Ripples in the Ether

I got a few pleasant surprises this weekend regarding ham homebrewing websites. First off, I received a very nice e-mail from Jonathan, KB1KIX. He stumbled upon my documentation for the Willamette transceiver (AKA the qrp-l.org Group Project), and took the time to do a very nice write-up about it on his blog. There’s a lot of excellent content on his blog, so I’ve added it to the blogroll. I hightly recommend that you stop by and take a look for yourself.

The other item that I found was the home of the projects of David Forsman, WA7JHZ. I had seen some of David’s projects highlighted on other sites (QRP Homebuilder, SolderSmoke, etc.), but I didn’t realize that he had his own projects site until I came across it randomly this weekend. It’s not strictly QRP (in fact most of the projects are not QRP), but there is a lot of emphasis on lower-power voice rigs (both SSB and AM). There’s a lot of great content to peruse, so get thyself over there right away and start browsing!

A few days ago, I built myself a simple 30 W dummy load out of Caddock power resistors. As I mentioned in the previous post, I was doubtful about the readings that I was getting from my LP-100 (not that I doubted the LP-100 itself, just my calibration of it). Tonight was my first night back to work after a long break, and I finally got a chance to put the dummy load on a calibrated VNA to verify its performance.

The instrument that I used was an Agilent N5230A. A very nice instrument, but one minor drawback is that its lower frequency limit is 10 MHz. Not to worry, as I suspect that if it does well from 10 to 30 MHz, that the lower bands are probably good enough for my home lab.

Caddock 30 Watt Dummy Load - S11 Plot

Caddock 30 Watt Dummy Load - SWR Plot

Caddock 30 Watt Dummy Load - Smith Chart

Sure enough, it turns out that the performance of the dummy load is great and that my LP-100 calibration is off. The return loss is excellent all the way up to the 2 meter band. There’s not much else to say since the plots speak for themselves. If you need a dummy load, you could definitely do worse than to pick up a couple of these resistors on your next Digi-Key order and slap one of these together in a few minutes. Caddock also makes 100 W resistors, which will probably be on my next order for goodies.

1. You can never have enough low frequency decoupling in an audio amplifier with 100 dB voltage gain.
2. You can never have enough voltage regulation on the VXO you are using to feed your diode mixer. Not enough and you get all kinds of nasty low frequency oscillations in the receiver.

30 Watt Homebrew Dummy Load

First off, I want to apologize for the sparse updating on the blog lately. The muse has not been my friend in the last few weeks. I try to keep the content mostly original, but perhaps I will have to turn to that time honored blog tradition of short blog posts repeating something cool I heard on the Internet.

Anyway, on to the good stuff. I remembered in my last Digi-Key order to grab two 100 Ω Caddock power resistors (MP930-100-1%) to make simple dummy load. These were mentioned in a QRP Quarterly article a little while ago as a great non-inductive resistor to use for RF dummy load applications. The datasheet looked good and the price was cheap ($3.51/piece), so I figured I would give ‘em a try when I got a few spare moments.

Dummy Load Close-Up

I wasn’t sure what kind of heat sinking was needed, so I used my most scientific method and took a wild guess. The resistors (in TO-220 packages with a ceramic contact pad) were mounted to a piece of copper clad measuring 2″ x 4″. A bit of thermal grease was smeared on the copper clad before mounting the resistors with 4-40 machine screws. I mounted the resistors so that one lead could be soldered directly to the center pin of the BNC bulkhead connector, while the other lead was soldered to the copper clad ground plane. I figured this should minimize stray inductance.

