Another K3 panadapter option?

Cross Country Wireless recently introduced a new product, a built, boxed and ready to use SDR receiver that is optimized to work with the bog standard sound card provided with every computer. It covers two 48kHz segments which may be on one or two bands using a switched local oscillator. At £49.95 it looks like something of a bargain.

I was looking at the product's web page this evening and the thought occurred to me that this might make a rather good inexpensive option for a panadapter for the Elecraft K3. It would need a different crystal to cover the K3 IF output frequency which is 8.215MHz, but that shouldn't be a problem. CCW might even offer this as a stock option if there was a demand for it.

As the receiver covers 40m and 30m everything else should work OK unmodified. The key point would be whether there is adequate isolation to prevent the local oscillator leaking into the K3 IF and desensitizing the receiver - the reason why most people use a buffer amplifier when using SoftRock boards for this purpose.

I don't have a great urge to have a panadapter display and I already have three sound cards (well, one and two USB sound modules) in use with my shack computer so this isn't something I'm planning on trying. But I thought it might be worth mentioning the possibility for other K3 users. Even if the idea is a non-runner, the Cross Country Wireless receiver still looks like a very nice product.

Change of resistance

I hate getting old! Last night my lower back was in agony from the couple of hours I spent hunched over the desk working on the SoftRock 6.2 Lite kit. So I didn't plan on doing any constructional work today. However, Maurice G4DVM read of my problem with the kit and asked me to send him some voltage measurements. It also occurred to me that I had tested the circuit with the power supply set to 9V instead of 12V. So I connected the kit to a 12V power supply, switched on the FT-817 (which was still set to 8.191MHz from yesterday) and this time a carrier signal was heard!

I started taking voltage measurements anyway and observed that the voltage on pin 2 of U2 was low: 0.79V compared to the 2.47V obtained by the author of the instructions. Pin 3 was 1.23V compared to 3.21V expected. More importantly I observed that the oscillator signal stopped as soon as the test probe touched the pad. Around this time I observed that oscillation also stopped if I touched the case of the crystal and it didn't always start up again afterwards. So I probably wasn't mistaken that I didn't detect the local oscillator signal yesterday.

According to the instructions, R6 (which has something to do with the bias of Q2) is not used in most versions of the SoftRock except for 40m. It is supposed to be used in the version for the K3 IF, and the value supplied is 22K. I thought it would be an easy test to remove it - just snip the exposed lead (since it is mounted on end.) If that didn't help I could always solder the ends back together again.

With R6 open circuit there was no local oscillator at all. I then decided to try a lower value. The instructions say a value from 12K to 22K may be used "as appropriate". I found a 15K resistor and tacked it across the pads on the underside of the board. I had a local oscillator again and this time it didn't stop when I touched the case of the crystal!

I absolutely hate replacing components in circuit boards with plated through holes. Removing the snipped in half 22K resistor from the circuit board was easy. But clearing the holes to allow the 15K resistor to be inserted in its place was a job I dreaded. Ironically it would have been easier if the SoftRock kit was all SMT - apart from the fact that I wouldn't have had a 15K SMT resistor to try.

I have a spring-loaded desoldering gun, but I didn't think I could use it as it was designed for larger boards with more space between the components and the holes. I have desoldering braid too, but most times I have used it I ended up lifting pads and traces on the board. I made a real mess of modifying the KSB2 board in my early model Elecraft K2. The board works, but it is a good job no-one can see it.

In the end I did the job using a couple of stainless steel needles from Olga's sewing kit. I heated up a hole and quickly pushed a needle through, then cleaned up the solder that was pushed through. It took a couple of attempts to clean the holes sufficiently for the replacement resistor's leads to pass through.

By the time I had finished the modification and checked that I still had an 8.191MHz signal my back was painful again. So much as I would have liked to finish the SoftRock and see if it works I shall have to pack it and my tools away and leave it for a few days. I hate getting old!

On the rocks

A disappointing day in the G4ILO shack. I did some more work on the SoftRock 6.2 Lite kit which I was hoping to make into a panadapter for my K3. But the project has hit the rocks as the local oscillator doesn't work and I don't know what to do about it.

I started off by adding the components for the regulated power supply. That checked out fine - not much to go wrong there, really.

Next I built the crystal oscillator part of the circuit. That also worked fine - I could hear a strong signal just below 32.768MHz on my FT-817 receiver.

