we explain how to use those features:

Feature 1: Direct Sampling HF Mode

This feature allows you to listen to HF signals between about 500 kHz to 28.8 MHz.

To use direct sampling mode

  1. Connect an appropriate HF antenna to the SMA antenna port (this is the same port where you connect your VHF/UHF antenna). 
  2. In SDR# select the Q-branch in the configure menu (the cog icon next to the play button). (If it is greyed out make sure you stop the SDR first, by clicking the stop button in SDR# first)
  3. Press Play and tune to 500 kHz – 30 MHz.


VHF antennas like small discones or short whip antennas will probably not pick up HF signals very well, if at all. If you have no such antenna you can try extend the large telescopic antenna to its maximum length of 1.5m, or use the screw nut provided with the antenna base to clamp on a long wire antenna. Ideally you should use a 9:1 unun with the long wire antenna for optimal reception. Even more ideally you’d use an antenna tuner, though this is expensive.

Other software like HDSDR can also support direct sampling. It may entail setting a device string, and for the Q-branch, the value should be 2.

To go back to listening to frequencies above 28.8 MHz remember to change the sampling mode back to “Quadrature Sampling”.

Note that this feature makes use of direct sampling and so aliasing will occur. The RTL-SDR samples at 28.8 MHz, thus you may see mirrors of strong signals from 0 – 14.4 MHz while tuning to 14.4 – 28.8 MHz and the other way around as well. To remove these images you need to use a low pass filter for 0 – 14.4 MHz, and a high pass filter for 14.4 – 28.8 MHz, or simply filter your band of interest. 

Feature 2: Software Selectable Bias Tee

V.1. and V.2. of our dongles included a bias tee which could manually be enabled by opening the case and soldering two pads on the PCB together. V.3. introduces a bias tee which can be enabled entirely in software.

WARNING: Before using the bias tee please ensure that you understand that you should not use this option when the dongle is connected to a DC short circuited antenna. Short circuiting the bias tee for a period of time could damage the bias tee circuit permanently. Only use it while connected to an actual powered device, like an LNA, active antenna or the Spyverter.

To make things clearer: DC Short Antenna -> LNA -> Coax -> V3(bias tee on) is absolutely fine. What’s not good and makes no real sense anyway is DC Short Antenna -> Coax -> V3(bias tee on). DC Short Antenna -> Coax -> V3(bias tee off) is fine.

To enable the bias tee in Windows:

  1. Download and extract all the files in this zip file to a folder on your PC. It contains two batch files that can be run.
  2. Make sure all SDR software like SDR#/HDSDR/SDR-Console etc is fully closed.
  3. Run the biastee_on.bat file to turn the bias tee on. It will run and open a CMD prompt that will briefly say “Found Rafael Micro R820T Tuner”. The CMD prompt will close soon after upon success.
  4. The bias tee is now on. To turn it off repeat steps 2 & 3, but instead run the biastee_off.bat batch file. Alternatively, simply disconnect and then reconnect the SDR to turn the bias tee off.

If you have multiple dongles connected you’ll need to edit the batch file to specify what dongle’s bias tee you want to activate. Open the bat file with any text editor, like Notepad, and add the dongle selector “-d” flag. For example to activate the bias tee on the dongle that was plugged in second you’d need to change it to “rtl_biast -b 1 -d 1″.

If you get a Smart Screen message, click on More Info, and then on Run Anyway. Also note that some versions of Windows may fail to run batch files due to misconfiguration or aggressive antivirus software. If you cannot fix these problems with Windows or your antivirus, run the command manually on the CMD line.

To run it manually on the CMD line first browse to the directory where the bias tee software is stored using “cd” (e.g. cd C:\SDR\bias_tee_folder), and then run:

  1. ON: rtl_biast -b 1
  2. OFF: rtl_biast -b 0
  3. If needed select a particular RTL-SDR device with the -d flag.

In Linux or MacOS download the source from git, compile it the same way you do the regular RTL-SDR drivers, and then run ./rtl_biast -b 1 to turn the bias tee on and ./rtl_biast -b 0 to turn the bias tee off. The procedure is:

git clone https://github.com/rtlsdrblog/rtl_biast
cd rtl_biast
mkdir build
cd build
cmake ..
cd src
./rtl_biast -b 1

If you want to be able to run the bias tee program from anywhere on the command line you can also run “sudo make install”.

