Nothern Utah WebSDR Logo - A skep with a Yagi Northern Utah WebSDR
Receiving equipment

Receiving equipment used at the Northern Utah WebSDR:

There are two general types of receivers being used at the Northern Utah WebSDR:

SoftRock receivers:

These receivers are the "High Performance" receivers used at the Northern Utah WebSDR and are based on the so-called "QSD" (Quadrature Sampling Detector) mixers (sometimes known as a "Tayloe" detector) 1  2  3  4 that use analog MUX switches to provide the "mixing" action.  This mixer topology is well-used in commercial amateur gear - notably more recent Elecraft gear - and it has the advantage of being both simple and capable high-performance as well as being easy to interface with conventional hardware - such as a standard audio-frequency sound card - for final digitization.  Essentially a direct-conversion receiver, the RF energy that is present +/- the center (local oscillator) frequency is digitized up to the Nyquist limit (half the sample rate).  The receiver itself has two conversion channels that are identical aside from the fact that the local oscillators are 90 degrees out of phase with each other ("I" in-phase and "Q" - quadrature) which, with a bit of math (addition and subtraction and phase-shifting) allow the entire spectrum above and below the center frequency to be represented in software.
Figure 1:
Three SoftRock Ensemble II receivers in service.
Click on the image for a larger version.
The three Softrock Ensemble II receivers on the RF shelf

There are two "SoftRock" receiver types in use on this WebSDR system:
For our purposes, the two kits function identically in their role in converting RF to audio as the mixer and audio amplification circuitry of the two are pretty much the same.  The "Softrock Ensemble II" kit has the obvious advantage of having a built in, tunable local oscillator - but it is a much more expensive and complicated kit.  The Si570 synthesizer, while convenient to use, has the disadvantage that it is not particularly stable with frequency - indeed, the Elecraft KX3 uses a temperature sensor and a computer lookup table to maintain frequency stability, but the Ensemble II lacks this so its absolute frequency stability can be affected by changes in ambient temperature.  On lower bands like 160 meters this is not much of an issue, but on a higher band like 15, 12 or 10 meters this can amount to several 10s of Hz change.  Because we are using an "outboard" synthesizer for the "Softrock II Lite" kits, we were able to substitute a TCXO (Temperature-Controlled Crystal Oscillator) to obtain excellent frequency stability - more on this later.

Initially, we used three Softrock Ensemble II receivers for 160, 80 meter phone and 75 meter phone coverage, but a later upgrade included the construction of a module with three Softrock II Lite receivers to cover the entire 80/75 meter band, freeing those two units for coverage of other bands as described below.

Both the "Ensemble" and the "Lite" receivers feed the RF directly into the mixer (after filtering, of course) which is then passed to low-noise audio amplifiers - but this also means that these receivers are a bit "deaf".  For low bands like 160 and 75 meters - where both noise and signal levels are quite high on a "full-sized" antenna like a dipole this isn't too much of a problem, but on higher bands where the intrinsic noise is lower - and the losses of the receivers' circuitry is higher - it is increasingly important that some sort of RF amplification be used:  This will be discussed later.

Receiver modules using the Softrock II Lite receivers:

Figure 2:
Inside the dual 40 meter receiver module.  In the center is power supply filtering (on the left) and a passive 2-way signal splitter (on the right) with the pair of Softrock II Lite receiver modules on each side.  In the center at the top is the ProgRock synthesizer, configured for two outputs.  The receivers have been slightly modified to accept an external frequency source in lieu of the original quartz crystals.
Click on the image for a larger version.
Inside the dual 40 meter receiver
As noted above, the "Softrock II Lite" kit uses quartz crystals and the frequency selection is rather limited, so if we wish to have flexibility in our frequency coverage we'll need to use something else to provide our local oscillator.  The device chosen for this role is the inexpensive ($18) "Progrock" kit sold by QRP Labs (link).  As the name implies, this functions as a sort of programmable crystal (e.g. "rock") and these devices, based on the Silicon Labs Si5351a (the same chip used to provide the local oscillator of the Elecraft KX-2) can easily cover from about 8 kHz to well over 150 MHz - and may be coaxed even lower/higher than that if one pushes it beyond its official specs!  What's more, these devices can output up to three programmable frequencies at once (with some limitations) which means that a single unit can provide the local oscillators for up to three different receivers.

As delivered, these devices come with inexpensive quartz crystals that are prone to drift with temperature, but for less than $3 they can be retrofitted with a TCXO (Temperature Controlled Crystal Oscillator) that "nails" down the frequency with part-per-million stability, giving them better frequency accuracy and stability than many commercial HF rigs.  The ProgRock kits are typically programmed using DIP switches and a pushbutton for frequency entry using binary-coded decimal (BCD) but newer versions may be programmed via an asynchronous serial port.  For our purposes, only the DIP-switch programming is used as there is no reason to be able to change the center frequency of a receiver remotely.