I gave the dummy load a test drive on the IC-718 set for 30 watts power output. A keydown period of 30 seconds showed a nice SWR on the the rig meter, although my LP-100 showed a reading of about 1.8. The dummy load was fairly warm, but could be handled. I wouldn’t want to key it for much longer at that power level, but I bet it could handle <20 watts quite easily. I expected to see a reading of nearly 50 Ω purely resistive on the LP-100, but surprisingly there was a fair amount of inductive reactance (hence the SWR of 1.8). Now I’m a bit doubtful that I’ve calibrated my LP-100 correctly since I wouldn’t expect such such a lousy reading. I’m going to wait to declare the LP-100 reading bad until I can get the dummy load on a calibrated VNA to make sure that it doesn’t really have this problem. Stay tuned for VNA measurements on the dummy load when I am able to make them.

Baxter Is Uncertain

I finally got a few days of decent sleep (decent meaning more than 4 hours), so I had a little energy to work on the simple DC transceiver. A few days ago, I got the remainder of the audio chain working. The emitter follower on one of the outputs of the differential mixer was yanked, and I connected a class-A audio amp directly to the mixer. Then I stuck the emitter follower on the output of the class-A amp to enable the receiver to drive low-impedance headphones. No, it’s not extremely efficient, but it is simple and it works. Best of all, no transformers are needed. As an afterthought, I added a simple shunt-to-ground mute circuit with a 2N7000. That might have to be tweaked a bit later

Unnamed on the Bench

The transmitter is also a simple design. The second output of the differential mixer is tapped with an emitter follower that will have its V_CC line keyed to control transmit. Directly following this is a 2N7000 class-C PA. After a bit of work tweaking the impedance matches to get the right amount of drive to the PA, I can easily get 2 watts out of the amp (before low-pass filtering). What’s neat is that the emitter follower puts out about +10 dBm, and it gets amplified up to +33 dBm in one stage. A very compact design that can generate a decent amount of power.

So in order to make this a true transceiver, I have a few things left to do. First thing, of course, is to get a low-pass filter on the transmitter. I’ll also need to provide a keying circuit and T/R switching. I’m still not sure what I’m going to do about a sidetone. I also think that I’ll put an RIT circuit in there and not worry about a fixed transmit offset (that would be very hard to get right in such a simple transceiver). Keep watching for another update, hopefully soon.

Unnamed Simple Discrete 80 Meter Rig

Yes, its a post about another simple, low-performing direct conversion receiver. However, I think that this one is slightly unique. I was inspired to give this a try based on the Flea minimalist transceiver that was introduced on the EMRFD Yahoo group. These little rigs are fun to build in an evening, but just how usable are they? Would you feel comfortable giving it to a new ham and believing that they even had a small chance of success? For me, these Pixie-class rigs are nearly unusable due to the horrible AM broadcast interference that blows right through the rig. While a minimalist rig is an admirable thing, they are only useful in limited circumstances. I figure that a few things have to be added to these rigs in order to make them more than a novelty. KD1JV also shares this viewpoint, and has created his own answer to the Pixie.

I’ve started with a similar philosophy, but built the rig around a different topology. The basic strategy is to use a differential amplifier as an active mixer. The rig is designed for the 80 meter band, which is probably the easiest for homebrewing. The LO is a Colpitts ceramic resonator oscillator, but is not separate from the mixer. Instead, the oscillator is built around the third transistor which acts as the constant current source. I know that this is certainly not a new idea; it’s used all of the time in NE602-based QRP circuits. However, I don’t think this topology is seen very much in discrete component use. It saves quite a bit of circuit space and is composed of very common components.

The rest of the receiver is very simple. I placed a standard double-tuned circuit bandpass filter in front of the RF port of the mixer to filter out all of the AM BCB crud. The output of the mixer feeds a dirt-simple emitter follower to transform the relatively high collector impedance of the diff amp mixer to a low impedance output. I haven’t designed the final AF amp yet, but I don’t think it will take much to get the signal up to headphone levels. When the emitter follower output is connected to my test bench AF amp, I have to have the amplifier AF gain control turned nearly all the way down, lest the whole thing start oscillating wildly.

