W6IWI

The call letters W6IWI were first used by Kauko Hallikainen in the 1930s. See the 1938 Amateur Radio Callbook. The 1930s QSL card was similar to that shown above (I may still have one of the originals somewhere). I acquired the call in 2016. Prior to that, I held the call WA6FDN, and prior to that, WN6FDN. The WA6FDN license was probably first granted in 1967 or thereabouts, with WN6FDN a year earlier.

WN6FDN started with a Heathkit DX-60 transmitter and a National NC-300 receiver running CW on HF. WA6FDN used a Viking Ranger transmitter running AM, CW, and RTTY on HF. RTTY used a Teletype model 15 printer, and a model 14 typing reperf and transmitter distributor. W6IWI now uses an SEA 245 running CW and SSB into an inverted V antenna in Arvada CO. VHF and UHF FM are covered with a Baofeng UV-5R and a Wouxun KG-UV-6X

W6IWI HF Activity

The plot below shows a historic plot of W6IWI HF CW activity.

Recent activity (in the past day or so) can be viewed here. These are both generated by the Reverse Beacon Network.

Power Line Interference


The image above shows the output of my SSB receiver tuned to 7.000 MHz LSB. The trigger is the AC line. The image is stable with this trigger indicating that the noise is synchronized with the AC line. Further, the noise peaks are about 8 ms apart, in sync with the half period of the AC line (8.333 ms). Note, the scope can save images to a 3.5 inch floppy disk, but I don't have a computer that can read them right now. So, you get cellphone pictures!

Below is a view of the noise waveform from the audio editing software Audacity.



and this is what it sounds like.

Here's more information on power line noise.

Here's a closer look at the power line noise at the receiver output. Most of the noise seems to be coming from a single phase of the power line though there may be noise from other phases causing some of the noise between the large noise spikes.

The first image above shows a capture of a couple of the noise bursts. The next image shows us zoomed in on the first noise burst. Note the peak amplitude is about 400 mV. The next image is the noise between the noise bursts. Its peak amplitude is about 50 mV. The final image above shows the second noise burst. Its peak amplitude is, again, about 400 mV. This puts the power line noise bursts about 18 dB above the background noise (daytime 40 meters). Calculations below put the power line noise at about S6 (AGC=118.2, input level = -92 dBm). If we could eliminate the power line noise, the remaining ambient noise would be -110 dBm or S3. The receiver equivalent input noise appears to be about -118 dBm.


The plot above shows the receiver (tuned to 7 MHz LSB) AGC voltage (actually, AGC data captured on the EIA 485 bus between the radio and the control head) over a 24 hour period (Sun Feb 04 16:06:43.103 2018 through Mon Feb 05 18:25:37.946 2018). At Mon Feb 05 15:43:41.673 2018, the antenna was disconnected to see what the reading was due to receiver noise. With the antenna connected, the average value was 118.2. With the antenna disconnected, the average was 10.2. See below for the relationship between AGC voltage and receiver input level. The fact that the noise is continuous over a 24 hour period seems to eliminate the source as grow lights or most household appliances. It could still be something that is powered continuously, such as a doorbell transformer, but current efforts are concentrating on power line insulators.

An ARRL power line noise mitigtation form is available here. A report of the noise was submitted on 2/10/18. On 2/26/18, I was visited by a tech from Xcel Energy. We rode around in his truck. He identified an improperly installed lightning arrestor on a pole a couple blocks from here. I could not really detect the noise with my MFJ-856 noise receiver, which receives at about 135 MHz, but he got it on his 300 MHz receiver. He could also hear the noise with his parabolic ultrasonic microphone. He could see that the ground on the lightning arrestor was improperly installed. A crew should be out to fix it in the next week or so. It's quite possible that this is only the first of several problems, so we'll see what happens to my noise level and if it is not fixed, we'll see what's next.

There's also a set of high voltage lines going through a park a few blocks from here. In the past, I could hear what sounded like arcing from these lines, but it appeared to be mid-span instead of at insulators on towers. I could not hear it yesterday. Both the MFJ-856 and the 300 MHz receiver showed strong signals in that area. We'll see what happens after the lightning arrestor is dealt with.

We talked about other problems he's found. One was a doorbell transformer. Turning off power to it made the noise go away. Turning it back on caused the noise to come back after about 5 minutes. Another problem is an electric fence. The owner would not let them turn it off for a test, so the FCC is being contacted on it.

Update 3/26/18 - Just got a call from the power company tech. They cleaned up two poles (at 53rd and Independence and at 52nd and Garrison) today including the one mentioned above with the bad lightning arrestor. Unfortunately, the noise level here did not decrease. He will be out again next week to check further.

Update 4/4/18 - The tech visited again and said he's found more poles with leaking insulators. Apparently road salt and other debris settles on the poles and insulators causing them to leak and, in some cases, catch the poles on fire . He'll do more investigation next week.

