W8WWV - QRP - The Small Wonder Labs DSW-II Kit (40 Meters)

Greg Ordy


I ordered my DSW-II kit on December 19, 2003. The kit was shipped on February 12, 2004, almost two months later. I suspect that this delay was due to the fact that the kit comes with several surface mount components already soldered on the board, and I assume tested as well.  The Small Wonder Labs web site includes a list of outstanding orders, as well as status information. I watched my order slowly creep towards the shipped status. This sort of customer feedback makes waiting somewhat easier. The SWL policies allow (unbuilt) returns and cancellations without hassle, so you can change your mind at any time.

Finally, on Saturday, February 14th, the kit arrived at my door via US mail. Here is a picture of the box contents, after I removed the foam peanuts that kept the contents from bouncing around. Please click on a picture for a larger view.

DSW-II-40 Upon Arrival
DSW-II-40 Upon Arrival

There was a note inside saying that the front panel would be shipped separately in about a week. That's fine by me, since I'm up to my ears in the N2PK VNA project.

The paper manual duplicates the manual found on-line, except that this is the 40 meter version, as opposed to 20 meters. I think it's a good manual. Nothing will be as detailed as the old Heathkit manuals, where every step seemed to have an exploded picture. But, as kits go, this one has very good documentation. A second manual with theory of operation and troubleshooting information is available on-line. Shortly after receiving my kit, I noticed that the Small Wonder Labs web site had all of the manuals on-line for all of the different band radios. You can really evaluate the entire kit by looking at the web site. Nothing is held as a mystery until you spend your money.

So, it's finally here, but I'm busy. Who planned that? I'll be putting the kit off to the side for a bit. When the dust clears on some other projects, I'll  assemble and test it, and document that on this web page.


This radio has also been reviewed on the popular eHam.net site.


Assembly took about 7 hours, and it was all fun. I like to solder. I don't like to collect parts. When I can pay somebody to collect parts, it's usually a good deal for me. Soldering is also somewhat  therapeutic, getting into a rhythm of part after part, and seeing each shiny, concave joint as a small victory towards a larger goal.

The circuit board is a high quality part with a good solder mask and complete labeling. Combined with the exploded view in the manual, it was very easy to find the right location for a part. The parts were organized into groups, which all made sense as you reached for a part. You can tell that the goal of the system was to reduce confusion, and I believe it does that.

A few specific comments.


I recently started to work with surface mount parts. The DSW-II-40 uses more traditional through-hole construction. Yes, there are a few surface mount components which are part of the kit, but they are large inductors that are easy to work with. I found that using the tools I had accumulated for SMT work made assembling the kit very easy. I used a desk magnifier, my smallest soldering iron tip, and 0.022" diameter rosin-core silver-bearing solder. Perhaps the only other tool which is nearly essential is a very sharp diagonal cutter, one which is good at making flush cuts.


Before I started, I photocopied the assembly instructions from the manual, and the parts list as well. I left the manual open to the (color) exploded view, and then checked off steps and parts on the other pages as I went along. This keeps the manual clean and I'm not always turning from page to page to find information. Although the manual suggests that if you are a big boy, you can assemble in any order you want, I don't know why anybody would ever want to ignore the instructions. The key part of the instructions is the order of assembly, which is designed so that you always have open access to the next part.


I also tried to orient any value markings on a part so that it could be read after everything was assembled.  If the part has a polarization, then there is no choice, but most capacitors, for example, can be installed in one of two orientations.

My kit included two 680 pf capacitors as opposed to one 680 pf and one 1000 pf. Small Wonder Labs has a no questions asked parts policy, so getting the missing part is not a problem. In addition, I was sent an extra 150 pf capacitor.  When you get to the end, and if you find some sort of parts mismatch, that's when it's important to be able to read the parts on the board. I rechecked my board, verifying that the 1000 pf capacitor wasn't used, and that all of the 150 pf capacitors were.


The enclosure and control mounting scheme is very clever and highly functional. That can be a messy aspect of construction when wires start to run all over the place inside a box. Here, everything mounts on either the main board of a small daughterboard, and taking apart the case is easy.


The daughterboard is connected to the main board with a 10-pin ribbon cable. The cable is soldered to the daughterboard, and plugs into the main board. Soldering the cable to the board is a little tricky. The problem is holding the stiff cable perpendicular to the the board while soldering the first few pins. Having a helper would make this easy, but  with only my two hands, I needed a better solution. What seemed to work well was to tape the cable to the side of a small box, pointing up, and then through the holes on the daughterboard, which was lying on the top of the box. Here's a picture. Please click on the picture for a larger view.

