Yes, a belated Happy New Year greetings! It’s hard to believe that 2013 is already well under way. I figured it was about time to give you a quick update on what’s going on in the shack right now.
First up is the discrete component grabber receiver for 14.141 MHz that I prototyped to be paired with the OpenBeaconMini project. The receiver itself consists of a roughly 2 kHz wide crystal filter on the front end, feeding into a single-balanced diode ring mixer, which drives an AF amp using 2N4401 and 2N4403 transistors. Because I’m not able to put up a proper outdoor antenna for the grabber right now, I decided to put the VE7BPO cascode active antenna on it instead. It seems to work well, but I don’t know for sure because there are basically no signals on this part of the band. I intended to use my Raspberry Pi with the receiver as a grabber, but I had no luck getting either LOPORA or QRSSVD to work properly and reliably. It may just be asking too much of the poor beast. So I’m going to try to appropriate another PC in order to get the grabber receiver QRV so that on-air testing of OpenBeaconMini can begin in earnest.
Next, I wanted to give you a very brief overview of my most recent purchase for the lab: a Rigol DS1022U arbitrary waveform generator. As far as I can tell, this appears to be pretty much the same as the DS1022A model that is sold in the US. But being a typical ham, I wanted to save a few dollars, so I purchased it off of eBay from seller who says he is an authorized Rigol dealer.
The DG1022[U|A] has two channels that can output a sine wave up to 25 MHz in 1 mHz (as in millihertz) steps. It can also provide square, ramp, pulse, noise, and arbitrary waveforms at lesser frequencies. It can modulate the waveform in a variety of ways, including AM, FM, PM, PWM, and FSK. It can, of course, also do sweeps of various parameters. The output amplitude into 50 Ω ranges from 10 Vpp on Channel 1 or 3 Vpp on Channel 2 down to 2 mVpp on both channels (or -50 dBm). The shielding on this AWG seems to be excellent. Using my HP 355C/355D attenuator combo, I can get a signal down to about -140 dBm (disclaimer: not a scientific measurement, made using my ear as a detector and listening on my IC-718). The dual outputs makes it very useful for a variety of two-tone receiver measurements, one of the big reasons driving my purchase. The Channel 2 output also doubles as a 200 MHz frequency counter input. Paired with the USB connectivity of the device (it seems to enumerate as a usbtmc device), that will be extremely handy for measuring oscillator drift. The DG1022 can also link the two channels together and give them a specific phase difference, as you can see below. This will make it very handy as a I/Q LO when I want to experiment with phasing and SDR rigs.
So far, I’ve been very pleased with my purchase. I don’t feel like I’ve had it or used it long enough to give you a full review, but I thought that this preview would at least be a bit helpful for those thinking about using it. One of my goals for the new year is to do a much better job of characterizing everything that I build. Since I intend to start selling transceivers in the near future, it’s doubly-important that I can make accurate measurements of my products so that I can properly state their specifications. To this end, I’ve decided to sell off a bunch of my unused or replaceable test equipment (please take a look at the for sale posting) in order to finance the new, calibrated test gear. Next up on my purchase list is a Rigol DSA815TG spectrum analyzer (just reviewed favorably in the February 2013 QST), but that’s going to require the sale of everything on that page!
Finally, I’ve got the CC1 prototype PCBs on their way from Seeed Studio right now. It looks like they just cleared customs in the US, so hopefully they will be in my hands in the next few days. With any luck, I’ll have the first one built by the weekend and will be well on the way to a new beta test. I’ll put up a quick post to show off the PCBs, and when the first prototype unit is completed. Stay tuned!
Sorry for the thin content on the blog once again. In the insufficient free time I have, I’ve been swamped with trying to keep OpenBeacon in stock and development of new products going. I’ve got a couple of questions for you, if you don’t mind chiming in.