If you have trouble running the bias tee please contact us at rtlsdrblog_AT_gmail.com.

Feature 3: Selectable Clock & Expansion Headers

This is for advanced users who need to daisy chain clocks together for coherent experiments, or need to access other ports. You can either bridge the clock selector the directly with a solder bridge, or solder on a 1.27mm 2×2 header pin jumper.

To add a jumper to the CLK selector header.

  1. Carefully remove the 0 Ohm resistor.
  2. Very carefully solder a 1.27mm 2×2 header onto the clock selector pads.
  3. You can now select your clock input.

How to connect the CLK jumpers:


 The first position allows you to output the dongles clock to the CLK pads. The second position allows you to input an external clock.


An example of CLK daisy chaining is shown below. One dongles TCXO is connected to two other dongles who have disconnected clocks.

CLK3Current Known Issues:

We’re constantly trying to improve our units and we always make note of what issues exist and how to fix them.

Please remember that these units do get hot to the touch when used in warm climates. This is not an issue and is expected. We have improved the thermal bonding and heat transfer between the chips and the metal case. This results in making the metal case hotter, but it keeps the chips much cooler, resulting in better performance. To lengthen their life keep the unit away from direct hot sunlight.

Known V3 Batch 1 Issues:

  1. Increased sideband noise on very strong narrowband signals. This should not be a significant problem as it only affects very strong signals. The hardware fix is to add about 100-220uF of capacitance on the 3.3V power line. Batch 2 will reduce this noise.
  2. The bias tee when turned on adds a large spur in direct sampling HF mode. This may be problematic only if you intend to use a bias tee powered HF LNA in direct sampling mode. This can be fixed by adding about 2.2uF of capacitance to the output of the LDO, before the inductor. Batch 2 will fix this.
  3. The bias tee can be damaged by accidentally short circuiting the output for a few seconds while it is on. This damage only occurs on USB3.0 and USB2.0 ports that can provide up to 1A or more or current. Batch 2 will add a resettable fuse to prevent damage.

Batch 2 and newer (Shipping Now):

No known issues. 

Posted in


  1. Erich

    Use this online tool

    The push pull amp that goes down to 70 kHz is nice. Or us an op amp

    Magnetic loop antennas are nice to use. The two variants are hi Q tunable and reduces the need of filter or a coaxial cable magnetic pickup loop that is wide band. Long wire antennas with balun so are impractical on LW. I recommend the coax cable pickup loop in fig. 4 with at least 4 m coax loop length.

    I did this before rtl-sdr was on the market and it worked perfect from VLF to HF.

  2. Altekalt

    If using Direct Sampling Mode; design a filter for the band of interest, include a preamplifier for this band, make your own bias tee, make a proper antenna and its impedance matching. If you do this you will get an excellent reception with an extra cost of under 50€. If you don’t then accept that you will receive a lot of crap.
    Antenna and balun are easy to make, preamp you can by, filter is a bit harder but not impossible to mend although this is the most important part to get rid of the manure.

  3. jean claude

    I have a RTL-SDR V.3 batch 2 It works fine on vhf uhf. I tried direct sampling ro receive MF/HF
    It works but dial frequency shown is completely wrong. E.G. if I want to receive an AM broadcast station transmitting on 900Khz, reception is good but on the dial the related spike appears around 200Khz, and no spike at all on 900Khz. Similar problems on HF freq,s. How can I cope with this problem?

  4. Loco

    There is any difference in use 50 or 75Ohm cable between antena and Your SDR? Of course antenna will be adopted to selected resistance.

