Sunday, February 28, 2010

Notes From SDR

I tend to always head down the Flex path with SDR, but Jim W4GB sent me a link that is worth looking at. It is a compilation of articles regarding the theory of SDR. I covered a lot of this early in my blog, but after close to 50,000 page hits it is worthwhile .

If you really want to understand the heart of SDR I suggest starting with Dan Tayloe's article

Just perusing the articles you can tell what is important in the SDR arena. Things like oscillator jitter and low low noise amplifiers

The other thing that is most important is to understand that SDR is done with DSP digital signal processing. This means the processing is going on in virtual time, not real time. The simplest analogue radio is a crystal radio aka a diode detector. A signal hits the antenna, hits the diode and audio is immediately produced. This is real time. In a commercial ham receiver you add amplifiers and detectors and filters and blankers into the signal path stretches things out a little bit between antenna and ear, but the signal is processed in virtual time with the SDR and that means you can manipulate it much more dramatically (aka mathematically) with far fewer components.

Much good stuff to consider on these pages and thanks to Jim for pointing this site out.


Saturday, February 27, 2010


Steve K5FR has been up to it again!! Steve of course is the author of the fine DDUTIL utility. He has added a memory feature to DDUTIL that expands the memory capability of the radio.

At the bottom of the form are the V<>M Recall, Save, Load and Clear buttons, and some boxes that readout freq mode and memory number. The memory system revolves around 5 tunable scrolling memory slots and a way to get in and out of the memory stack using the V<>M button, and I think this is one of the best features of this applet.

You save memories simply by tuning your VFO to what ever freq you want to save, and hit save, for example

I'm on 7020.308 CWL at 25hz band width all I do is hit save and this is memorized including VFO B data into the next available memory in this case memory 1. The memories are volatile and first in first out, with the idea that you only want to scratch pad the data, not remember it forever, and the oldest data is the least desirable. If you want a more permanent memory you can use the system that already exists in PSDR I can return to this memory at any time simply by queuing up memory 1 by hitting recall to scroll me through the 5 memories and then hitting load to send the chosen memory into the Flex radio. If for example you are listening to a pileup but the pileup is too big, or the propagation still needs to build, you can set up your station, mode filter offset and all that and simply memorize it. You can then continue to tune the band. If you hear another pile up you can memorize that one also and toggle between the 2 memories.

Here is the nice feature, you can also toggle in and out of the VFO. To enter the memory stack you hit the V<>M button. When you do that your current VFO freq is memorized into a hidden 6th scratchpad memory, and the last memory slot you used is then loaded into the Flex radio. This means you can continue to tune up the band, or tune another band and keep checking back on your pileup with just one button click. If you are not ready to pounce on your prey, you simply hit V<>M again and you are placed exactly where you started before you entered the memory stack. Mode and filter etc is memorized so you can be listening to a SSB QSO up the band and instantly check your pile up with a single button click VERY COOL

Here I am listening to a SSB QSO on 7238 listening to some guy give bad advice on antennas

I hit the V<>M button and I am immediately transported to 7.020.308 CWL, the contents of memory 1.

Notice how 7.238 is now displayed in the window next to V<>M which has turned yellow. This informs you that you have entered the memory stack. If you hit the V<>M button again you are back at 7.238 ready to tune up the band some more.

Slicker 'n ______ (you fill in the blank)


Sunday, February 21, 2010


I decided to crank the F5K up in the ARRL CW contest this weekend. I am by no means a die hard contester. In fact I was on call Friday and I was up all night in the operating room and not on 40M. I have spent so much of my life awake all night that the idea of doing it on purpose just to work a bunch of EA's and LZ's going 50wpm no longer appeals to me.

There is a lot of FUD broadcasted regarding the F5K's contesting capability. This is a FINE contest radio, in fact better than my Orion was

Here is the setup at W9OY

I'm using the latest PB-PAL code and drivers, DDUTIL, CW skimmer, and N1MM. I use CW skimmer for the spotting function, and this is the first time I have tried this particular setup. Skimmer can act as a Telnet server and it serves up what it copies to the N1MM band map, so all those stations in the band map were populated in the map by skimmer. You can also have your band map populated by a DX cluster if you like, and this is the way I have run the station in the past. One thing to be sure you HAVE to copy the station before you enter the callsign into the log entry screen in N1MM. Skimmer just decodes what ever it hears, and what it hears could be the other side of a DX contact. For example if UT9DX was on 14.020 and was being called by K3RR you may just as easily have K3RR in your band map as UT9DX, so this is assisted, but not very assisted. What it does do is give you a clear picture of what is on the band at any given time. One thing N1MM does is keep you abreast of new multipliers and that is a good feature especially for SnP and SO2R operations. I decided to run this contest at 100W instead of full power. I don't have great antennas. I have a 43ft vertical to use on the high bands 20, 15 and 10, and I have a 135ft open wire dipole to use on those bands as well, and a 160M 80M, 40M and 20M dedicated verticals. On 20 the flat top is almost always better than the verticals, so I figured running this contest with this antenna compliment and 100W would be a challenge and would be a good example of what the average guy could do with simple antennas.