First, I tried asking a question similar to this on the KnightsQRSS mailing list, but it rapidly devolved into a flamefest and I never really got much good, constructive feedback. So I put it to you. I’m interested in putting an 80 meter version of OpenBeacon on the market. There doesn’t appear to be much 80 meter activity in North America, but what there is seems to be located just above 3.500 MHz. The issue that I’m seeing is that choice of frequency excludes all American non-Extra Class hams. With 80 meters being such a large band, I don’t see any reason why another frequency could not also be used. I’m proposing to put the 80 meter OpenBeacon on 3.582 MHz. If operation was kept between 3.581800 MHz and 3.582000 MHz, I don’t believe it would interfere with any current informal band plans, but I’m not certain about that. I have a very large stock of 3.582 MHz crystals, which obviously also plays a factor (I would be willing to sell them individually to anyone who wanted the for their own homebrew endeavors). So my questions are: does this look like a decent frequency and is this something that would interest you?
The second query is in regard to a potential new product. I’m giving consideration to bringing to market a sort of companion receiver to the OpenBeacon. It could be used for a QRSS grabber or a dedicated monitor receiver for any of the digital modes with automatic propagation reporting such as WSPR, PSK31, or JT65A. I envision it being paired with a small SBC such as Raspberry Pi so that it could make a complete, stand-alone, efficient HF monitoring solution for around $100 total cost (Raspberry Pi currently costs $35). In my opinion, there is a lack of QRSS grabber stations in North America, and using OpenBeacon or other MEPT transmitters will be a lot more fun when there are more stations that can listen for your signal. If you use the receiver for the automatic reporting modes, you can build up a very nice set of data about propagation to your QTH. Here is a list of preliminary specs:
DDS or Si570/Si514 LO for wide tuning range and stability
PC tuning and control via USB (similar to OpenBeacon)
Single-signal reception (probably filter method, but maybe phasing)
Line-level output for PC consumption
So I ask you: is this something you would be interested in? Is there anything feature-wise you would like to see included?
Thanks for letting me pick your brains. I hope you stop by in the comments and leave some feedback!
I’ve got a quick grab bag of OpenBeacon updates for your reading pleasure tonight.
First off is the wonderful find and awesome mechanical construction skills of WA4KBD. He posted a message on the Etherkit forum about an extruded aluminum enclosure that he found on eBay that works perfectly for OpenBeacon. He brought pushbutton S1 and the TX and FSK indicator LEDs out to the same panel as the connectors, leading to the cleanest and best build of an OpenBeacon that I’ve seen yet. Bill also reported much greater frequency stability once OpenBeacon was housed in the enclosure. FB Bill!
There is also some new OpenBeacon firmware available for testing to those who have the ability to in-system program AVR microcontrollers. This update will correct some minor bugs, including a bug in the msgdelay function in CW mode. Importantly, there is also the addition of CW ID mode in the non-CW modes to give better compliance with FCC Part 97 ID rules. All of the details can be found on the Etherkit blog.
Finally, due to some unexpected and unsolicited blowback that I received on the KnightsQRSS listserv regarding the suitability of crystal oscillators in QRSS applications, I decided to look into methods of increasing frequency stability for OpenBeacon. To that end, a crystal heater seemed like the best bet, but they don’t seem to be manufactured anymore (at least to my knowledge). Some investigations let me to discover that one type of heater was simply a thermistor mounted to a metal clip which slipped over a HC-49 crystal. So a bit of research at Mouser led me to a candidate thermistor which gets to about 80°C when connected to 13.7 VDC. I’ve mounted it using epoxy (JB Weld, to be exact) to a heat sink (rumor is that it might be a coin…but that might be of questionable legality). Then the heat sink/thermistor combo was secured to the side of the crystal with 3/4″ diameter heat shrink. I’m in the middle of running tests right now, but initial results look promising. If I have a winner, I’ll post instructions on how you can build your own cheap crystal heater, and might even offer a “kitlet” for sale.