Figure 2 shows the dual 40 meter receiver module.  On either side is a Softrock II Lite module built for 40 meters with the ProgRock synthesizer at the top, two outputs used to feed the receivers which have been slightly modified (the addition of a single capacitor, the removal of a different capacitor) to accept an external input in lieu of the original quartz crystal.  The ProgRock itself has been modified to use a TCXO as its frequency reference to provide a stability of approximately 1ppm over a wide temperature range which translates to an on-frequency stability of about +/- 7 Hz.  It's worth noting that the local oscillator frequencies being fed into the receiver modules operate at four times the center frequency of the receiver because there is digital divide-by-four circuitry to provide the quadrature local oscillator feeds for the QSD mixer.

In the lower-center, constructed "Manhattan" style - is a power supply filter on the left and a 2-way passive RF signal splitter on the right, the latter sending equal amounts of received RF to each of the two receivers.  As noted above, additional RF amplification is used to bring the signal levels up a bit and this will be discussed in greater detail later.

Potential issues with spurious signals:

As mentioned above, the ProgRock is capable of producing up to three independent frequency outputs at once, all from the same (tiny!) Si5351a chip - but there is basis for some concern if one does this.  Using just a single output, the signal is actually quite "clean", spectrally - far cleaner in non-harmonic content than that typically obtained using a typical DDS synthesizer module - which is why this same chip is the basis for many commercial and kit radios these days.  If more than one output is used, the spectral purity of the Si5351a chip suffers - but how much?

In the case of the dual receiver - where two outputs are used - if a single, strong CW signal were injected into the receiver (say, -20dBm - a "50 over" signal) a number of low-level spurious responses could be seen in the receiver's waterfall, the worst being on the order of 70dB (or more) down.  What this means is that if a "20 over" signal were present on the input (about -53dBm) a signal of S-1 or lower would result - but this would be at or below the noise floor on nearly any HF frequency, anyway.  Whether or not you might think that this was poor performance it's worth pointing out that many HF receivers do have similar spurious signal performance numbers, these "deficiencies" going unnoticed by the casual user - particularly on a busy band.
Figure 3:
Top:
 Close-up of the 2N5109 RF amplifier - one of three
 contained within a single amplifier module.  A later
modification of these amplifiers included the addition of a 2dB resistive pad on the output to assure unconditional stability
with reactive sources/loads.
Bottom:  Schematic diagram of the amplifier module
Click on an image for a larger version.
Close up of one of the broadband RF amplifier sections
Schematic diagram of the amplifier module

RF amplification:


As noted previously, these "SoftRock" receivers - with no active devices in the signal path prior to the mixer - tend to be a bit "deaf".  In theory, their audio outputs are pretty low-noise with microvolt-level RF signals appearing nicely at the output, but the problem comes about when interfacing these same low-level audio signal to sound cards.  A typical good-quality sound card (such as those in the Asus Xonar series) is able to "see" weak signals like this - but there are two other issues that tend to show up:
By boosting the RF signal a bit the two noise sources can be submerged by RF noise coming in from the antenna - but there is a delicate balance:  Too much RF gain and the high-signal performance of the receiver will suffer, and too little gain, weak signals are lost in the noise.  "Barefoot" (e.g. without any amplification) these SoftRock receivers will start to saturate/clip at about -12 to -17dBm (a signal level of about "60 over") which means that one could "safely" add another 10-15dB gain without much worry about a few very strong signals, or the cumulative RF power of many signals on a band, causing overload.  For example, if the receiver were to start to overload at -25dBm (about "50 over") it would take about 100 "20 over" signals on the band (not including overall noise) to attain this much signal power:  While not impossible, this is unlikely to happen - even during contests.

To be "safe", one must keep in mind the following for any RF amplification that is to used:
A useful article on this topic (among many) is one written by Gary, WB9JPS 5  where he discusses various requirements of signal amplifiers, including gain and noise figure.   (An article that comes to similar conclusions is one written by AB4OJ - see reference 8.)   Gary's conclusion - which is not unique to this paper - indicates that conservative system noise figure requirements of HF receive systems are modest and along the lines of:
The amplifiers discussed in the article by WB9JPS reminded me of a November, 1984 article in Ham Radio magazine by Joe Reisert, W1JR 6 , where various topologies of amplifiers using the venerable 2N5109 transistor are discussed- a device designed specifically for broadband, low noise, linear operation and is, more importantly, still readily available!  While both the W1JR and the WB9JPS articles describe amplifiers with better signal-handling performance and lower noise than the common-emitter configuration that I used (see Figure 3 in the WB9JPS article, which references a design by W7ZOI) the performance of the amplifier is quite good and more than adequate for the task at hand.

Figure 4:
Top:  The dual 20 meter receiver.  This is very similar to the 40 meter
receiver except that there is an RF amplifier for each receiver to increase isolation of the LO bleedthrough between the two.
Bottom:  The schematic diagram of the dual receiver module.  This
module could be used on any HF amateur band -
it is only the receivers and the programming of the LO that
dictate the frequency of operation.
Click on the image for a larger version.