The stage after that was the divider circuit, which is supposed to divide the oscillator frequency by 4 to give the 8.192MHz local oscillator required by the K3 version of the SoftRock. I didn't need to add any components for that, as I'd already soldered in all the SMT components and this stage just involved adding the divider ICs. So I tuned down to 8.192MHz and heard nothing.

The current consumption of the SoftRock was correct for this stage, but some of the voltages on pins of U2 were not correct. There was zero volts on pin 2 where there should have been a reading. I checked the board carefully but I could not see any solder bridges or anything else I have done wrong. I don't have an oscilloscope so I can't see what is happening.

I don't know what is wrong, and I don't know what to do. Perhaps it was a mistake to solder all the SMT parts first, but I really would not have liked to do it after the through-hole components were added, making access difficult. Perhaps my anti-static precautions weren't good enough. I did use an anti-static mat and wrist band, but I'm getting pretty absent minded these days and there were times I forgot to connect the ground strap to the wrist band before handling the board.

I don't have any replacement divider ICs to try, and to be honest I find the board too small to work on. I doubt that I could remove a chip without lifting the PCB pads at the same time. I don't know why the SoftRock couldn't have been designed using all through-hole parts with socketed ICs, on a larger board. It doesn't appear to use any components that aren't available in leaded versions, and it would have made the project easier to build and easier to troubleshoot for most people.

The SoftRock has gone into the box of projects that didn't work, along with my short-lived enthusiasm for SMT construction.

Breaking my SMT duck

It was nagging at me like a persistent toothache. The SoftRock kit was sitting in an envelope with all the parts sorted and the instructions printed, waiting for me to start building. But the anxiety about soldering the SMT parts made me put off starting. Finally, this afternoon, I decided to bite the bullet and make a start.

The first snag was my soldering iron. The regular sized bit had seized on to the shaft of my Antex TCS. I could not get it off, so I could not replace it with the small one I had bought for just this task. In the end I decided to use my 40 year old Antex Model C. Years ago I bought a fine tipped bit for it which I had always found to be too feeble for any projects I had built. Miraculously I had not managed to lose it during all that time, so at long last it was to be of some use!

I made the decision to mount all the SMT parts first, instead of building the board stage by stage. I thought that would make the task easier by keeping the board flat and not restricting access to the SMT pads in any way. I decided to start with what I thought would be the easiest parts: the 0.1uF SMT capacitors. There are 10 of those in the SoftRock 2, and 11 were supplied, so I had one spare!

Things did not start well. For the first one I made the mistake of deciding to tin (apply solder to) the two pads before soldering. This made the pads uneven which made it even more difficult to hold the capacitor in position for soldering than it normally is. At the first attempt the capacitor was standing a bit proud of the board at the other end. I applied some pressure with tweezers to try to level it out. Only after I had soldered both ends and then inspected my work did I see that I had cracked the component. So I had to remove it and start again - my one spare lost before I had even started!

I nearly gave up at this point. I had spent about 15 minutes trying various ways to hold the component still so I could fix it with a dab of solder at one end. "OK", I thought, "SMT is not for me. This is no fun at all." I was close to packing the SoftRock all away and giving the kit to someone else. But then I thought "what the hell." There was nothing to lose by trying. The kit hadn't even cost me anything, thanks to the generosity of Craig VK3HE. So I tried again.

Eventually I hit on a technique that worked. I used a bronze bladed trimmer tool with BluTack on the end to pick up an SMT part and hold it in position. The BluTack was necessary because otherwise the slight tremor in my hands would jiggle the capacitor out of position. Then I would fix one end in place with a blob of solder from the fine tipped bit. Next I would solder the other end of the part. Then finally I would go back to the first end and try to make a better job of it.

I should point out that this was only possible with the aid of a headset with magnifying lenses that I bought on eBay several years ago. It took nearly an hour to solder in all 10 capacitors, and my back was protesting a bit at all the bending close to the desk to get the board in focus with the high magnification lenses I was using. Only one capacitor pinged off into my lap and fortunately I immediately saw it. If I had lost one in the carpet that would have been that.

I was going to quit while I was ahead but I was fired up and wondering how I would manage with the SMT ICs in the kit. There are four of them. The instructions say to use electrostatic precautions so before I could carry on I had to unroll the electrostatic mat and ground it using the negative terminal of my shack power supply.

Amazingly, I found the SMT ICs easier to install than the small capacitors. Contrary to all the advice found on the web I did not use flux and desoldering braid. I soldered each leg of each IC individually, just as I would do with through-hole components. The fine-tipped soldering iron bit made this possible, as did use of some 0.2mm diameter solder that I had purchased on eBay. Thanks to the fantastic macro facility of my new £25 digital camera you can have a good look at the result.