Even though I’ve been insanely busy with home life and running Etherkit, I felt like I needed to get out and do a bit of operating to get back in touch with that aspect of amateur radio. In the past few years, Dave W8NF has invited me to come up to the OTVARC Field Day site at Stub Stewart State Park, but I’ve avoided it due to the fact that late June is usually the time of year when my grass allergies peak here in western Oregon. Fortunately, this year has been a bit of a La Niña year, so it has been unusually wet and mild, which means that the pollen is under good control after a nice rain. A few days before Field Day this year, and my allergies had been pretty mild, so I decided to invite myself up to the public site to check it out and maybe do a little bit of operating.
I arrived at the park at about 4 PM on Saturday. As you can see from the photo above, the weather probably wasn’t to the liking of most people, but it was perfect for me: dry, having just previously rained. OTVARC had four operating positions set up: one CW/digital tent (two K3s), one phone tent (two IC-756IIIs), one VHF (where the above photo was taken), and the GOTA station in the RV you can see in the center of the photo. It’s hard to tell from the photo, but the antenna farm ran in a straight line from where I was standing down towards the pavilion in the background. There were 40 foot masts roughly equally spaced out with fixed beams on them, then wire antennas for the lower bands strung between the masts.
After a bit of chatting with W8NF and some introductions to club officers and members, I partook in the grilled hamburgers which were offered (wasn’t going to eat OTVARCs food, but thanks for the invite!) and then Dave and I sat down in the phone tent to try to work some 20 meter SSB. I was at the logging PC and Dave was manning the mic (each phone station was equipped with Dave’s brilliant Logiklipper, natch). We didn’t have much success for some reason on 20 phone, probably due to the antenna we were stuck with (the G5RV, if I remember correctly). I ended up moving over to the other phone station, which was on 15 meters. I actually had a decent run of search & pounce operation, for a non-contester like myself. I wanted to try to park on a frequency and CQ, but I noticed it was 8 PM by this time, so I needed to get back home.
Both before I left for Stub Stewart and after I returned, I also used Field Day as an opportunity to test out the “mainframe” of a new SSB transceiver design I’ve been working on (meaning the RF stages, minus the microcontroller/DDS/LED frequency display). A breadboarded DDS-60 was used as the temporary VFO for the radio, and I connected the whole works up to battery power to work as 1E OR from my own station. This iteration of the radio is monoband (20 meters on this unit) and QRP (power output is about 7 watts max with a IRFIZ16G final), so I knew I would be a little guppy in a big pond, but figured it was worth a try just to see that it was working properly. I actually ended up doing better than I expected. In about 1.5 hours of casual search & pounce operation, I was able to make 11 QSOs with stations in the sections NE, ID, AK, PAC (x3!), KS, AZ, and NM. The PAC stations were all in Hawaii and were booming in easily 20 to 60 dB over S9. Two of those three PAC QSOs were snagged on first call. It’s a little spooky having such an easy QSO only using 6-7 watts SSB over such a long distance. Yes, all of the heavy lifting credit goes to the other station, but QRP SSB can work if you take care to know your propagation and try to work the stations which are loudest. Given a QTH from a peak (such as a SOTA activation) and a decent antenna, I don’t doubt that it could be quite effective.
So the rig seems to work, and I don’t even have a name for it yet, but it’s shaping up quite nicely. The microcontroller/DDS/LED module is nearly complete, then I’m going to package the rig in a WA4MNT-style copper clad chassis so that I can take the rig with me to Salmoncon in a few weeks. With any luck, beta testing will begin in no more than two months, and hopefully a quick entry to the market after that. On a side note, as much as it pained me to set aside the CC-Series to develop something else, I think it was vitally important for me to do so. I was too stuck in a rut with the CC-Series design and needed a mental breakout to something different. I’ve learned some good circuit design ideas from this radio, which should translate into vast improvements in the next iteration of the CC-Series. I do intend to give my CC-Series beta testers a worthy radio in this next round of testing.