Inside the dual 20 meter receiver
Diagram of the dual receiver with amplifiers
A set of three of these amplifiers were constructed and housed in a Hammond 1590D die-case enclosure using the circuit depicted in Figure 3 on this page.  Between each amplifier is a "wall" of double-sided, glass-epoxy circuit board material and each amplifier was built on its own, private board using "Manhattan" ("dead bug") techniques using "Me Squares" sold by QRPME (link) that were (literally!) glued down using high-quality cyanoacrylate adhesive (e.g. "super" glue.)   The usable frequency range of these amplifiers is on the order of 50kHz through 200 MHz, but they are flat to better than 1dB over the range of 1.5-30 MHz.  In testing these amplifiers they maintained very linear output (e.g. negligible intermodulation distortion) at power levels over +20dBm (100 milliwatts).  Even though there are three amplifiers in the same enclosure, the isolation between them was around 100dB at 10 MHz degrading to around 80dB at 30 MHz. 

The idea behind three amplifiers in one enclosure was that they could be used as general purpose gain blocks:  If extra gain was needed somewhere, these would be available to provide it - and upon installation of the WebSDR, one section was used with the 160 meter "Softrock Ensemble II" while another section was used for the dual 40 meter receiver module depicted in Figure 2:  This added gain (13-15dB) was about right to allow the receiver to "see" the noise floor during daylight hours, but not so much that receiver system performance was compromised even when there were a lot of "big" signals during contests.

Additional receiver modules:

When the WebSDR system was first brought online the three Softrock Ensemble II receivers were used to cover a portion of 160 meters and most of the phone portions of the 80/75 meter band.  At this time additional equipment was in the works that would be used to provide coverage of allof 80/75 meters in three chunks and the entire 20 meter band in two - all of these using SoftRock Lite II modules.  As new modules were constructed using the Softrock Lite II receivers, the Softrock Ensemble II, being capable of being tuned anywhere, became available for general use - such as providing coverage for "new" bands or being used as a spare receiver in case of some sort of equipment failure.

20 meter coverage:


By observation, it was known that amplification would be required for the 20 meter SoftRock receivers so it was designed from the beginning to include it - but there was a bit of a twist:  One issue related to any receiver using a QSD mixer is that it can have a "significant" amount of local oscillator energy appearing on the antenna port, and on the 20 meter modules this signal level was on the order of -33dBm, or about "40 over" S-9.  To the receiver itself, this amount of signal is irrelevant as it cancels out and doesn't appear as a strong "zero Hz" component, but on the dual receiver modules the local oscillator for one receiver appeared in the other and this "big signal" could have potentially degraded performance - mostly in the form of a strong, off-frequency signal that could mix in various ways with low-level local oscillator spurious signals and the myriad of signals that might appear on a "busy" band.

A degree of isolation (15-20dB) between the receivers was provided by the passive 2-way splitter, but it was decided that each, individual receiver would sport its own, private RF amplifier, adding another 20dB or so of isolation, the end result being that one receiver would "see" a signal of -60dBm (about "10 over") or less from the local oscillator of the other.  Because these amplifiers are relatively simple - and the parts cheap - the construction of the added circuitry was not much of a burden.

Figure 5:
Top:
 The "triple" 80/75 meter receiver module.  Because there
are three receivers, the physical layout is different from the
dual receiver modules and like the 20 meter receiver, each
receiver has its own, private RF amplifier - both for gain
and LO isolation.
Bottom:  The schematic diagram of the triple
receiver module.  Like the dual receiver, it's only the receiver
itself and the programming of the synthesizer that determine
the HF band on which it operates.
Click on an image for a larger version.
The triple Softrock receiver module with RF amplifiers.
Schematic diagram of the triple Softrock receiver with amplifiers
Figure 4 shows the completed receiver module.  It is nearly identical to the dual 40 meter receiver module in that it uses a single ProgRock synthesizer to to provide the local oscillator signals for both receivers.  If you look closely, you can see that there are two transistor amplifiers on the right-hand side of the copper-clad board in the bottom-center, following the 2-way splitter and each one feeding a receiver.

In testing on the workbench, the "MDS" (Minimum Discernible Signal) of each 20 meter receiver was better than -127dBm (e.g. 0.1 microvolts in a 50 ohm system) indicating that they were as sensitive as they needed to be:  In other words, this receiver was more than capable of hearing its fair share of ionospheric noise when connected to an HF antenna and as such, more sensitivity would not improve the ability to "hear" weak signals!

80 and 75 meter coverage:

It is somewhat inconvenient that most amateur bands are sized in multiples of 100 kHz but audio sound cards have sample rates of 96 or 192 kHz.  In the case of the U.S. 40 meter band we would need two 192 kHz sound cards to fully-cover the 300 kHz-wide band.  A somewhat similar situation exists on the U.S. 80/75 meter band, which covers from 3.5 to 4.0 MHz where we would need three receivers to cover the entire band.  It is convenient that the ProgRock can output three simultaneous outputs, so another receiver assembly was constructed using three Softrock Lite II modules.

The picture in Figure 5 shows the layout.  In the upper-left corner is the ProgRock synthesizer and like its counterparts, it, too is equipped with a 1ppm TCXO.  Below it are the three, identical Softrock Lite II modules and to the right of those, one for each receiver, are three RF amplifiers.  In the upper-right corner is a three-way splitter that divides the signals to the receivers equally and provides a bit of additional LO isolation between the receivers and on the other side of the divider is the same type of power supply filtering found in the other receiver modules.

