Over on YouTube channel TheSmokinApe has uploaded a video about using RF filters with an RTL-SDR. In the video he first explains why FM bandstop and AM high pass filters might be required when using a software defined radio in order to avoid overloading the SDR with very strong signals. He goes on to test and review our RTL-SDR Blog FM Bandstop and AM Highpass filters, by testing them on a spectrum analyzer.
Over on YouTube user ElPaso TubeAmps has uploaded a video showing his tests on our broadcast AM (BCAM) high pass and broadcast FM (BCFM) band stop filters. These two filters are designed to block broadcast radio signals which in some locations can be extremely strong. If they are very strong then they can overload your SDR which causes very poor performance, even on other frequencies.
Some possible solutions for reducing overloading include:
In the video he uses his signal generator and a spectrum analyzer to analyze the output of the filters. His results closely match our VNA results which are posted on the BCFM and BCAM filter product release posts.
Back in October we released a broadcast FM bandstop filter for removing strong signals in the 88 – 108 MHz region. Today we’re releasing a new broadcast AM high pass filter (BCAM HPF) with a 2.6 MHz cutoff. The cost is the same as the BCFM bandstop filter at $14.95 USD including free international air shipping. Faster shipping options may also be chosen if desired. We’ll eventually have this product on Amazon USA in a few months too, but for now it is only available from our Chinese warehouse.
The filter comes in a 2.8 cm x 2.8 cm x 1.3 cm aluminum enclosure and uses female SMA connectors on each end. Included in the package is also a SMA male to SMA male straight barrel adapter.
This filter is designed to eliminate broadcast band AM (BCAM) stations by attenuating (blocking) any signals below 1.7 MHz. In reality due to roll-off the filter is usable from about 2.5 – 2.6 MHz and above.
The broadcast AM band exists at around 525 kHz to 1.705 MHz. These signals are usually local, and if you live close to a transmitter they can sometimes be extremely strong. Broadcast AM signals that are too strong can overload your SDR or radio, causing poor reception in other HF bands too. The filter also helps attenuate any other strong VLF/LF/MW interference. Note that this filter is a high pass and not a bandstop, so it will also block VLF signals. Specifications are shown below:
Filter Type: LC High Pass Filter
3 dB Cutoff: 2.5 – 2.6 MHz
Attenuation: ~60dB
Pass band I.L: Typically well below 2 dB
Power Levels: RX power only, cannot pass DC
We’ve also uploaded a video below that shows a demonstration of reception when using an RTL-SDR.com V3 dongle in direct sampling mode together with the BCAM HPF. In the video we first compare reception against an upconverter (the Spyverter). It’s worth nothing that the upconverter can receive signals well even without the filter in place. Using the filter does still help the upconverter receive a little bit better but the effect is not shown in the video. Then we simply scroll through the spectrum and listen to a few signals.
Over on YouTube channel TheSmokinApe has uploaded a video about using RF filters with an RTL-SDR. In the video he first explains why FM bandstop and AM high pass filters might be required when using a software defined radio in order to avoid overloading the SDR with very strong signals. He goes on to test and review our RTL-SDR Blog FM Bandstop and AM Highpass filters, by testing them on a spectrum analyzer.
Over on YouTube user ElPaso TubeAmps has uploaded a video showing his tests on our broadcast AM (BCAM) high pass and broadcast FM (BCFM) band stop filters. These two filters are designed to block broadcast radio signals which in some locations can be extremely strong. If they are very strong then they can overload your SDR which causes very poor performance, even on other frequencies.
Some possible solutions for reducing overloading include:
In the video he uses his signal generator and a spectrum analyzer to analyze the output of the filters. His results closely match our VNA results which are posted on the BCFM and BCAM filter product release posts.
Back in October we released a broadcast FM bandstop filter for removing strong signals in the 88 – 108 MHz region. Today we’re releasing a new broadcast AM high pass filter (BCAM HPF) with a 2.6 MHz cutoff. The cost is the same as the BCFM bandstop filter at $14.95 USD including free international air shipping. Faster shipping options may also be chosen if desired. We’ll eventually have this product on Amazon USA in a few months too, but for now it is only available from our Chinese warehouse.
The filter comes in a 2.8 cm x 2.8 cm x 1.3 cm aluminum enclosure and uses female SMA connectors on each end. Included in the package is also a SMA male to SMA male straight barrel adapter.
This filter is designed to eliminate broadcast band AM (BCAM) stations by attenuating (blocking) any signals below 1.7 MHz. In reality due to roll-off the filter is usable from about 2.5 – 2.6 MHz and above.
The broadcast AM band exists at around 525 kHz to 1.705 MHz. These signals are usually local, and if you live close to a transmitter they can sometimes be extremely strong. Broadcast AM signals that are too strong can overload your SDR or radio, causing poor reception in other HF bands too. The filter also helps attenuate any other strong VLF/LF/MW interference. Note that this filter is a high pass and not a bandstop, so it will also block VLF signals. Specifications are shown below:
Filter Type: LC High Pass Filter
3 dB Cutoff: 2.5 – 2.6 MHz
Attenuation: ~60dB
Pass band I.L: Typically well below 2 dB
Power Levels: RX power only, cannot pass DC
We’ve also uploaded a video below that shows a demonstration of reception when using an RTL-SDR.com V3 dongle in direct sampling mode together with the BCAM HPF. In the video we first compare reception against an upconverter (the Spyverter). It’s worth nothing that the upconverter can receive signals well even without the filter in place. Using the filter does still help the upconverter receive a little bit better but the effect is not shown in the video. Then we simply scroll through the spectrum and listen to a few signals.