My apologies for being a bit neglectful of the blog. Since a few weeks before FDIM, I’ve been in a mad frenzy to get OpenBeacon kits into production, get the Etherkit website up and running, do a better job of completing OpenBeacon documentation, supporting the inevitable hiccups that come with a new product release, and start working on development of my next kit. Between that and taking care of two little boys during the day, I’m sure you can imagine that something has to give.
So I wanted to let my loyal blog readers know that the OpenBeacon QRSS/DFCW/Hell/CW transmitter kit is in full production and is available for you to order. Currently the kit is available on the 30 meter frequency of 10.140 MHz, but I am working on getting a batch of crystals ordered so that I can start to expand the band offerings. The slow-speed CW modes are an excellent way to experiment with propagation, and it’s a lot of fun to see how far a QRPp signal can go with these modes.
Thanks, hope to be able to give you more content soon!
After a marathon session of hacking firmware, writing documentation, fixing bugs, and setting up my Etherkit shopping cart software, I was finally able to get the OpenBeacon project to a ready-to-sell state. I had enough parts on-hand to make 10 kits, so I spent a good portion of Sunday night getting them all packaged up and ready to go. After a bit of nervous testing, I flipped the switch to the Etherkit store from “sandbox” to “live”.
Once I posted a promotional message to QRP-L and qrp-l.org, I sold all of the kits pretty quickly (within about 12 hours, I think). I wanted to post to the KnightsQRSS listserv as well, but I never did get a reply to my email to the list owner to get permission to post my message. They probably would have went quite a bit faster if I had been able to post there.
There were a few folks who missed out on the sale. I apologize for that, I wish I had more kits to sell now. I did re-invest my sales into a large batch of OpenBeacon PCBs, which should be here in about 3 weeks. There’s a small chance I might be able to sell some more before I leave for FDIM, but if not, I will have them back in stock when I return.
In the meantime, I am working on the component changes needed to get OpenBeacon on 80 meters. I also have a request for a quote on custom crystals out to a vendor, so hopefully I can start expanding the bands fairly soon. I’m looking forward to trying to adapt OpenBeacon to 500 kHz, it should be a perfect kit for the band once it’s available to us US amateurs. Stay tuned for more updates soon!
My single HF antenna (a ZS6BKW) is pretty good for multi-band use, but it is fairly lacking in the 30 meter department. So any 30 meter OpenBeacon captures that I’ve been getting have been pretty exciting to me, given that I’m running 300 mW maximum power out of the transmitter. God only knows what my EIRP is. So far all of my captures have been from North Amercian grabbers, so I’ve been craving that magical “across the pond” capture from a far away land (as a side note, why does there appear to be virtually no functioning grabbers in Asia?).
Tonight, I finally got that capture. Two days ago, I implemented QRSS and DFCW (dual-frequency CW) modes with a very long dit time of 120 seconds. This is to allow a very weak signal to be integrated over a long period, such as the 6-hour grabbers available from the few fine hams who provide them. Last night, wasn’t able to get any coherent signals from the most reliable grabber for me, the W4HBK “Pensacola Snapper”, but I could see traces of my signal present. Tonight I turned on my OpenBeacon a bit earlier in the evening and was rewarded with a very nice capture from ZL2IK after waiting about four hours.
You can see my DFCW signal right in the middle near 10.140.000 Hz, with a bit of upward drift as I opened the door to my shack midway through the transmission. As you can probably tell, all of the other visible signals are meant for much shorter capture periods, so you can’t distinguish any callsign information from them in this long capture.
For those folks with actual decent antennas for their band of choice, this mode will allow them to really push the limits of QRPp. The OpenBeacon output power can be adjusted to somewhere around 20 mW, and with in-line attenuators, the noise floor is the limit. It will be fun to see if some people take up the challenge of very QRPp operation when OpenBeacon gets to market soon.
30 Mar 2012 Edit: Here are some more captures from today. The first is from N9VN (thanks Vince!) and the second is from the Pensacola Snapper.