Figure 5 also shows the diagram of the triple receiver module, the circuitry being representative of that in the other two modules.  If you look carefully you will notice that the RF amplifiers in the receiver modules are slightly different than that depicted in the WB9JPS article - and is, in fact, a direct "quote" of one of the amplifiers discussed in the November 1984 W1JR article.  The main difference is that these amplifiers lack the output balun/transformer which somewhat reduces the large-signal performance, but because these amplifiers are placed "downstream" bandpass filtering that is specific to an amateur band, they will not be "seeing" much of the HF spectrum and will be dealing with fewer signals, overall.


RTL-SDR dongles:

The "other" receivers - the ones that are not considered to be "high performance" - use the so-called RTL-SDR dongles.  These USB dongles are ubiquitous and versatile:  They can cover (more or less) from a few hundred kHz to over 1.3 GHz using various on-device signal paths - but all of these signal paths have in common one important limitation - The A/D converter is only 8 bits.  Despite these limitations, they are attractive because they are cheap - from $4 for the "bottom end" and cheapest devices (which are far noisier than they could be) to over $50 for units with frequency converters and a few other bells and whistles - including band-pass filters.  The devices that we are using are just $20 and are the RTL-SDR dongles sold by "RTL-SDR Blog":  These units have thoughtfully-designed circuit boards that minimize extraneous, spurious responses and include 1ppm TCXOs for decent frequency stability as well as providing separate signal branches for "direct" and "quadrature" signal paths - but more on that later.

Ideally, the maximum range represented by an 8 bit A/D converter is around 48dB - and this is approximately what can be expected from these devices, but as with most things in the real world, the actual answer to the question "what is the dynamic range" is more complicated.  In reality, noise considerations of the device reduce the number of usable A/D bits and thus the dynamic range, this noise coming from the device itself and other devices in the signal path.  When used "on air" their effective dynamic range can often "seem" to be greater than the 40-50dB that one might expect, and this can be due to several factors:
Even with all of these effects, their useful range is quite limited which means that if there are both very strong and weak signals being digitized by the dongle's A/D converter, you are faced with a choice:  Decrease the gain to prevent the strong signals from badly overloading it or increase the gain to allow reception of the weaker signals, but suffer the effects when strong signals appear.  If one uses these dongles it is imperative that one avoid slapping it on an antenna, but include in the signal path a band-pass filter that limits the signals getting into it to those frequencies around the range of interest.  In the case of a WebSDR we do this because we set up a receiver to cover a specific range and it never needs to be tuned anywhere else.

The two signal paths within the dongles:

The RTL-SDR blog dongles that we are using have two entirely separate signal paths:
Some RTL-SDR dongles include a frequency up-converter that takes the HF frequency range and presents it to the dongle in the 125-155 MHz range, but the RTL-SDR dongles that we are using do not have this feature, so we are using the "direct" branch which has certain frequency limitations due to the 28.8 MHz sample rate.  As noted earlier, the Nyquist frequency - the maximum frequency where we can faithfully digitize the input signal - is 14.4 MHz and this means that we can use it to directly "receive" bands up to 30 meters:  20 meters is problematic because the top end of this band - 14.35 MHz - is only 50 kHz away from the 14.4 MHz Nyquist frequency and making a practical filter to remove the images at 14.45 MHz and above that would appear in the 20 meter band is very difficult to do!

At lower frequencies, such as the AM broadcast band, we can receive signals directly - but the problem of dynamics rears its head again:  If one is located near an AM broadcast band transmitter - or near a metro area where there are several of these transmitters - the signals from these AM stations can vary over 60dB, from the weak "nearby" stations to the very strongest - a range entirely outside the capability of the dongle itself unless certain heroic measures are taken.  (For an article on this, see reference #7, below.)    This article describes a means of attenuating signals within the AM broadcast band - with additional "notching" of the strongest signals - while preserving sensitivity on the adjacent 160 meter band, generally keeping all signals within the usable dynamic range of the RTL-SDR dongle.

At higher frequencies, things are a bit easier to manage in that one simply constructs a band-pass filter for the frequency range of interest.  Using a filter design program like Elsie (link) which has a free (limited)"student" version, one can input the desired center frequency and bandwidth to yield simple - but adequate - designs that can be realized using standard-value capacitors and easily-wound toroidal inductors.  Alternatively, you can get band-pass filter kits from QRP Labs (or simply "borrow" their published designs) that can be easily adapted to nearly any HF frequency range.  In this case it is useful to precede the RTL-SDR dongle with a bit of excess gain and provide an attenuator (typically post-filter) that can be adjusted to find the "sweet spot" where the probability of overload from strong shortwave broadcast stations is minimized and weak signals can (usually) be heard.  This method is used for 60-49 Meter, 31-30 Meter, 25 Meter and 19 Meter coverage on the Northern Utah WebSDR with reasonable effectiveness.