  5. Michel Tossaint

    Dear Lucas,
    I tried to compile the software for the v3 dongle from Github on Mac OS Sierra, but several errors appeared in rtl_test.c. The errors were corrected when I replaced -> by . in some assignments.
    Now I can use the biastee function to drive an active GPS antenna.
    Maybe it will help others…
    gr Michel

  6. Matthias

    I ordered my V3 last week and it is being shipped from China. Will it be Batch 2?
    I want to use the V3 with a LNA4all on a T-Piece and a R2FU (1 antenna for 2 devices). The R2FU can provide Bias-T as well. Could the V3 get damaged when I turn on Bias-T on the R2FU while not using the V3? The DC stop of the LNA4all will not work because the LNA is located on the incoming side and the V3 an R2FU are on the outgoing side of the antenna. Thanks for your advice.

    • admin

      The V3 won’t get damaged from voltage at the input as there is a blocking cap, but to be safe limit the voltage to 5V, as the cap is not rated for much higher voltages. If its higher voltage use a DC block component on the V3.

      Only b2 is shipping now.

  7. Hank

    I had initial problems getting the V.3 to work. Remembered that another piece of gear did not work with ANY usb extension cable – (forgot this time). After plugging directly into a computer port, all is well now….

    • admin

      If the USB cable is of poor quality it can impact the power resistance, and the signal lines. SDRs are especially susceptible to not working with low quality USB cables because they are a high data rate device. Use a high quality cables with large power wires (lower AWG rating).

  8. Lucas Pereira


    In the last month i bought two units of version 3 and today when i received them, one of them does not work! What should I do.

    Windows warns that it does not work properly and can not be activated. This only occurs with one of the units, the other works perfectly.


    • admin

      Hi Lucas, all the units are fully QC checked and known to be working at the factory, but I guess sometimes after bumping around in transit weak links get exposed and some fail. Please contact us at rtlsdrblog_AT_gmail_DOT_com for a replacement.

  9. Petr

    WARNING! If you are using TV drivers, the BIAS is permanently ON – 4,5V in the connector. My dongle was still very hot! So, which componnent desolder?

    • admin

      I’ll check it out. To completely disable the bias tee you could just remove component RF2 (its on the edge of the PCB), and that would break the circuit. But as long as your not using a shorted antenna the bias tee should not be drawing power, even if its on.

      Remember that these dongles get naturally quite hot, because we’ve improved the thermal interface between the hot components and the case. The hotter the case, the cooler the chips inside.

  10. Orrin Winton

    As of today (Sept 29) who has the version3 in stock? Or do i need to wait a while?

    I checked eBay and Amazon, but did not see it.

    I believe i damaged my v3, by my own error, and want to buy another.

  11. Pedja

    Just got RTL/SDR V 3. On VHF it works as expected, but I cannot use it on HF (which is why I bought it).

    After i set it to Q branch and click play it just reports it cannot connect to RTL/SDR device. When I set back to quadrature sampling it again works fine on VHF.

      • Pedja

        Thanks, that seems to be it. I already had driver installed for other SDR dongle but after I reinstalled driver and dongle now works in HF mode. I hear nothing but I guess I will have to put up some better antenna.

        • Daniel

          I haven’t been able to pick up any HF signals unless I plug the rtl up to one of my ham antennas outside. About the only things I’ve been able to pick up with the stock antennas are very strong nearby signals (such as me checking to see what my transmitted signals look like) and FM radio stations.

    • Errol

      I have had my RTL-SRD Tuner, for the past 3 weeks – works great on FM for local stations – but I cannot get HF to work. Please note the tuner has a VHF and an HF SMA input on it.
      For HF I use the IF output stage on my Yaesu FT1000D – which I thought would work well with the tuner, but no go – no joy.
      I would love to hear from someone who has the same type of tuner and the FT1000D – just to exchange what settings to use with SDRSHARP software – please any assistance I really would love to get the unit working on 40Metres my favorite HF section to date.
      Thanks and 73 from South Africa>

      • admin

        If you have one of those units with two inputs, the HF input is probably direct sampling. You need to select the Q-branch in whatever software you are using.

  12. Derek

    Just getting started with SDR and just got my RTL v3 yesterday. Using SDR# I cannot get direct sampling to work. SDR tries to connect but automatically stops. Not too sure if this is a SDR issue or defect with my unit.