Over on the SDRSharp Yahoo group, HB9AJG has posted an interesting report in a PDF file containing some measurements (Note you will need to be a member of the group to download the file titled "Some Measurements on E4000 and R820 Tuners.pdf". Here is a Direct Mirror of the file.) quantifying the performance of both the E4000 and R820T RTL-SDR DVB-T dongles. See the discussion in the Yahoo group here.
These results confirm the feeling that many RTL-SDR users have had: that the E4000 is more sensitive in the lower frequencies, and that the R820T is more sensitive in the higher frequencies, which is why it is recommended for ADS-B. The results also show that the R820T is better in terms of Image Rejection and Internal Signal Birdies.
He comprehensively summarizes his results in the following
Image Rejection
Because the E4000 is a Direct Conversion Receiver, it has an Image Rejection problem. By switching on Correct IQ in SDR# a more or less acceptable 50dBs are reached. For the same reason, a "hump" shows in the center of the spectrum display. By using a well filtered external power supply (not from the USB connector) the hump might be reduced.
Internal signals
The E4000 shows many signals actually not present at its input ("birdies"). Birdies are easy to recognize: most of them (except the harmonics of the clock) vary their frequency when moving the spectrum window in frequency. Many of them even move up if you move the window down in frequency.The R820 is much cleaner in this respect: besides the harmonics of the clock (28.8MHz) only few birdies show up.
Sensitivity
Both dongles have a very high sensitivity. Between about 50 and 450MHz the E4000 is about 5dB better than the R820 (-139dBm vs -134dBm). At 1000MHz the E4000 is about 8dB less sensitive (-129dBm vs -137dBm). No measurements could be made above 1040MHz.Overload and 1dB Compression
If a signal is strong enough, it may cause overload, i.e. many (unwanted) signals show up on the spectrum display that are not present at the antenna input. Also, if we listen to a desired signal, another signal (if strong enough) may cause a reduction of the S/N ratio of the desired signal.Both dongles have a digitally tuned RF filter after the preamplifier that (together with the following digital signal
processing) improves the overload/1dB compression limit considerably.- The filter of the E4000 is about +/-0.8MHz wide, but less steep than the filter of the R820.
- The filter of the R820 is about +/-3MHz wide, but steeper than the filterof the E4000.For the E4000 the overload/1dB compression limit is not linearly dependent of the gain set in Configuration of SDR#:
if the gain is reduced by 13/20/30dB, the overload limit is improved by only 7/14/25dB (measured on 145MHz only).For the R820 the overload/1dB compression limit is quite linearly dependent of the gain set in Configuration of SDR#: if the gain is reduced by 11/20/30dB, the overload limit is improved by 12/20/30dB (measured at 145MHz only).
For both dongles it seems there is nothing to be gained from activating RTL AGC or Tuner AGC.
Intermodulation
Intermodulation products in general show up close to the overload/1dB compression limits. However, if the strong
signal is on the roll off of the filter, they appear well before this limit.Aliasing
Aliasing always occurs if an insufficiently band limited signal is sampled, i.e. if the signal to be sampled contains frequencies above half the sampling frequency. Thus, aliasing is an effect showing up in many SDRs, not only in these dongles. In both types of dongles there is not much space for brick wall filters. Therefore, aliasing effects are well visible with both dongles.What do we learn from these tests?
- Both types of dongles are very sensitive. The choice depends on which frequency range you are most interested in.
- Considering internal signals and image rejection, the R820 is much cleaner than the E4000.
- Set the spectrum display of SDRSharp to show a range of not more than 60db above the noise floor. If a signal is close to the top, you know you are close to overload.
- Both types of dongles are prone to overload by strong signals within their filter bandwidth: +/-0.8MHz for the E4000, +/-3MHz for the R820. Therefore, keep signals within this bandwidth to not more than about 60dB above the noise floor
by reducing the gain. If increasing the gain does not audibly increase the signal to noise ratio of the desired signal any more, reduce the gain by one step. Do not switch on RTL AGC or Tuner AGC, as it seems there is nothing to be gained.- Outside their filter bandwidth both types of dongles can live with much higher signals without showing serious degradation. Use the gain control as explained above to check a possible reduction of signal to noise ratio of the desired signal or the appearance of "new" signals not present at the antenna
input.- Intermodulation occurs if several strong signals are present within the bandwidth of the dongle. Their individual power adds up (add 3dB per equally strong signal). Therefore, in frequency bands with many strong signals, e.g. broadcast bands, the gain must be reduced even further. Watch for "new" signals appearing when increasing the gain, and then reduce the gain by one step.
- If very strong signals are present at the antenna input >-40dBm), they should be attenuated by bandstop or notch filters.