Tonight, I connected the RX to the bench AF amp and the antenna to see how it would work. Tonight was an excellent night to try, since we are right in the middle of Sweepstakes. Pleasantly, the receiver immediately came to life with a cacophany of CW signals in the unfiltered audio output of the receiver. I’ve attached a recording of the receiver output so that you can get a feel for how well it works for such a minimalistic design. The ceramic resonator osc tunes from nearly 3.500 MHz to 3.580 MHz, and I tune across the entire band in this clip.

All I have to do to finish the receiver is to add on a discrete component AF amp. I think that a single class-A stage of amplification will be enough to get the audio up to headphones level. After that, I’m going to try to tack on a transmitter by picking the VFO signal off of the other unused collector port of the diff amp. I think that I can get away with another emitter follower as a buffer, followed by a class-C PA. I’m shooting for around 1 watt of output power, which is enough to snag QSOs without too much difficulty. I think this could be a lot of fun to build as a kit. It’s will be quite a bit more complex than a Pixie or Flea, but also quite a bit more usable. Stay tuned for further developments on this rig.

Code Practice Oscillator Schematic

Here’s quick update to post the schematic of the code practice oscillator that I mentioned previously. As you can see, I just paired a twin-T sine wave oscillator with a buffer amp that feeds directly into headphones. The twin-T provides plenty of voltage, so the buffer is all we need to provide enough current to drive headphones. You can also download a PDF file of the schematic here.

Thanks to a suggestion from David KB0ZKE, I’ve decided to rework the layout to fit in an Altoids tin (an excellent idea!). I’m also going to try to come up with an easily reproduced straight key made from a paper clip, knob, wooden base, and wood screws. This idea was inspired by KE6GS, who has a great example of such a key right here next to his Willamette. I’ll try to get the layout changes finished real soon. The detailed documentation will have to wait until after I finish the Willamette documentation (but I have started on it again, so it will come in the near future).

I know that the blog updates have been a bit light over the last week or so. Although we have been in our new house for three weeks now, it seems like the chores just keep piling up. However, I have done a little bit of work in the homebrewing department. Inspired by messages from WB9VTB and KB9BVN, I decided it might be nice to create a simple code practice oscillator based on the twin-T sidetone oscillator from the Willamette. The discrete component CPO published by the ARRL is really neat because it is very simple and has a unique build method. However, the circuit is your traditional astable multivibrator, which produces a near-square wave. I guess I’m spoiled, but I like listening to a clean sine wave. It certainly doesn’t take any more components to build a twin-T compared to an astable multivibrator.

I experimented with a few different ideas for simple, discrete component audio amplifiers to pair with the oscillator, but settled on perhaps the simplest of all: an emitter follower. The twin-T oscillator puts out a waveform with a fairly large voltage, so all I really needed to do was tack on an emitter follower to provide some high impedance buffering for the oscillator. The entire circuit is extremely simple and produces a pleasant tone at 600 Hz, which you can sample here. The circuit can easily drive low impedance headphones, but if you wanted to listen on a speaker, you would need more amplification. An easy solution would be to plug in a set of amplified speakers, but it wouldn’t be hard to add another stage of amplification.

I’ve created a Manhattan layout for the circuit, and I think that I would like to develop a complete kit with full build instructions at some time in the future. Something that would make it easy for the complete homebrewing novice to successfully build. I know that CPOs are a dime-a-dozen, but I think that the simplicity of this design (2 NPN transistors, handful of resistors and caps, a few 1/8″ phone jacks) is a bit unique.

Today I finished construction on the circuitry for the RMS-to-DC converter described on the October 1992 QST article by W0IYH on measuring receiver performance. This is an article that I found on the EMRFD CD when I was searching for information on measuring the noise figure of amplifiers and receivers. I didn’t find the article directly, but was directed to it because it was referenced in another paper (on the EMRFD CD) by W0IYH on homebrew noise sources. Thankfully, the authors were kind enough to include both papers on the CD, since the October ‘92 article gave a lot of the background needed to fully grasp the noise source article. The detector is a fairly simple circuit, since all of the hard work is being done by the Analog Devices AD636JH RMS-to-DC converter. This device measures the true RMS value of DC and AC signals (up to 200 mVrms full scale) by converting the input signal RMS value to an equivalent DC voltage. The rest of the circuit basically consists of op-amps providing plenty of buffering and one stage of gain. The gain stage is present so that you can use a 10:1 scope probe on the input.