Aligning the Ribbon Cable to the Daughterboard
Aligning the Ribbon Cable to the Daughterboard

Here are some pictures of the start and end of the assembly process.

Starting Out Finished Main Board Completed Kit
Starting Out Finished Main Board Completed Kit

I'm very happy to report that after adding the final 1000 pf capacitor, which was in the transmitter output filter, the radio jumped to life, and is fully functional. By the way, the missing capacitor showed up 2 business days after the missing part was reported - very good service.


When I first saw a knob labeled Gain on the front panel, I naturally though it was audio gain. It's actually RF gain, and that makes a lot of sense when you think about it. The receiver does not have an automatic gain control, it's set to maximum amplification, all of the time. This is great for weak signals, but strong signals are then positively ear shattering in loudness. By reducing the RF gain, the volume does indeed drop, as does any distortion generated within the receiver. The gain control is a variable resistor located near the antenna input. Back-to-back  serial diode pairs protect the receiver against damage.

The audio level has been preset to a nominal value which I found to be quite acceptable. If you need an audio volume control, you can use a headset with an integrated volume pad. They are quite common, especially for computer use.

Calibration or alignment consists of two steps. In the first step the receiver input transformer is adjusted for maximum signal strength.  The claimed 3 dB bandwidth is 150 KHz, which completely covers the US 40 meter CW segment. I peaked my receiver for the power-up frequency of 7.040 MHz, a common QRP frequency. You could certainly decide to peak the input at some other favorite frequency.

The second alignment step adjusts the BFO (beat frequency oscillator) so that the receiver and transmitter operate on the same frequency when the received CW tone matches the keyer sidetone.  Both of these tones must be 800 Hz. One is fixed, the other is adjustable. As suggested in the manual, I used DigiPan to display the tone, and adjusted a single control for a 800 Hz display on the screen. This step can be done by ear, if you are willing to press the key, and adjust the control until the two tones match.

I made all adjustments with a 12 volt DC power supply.

Test Results

I decided to make a few measurements before closing up the enclosure.

In case I ever used the radio with battery power, I wanted to check the receiver current consumption. Connected to a dummy load, the receiver consumed 55 mA with a 12 VDC supply. The key down current draw was 588 mA at 12 VDC (7.06 watts), with the power output control set to maximum.

While keying the transmitter with a 12 VDC supply, drawing 588 mA, the RMS RF voltage, as measured on an oscilloscope was 13.43 volts into a 52.2 Ohm dummy load. I measured the peak to peak voltage, and converted to RMS by formula. The RF power output, at 7.040 MHz, by calculation, was 3.46 watts (with a 12 VDC supply). The RF output can be set with a variable resistor. It's minimum setting produces very little output. I have no doubt that you could set a very precise power level if that was of interest.

The radio is claimed to produce 5 watts output with a 13.8 volt supply. Given 3.46 watts output at 12 volts, the claim is reasonable. By the way, 3.46 watts produces a signal which is 14.6 dB lower than a typical 100 watt radio. If you can make a contact with a 100 watt power level, ask yourself if you can lose 14.6 dB of signal strength and still be heard. Odds are that if a 100 watt signal produces an S5 level on your S meter, you should still be able to hear the signal if the power is dropped to 3.46 watts. There are a lot of assumptions built into that estimate. Band noise and QRM/QRN clearly are important. The first assumption is that there is no band noise or interference. The second assumption is that S0 signal levels represent the minimum detectable signal level. With most receivers, you can still hear signals that are under S0 in strength. Finally, there is the relationship between S Units and dB. Most of the S meters that I have measured are very inaccurate, and very nonlinear. One S Unit is not 6 dB. The general trend is that an S Unit may be equivalent to only 1 or 2 dB by the time that you get near S1. Let me say it this way, for many popular radios around the year 2000, there will be at least 14 dB under S5 before a signal drops into the noise of a quiet band. My point is that there are lots of contacts which can be made with QRP power level.

I have left the radio on (receiving) for days, and I can detect very little heat generation from the board. The receiver, at least, runs very cool.

Wall Warts

It's very tempting to power the radio with a wall wart power supply. I've collected a whole bag of wall warts over the years. The modem, answering machine,  cell phone, or gizmo is long gone, but I saved the wall warts, and when I need one, I search the bag. Since the radio is specified to accept a maximum input of 15 volts, a 12 volt wall wart looks like a good idea.

Be a little careful when doing this. Most 12 volt DC wall warts seem to have a no load output of 18 volts. A load will bring them down, but since the receiver draws so little current, it might not be enough to bring the supply voltage down under the specified maximum.

I happened to have a quality wall wart from an HP paper scanner. It's no load output was 12.08 volts DC, and under full transmit (588 mA), the voltage dropped to 11.97 volts. Very little sag.

I had another wall wart marked as producing 6 VDC, and it's no load output was 9.19 volts, above the radio minimum voltage of 8 volts. When the transmitter was keyed, the voltage dropped to 7.07 volts, and the sidetone  quality was clearly suffering. The moral of the story is to select a wall wart with good regulation which does not supply too much voltage, nor have too much sag on transmit.

Be sure that your wall wart has the matching polarity, and produces DC, not AC.

In many cases, QRP radios are powered by batteries. Twelve, 2200 mAh NiMH batteries would power this radio for a long time, and fit into a small package.

Receiver Comparison

My main station radio is the ICOM 756PRO. I wanted make a quick comparison of the PRO receiver against the DSW-II. The two traditional characterizations of a receiver are selectivity and sensitivity.


Selectivity is achieved in the DSW-II by running the received signal through a cascade of discrete crystals. The claimed bandwidth is 500 Hz, without any specification of shape factor. Although somewhat crude, one of the measurements that can be easily made is to display a spectrum analysis of the receiver audio while listening to uniform noise. I use Spectrogram for this function. Here are three screen captures of Spectrogram data. Please click on a picture for a larger view.

DSW-II-40 756PRO 500 Hz CW 756PRO 500 Hz SSB
DSW-II-40 756PRO 500 Hz CW 756PRO 500 Hz SSB

All of these screens were captured while listening to a 40 meter CW frequency in the evening (more noise), where there was no signal in the receiver passband. The 756PRO filter selection is described on another page. The absolute signal levels should not be compared, what matters is the shape of the response.  I adjusted the PRO passband to center over 800 Hz, to match the DSW-II. I normally use a lower frequency sidetone.

The spikes on the left appear to be 60 Hz hum. This hum is part of the lash up, not part of either receiver.

The DSW-II filter is not as sharp as the PRO. Still, once you get several hundred Hertz away from the center frequency, signals are attenuated by 35 to 40 dB.


The sensitivity test I performed was to tune around the band with the DSW-II, looking for weak signals. When I found one, I tuned the 756PRO to the same frequency, and tried to determine the relative sensitivity ranking between the two radios.

I need to do more testing, but after a few listening sessions, the sensitivity seems very similar. The DSW-II sensitivity was slightly greater than the 756PRO, without preamp. If either preamp 1 or 2 was turned on, the PRO was more sensitive. This was my subjective perception. On 40 meters, with a typical antenna,  it's very rare that the preamp is helpful.

It's clear from these simple measurements that the DSW-II receiver is hot, and that reception is not a weak link. For all but the weakest of signals, the gain will be reduced to keep the receiver from overloading. There is a lot of gain there.


The kit was fun to assemble, and works as advertised. Now, it's time to move from construction to use.

I made my first QRP contacts on February 27, 2004. I worked an EA7 (Spain) and ZP6 (Paraguay) a little past my local midnight. I received good (579) signal reports. Having a 6-element vertical array does help. Still, doing it with 3.5 watts, and a radio about as big as my straight key base is fun.

Epilogue 1

I started off using my normal (Bencher for the ICOM 756PRO) paddles with the QRP radio, but the wiring on the desk was a mess. I looked around for a small, possibly portable paddle. I ended up purchasing a Mini-Paddle from Morse Express. I picked the gray color to match the case of the DSW-II, and also purchased the optional magnets so that the paddles would stick to the radio. The paddles are apparently made by Palm Radio.

I have the paddles off to the right side of the radio. It's a very convenient location, and seems to work out well. Here's a picture. Please click on the picture for a larger view. It's held on with magnets, so it can be easily removed, or moved to the other side or the top.

DSW-II with Mini-Paddle
DSW-II with Mini-Paddle

Another interesting keyer alternative is the P1W paddle/keyer, made by CW Touch Keyer. I have not used this product, but the lack of delicate mechanical parts, and the presence of  message memories, which could be great for calling CQ on QRP, make it well worth checking out. It's small, and can be battery operated, which is often times very desirable for QRP operation.

Back to my QRP Page

Back to my Experimentation Page

Back to my Amateur Radio Page

Last update: Sunday, December 11, 2005 11:43:19 AM
Back to my Home Page