    • admin

      If it works on VHF mode, then it is probably a SW problem if it just stops in direct sampling mode. Check that you’re using the default drivers as some of the experimental ones don’t support direct sampling.

  13. George

    Forgive my English. This google translate
    I received from you last third version with dongle rtl-sdr.com. To be honest, I did not expect from him such a good work on HF. On 40 meter Delta Loop needed only attenuator. One drawback – the overall low-pass filter 24 MHz, and thus above 14.4 MHz – the frequency of the mirror. I will make a simple preselector, I think everything will be fine.

  14. Max

    What is the correct wiring of the usb cable’s shield considering that some modification has been done on the new V3 dongle for better filtering? According to the old school the shield should be interrupted before reaching the connector on the dongle and in case the dongle is inserted in a metal enclosure the cable’s shield should touch the enclosure but not the usb connector. Is this still valid?
    My impression is that in this field we are navigating between myth and legend :)

    • admin

      Yes no one can agree on how to connect the shield anymore. We’ve left it connected as it seems to give lower noise, as well as providing better grounding. Your results may vary though.

  15. Mick Cooper

    Has anyone had any tried feed in an external clock and daisy chaining several dongles ?. Do you need any additional components to couple and match impedance etc etc ?

  16. Alan Clark

    Hi. I got my V3 this morning and set up as described with the Q Branch and 2.048mhz sampling rate with a standard long wire HF antenna, but I am absolutely flooded with AM echoes throughout the HF band. I’m unable to adjust gain when I choose the Q Branch. Anyone got any ideas about how I can fix this?

    • admin

      If you have very strong BCAM in your area youll probably need an attenuator. The DSM isn’t designed to handle very strong signals, and it that case an upconverter would be the better choice.

      • Alan Clark

        Hi. Thanks for your response. I’m just North of London in the UK and there are two major MW AM transmitters close. One is 7 miles and one is less less than 2 miles away. I guess that’s the problem! I did buy an attenuator for another project and tried it out but no luck so far. I have an SDRPlay which gives me OK HF stations, but I wanted to see how the new RTLSDR worked out. Withe the SDRPlay, I don’t get good HF signals from about 12-30mhz. Part of the problem is a very noisy Toshiba TV screen.

      • Dirk

        I just want to let you know their is a linux distribution called skywave that is supposedly plug n play for those that know whats what give it a try so you can save me the headache of following all those steps for windows

  17. Tomo

    I have problem in receiving HF and MW with the V3. VHF such as FM radio works fine. I connected the HF antenna to the V3, but the noise level does sound the same as if nothing was connected. I did the same while receiving FM radio, it made a big difference even with an HF antenna. The HF antenna works fine with my some other conventional HF receivers. I think I set the RTL-SDR controller of the SDR# as instructed on this page. Any help will be greatly appreciated.

  18. AM909

    I look forward to being able to receive WWV as a calibration source. If I need to tune to 20,000,016 Hz to get the WWV carrier centered, does that same 0.8 ppm error apply to higher frequencies (e.g. +748 Hz @ 935 MHz), or is there a difference between the direct sampling mode and the IQ mode?

    • admin

      Yes the same PPM will apply on any frequency. As long as you use the PPM offset setting in the program you are using it will map the correct frequency offset to different frequencies.

      • AM909

        Thanks for the reply. Is there any other way of correcting an initial calibration error in the dongles (i.e. permanently)? I have a v1 that is almost exactly only* 0.5 ppm off, which is fine for most applications. However, this produces ~470 Hz error at 935 MHz. SDR#, unfortunately, does not support fractional PPM correction, so I’ve no way to get it any closer (i.e. 0 ppm is too little correction, -1 ppm is too much), other than to manually calculate and enter a shift frequency all the time (and SDR# devs don’t believe a decimal place is necessary, even though it _does_ affect decoding in this narrowband environment).

        * Don’t get me wrong. I’m amazed/thrilled that it’s this close – it gets to that number and stays within ~0.05 ppm of it after just a few minutes’ warmup.