As you can see in the photo to the right, there are some external controls for the circuit. At the far left is a pot which controls the signal level from the first buffer into the second buffer (which preceeds the gain stage). This allows you to accurately calibrate the detector to an on-board 200 mV reference. The rotary switch to the upper-right of the pot is the input coupling selector. You can choose from DC, AC, or GND, just like an oscilloscope. In the middle of the photo, you can see the 10:1 Tektronix scope probe that I’m using with this detector. Finally, in the upper-right corner is a rotary switch that allows you to switch in three different levels of signal attenuation: 0, 3, and 10 dB. The 3 dB and 10 dB levels are set with the little blue trim pots to the right of the probe tip. By having this attenuation built-in to the detector, the process of making NF and S/N measurements is a lot simpler.

Unseen in these photos is the simple dual polarity power supply that I built for this detector. Wall current is fed into a 120 V:12 V transformer, which then is recitified by a 1N4007 and regulated by a 10 V zener for both the positive and negative rails. I’ve still got to get the thing in it’s enclosure, but of course I had to smoke test it first before I went any further. After trimming the offsets on the gain amp and the RMS-to-DC converter, the circuit performed pretty much exactly as I expected. No great surprises, fortunately. I’ve still got a bit of copper clad to trim off of the top edge of the board, then I should be able to get the whole thing mounted in the enclosure and ready to go by this weekend.

This morning we had the first Portland-area version of the Pie & Coffee, which we have informally dubbed the Eggs & Coffee since we are meeting at breakfast time. The meeting (and I use the term informally) was at the Peppermill restaurant in Aloha. In attendance was Dave W8NF, Carl WS7L, and myself. Not a huge gathering, but it’s a good start.

Dave brought along his very impressive solid state kilowatt RF amplifier. He designed this amp while working for Erbtec Engineering, and happened to find one for sale at the Dayton flea market. The amplifier has about a one square foot footprint, but most of the bulk of it comes from the substantial heat sink on the underside of the PCB. You can see the amp right in front of Dave in the photo below. Not QRP by any means, but very cool regardless.

We also got a chance to see some prototype PCBs (you can see them above just to the right of the big amp) and documentation for Dave’s upcoming LogiKlipper RF clipper (to be released soon by Idiom Press). It was extremely interesting to learn about the design choices and trade-offs that went into the design.

Carl didn’t bring any projects, but did bring along an obviously extensive knowledge of ham radio and other technical topics. This was my first introduction to Carl, and I found out that he’s pretty involved in the local ham radio community, including his role as the local VE coordinator. I think we were able to share some of that QRP/homebrewer excitement, as he seemed very interested in the projects that we brought along and discussed. He also gave us a very nice overview of his impressions of operating the Elecraft K3.

I came with an armload of projects and components, as I threatened that I would. I got a chance to show off the pQRP LC Meter, the Willamette DC transceiver, the Tualatin superhet transceiver, my QRP EFHW antenna tuner, and an AVR-based SWR meter. The hit of the E&C seemed to be the LC meter, so kudos to NB6M for producing such a great kit. I also brought along a huge grab bag of components that I tried to pawn off, but the other two were too smart to take any of it off my hands. I did get a chance to give Dave some copper clad to use for Manhattan pads, so I’m happy that they will go to good use.

We chatted for about 2 ½ hours, but it went very quickly with all of the interesting discussion. Our table was smack dab in the middle of the restaurant with radios and parts scattered about, so we got our share of curious looks. Fortunately, the restaurant is very ham-friendly since they already host some other meetings here. I had a great time, and I think that the other guys did as well. I hope to make this a regular occurance, and that we can convince some more of you Portland-area hams to stop by next time.