Update 4/20/18 - The tech was working on this yesterday and today. He identified an area around 51st and Iris that seemed noisey. He turned off the power to two houses that were suspect, but the noise here remained. There is possibly a problem on the pole feeding these two houses. Connecting his EMI reciver to my antenna, he was easily able to receive the noise on 7 MHz and capture a trace. However, connecting it to his mobile antenna, he was not able to receive the noise. Using his 300 MHz receiver, he did find noise at the intersection mentioned above. I have not really been able to find the source using my MFJ-856. The antenna on the MFJ-856 is really broad, so you have to look for a null off the side. The direction of the null seems to be different on the left and right side of the antenna, further adding to confusion. I took a VHF AM receiver (aircraft receiver) around trying to determine where there was the most noise. However, in many locations it was overloaded by FM broadcast stations. I did find high noise levels at some high voltage towers with broken gound wires. But, I don't know if missing grounds on these towers would cause RFI. You can hear the noise at test.hallikainen.org. This is the output of my HF receiver which will be tuned to 7 MHz LSB unless I am using it for something else.

I am very impressed with dedication of the power company tech to solving the issue. It's a difficult problem, and he's not giving up.

Receiver AGC vs Input

To get an idea of signal strength, data was gathered on AGC voltage (actually AGC count captured from the EIA 485 bus between the radio and the control head) versus receiver input level at the center of each band. A SARK-110 (thanks to Jack, KE0VH for the loan of this excellent instrument) antenna analyzer was used as a signal generator to drive the SEA 245. The raw data is shown here. A plot of the data for 40 meters, as generated by https://mycurvefit.com/ is shown below. This was generated with an auto-smoothed spline fit. It is extended down to the measured noise level on the receiver (AGC count of 10 corresponding to an input level of -118 dBm. From this curve, the 40 meter noise level measured above (AGC = 118.2) corresponds to a receiver input level of about -92 dBm which is 26 dB above the receiver internal noise. S9 is defined as -73 dBm with each S unit being a change of 6 dB. On 40 meters, an input level of -73 dBm gave an AGC count of 184. The power line noise of -92 dBm is 19 dB below -73 dBm or 3 S units below S9. The power line noise is, therefore, about S6.

Antenna Analysis

Since I borrowed the SARK-110, I may as well look at how my antenna and tuner work. The antenna is a multi-band inverted V with the center up about 25 feet. It has wires cut for 40m, 20m, and 10m. The 40m is oriented east-west. The others are pretty much north-south. They are all fed with a single 50 ohm coaxial cable with no balun. The coax is about 30 feet long with a lightning arrestor about 5 feet from the antenna tuner at the base of the mast. The lighning arrestor is at the top of a ground rod that goes down 8 feet. A wide braid connects the ground rod to the antenna tuner.

Below is a Smith chart plot of the antenna (at the bottom end of the transmission line, where it connects to the tuner output) from 1 MHz to 29 MHz. The minimum VSWR is at about 13.97 MHz where the impedance is 50.6-j .09 ohms. The green portion of the plot is for frequencies below the mid-frequency of the scan (15 MHz), and red is for frequencies above. The red circle in the middle corresponds to a VSWR of 2.0.

The table below shows an R+jX and VSWR sweep of each band from 160m to 10m. The left columns show the impedance and VSWR at the radio (before the tuner) after the tuner autotuned to the center of the band. The right columns show the antenna impedance as measured at the output of the tuner (before the transmission line going up the mast about 25 feet).
BandImpedance at RadioAntenna Impedance at Tuner Output
R+jXVSWRVSWRR+jXVSWR
160m
80m
60m
40m
30m
20m
17m
15m
12m
10m

.

Web SDR

Until the power line interference issue is resolved, most HF receiving is done using Web SDR. A truly amazing project that lets you listen to receivers around the world. See here for a history of early web SDR hardware. In its basic form, a web SDR is a high speed ADC (for example, the LTC2216 16 bit ADC running at 77.76 MHz) driving an Ethernet interface to a server computer. The server provides a user interface to multiple users, demodulates the user chosen frequency, streams the resulting audio, shows a waterfall plot of the surrounding spectrum, and many other features. To me, this is truly amazing! The web SDR may also decrease the Ethernet bandwidth requirements between the ADC and the server by only sending selected frequency ranges (bands). In this case, digital down converters are included in the FPGA between the ADC and the Ethernet PHY. Just as in analog, a digital down converter multiplies the incoming RF by a "local oscillator" and filters the output to the desired spectrum (and removing the image). The local oscillator is a direct digital synthesis sine wave generator (a phase accumulator determines the phase of the local oscillator at each clock edge. The phase is passed to a sine lookup table to generate the sine wave local oscillator signal). "Mixing" is just multiplication of the sine wave local oscillator with the incoming RF. The resulting product is filtered (often just a low pass filter) to remove the image and define the received band. Resulting samples can now be down-sampled since the highest sampled frequency is lower than with the incoming RF. This reduced bitrate signal is sent over Ethernet to the server for further processing. Again, truly amazing!

Fun Stuff

Contact

Contact me with any comments at harold@w6iwi.org.