I used a couple of small balls of BluTack to anchor the PCB to the work surface, then picked up the IC with tweezers and dropped it on to the board. Then I nudged it into the correct position using one end of the bronze tipped trimming tool. I held the IC in position using the other blade of the trimming tool which had some BluTack wrapped round it, and quickly tacked one corner leg to the board with solder.

Next I rotated the board so I was looking at the other side of the IC, checked that all the pins still lined up with the pads, and tacked the opposite corner leg into position. I then soldered all of the pins individually by applying the fine tipped bit and fine 0.2mm solder.

I mounted all 4 ICs in less than an hour and did not create a single solder bridge, which is better than I usually manage soldering regular sized through-hole parts! I am over the moon to have broken my duck and overcome my fears of working with SMT components, though my back is telling me that two hours of this in one session is more than enough!

Preparing to build the SoftRock

I spent a couple of hours yesterday evening preparing to build the SoftRock 6.2 Lite kit that I hope to use as a panadapter for my K3. I started by printing out all the detailed build notes at WB5RVZ's SoftRock page. This describes building the receiver in seven separate stages, starting with the power supply and ending with the external connections. Each stage has its own bill of materials or components list, so I then checked the supplied components and put them into individual small polythene bags, one for each stage.

When I got to the local oscillator stage I realized that I didn't have a crystal. An email to the Elecraft reflector produced the information that the crystal used for the K3 panadapter version is a 32.768MHz. Even better, it produced an offer from Alan G4LWA to send me one he had that was surplus to requirements. Aren't hams a great bunch?

I'm also missing an unusual component, a 4.7uF ceramic capacitor listed in the operational amplifiers stage as Cxx with the note "for audio test". I can't see where on the board that goes and on a scan through the instructions I couldn't see any mention of it. I'm sure I have a 4.7uF electrolytic capacitor in my parts box which will hopefully do, but I think I'll have to solve this problem when I come to it.

I'm a bit apprehensive at building this kit. I'm going to take it very slowly, in stages, no more than one every few days, not least because I can easily set off my back problems by spending too long hunched over the desk which is necessary to see what I am doing using my magnifier headset.

There are about 4 SMT ICs to be soldered in, and ten SMT capacitors. At the moment, the capacitors look more scary than the ICs. I can't even see how to get them off the backing strip without them pinging off to be lost forever in the carpet. At least the ICs are big enough to hold on to.

Although building the receiver in stages makes a lot of sense, I'm wondering if it would be easier to solder all the SMT parts on first. I think it would be easier to do while I can put the board flat on the bench. As soon as any through-hole parts are on there they will restrict access and make it harder to keep the board in a stable position. I need to think about what is the best way to proceed before I can actually start.

Between a SoftRock and a hard place

A couple of weeks ago I posted a message on the Elecraft reflector to the effect that if anyone was using a SoftRock SDR kit as a panadapter for the K3 and was planning on replacing it with the Elecraft P3 panadapter that is shortly going to be available, I was interested in experimenting with it and would give the SoftRock a good home. Craig, VK3HE replied that he had a kit which was mine for free if I would send my address. I thanked him very much, sent my address and a small jiffy bad with an Australian stamp plopped on to my doormat this morning. Thanks again, Craig!


What I hadn't expected, since I'd anticipated that someone would be replacing a SoftRock they'd built and used, was that "kit" meant it was an unbuilt SoftRock 6.2 Lite kit! Inside the jiffy bag was a polythene bag containing one of the most densely packed printed circuit boards I'd ever seen. There hardly seems space for all the components, which include a few SMT parts. This is going to be a baptism of fire for sure!

However, no building instructions were included. I guessed there was a PDF file on the web that can be downloaded and printed out, but so far I have been unable to find it. I went to the SoftRock Radio website and clicked on the Documentation link and bizarrely, it opened some pages about installing WordPress! This reminded me that during a couple of other wet weekends in the last couple of years I had toyed with the idea of building a SoftRock kit and couldn't even find on the website where to order one!

Google found some pages by Jack Smith that describe building a Softrock and even using it as a panadapter which are going to be required reading, but he doesn't mention where he found the instructions on how to build it. Hopefully one of my more savvy blog readers will help me out here. Otherwise I'll probably have to send Jack an email and ask where he found them.