Using "direct" mode, these dongles effectively have a "hole" which excludes direct coverage of 20 and 10 meters owing to the aforementioned Nyquist and filtering limitations - and as is the case with lower frequencies which means that if we wish to cover the 20 and 10 meters without being plagued with images, a frequency converter (with appropriate filtering) is required.  If a more expensive dongle with a built-in frequency converter were chosen, one would want the type with TCXOs to minimize frequency drift, which can be greatly magnified because these converters - and the tuner itself - operate in the 100-200 MHz range.  It is quite practical to build a frequency down converter to yield comparable results as described later in this page.  For example, if one were to mix 10 meter signals with a 20 MHz oscillator the result would be a conversion of the 28.0-29.7 MHz range down to 8.0-9.7 MHz:  By down-converting, frequency drift of the various oscillators is dramatically reduced as all of the frequencies involved are about an order of magnitude lower than they would be with a 100+ MHz up-converter.


Figure 6:
Block diagram of the HF splitter/AM BCB filter/AMP and the "Low" and "High" splitter modules.
Click on the image for a larger version.
Block diagram of the AM BCB splitter-amplifier and low HF splitter
RF Filtering:

RF splitting and filtering:

To achieve our goal, we decided from the outset that we should make the receive system capable of receiving on every HF band, but to do this we'd need a lot of outputs, as in:
  1. 630 meters
  2. 160 meters
  3. 80/75 meters *
  4. 60 meters 
  5. 40 meters *
  6. 30 meters
  7. 20 meters *
  8. 17 meters
  9. 15 meters *
  10. 12 meters
  11. 10 meters *
* - Multiple antenna connections needed if narrow-band "Softrock" type receivers sound cards are used.

One way that we could have done this would have been to use conventional transformer-type splitters to divide the signal and the simplest way  - to divide it by 16 - would have yielded about 20dB of insertion loss - and the above doesn't take into account that for many of the HF bands (those marked with an asterisk) we'd need several antenna connections to feed enough receivers to cover many of the bands if we use "high performance" receivers that can provide only up to 192 kHz of coverage.

With this method there are two other problems with which one must contend:
While it is certainly possible to make this scheme work, there's another method:  Use a "diplexer" type splitter.

Diplexer splitting:

A "diplexer" type splitter minimizes the insertion loss by selectively "picking" various bands from a common bus.  By having a filter that pulls only narrow ranges of frequencies of individual amateur bands - but leaves the other frequencies alone - we can put several of these same filters on the same bus and instead of 15-20dB of insertion loss from cascaded splitters we can easily keep the loss down to single digits of dB.  The idea is simple - but we decided early on that this wasn't going to be our only approach.

The receive signal path (from the antenna) was designed from the outset to be both versatile and high-performance with the following goals in mind:
To accomplish this several modules were built, depicted in Figure 6:
Figure 7:
Top:
 The schematic of the "Splitter/AM BCB Reject/Amplifier" module.
Upper Middle:  The schematic diagram of the "Low HF Splitter" module.
Lower Middle: The diagram of the "High HF Splitter" module.
Bottom:  The diagram of the splitter/low-pass/BPF module for RTL-SDR receivers.
Click on an image for a larger version.
The AM BCB filter/splitter module schematic
The low HF splitter schematic
The High HF splitter schematic
The BPF/LPF/Attenuator for RTL-SDR receivers
Also depicted in Figure 6 is another module, connected to the output of the Splitter/BCB filter module, that feeds two RTL-SDR dongles.  As required for best performance, these devices should have their inputs filtered to pass only the frequency range of interest and the diagram shows this being done:  A 3 MHz low-pass to accommodate the receiver that tunes 630 through 160 meters (including the AM broadcast band) and a 4.5-7 MHz band-pass filter for the receiver that tunes the 60 Meter SWBC and amateur frequencies and the 49 meter SWBC bands.  This module also has adjustable attenuators that are set to the "sweet spot" - that is, just enough attenuation to prevent serious overload by strong signals and not so much attenuation that weak signals cannot be heard.

At first glance it might seem that placing a splitter at the input of the system and losing 3dB "off the top" would be a bad idea, but this ignores a fundamental truth about HF signal reception:  As noted above, the HF frequency range is very noisy, which means that we can tolerate quite a bit of loss (and incur a rather high system noise figure) in front of our receivers without actually degrading overall system sensitivity.  This simple fact can be demonstrated by connecting a highly-sensitive receiver to a full-size receive antenna and experimenting with a step attenuator and noting the amount of attenuation required to quash the atmospheric noise.  Typically this value, on an antenna devoid of man made noise under normal "quiet", HF conditions, implies that an acceptable system noise figure ranges from about 45dB at 160 meters, decreasing to 24 dB and 15 dB and 20 dB and 10 meters, respectively 5.  What this means is that even if we end up with 6 dB of added loss in our HF signal path through splitters and filters, it is still possible to recover the natural noise floor on HF without requiring any sort of exotic, low-noise amplification.