Now that we are starting to get settled into a routine life with our new baby boy Eli, I’ve been able to sneak in some more work on the OpenBeacon project. The beta test is going fairly well, but I made a poor design decision in choosing a USB Micro-B connector and also made some schematic errors which needed to be patched for the beta PCBs. After we got home from the hospital and while Jennifer had her mom around to help, I was able to get back to Kicad, make the necessary changes and fixes, convert the USB Micro-B into a plain vanilla B and resubmit the files to Seeed Studio for rapid prototyping (also converted my EtherProg AVR programmer side project to USB B as well). Slightly less than two weeks later, and I had the nice circuit boards in hand!
First, I put together the EtherProg board to make sure that my new USB connector footprint was OK. A very quick assembly showed that everything was actually good this time! This gave me a bit of hope that maybe I completed the OpenBeacon PCB correctly this time. A bit nervously, I assembled the power supply and digital side of OpenBeacon, which rewarded me with some nice blinkenlights and proper USB enumeration on the PC. The analog side also went together quite nicely, although I needed to make a few component value tweaks in order to get the desired output power (about 300 mW) and enough harmonic suppression (maximum harmonic content of -45 dBc). A joyous day! Two successful PCB spins!
Now that the hardware is pretty much 100% nailed down, it’s time to turn my attention completely to finishing the firmware. The basic beacon stuff is already in place, such as the QRSS, DFCW, Feld Hell, and CW modes. I still need to add extras such as multi-mode operation, custom glyphs, and multiple messages. But something has been whispering in the back of my mind lately. All of the previously mentioned modes are cool, but they lack the automatic reporting of some of the newer modes. It’s particularly aggravating right now that there aren’t many operational 30 meter grabbers in North America. It would be really cool to be able to add WSPR to the OpenBeacon repertoire so I can just set it and forget it. That seemed like a big challenge, but I have been following WA0UWH and KO7M having all kinds of beacon fun with their Propeller boards, and their efforts make it seem workable.
Thanks to an excellent blog post by KO7M, I was able to suss out the basics of the WSPR protocol and how to implement it in the relatively simple OpenBeacon hardware. The OpenBeacon uses non-linear varactor tuning of a VXO, while KO7M’s Propeller beacon uses very precise frequency synthesis. I wasn’t even sure if it would be possible to fake the necessary phase continuous 4-FSK modulation with the OpenBeacon, but I figured it was worth a shot to at least try to fake it.
Long story short, due to the robustness of K1JT’s protocol and decoding software, I managed to pretty easily get a pre-generated message to transmit correctly with almost no tweaking of the transmitter. In fact, getting the transmit interval timing correct proved more challenging to me than the actual sending of the WSPR message symbols. The firmware is currently very bare-bones, with a hard-coded WSPR symbol string, hard-coded transmit interval (every 10 minutes) and the necessity to turn on the WSPR mode at precisely an even minute interval. Finishing out the firmware will require adding in the ability to change the WSPR message just like the standard message buffer, access to all of the WSPR parameters via the PC client program, and the ability to start the transmission with the pushbutton instead of doing it via the client program. Configuring the WSPR parameters will be a bit manual, but the beacon should be able to just sit there and do its thing once you’ve got that all setup.
So now the goal is to finish the firmware soon and get the Gerber files sent off to my PCB production house for a real production run. And get ready for my talk at FDIM! I know that these last two months are going to go awfully fast.
In the mean time, I’ll be running the WSPR beacon for a while to see what captures I can get off 300 mW on 30 meters. It will also be a good test to see that the firmware can keep the transmit intervals synchronized over long periods of time. If I get any spots in the WSPR DB, I’ll post them here as an update.
Edit: Here are my spots as of 0400 – 19 Mar 2012:
The power is a lie, I’m actually at 200 mW, not 20 mW. Need to fix my WSPR symbol string.
25 Mar 2012 Update: I updated the firmware and client software to allow a WSPR transmission to start on command from the client. This allows me use the much more accurate PC clock to sync the transmissions. When only using the ATtiny85 timer, the best I could do was keep the beacon in sync for about 6 hours before it would drift fast or slow too much. With the PC tethering, I’ve been running overnight and all morning, and have managed to pick up a bunch of spots with my 300 mW.
