Elecraft K2 Part 3

So after a bit of a delay in things we eventually have an oh so nearly completed K2. The building phase was a considerable amount of feeding capacitors into holes and winding toroids but the sense of achievement is pretty good.






This is the radio prior to the Part III testing and alignment. The rubber collar on the VFO main tuning dial is another Dave Richards suggestion. The absence of a finger dimple on the dial was bugging me and I was going to try and get one 3D printed and stick it on the dial but this does the job very nicely and makes things a bit simpler.


The transmitter alignment was a bit of a head scratcher in places in that you need either a signal generator to inject a signal at the desired range of frequencies but with a bit of jury rigging the old Yaesu FT-857D on a dummy load was used to generate a CW signal on all the desired frequencies.

With that all sorted, we really were into the final furlong. A few more minutes build time later and we have . . .



Job done – just in time for Santa to arrive!



Elecraft K2 Part 2

So, after a considerable amount of construction work you get to play with your creation and get it another step closer to being a radio.

Calibration and alignment hinges around the 4MHz crystal situated on the control board. This is used as the basis for all the later steps and getting this set to bang on 4MHz helps a great deal. In an ideal world you’ll have a nice calibrated external frequency counter to use. I haven’t and plumped for the idea of using my Yaesu FT857D  to zero-beat the 4.000MHz signal.

This became a bit of a head scratcher which boiled down to the fact I was trying to run the FT857D off of a battery pack on my work bench and things weren’t quite as they should. Thankfully, my good chum Andy G7UHN came up with a stonking suggestion which I share here with you as that’s what this is all about – sharing ideas and knowledge to make things easier for those foolish enough to follow suit!

How about using my existing setup as a spectrum analyser? What a great idea!


So by moving indoors onto a proper PSU, firing up Fldigi, tuning the FT857D to 3.999MHz and turning on the K2 you get this.


Our crystal is producing the nice signal just to the left of 900Hz. With a little adjustment of C22 with a ceramic screwdriver we get that signal to sit exactly on 1kHz


and when the FT857D is tuned to 4Mhz, the signal is lost as we achieve zero-beat.


Now that was so easy it was untrue! Wish I could claim the credit for thinking of it.

The next series of tests are simple enough but a word of warning when you get to the VCO alignment. Do NOT allow any other members of your household or family anywhere near you when you do this. You need to attach an alligator clip attached to your DMM probe to one end of R33 which is a single resistor in amongst a veritable sea of components on the RF board. There’s very little room in there and you don’t want it shorting onto anything else nearby. So, while merrily recording the VCO control voltages, someone who will remain nameless thought it’d be fun to slam a drawer shut on the desk I was sat at. Said alligator clip duley sprung from its point of attachment and hit every nearby component to Hollywood slow motion sequence gasps of “NOOOOOOOOOOOOO!!!”. Oh how we laughed and saw the funny side of that one as I franticly killed the power and grabbed the probe out of the chassis! We didn’t release any blue smoke but trust me the air was blue!

I sat with my head in my hands for hours as Mr K2 decided he didn’t want to work anymore. Nothing I did could get the radio to work. Bearing in mind I was mid VCO Alignment, my readings which had been brilliant on the tested bands were spot on. The radio would power on, there was functionality but the Alignment readings were far from correct. Having visions of having fried something on the synthesiser side of the board I started painstakingly trying to work out what was going on, or not as the case may be! A day and a half later and a light bulb moment occurred when I realised that in the panic to kill the power something on the front panel had been knocked and I’d somehow managed to change band or alternatively whatever had got shorted inside the K2 had achieved the same result. Either way by rewinding considerably and starting again we were in business.

The alignment of the K2 takes a bit of time and has you measuring and testing various parts of the transceiver. I ran into a wall with the BFO test and couldn’t get the required range of >=3.6kHz between the BFO High and Low frequencies. The manual gives a series of things to check, which revolve around the BFO section of the RF board and primarily L33 and R116. After a lot of head scratching, as everything was as it should I took a long hard look at L33 and R116 and realised I had been a victim of interpretation of the manual. When you’re fitting these components the manual said “the resistor’s body should be partially recessed into the well left in the center of the toroid” and I had gently pushed R116 down a bit so it was. It transpires this isn’t what you need and by gently pulling R116 back into a more horizontal arrangement my BFO range shot back up to something far more respectable. With a desired range of 4 to 6 kHz, I can live with 5.35!

I spent a bit of time calibrating the filters as per the manual and then took a time out just to play with the K2 to see what it would do.

CW reception

SSB reception

Anyway, enough playing around. Part of my reading involved the advice regarding using a noise source to carry out the filter alignment rather than relying on background noise. I’d been on a mini mission to track down a suitable circuit and both Elecraft and N0SS provide a lot of helpful material. In the end I decided I was going to build myself a noise generator based on


Just as I was about to sit and work out a Veroboard layout I had a flash of inspiration. I’d got a pile of MeSQUARES laying around from my Rockmite build. Time to go freestyle! Or Manhattan style as the case may be. Then I remembered I was reinventing the wheel here. That all so clever person Dave Richards had been there before and built one of these back in 2012 when he was playing with his K2. So, in a blatant act of plagiarism I had a look at Dave’s build and duly copied it. Sorry Dave, but it seemed daft to try and change simplicity and perfection.




That’s why I needed those mint tins people! In the absence of Altoids, M&S mint tins do the same job quite nicely. Say hello to the “test tin” as it was dubbed by my little helper!

I can hear the obvious question. Why is there a need to build something in a mint tin to set a radio up? Good question. When you’re trying to align and calibrate something you need a reference source. Using background noise is all well and good if you’ve got enough of it to be usable. That reference source needs to be constant. Bearing in mind the audio spectrum for a quality track by AC/DC looks like this


you could do with giving yourself something reliable and repeatable as your frame of reference, hence the Broadband (HF) Noise Generator for Filter Alignment circuit.


So on an increasingly cluttered desk we now have the Noise Generator feeding its signal into the K2 antenna jack, the headphone out from the K2 hooked to my trusty Creative Labs Sound Blaster USB and my laptop running Spectrogram to give a visual representation of how the K2’s filters are behaving.


So after a little bit of playing around based on Don Wilhelm’s site along with Elecraft’s own input and a little bit of Dave Richard’s skills I had filter bandwidths I was happy with. So with nothing left to do at this stage I took it all apart and carried on soldering a pile of resistors and capacitors onto the RF board.