Band-pass filter/attenuator modules for the RTL-SDR dongles:

If you've been reading along you'll already know that it is imperative that RTL-SDR dongles used on HF (or anywhere else) MUST have filtering of some sort on their RF input:  It's not just the signals in the frequency range of interest that are "seen" by the A/D converter when operating in "Direct" mode, but all signals at all frequencies.  In order to maximize what (little) signal handling capability these devices have, it is required that effective filtering be used.

As mentioned previously, one must also provide a means of adjusting the RF single levels being applied to the input of an RTL-SDR dongle, trying to find the "sweet spot" where there is enough attenuation to prevent overload by strong signals yet there is enough overall system gain to receive weak signals.  This balancing act can be quite tricky - particularly when one considers the number of signals and that the strength of those signals vary dramatically between day and night.  At the Northern Utah WebSDR, we are "fortunate" in that there are no strong shortwave broadcast stations "nearby" that beam their signal in our direction - but in Europe and eastern North America, the story can be quite different, with multi-hundred kW stations being beamed in your direction and only one "hop" away!

The diagram of the filter module is shown in the bottom of Figure 7 and enough information is provided for several options.  A two-way splitter is depicted on the diagram to allow the feeding of two separate RTL-SDR dongles and their filters while off to the side, a 3-way splitter is shown.  If a 4-way splitter were required, one would cascade a pair of 2-way splitters after a single 2-way splitter (for a total of 3 splitters) - but as noted on the diagram, each set of splitters would incur a loss of about 3.5dB.  If no splitting is required, these would simply be left off.

The upper portion of this diagram also depicts a filter suitable for use on the AM and 160 meter bands.  The left-hand portion is a 500 kHz high-pass filter that removes potentially strong LF signals and noise while the right-hand portion cuts off signals above approximately 2.5 MHz.  On the output of the filter is a very simple attenuator that is used to adjust the signal levels being fed to the RTL-SDR.  Using a single potentiometer, this attenuator is not a "constant impedance" device, but it does provide an "approximate" load for the filters to preserve their general characteristics.  In reality, the RTL-SDR really doesn't care about its input impedance, and at HF frequencies with fairly short cables, it's not all that important, either!

Also depicted in the diagram is a band-pass filter along with the same attenuator seen in the low-pass portion.  The design of this band-pass filter is one that is "borrowed" from the QRP Labs web site, from their "Band Pass" filter products (a link to that page is here).  In the assembly manual - which may be found on that web page - you will find a technical description of the filters (along with some representative band-pass plots) that provide enough information for you to build your own filters.  If you wish, you may buy these modules in kit form (and I can recommend that any of the kits sold by QRP Labs are worth getting!).  If you plan to cover a frequency range that isn't shown - such as a shortwave broadcast band - these filters can be tuned/modified from the nearest amateur band.

As noted previously, these RTL-SDR modules are somewhat "deaf" so it is likely that some sort of RF amplifier will be required - particularly to provide the bit of "excess" signal that one would need to be able to adjust levels downward again:   Any of the 2N5109-based amplifier modules described earlier in this page will fit the bill nicely.

Finally, remember that RTL-SDR dongles in the "direct" mode aren't really all that well-suited for covering the 20 or 10 meter bands owing to the Nyquist limitations - and reception on frequencies between these bands (e.g. 17, 15 and 12 meters) will suffer a bit owing to decreased sensitivity and the increased tendency for spurious signals to appear.  On 20 and 10 meters one would be better off using a dongle that includes an "up converter" - or build a simple "down converter":  In any case you will always want to use a band-pass filter in front of the RTL-SDR dongle's receive system to maximize its performance!


A downconverter for RTL-SDRs:

It is quite common for HF coverage via an RTL-SDR dongle to be achieved through the use of an upconverter  - a device that takes the HF spectrum and shifts it up by 100, 125 or even 200 MHz.  The reason for doing this is easy to understand:  Continuous coverage across the HF spectrum is afforded, which is very convenient - but one must still take care when doing this:
Another method of providing HF coverage is with the use of a frequency downconverter.  This is fundamentally different from the upconverter in that instead of shifting the entire HF range upwards by some amount, one takes the narrow slice of interest and converts it down to a lower frequency.  Doing this solves/minimizes several problems, such as:
If the RTL-SDR can receive HF in its "Direct" mode, anyway, why would you use a downcoverter?  As mentioned previously, the A/D sample rate of the RTL2832 chip is 28.8 MHz, which means that both 20 and 10 meter coverage via this mode has the problem of being too close to the Nyquist limits:  In the case of 20 meters, half of the sample rate (14.4 MHz) is just above the top of the band at 14.35 MHz and images cannot be easily filtered.  In the case of 10 meters the 28.8 MHz sample rate lands right in the middle of 10 meters which means that even if you could build an effective filter to suppress images, you'd only be able to effectively cover a small-ish portion of the band.  At the various bands in-between 20 meters (17, 15 and 12 meters) coverage is possible, but good filtering is still required - and the undersampling means that the sensitivity will be a bit worse than it would be at lower frequencies, presuming that the low-pass filtering in the dongle on the "Direct" branch weren't an issue.  In short, in "direct" mode the RTL-SDR dongle works best from around 1 MHz up to around 12 MHz:  Above this, it becomes increasingly difficult to construct anti-aliasing filters that are both simple and effective.