I ran all bands from 160 to 10, but I only put in about 3 hours total doing SnP. The radio performed flawlessly. Here are some shots of the band noise

Here is 15M

and here is 20M

This is about what my band noise runs on 20M contest or not. There was not ONE beep or boop in the RX passband that I did not expect to hear (as in overload) even though the band was chock full of super strong stations.

Later in the contest I decided to switch from Skimmer feeding my band map in N1MM to the W9AZ DX cluster server. I think over all I prefer the DX cluster feeding the band map and Skimmer over to the side doing its own thing. With the cluster feeding the band map it was super easy to just go down the list one after another picking up multipliers. Using DDUTIL there was no problem controlling the rig either from N1MM or from Skimmer so I just hit the call I wanted to work and BAM I was on freq. For my keyer I used my K1EL USB version, but I ran it from push buttons instead of N1MM

I later switched to the F3K using a laptop, and my 43ft vertical on 20M and used N1MM and skimmer to do the duties. I had this radio set up on a separate power supply as well essentially a completely separate station. Running the 3K single band was basically just like running the 5K except I didn't have all the auto band switching of antennas set up on the 3K like I have for the 5K. My 43ft vertical has a MFJ auto tuner set up at the base of the vertical so I just tune that on any band I want to work.

The F3K however does have the capability to seamlessly integrate completely into my auto band switching station. All I would have to do to integrate it is go unplug the F5K, plug in the F3K into my station computer, change RF output cable on my patch panel from F5K to F3K and plug in some headphones to the F3K For the average guy the 3K is a very good choice for a contest radio as well and would be a good choice to throw in the bag and take on an island DX pedition. The receiver in that radio is superb.

I was only able to operate a scattered 3 hours 10 minutes total time during the contest but it was a lot of fun. It was great to see activity on 10 and 15. There was an amazing number of 160 and 80M stalwarts holding the fort and 40 and 20 were just nuts

73 W9OY

Tuesday, February 16, 2010

Orlando Hamcation 2010 plus F3K post script

I attended the Hamcation this past weekend and had a great time. The attendance seemed a bit down, but I was speaking with some vendors who said business was pretty good so maybe that's a good sign for the economy. Everyone of course was there with their version of some kind of SDR

Yaesu was there with its latest 5000 radio, my impression HO HUM

Elecraft had their little answer to the pan adapter, what a joke. It was a little screen maybe 9" and made the radio look like a toy. The quality of the display was horrible.

I swung by the Flex booth and met Greg Jurrens and Steve Hicks and they had the Flex radios set up on 40" 1080p monitors, one on the F5K and one on the F3K and it totally cracked me up to see the difference in performance. The displays were beautiful (of course no prejudice here :)

They also had the F-1500 on display. Man what a cute little bugger. It was a little bigger than I anticipated but still quite small. I think this radio will definitely be a winner. I'm not a big QRP kind of guy but I still can't wait to get one and give her a try. This will make a very nice TINY little rig to hop on a plane and use for a contest. You can see the block diagram of the radio here

When I got home from the hamfest I was thinking about the contest aspect of the Flex line and decided to see how small a set up I could come up with. I took my F3K out of the shack and grabbed a Samlex 25 amp switching supply I had in the drawer and unplugged my Begalli Graciella paddle a mic and some headphones, and decided to set up a radio station.

I set up the F3K on a couple of hook up wire spools to get it off the ground.; For an antenna I used my 45ft end fed wire vertical I had hanging out of a tree in my front yard

You can see the ATU sitting at the base, here is a closer shot

In the Ammo can is a MFJ-929 auto antenna tuner. I run power to the tuner over the coax and I initiate a tune by applying 10 watts to the tuner from in the shack. The tuner will tune this wire from 160-6 meters and I have made contacts on all the bands (except 60M, NO CW). I turn down the power some on 160 because you can generate a LOT of reactive voltage at 100W, but I have run 100 W with this setup on 160 with no problem. This could easily be an antenna you could take out to the field or on a DX pedition

For a computer I used my kids Lenovo R61i laptop. I bought it from NewEgg a couple years ago for $450. It is a 1.7ghz dual core mobile computer with 1mb L2 cache and 1gB of memory with a 15" screen. Their computer was virginal as far as SDR software is concerned and I wanted to see how hard it would be to get this system running.

First was the .Net frameworks that had to be installed. I googled .Net 1.1 and and .Net 3.5 and installed both and the SP updates. Next I downloaded the latest driver, F3K firmware and Release version of PowerSDR from the Flex download page. I decided to time how long it would take to set up the radio, from driver, firmware, PSDR, to first contact. I started the setup at 18:15. I loaded the driver, which required a reboot. I next powered up the radio and the computer recognized the F3K and intalled the usual 3 parts of the driver. Once I was talking to the radio I updated the firmware, and then installed PSDR. Once installed I went to setup and setup my usual preferences. I often download the alpha code which sometimes requires a new database so I don't have a lot of customization in setup, but I do like to optimize CW so I set the audio buffer to 48khz and 512, and the DSP buffers to 4096 and 512. I measured the DPC latency at about 200 for this computer. After I closed setup I turned on the radio and switched to 40M and heard EA6UN on the Balearic Islands calling CQ. I worked him first call with a 599, 100W to my trusty wire vertical. He was in the log by 18:29 14 minutes from setup to completed contact. Not too shabby.

