"Build to Learn" Q & A

The November 2007 QST carried an article I had written entitled Building Kits to Learn. It's thesis was: "If you want to maximize the learning experience from building a kit, you should build a kit one testable group of components, or stage, at a time." The Pixie 2 was my demo kit. I think you can get a pdf of the article here . Maybe you have to be an ARRL member. The following is an emailed question that I got from that article and my response.

QUESTION:

I just like to tinker, always have, and now I want to build my own QRP CW rigs. Any kits you can recommend? I prefer to learn something and not just be a robot as you put it.

ANSWER:

What kits would I recommend as build to learn? That's a great question and really got me thinking. The Pixie 2 is about the simplest kit I've built and can be gotten from HSC Electronic Supply search for Pixie 2. It's a great kit for making incredible QRPp contacts, but not something you would use every day. On the other end of the spectrum I am currently building an Elecraft K2 and last summer built a KX1. I'm planning on using the K2 as my base station at home and the KX1 will be my main travel rig. But they seem too complex to be good learning radios. I have been following their construction manuals and when you populate 30 or 40 resistors at a time you won't learn much about the circuits they make up.

So what lies in between? The first place you could check out is QRPme That's where I got the Tuna Tin 2 that made up the NeoFish mentioned in the article. You could get a Tuna Tin 2 and the matching Sudden receiver and have a QRPp station that would be simple enough to understand the circuits, but a step up from the Pixie 2. But you would be bound to the frequency of your crystal. And you would have put out about $65 with shipping.

Another choice is Small Wonder Labs Dave Benson is the guy who designed the NN1G I mentioned in the article. Also this is where you can get the Warbler PSK31 that was my first attempt at "build to learn". That kit is not in production but sounds like it will come back. He also makes the RockMite which is fixed frequency tranceiver for $29 with shipping and puts out 1/2 watt. That might be a neat way to go. His SW+ is a much more useful transciever. It's a small superhet. I built one over a decade ago before they had the +. Those are only $55 with shipping and put out about 2.5 watts. A great friend of mine pointed out that the Elmer 101 course is based on the SW+. So it should be easy to "build to learn". But right now shipping is backed up on them. It probably would take several months to get it.

Hendricks QRPKits also makes some nice kits. Their DC40A tranceiver is kind of like the RockMite. It is fixed frequency, and will cost you $39 with shipping. The construction manual is already "build to learn", you build it by stages, testing each. So you wouldn't have to figure out the circuit. Their MMR 40 looks very interesting. It also does SSB and puts out up to 6 watts. It costs $110. I have their Firefly, which is a hybrid Software Defined receiver with a VXO transmitter. It also does a stage by stage build and is $65 or $70 depending on surface mount or thru hole.

If you are interested in Software Defined Radio, Tony Parks makes the SoftRock series of kits, $10 for the SoftRock Lite (receiver) and $32 for the SoftRock RxTx (transceiver). I have a "build to learn" presentation using the SoftRock Lite on this site. Build to Learn - SDR Hardware

The ultimate "build to learn" kit in my opinion is the NorCal 40A David Rutledge used it in his undergrad RF course. I have that kit and Rutledge's book, The Electronics of Radio, on a shelf waiting for my oscilloscope skills to improve and my math skills to be honed. That kit could be used for normal "build to learn", but I'm saving it. Also it costs $145 with shipping.

My recommendation for a "build to learn" kit is the SW+ from Small Wonder Labs. It is simple enough to learn from yet robust enough to be used regularly. With the Elmer 101 as reference it should be straight forward to build and test it in stages. It should be worth the wait. And at $55 it seems to be a bargain. I have ordered one and hope to document it with pictures as I build it.

QUESTION:

Does anyone make a transceiver kit that receives in voice or do they only send and receive in CW-Morse?

ANSWER:

Great question. This one came by snail mail. When we talk simple kits we usually think Morse Code because they are simple and straight-forward. Also at the low power these kits tend to run, SSB or voice is not as effective. Lets expand this to talk about two non-CW modes, SSB and PSK31.

PSK31 is a computer mode that is very effective at low power. My first attempt at the build to learn approach to kit building was Small Wonder Labs' Warbler, an 80 meter PSK31 rig. I made several contacts with it. When that comes back into production it will be an ideal non-CW build to learn kit.