A converter has the obvious disadvantage that it isn't as convenient:  You probably can't just go out and buy a downconverter like this - and the design considerations require that one pick the frequencies of the local oscillator and down-converted frequency range a bit carefully with respect to the intended coverage.  For example, the local oscillator's harmonic(s) should not land in or very close either the input frequency (the one to be down-converted) as this will result in a very strong signal that could result in intermod products (e.g. birdies.)  The other issue - the down-converted frequency output - should be carefully chosen so that its harmonics are several MHz away from the input frequency coverage as this, too, would result in "birdies" or other undesired responses.

A downcoverter for 15 meters:

As a matter of convenience, I chose a 10 MHz local oscillator frequency because inexpensive TCXO devices are readily available, and it adequately meets the design criteria:
A diagram of the as-built downconverter may be seen in Figure 8.

Figure 8:
Top:
 Inside the 15 meter downconverter for RTL-SDR devices, using the "Direct" branch input
Bottom:  Schematic diagram of the converter
Click on an image for a larger version.
Close up of one of the broadband RF amplifier sections
Schematic diagram of the 15 meter downconverter

Circuit Description:

As with the diagram depicted in the bottom frame of Figure 7, above, the first input stage uses a bandpass filter "borrowed" from the QRP Labs web site, from their "Band Pass" filter products (a link to that page is here).  In the assembly manual - which may be found on that web page - you will find a technical description of the filters (along with some representative band-pass plots) that provide enough information for you to build your own filters.  If you wish, you may buy these modules in kit form (and I can recommend that any of the kits sold by QRP Labs are worth getting!).  This filter does the job of both attenuating the receive image that would otherwise be present at the sum of the local oscillator and the 15 meter band (e.g. 31.00-31.45 MHz) as well as reducing the overall amount of energy impinging on the mixer from signals outside the 15 meter frequency range.

Following this bandpass filter is an amplifier based on the 2N5109 transistor - chosen in this application due to its superior performance over a standard 2N3904 transistor:  The latter will work, but both the gain and noise figure will be somewhat higher.  This amplifier overcomes the losses related to U1, a diode-ring mixer which has an intrinsic loss of about 7 dB.  Following the diode-ring mixer is another bandpass filter - another one modified from the 30 meter QRP Labs design by reducing the inductance of the "larger" windings by removing a turn or two:  This filter will be a bit narrower than the input and further-reduce off-frequency signals - and it also eliminates the image response from the output.  Ideally, a "diplexer" would be included on the output of a mixer to assure that it was terminated at the image frequency, but this was omitted because of the relatively narrow passband filter on the input that reduces the number and breadth of signals and because this is a "non high-performance" application.

Providing the 10 MHz local oscillator is an inexpensive (approx. $2.50 in single quantities) TCXO.  This device provides just enough drive for the diode ring mixer and it will be stable to within a few Hertz at 10 MHz, assuring frequency stability that is around an order of magnitude better than what would be experienced if one were to do "upconverting" to >100 MHz as is commonly done.  In our application the temperature will very wildly - from below freezing to as high as 140F (approx. 60C) so this degree of stability is important.  The down-side of this devices is that it is very tiny - about 2.5x3.5mm - so it is "super glued" to the board bottom-up and tiny (30 AWG) wires are soldered to the surrounding connections.  This device requires a 3.3 volt supply so a standard 5 volt regulator was used with a normal "dim" red LED in series to provide a 1.6-1.8 volt drop.

Following the output bandpass filter is another amplifier - this time, using the generic 2N3904 - which boosts the signal a bit more to overcome the intrinsic (relative) insensitivity of the RTL-SDR dongle:  At this lower frequency, the gain will be higher and the noise figure - set by previous stages - is less important, allowing the use of a general-purpose device.  Hanging on the output of this amplifier is a series L/C network tuned to the 10 MHz local oscillator frequency to reduce its energy by 15-20dB:  This signal is otherwise a bit strong and as we know, it's a good idea to keep the extraneous signals entering an RTL-SDR dongle to a minimum!

The final stage is a simple potentiometer-type attenuator - nothing more complicated being required as constant impedance is not very important for the input of an RTL dongle, particularly if a short cable is used.  As we already know, when dealing with RTL-SDRs on HF, it's best to start out with a bit of extra signal - and then tweak the levels downwards as necessary to find the "sweet spot" between being able to hear weak signals and prevent overload from strong signals.  In the case of 15 meters, at this time of the sunspot cycle when band openings are a bit rare and signals are on the weak side it's probably best to adjust the overall system gain to just be able to hear the background ionospheric noise - and no more!

Testing:

When tested using the "SDR Sharp" program, the combined sensitivity (with R10 set for maximum signal) of the downconverter and RTL-SDR dongle running in "direct" mode was on the order of -130dBm in an SSB bandwidth - less than 0.1 microvolts, and more than enough sensitivity to "hear" anything that would fall on any decent HF antenna in an RF-quiet environment when the band was dead.  When installed on site, R10 was adjusted so that, with a "dead", quiet band the background ionospheric noise was be just registered on the S-meter and A/D converter by several dB to maximize its signal handling capability.