The radio prompted me to do a re-calibration after the new firmware install, so I did a CTL-SHFT-P and brought up the calibration screen. I put a dummy on the antenna, checked off the recal tests I wanted to do and, hit the start button and went and got a coke. In about 20 minutes I had everything recalibrated and all screens were green and the new data was saved in the EPROM.

With my newly calibrated radio I next set up SSB using the metering choices provided. I set up the mic gain to 0dB and set up the EQ and the EQ gain, checked the leveler and the ALC and set up the compander to +3. I looked at the waveform on the scope and found it to be acceptable and listened to the audio in the headphones. I switched up to the phone band on 40M and there was YU7ZEX over in Serbia. He had a little pileup going, and I gave him a call. He came right back and gave me a 59 and good report on the audio. I next switched to 80M and worked US7WW on CW, 559 both ways. By now it was closing on 19:00, so in 45 minutes I had 3 DX contacts in the log basically starting from nothing except having the .Netframework already loaded.

I next installed VAC (virtual audio cable) and VSP (virtual serial port) and configured these and linked them to PowerSDR. I next installed DDUTIL, CW skimmer and the entire DX lab suite. I setup winwarbler and made a PSK31 contact. Instructions for setting up winwarbler with PSDR are here in the Flex knowledge base Note that there are no added physical modules, weird little panadapter boxes or extra cables to get all this going. It's all done in software.

I found that as I added software my CPU usage started to climb. It was around 15-20%, went up into the 30% with Skimmer and went up into the 60's with DX lab with SpotCollector running and around this level is where you start getting dropouts. So my kids little $450 computer does have its limits. I did nothing to optimize this computer just loaded the software and hit start so I could probably get some improved performance by shutting unneeded stuff off, but since this is my kids school computer and they use it online for classes I decided to leave well enough alone. With the addition of DDUTIL this radio becomes frequency agile and you could use the LPT port to drive antenna switches and Steppir antennas and amplifiers to automatically band follow you as you change freq. The receive quality was marvelous, basically as good as my F5K in this environment. There has been a lot of improvement in CW switching as well. The weight of the entire station comes in under 20lbs excluding coax and the ammo can. If I was taking this station to field day or something I would simply wrap the tuner in plastic and save the weight, but I have found the ammo can to be basically indestructible and bone dry even in the wildest FL deluge so it has worked out to be a good choice

For fun I decided to see how the station would look with the radio in the cabinet next to the desk. I occasionally see people requesting info on a small radio they could use in the office for example, HERE YA GO high performance radio with virtually no footprint


Monday, February 8, 2010

End of an Era

This entry is not about ham radio. I live and work in a town that is scattered with remembrances of manned space flight. One of the local high schools has a 2 stage rocket in front of it. The other high school is named "Astronaut". Many of the grade schools hold names like Apollo or Challenger. We have parks with monuments to Mercury and Gemini.

If you look across the river from the hospital I work in, you can see the VAB the vehicle assembly building, where the shuttle is readied for launch.

The VAB is 525 ft tall. The Statue of Liberty is 305. The volume of the VAB is 129,428,000 cub ft. The volume of the Pentagon is 77,025,000 cu ft. The VAB equals 3.5 Empire State Buildings. It covers 8 acres. To paint the flag on the building required 6,000 gallons of paint. It is hardened to beyond a cat 4 hurricane, something greater than 140mph sustained winds

The Space center sits on a little jut of land called Merritt Island that sits next to the Atlantic:

You can see launch pad A and B next to the ocean I live out in the sticks in the upper left corner of this picture in a place called Turnbull and I work in Titusville a few miles south.

At a little past 4 this morning I witnessed the letter perfect liftoff of the last night launch of a manned space craft this country will likely ever perform, and it made me feel very very sad. The temperature was about about 45 F and 63 seconds after the candle was light, the tremendous wave of noise that is put off by the explosive power of such a device hit my house. Because of the cool temp the air was particularly dense and the rumble was very loud, and in that instant I realized I would hear that rumble only a couple more times until it would go silent likely forever. (63 sec*1100 ft per second at sea level =13 miles = the distance between me and the candle)

What started with Kennedy will end with Obama because there will be no political will to start up the program once it has closed. The energy barrier to reignite manned space flight will be too high. The thousands of trained technicians will disperse into the world, perhaps to places like China and India and Russia to man their space programs. The engineers and scientists will disperse as well and the whole superstructure of human capital will be laid to waste, and the expenditure of energy required to reignite manned space flight will never be expended. Other countries will take up the gauntlet, and other countries will reap the rewards, and yet another thing in which we presently lead the world will be no more. I think it's very short sighted.