In the SSB area Hendricks QRPKits mentioned in the first question has two interesting kits. The BitX20A is a SSB only 20 meter kit that runs up to 10 watts. If you look at the manual it takes you through the build stage by stage. It costs $82 with postage and looks like it has begun to ship. The MMR-40 is a 40 meter CW and SSB kit, which I mentioned in the first question. It puts out 6 watts. The manual for this is not laid out in a build to learn format. It costs $110 with shipping.

Don't forget about the Software Defined Radio Kits. If you go through my building process, the final stage shows the SoftRock Lite running on Rocky and printing out PSK31. You certainly can listen to SSB with it. If you get a RxTx kit you will certainly be able to transmit with PSK31. If you run the right software you will also be able to use SSB, although at 1 watt it may be tough to make SSB contacts.

QUESTION:

I really enjoyed your article in the November 2007 QST and am in the process of building the Pixie 2. Would you kindly provide me with the antenna filter values for the following: 30m, 20m, 17m, 15m, 10m

ANSWER:

I have been looking around the web for this kind of info. I did find an interesting article for the Tiny Tornado a Pixie 2 variant. It uses toroids for its inductors and had values for 40m, 30m and 20m. It's output filter is slightly different from the Pixie 2 in that it also has a capacitor paralleling the inductor. Here are its values: L1 is 28 ga wire wound on T37-6 core. C2 is the additional capacitor paralleling the toroid.

• 40m L1 = 17T, C1 and C3 = 470, C2 = 150
• 30m L1 = 15T, C1 and C3 = 300, C2 = 91
• 20m L1 = 12T, C1 and C3 = 220, C2 = 75

Here is a note I just received from David Joseph with more values for various bands.

Here are some starting values I wrote down some time ago.  The Q for these are 15.  The capacitor values are rounded to the closest standard value.
Using the kit's pi-net capacitor values and just changing the inductor for 40 does work(made plenty of contacts that way), although the cap values listed below might work better.

Band     Cap(pF)    Coil(uH)
40         420            1.1
30         330            0.8
20         220            0.6
17         180            0.4
15         150            0.4
24         120            0.3
28         110            0.3

NOTE: These values are NOT optimized.  They are ballpark ONLY.  For those who actually calculate the collector impedance of Q2, capacitor values would be adjusted for best efficiency.

QUESTION:

Rich, you authored a recent article in QST about "Build to learn". Tell me what a "soldering robot" is.

ANSWER:

This was a prep question for a pod cast and it really brings up an important point. Is it bad to be a soldering robot? Not really. To my way of thinking there are 3 levels of kit building:

• Learn to build - this is where you become a good soldering robot.
• Build to learn - now you begin to understand what circuits you are building. That's what the article was all about.
• Build to experiment - here you know what you are building and you want to change it, tweak it, improve it. Or maybe you want to build you own design from scratch.

QUESTION:

I built and Rockmite it has a direct conversion receiver with a mixer that mixes a 7040 local oscillator with 7039, 1kc is passed to the audio amp. NOW your pixie also had a local oscillator with 7040 coming in by way of C3. BUT what is it being mixed against??

ANSWER:

Great question, I had a very similar one a year ago and sent it to David Joseph. Here was my question:

Also since I am not an EE I am a bit fuzzy on how the second transistor works as a mixer when the key is up not grounding the emitter. Would love some more light here.

Here was his answer.

The mixer/amplifier concept works like this. As you know, when the key is down, Q2's emitter is grounded, and the second stage is an RF amplifier, boosting the level of the oscillator. With the key up, the emitter of Q2 now passes the signal from the oscillator, injecting it's signal into Q2's base, onto signals appearing on or near the crystal frequency on it's collector. The difference between the oscillator's signal and any signals close to it appear as audio at the emitter. This is called 'direct conversion'. The 10K resistor and .047uF cap on Q2's emitter keep the signal off ground at Q2's emitter during key-up, and help provide some basic RF bypass filtering. The .1uF cap, also attached to Q2's emitter, couples this audio into the LM386 audio amplifier.