A downconverter for 10 meters:

A similar downconverter was constructed to provide full 10 meter coverage, also using an RTL-SDR dongle.  This converter is nearly identical to the 15 meter converter, except as follows:
The RTL-SDR dongle is configured to run at 2.048 MSPS with the center frequency being 28.676 MHz, providing coverage from 27.652-29.700 MHz.  Because of the RTL-SDR's internal 28.8 MHz clock, there is a moderately strong carrier at this frequency that cannot be avoided.

Addendum:

It was later found to be necessary to add yet another RF amplifier after the one on the downconverter's output to bring the signal level up enough to make the RTL-SDR dongle capable of hearing the background (thermal) noise on the 10 and 15 meter amateur bands on the respective downconverters.  This amplifier was placed after the output level adjustment potentiometer (R10 in Figure 8) to minimize the amount of signal that this amplifier could see.  Because of the comparative deafness of the RTL-SDR in "direct" mode, the apparent increase in noise figure by placing the gain control at that point is irrelevant.

Because of the limited dynamic range of these dongles (no matter whether you use an upconverter, downconverter or "direct") having such a level control on the input is absolutely necessary



The 630 meter receiver:

The 630 meter band (472-479 kHz) is the newest "MF" band to U.S. Amateurs - the other MF band being 160 meters.  Like 160 meters, it is mostly a "winter time" band when noise - a significant portion of which is lightning static - is lower and the nights are longer and deeper - both of which are beneficial to reception at these frequencies.  Like 160 meters, there is a significant challenge with transmitting:  Full-sized antennas are out of the question which means that overall transmit efficiency is quite poor meaning that signals are generally weak.  Because of the comparatively weak signals, the high noise levels and the fact that the band is only 7 kHz wide, voice modes are rarely used with most operation on CW, WSPR, JT-9 and similar weak-signal modes.

Figure 9:
Diagram of the 630 (and 2200) meter dual receiver system detailing the modifications to the Softrock Lite II receiver modules.  Even though the lower receiver is marked as being intended for 2200 meter use, the designed frequency coverage of the input filter is for the range of about 125 kHz to 215 kHz, easily including the 1750 meter "LowFER" band.
Click on the image for a larger version.
Schematic diagram of the 630 and 2200 meter modified Softrock Lite II receivers
The receiver is a modified "Softrock Lite II" - the same receiver used for many other bands as described above.  Designed primarily for operation on HF, it was necessary to slightly modify the receiver, including:
In the end, the completed receiver looked very similar to that in Figure 4, above - with a few key differences:
As can be seen in the block diagram of Figure 6 and the schematic in Figure 7, the "Low-HF" signal splitter has a port labeled "<=500 kHz" that was designed to accommodate precisely this type of receiver.  At the time that the splitter module was constructed it was known that the antenna did work at least somewhat at 630 meters, but the practical low end usable frequency was unknown - but this has since been determined as mentioned below.

What about the 2200-1750 meter receiver that was mentioned?

There are tentative plans to add a receiver that will include the 2200 meter amateur band (135.7-137.8 kHz) and the so-called "1750 Meter" band (see FCC Part 15 §217) that covers from 160 to 190 kHz - the two bands being comfortably covered using a receiver with 96 kHz of bandwidth.  Unfortunately, the HF antenna at this site works miserably below about 250 kHz which means that another antenna - perhaps an E-field whip or an H-field loop - will be required to cover this frequency range.

Preliminary tests indicate that this antenna cannot be located at the building housing the receiver gear owing to its proximity to the above-ground power line that feeds the site and low-level leakage from the various power supplies, but it may be possible to place it on a tower some distance away, feeding it via coaxial cable.  When this might be done is unknown as it is not very close to the top of a very large "to do" list!




References:

  1. Youngblood, Gerald (July 2002), "A Software Defined Radio for the Masses, Part 1" (PDF), QEX, American Radio Relay League: 1–9
  2. Youngblood, Gerald (Sep–Oct 2002), "A Software Defined Radio for the Masses, Part 2" (PDF), QEX, American Radio Relay League: 10–18
  3. Youngblood, Gerald (Nov–Dec 2002), "A Software Defined Radio for the Masses, Part 3" (PDF), QEX, American Radio Relay League: 1–10
  4. Youngblood, Gerald (Mar–Apr 2003), "A Software Defined Radio for the Masses, Part 4" (PDF), QEX, American Radio Relay League: 20–31
  5. Johnson, Gary, "Measurements on a Multiband R2Pro Low-Noise Amplifier System, Part 2" (PDF)
  6. Reisert, Joe, (November, 1984), "High Dynamic Range Receivers, Ham Radio.  An English translation of part of this article from a Dutch web site may be found here.
  7. Turner, Clint, (March, 2018), "Managing HF signal dynamics on an RTL-SDR receiver"
  8. Farson, Adam, "Antenna and Receiver Noise Figure"


Additional information:
 Back to the Northern Utah WebSDR