Building the Elecraft K1

Elecraft K1 and Palm Mini-Paddle in operation at Maclear’s Beacon (Pic: Stacey ZR1SC)

A couple of months ago I became interested in the Summits on the Air (SOTA) programme and with the assistance of the SOTA Management Committee and the SARL initiated the process of establishing SOTA in South Africa. SOTA offers awards for amateurs who activate or contact a certain number of mountain summits (see http://www.sota.org.uk). Although VHF contacts are accepted for SOTA I wanted to be able to operate HF, both because CW is my preferred mode and so that I could link up with SOTA members in Europe. So I set out to find a QRP rig that would best meet my requirements for portable operations.

The key criteria were:

  • It should be small and light, as I would have to haul it up mountains;
  • It should have minimal power consumption on both receive and transmit, so that batteries would give a reasonable operating time;
  • It should offer a reasonably wide tuning range, and multiple bands would be a definite bonus so that the appropriate band could be chosen according to conditions;
  • It should have good support for CW, since this is my preferred operating mode and also the most efficient mode for QRP DX.

I considered a number of possibilities, including the Yaesu FT-817 and several kit radios. Finally I chose the Elecraft K1, a CW-only transceiver that is available in a choice of 2 or 4-band versions. The K1 differs from most QRP transceivers in having full microprocessor control which makes it exceptionally versatile. It includes a built-in CW memory keyer and provides completely solid-state full break-in keying (QSK) without a changeover relay. Power output is 5W on all bands, although Elecraft recommends reducing this to 3W to maintain maximum efficiency when using low-voltage supplies such as the internal battery pack option, which delivers 9.6V when fitted with eight 1.2V rechargeable batteries.

Best of all, current drain is only 50mA on receive, and about 650mA on transmit (assuming 3W output from a 9V supply). This means that during typical operating, with 10% “key down” time and the rest on receive, 2000 mA-Hr AA batteries should last for about 18 hours! In comparison, Yaesu’s specification for the FT817 is 450 mA on receive, and 2A on transmit (5W output at 13.8V), giving a comparable battery life of less than 3 hours.

As well as the basic K1-4 kit, I ordered the internal automatic antenna-tuning unit; internal battery pack and wide-range tilt stand options. I also purchased a Palm Mini-Paddle from Palm Radio (www.palm-radio.de). The result is a complete QRP HF station in a tiny package – at 133mm wide, 56mm high and 144mm deep the K1 occupies only one sixth the volume of my main station rig, which does not include the power supply, ATU or keyer! The total weight including the tilt stand, paddle and batteries is only 1.1 Kg.

I ordered my K1 via the Elecraft web site (www.elecraft.com) and it arrived three weeks later. I already had all the basic tools required for assembly – a temperature-controlled soldering iron, de-soldering tool, needle-nose pliers, miniature side-cutters and assorted screwdrivers – but in addition purchased a “field service kit” containing an anti-static mat, wrist strap and earth lead. In retrospect I could probably have managed without this, as long as I remembered to ground myself before handling ESD sensitive components such as the PIC micro-controller, but better safe than sorry!

I started by reading through the excellent manual that gives very detailed step-by-step instructions, including photographs of most components, as well as complete circuit diagrams and a brief circuit description. The analogue circuitry in the K1 is a fairly conventional single-conversion superhet design using NE612 mixers and a ladder filter made from discrete crystals with audio-derived AGC and an output stage based on a LM386/380 combination. The local oscillator uses a varicap-tuned VFO, combined with a band-switched crystal-controlled premix oscillator.

What distinguishes the K1 from most other QRP designs is the use of a PIC micro-controller which controls the band-switching relays and T/R switching, manages the multipurpose display, operates the menus, generates the side-tone and implements the automatic level control and the memory keyer. I was particularly enthusiastic about the use of an I2C control bus to route control signals from the main micro-controller to auxiliary controllers on the band and ATU modules, resulting in a great deal of flexibility for future expansion. Another innovation is the use of varicap diodes as the inter-stage shunt capacitors in the crystal ladder filter, which allows the micro-controller to change the filter bandwidth by altering the bias voltage across the diodes. The result is three filter settings – wide, medium and narrow – with no additional crystals. Now that’s smart!

Construction was relatively straightforward thanks to the very detailed instructions. I started with the 4-band filter board, followed by the front-panel board which contains the micro-controller. (After completing the front panel I started muttering “oh so cute!” to myself, much to the amusement of my YL Stacey, ZR1SC!). The final board is the RF board, which is constructed in two stages – first the receiver components, then the transmitter components. This has the great bonus of allowing you to align and test the receiver as soon as it is completed, without having to wait to finish the transmitter section.

After each stage there are resistance and/or voltage measurements that give an initial indication as to whether everything is all right. These picked up the two errors I made – the first was to leave one of the pins of an integrated circuit unsoldered, and the second was to forget to include an errata entry which required a link to be soldered in place of an RFC on the RF board, despite having conscientiously written the errata entry into the manual at the appropriate point! In both cases my mistakes were easily located by referring to the circuit diagram.