  19. Dave

    I got my V3 dongle yesterday, and started playing with it after the nightly honey do list was done ;-). Anyway, setting up SDR# I was astounded by the HF receive quality without having to use a upconverter as I have to use with my AirSpy Mini.

    After about 30 minutes of listening to ham and SWL bands, I went to disconnect the V3, nearly burnt my fingers. This thing got extremely hot. My other dongle never got as hot as the V3 did last night.

    Now I hoping I didn’t scorch anything on the board. Has anyone else experienced any overheating issues with the V3?

    Dave W8EIR

    • admin

      Hmm it should not be getting so hot as to be painful to touch! What’s the ambient temperature in your location? The dongle actually uses less current in direct sampling mode than in quadrature mode, so this makes me think that there might a a tiny short circuit somewhere, causing excessive current draw. Please contact us rtlsdrblog_AT_gmail.com for some more troubleshooting.

  20. Clifford

    I do not see a way to use q branch in WebSDR software. I am putting up a WebSDR station 3000′ above Los Angeles and bought 2 of the V3 to see if they would work. I cannot afford a bunch of SDRPlay units right now. Anyone have ideas? Thx

    • admin

      Most software has the option to enable direct sampling, but i’m not sure how WebSDR works. You might need to contact the author and ask him to implement a direct sampling option.

  21. Bin Kenney

    I am pretty certain I have an HF antenna, but in case my doesn’t work as expected (18 AWG stranded wire, ~20′), the 1.6m telescopic should work fine, eh? I am thinking about buying/building a 9:1 unun for the HF antenna and the v.3 dongle, and my other dongles using the mod driver. Thanks!

    • Daniel

      Hi, I’m thinking 20′ is about a third of what you might want to use as a longwire antenna. I had one around 100′ ran up into a tree and about a 30 degree angle and it worked great, but any shorter and I lost a lot of the receive capability. They also say you really should have a 9:1 unun, available on amazon. Also, if you haven’t done this already, try running coax outside and up a far as you can before connecting the wire to the end, that might help alot.

  22. Dex

    Hi guys!

    When you plan release second batch (with small bug fixies)?

    I own V.2 but HF capabilites are very interesting for me and of course I want support this project ;-)

    Thanks for your enthusiasm – it is very impressive what can be “finally” done from cheap TV dongle!


  23. Lucas Teske

    Btw admin, do you have the datasheet of the Realtek chip? Can you check what is the latency from clk in to clk out?

    I’m planning to make a bandwidth expander with the RTLSDR Dongles, but this can be broken if the latency is too high in daisy chain.

  24. John AE5X

    I received my V3 RTL dongle yesterday and had no problem getting it going and receiving VHF and up by following the quick start guide. Then I tried to configure it for HF via the instructions here. Device and Sample Rate can be set to indicated options but the Sampling Mode menu option is grayed out. Any ideas why this might be the case? I’m using Windows 8.1.


    John AE5X

    • Craig

      I screwed up the block quote.

      On the issue of increased sideband noise on very strong narrowband signals, how strong are we talking about?

        • Craig

          I have a v3 on order. I do a lot of decoding of local nxdn and p25 signals some of which are very very strong. Can you provide any more info on the hardware fix? I’m hoping that I won’t need it but just in case.

          • admin

            Have a look at the image on feature 3 of the clock daisy chaining on this page. Solder a 220uF cap on the left most 3.3v and GND pins. If all the signals are strong there shouldn’t be a problem for you anyway. A problem might occur if there are very strong and very weak signals right next to each other.

  25. Seasalt

    If you want to run Q Branch direct sampling in Linux GQRX here is a reply I got from the GQRX forum written by Alexandru Csete.

    “Gqrx supports the direct sampling mode by adding “direct_samp=…” to
    the device string:

    Enable direct sampling mode on the RTL chip. 0: Disable, 1: use I
    channel, 2: use Q channel

    So, for using the Q channel the device string would be something like:


    You may have to check “No limits” to allow tuning below 24 MHz, I
    don’t remember.


    This works great on my Soft66Q HF RTL from Japan in Q mode.

Post a comment

You may use the following HTML:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>