I’ve got another grab-bag of miscellaneous news for this post, but I’m going to lead off with the big one: I’m going to be a presenter at the world’s preeminent QRP convention: Four Days In May 2012. The tentative topic for my presentation will be about the free and open source tools that I use in the development of my products and how you can put them to use in your own homebrewing endeavors. This will be my first time speaking to an audience larger than about 25 people, so I hope that I can provide an entertaining and informative talk at such a prestigious event. I’ll be speaking in front of a lot of people who I consider to be much more capable than I and some who I consider my virtual Elmers. It is my sincere desire to not disappoint.
I am very excited for the opportunity to go back to Dayton so soon after my last trip. I really didn’t expect to have the chance to go again for quite a few more years, so the ability to get back to the convention after only two years is a great blessing. I owe a great debt of gratitude to Jennifer, who didn’t hesitate to encourage me to go, even though she will be dealing with a 2-month-old baby and a near 2-year-old by herself for a few days while I’m away.
In other news, I feel like I’ve gotten over the steep part of the learning curve with Kicad, having successfully made PCBs for my little Twin-T code practice oscillator. You can see a short video of it in action above. The output level is suitable for modern, sensitive headphones, but if you want room-filling audio such as in my video, you’ll need to connect it to an amplified speaker. The PCB is designed to fit in the ubiquitous Altoids tin, with room to spare for a 9 V battery. I expect that this will eventually make it to my stable of products, but it’s low priority considering the long delay on the CC-Series and the need to get it ready to sell by May. If you are really interested in the project, write a comment or shoot me an email (milldrum at gmail) and I’ll see if I can’t work something out to get you hooked up with a kit early.
The OpenBeacon project is cruising right along. Now that I know that I can successfully make a PCB with Kicad, I’ve taken the plunge and decided to migrate all of my workflow there (I think this will include the next board spin of CC-Series, since there are so many changes to be made there will be no real advantage to staying with TinyCAD/FreePCB). The OpenBeacon PCB design is nearing completion. Once I get a shipment from Mouser in the next few days to verify that my newly-created PCB footprints match the actual physical components, I’ll be ready to submit my CAM files to Seeed Studio for prototype boards. With any luck, I’ll have them back within about two weeks. (Protip: it’s worth taking the time to place your component against a 1:1 printout of your Gerber to make sure it will fit. Don’t ask me how I know this.)
Once those CAM files are off to China, it will be full-bore on the CC-Series. With the deadline of mid-May staring me down hard, I figure I will have to get those CAM files out within no more than three weeks. That will put me into mid-March for getting the PCBs back, which will give a pretty slim margin of time to beta test and prepare the kit for final sale. Going to be pulling some long, late-night shifts…that I can already see.
I’ve also got a few more projects in the pipeline for after FDIM and the deployment of CC-Series and OpenBeacon. The first is a fairly simple and inexpensive VXO DC transceiver that I hope to initially kit for the high bands of 10, 12, and 15 meters. It uses a topology which is somewhat unique. The other is an extrapolation of the receiver circuitry of this rig to use as a dedicated QRSS grabber receiver. But I may be getting a bit ahead of myself. Let’s get this CC-Series launched, then see where the winds take us.
Here’s a quote from Wes describing the equipment that he was using on his end:
I hope that my signal was OK when we worked. I was in the midst of wrapping up a frequency synthesizer project and had it running on the rig for the first time. When I heard you on 20, I could not resist calling. You were the first contact using that source. But I then discovered that the PLL was oscillating. It was a low level oscillation and didn’t present an obvious problem with regard to what I heard on the air. But it was there. I have since then changed the phase/frequency detector circuitry and have eliminated the oscillation. I am not thrilled with the 74HC4046. I get much more repeatable performance from a dual D FF with a NAND gate.