Construction includes the winding of a number of torroidal inductors and transformers. Although I had never attempted this before, it turned out to be simply a case of threading enameled wire through the cores the specified number of times, and then stripping and tinning the leads. The manual gave instructions for heat-stripping the enamel using a soldering iron but I wasn’t satisfied with the results achieved this way, I suspect because my soldering iron was underpowered for the task. So I preferred the old-fashioned method, using a needle file to remove the enamel and carefully testing the leads for conductivity before tinning them.

One of the last components to be installed in the RF board was a 10nF capacitor, but by the time I got there I had only a 1nF capacitor remaining. I checked all the locations where 1nF capacitors had been installed, suspecting that I may have installed the wrong value in error, but they were all correct, so I presume that the error was made when the kit was packed. Fortunately I had a suitable 10nF unit in my junk box, so I used it instead. (I guess you could say that Elecraft now owes me 9nF.) Other than this the kit was complete, including everything required except for the solder.

Once the RF board is completed there is only final assembly and alignment, which is relatively simple and mostly consists of tweaking various trimmer capacitors and setting a couple of parameters in the menus that are accessed from the front panel switches. The shaft of the ten-turn tuning potentiometer was slightly too long, resulting in a 4mm gap between the tuning knob and the front panel, which detracted from the aesthetics. A few minutes with a hacksaw (being oh so gentle so as not to damage the potentiometer) quickly solved the problem. From start to finish the basic kit excluding options took me about 27 hours of construction time spread over a few days and a weekend.

Then the moment I’ve been waiting for – plug the antenna into its BNC connector, connect the DC power, and turn on the K1. The self-test completes successfully in about a second. Select the 20m band, use the menu to set the output power to 5W, set the keyer speed to 20 wpm and listen for a signal. There’s N1DT calling CQ on 14010. I reply “N1DT de ZS1AN ZS1AN kn” and… he hears me and comes back! Brilliant, only 5W and I’m in chatting with Don in Massachusetts through my lowly apartment-block-limited dipole! Several other QSOs follow, including one with Rudi ZS6DX, another SOTA stalwart who tells me he’s also just ordered a K1. (It must be contagious since Ian ZS5IAN, who is also active in SOTA, has already built one!)

My first impression of the K1 were:

  • The “fit and finish” is superb with professionally silk-screened front and rear panels, giving the end product a very professional look (and it’s oh so cute!).
  • The ergonomics are excellent. The controls I use regularly (band switching, RIT, keyer speed, message playback and filter selectivity) are accessed directly from the front panel, without having to use the menus. The menus provide a convenient way to set parameters that are changed less often like power output, keyer emulation (straight key, or mode A or B iambic), sidetone level and pitch (even my TS680S won’t let me set the sidetone pitch) and so on. I think my only complaint here is that there is no way to determine the current filter setting without changing it, but that’s a very minor niggle – perhaps a future release could make the first press of the filter button display the current setting, and subsequent presses alter it.
  • The keying characteristics are superb. Full-break-in keying is smooth and silent, there is no noticeable shortening of dots, and no chirp or key clicks. The memory keyer works well and includes an auto-repeat function so I can watch cricket (with the sound turned down) while calling CQ on 40m!
  • The variable-bandwidth crystal filter works very well. On the widest setting, SSB transmissions are just intelligible, while the narrower settings are effective in reducing both QRM and noise levels. (This is unlike narrow filters on some “big” rigs where the high insertion loss means that the noise-reducing properties of a narrow filter bandwidth are not realized in practice).
  • The AGC does work, but the combination of the slow attack time characteristic of audio-derived AGC and the very capable audio output stage means that, when listening through headphones, strong signals can start off unpleasantly loudly before decaying to a more acceptable volume as the AGC takes hold. I think the single biggest improvement in this rig would be to substitute IF derived AGC.

As well as the basic K1, I also purchased the internal battery option, wide-range tilt stand and internal ATU, which was built by my YL Stacey, ZR1SC. Since she had very little prior electronics experience I gave her the soldering notes from the Elecraft site along with a piece of Veroboard and some resistors to practice on. Needless to say she did a great job and the ATU worked perfectly first time.

The first full field test was on 1 January when I activated Maclear’s Beacon, the highest point of Table Mountain, for Summits on the Air. Everything worked perfectly and I had a great afternoon making contacts halfway around the world with only 3W – that’s about the power of a small torch! Best of all, my complete HF station – including batteries, antenna and mast – weighs less than 5 Kg so it really is portable.

Overall I am very happy with my new baby, which will make an excellent station for portable operations. It will also be a great “take along” rig for holidays and travel when I don’t want to have to pack the 70 Kg or so that my complete station weighs (I did that for field day and for a recent trip to Dassen Island – what a mission!) All this, plus the satisfaction of knowing I built it myself and understand it well enough to make repairs should that ever be necessary.

I would not recommend the K1 for multi-station DXpeditions as the low power consumption requirement has led to an inevitable compromise on dynamic range. The ARRL lab test results show its 3rd order IMD and blocking dynamic range results to be on a par with other small rigs like the Yaesu FT817 and Icom 706 Mk 2, neither of which is suited to very demanding RF environments.

No, for my next DXpedition I’m going to beg, borrow or steal a K2 – but that’s another story!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: