As described in my other BLOG posts, I have two KD9SV reversible BOGs (Beverage On Ground) oriented along NE-SW and NW-SE directions. They are both 300′ long (rabbit proofed 12 AWG wire with two parallel conductors) and lay directly on the ground (rocky soil; low conductivity). On 160 m band I use KD9SV preamplifier to bring the signal to approximately the same level as my inverted L antenna (inv-L). I do not have such preamplifier for 80 m at the moment (but use the built-in preamp2 setting on the K4). In addition to my BOGs being perhaps a bit too long for 80 m (but note that they still work on 60 m), this lack of external preamp may have reduced BOG performance on 80 m when the band is quiet (as it was during this test). Naturally ionospheric conditions (arrival angles, polarization) have also contributed to the results. There were no major thunderstorms nearby during these tests so QRN was minimal. If storms were located away from the direction of BOG, its RX performance over inv-L (approximately omnidirectional) would improve significantly. Since FT8 uses 15 second RX/TX window, the results obtained represent average signal to noise ratio (SNR) over that period. Note that inv-L antenna is mostly vertical polarization whereas the BOGs are horizontal and collect signals very efficiently at low arrival angles. Here are the results collected using the multi program (described in other BLOG posts) for 80 m band (EU direction / NE; only EU stations included):

Note that the multi program reads the audio data simultaneously from the two RXs in my K4, which are connected to inv-L and BOG antennas. The audio data is processed by two separate JT9 processes and aggregated into single WSJT-X instance. The X-axis represents SNR of the decodes obtained on inv-L and the Y-axis the corresponding SNRs on BOG. Each point represents a simultaneous decode from the two antennas. With this axis choice, the points above the straight reference line indicate decodes where BOG SNR was better than inv-L and vice versa for the points below the line. If the given antenna could not decode the signal, this was arbitrarily chosen as SNR -26 dB (below the -25 dB FT8 decode level). The colors indicate how many decode pairs landed on the same point (scale shown on the right; see “Many points” on the graph). Interestingly BOG was able to decode everything that inv-L could but with only one exception (inv-L only). On the other hand, BOG was able to receive many decodes that inv-L could not (BOG only). The intermediate group of blue points below the straight line (where inv-L had better SNR) mainly consisted of one station (calling CQ forever) that was somehow always received better on inv-L – possibly due to the band conditions. While not always giving best SNR, BOG was able to decode significantly more stations than inv-L could. Note that there have been situations where a QSO could not have been completed by using just one antenna: one decode was obtained only on inv-L and the other only on BOG. This can be achieved with the multi program, which automatically picks up the strongest decode from the two receivers / antennas.

Here is the corresponding data collected on 160 m band:

While this dataset is smaller than for 80 m band, the results here show a much clearer trend. A large number of stations could not be decoded on inv-L (BOG RX only) and those that are received on both give better SNR on BOG (i.e., points above the reference line). Only two decodes had slightly better SNR on inv-L than BOG and inv-L did not hear any stations that BOG could not. The RX performance of BOG on 160 m band is excellent!

I have KD9SV 160 m preamp that I would like to use on 160 m only and have it turned off for all other bands. This unit does not have any computer control – just the old style rocker switch to turn it on/off. So, I decided to keep this manual switch in the on position all them time and install USB relay to control the 12 V power to the unit. This cheap USB relay switch seems to work well (model DSD TECH SH-UR01A):

It sells for $15 on Amazon. It appears as a RS-232 device on the computer (Linux system in my case), which requires baud rate of 9600, 8 data bits and no parity. To turn the relay on, use:

AT+CH1=1

and to turn it off use:

AT+CH1=0

Note that newline or carriage return should not be sent. I use this in my antenna control program (ant.c) where I now have a button to toggle the preamp state.

I have had terrible time with rabbits chewing on my BOGs. In order to rabbit proof these antennas, I have replaced the original construction with 12 AWG two conductor wire (300′ long), which is embedded in a 1/2″ split conduit:

I used electrical tape at several points to make sure that the wire cannot come out of the conduit split. I have a nice tool for putting the wire into the conduit which made the it easy! It would be hard to get the wire in there by hand.

It is supposed to be Fall but it is still well over 90 F outside and humid. Not the best weather for antenna work!

I gave FT8 intro presentation at our ham radio club: http://aa6kj.hopto.org/ft8-intro.pdf

and a quick demo by remoting to my station.

I have had S7 power line noise (RFI) on 160 m for several months now (as measured on inverted-L antenna). I spent almost a month collecting observations and search for possible RFI sources nearby.

The first step was to characterize the noise source. I first observed that the noise goes away during rain, which is expected for arcing power lines. Otherwise it was constantly on 24/7 but its intensity varied slightly from day to day. During heavy winds the noise sometimes showed erratic behavior, like a bad contact. Setting my radio to AM mode on 160 m band and checking the WSJT-X water fall (starting from 0 Hz rather than the usual 200 Hz), one could clearly see peaks at 120 Hz plus its multiples. Occasionally weaker 60 Hz and multiples were also seen. All this is consistent with power line arcing. So, now I knew what I am looking for.

To locate the source, I first used my 160 m receiving antennas to get the approximate direction and then concentrated the search in that approximate direction. Since it could only be heard on the 160 m band, it is likely not very close to me (> 1 mile distance). Google Earth is a great help in mapping out the search area. When I eventually drove near the RFI source, the noise showed up on my car AM radio but then I failed to observe it on portable 120 MHz AM radio, which indicated that I could not get very close to the troublemaker (private properties with barbwire fences, animals etc.). Unfortunately, AM radio observation is often not very reliable because the noise can be conducted through the power line wires (antenna) and radiate over a large area. Fortunately, in this case the information I provided turned out to be sufficient for the power company to locate the problem.

I detailed all the findings (including proof that the RFI is not from my property) and sent a letter to the power company (Craighead Electric). They received the letter and responded the same day (9/28). Even though I was not able to pin point the exact source of the RFI, they found the problem & fixed it in just a couple of hours (9/30)! The actual broken equipment was located using ultrasonic microphone equipped with a parabolic dish. The RFI was caused by arcing of this jumper (see the brown piece on the left where some wire has evaporated):

Over time this would have led to power line failure since the wire was getting evaporated slowly. When it gets thinner (increased resistance), it begins to heat more (on top of the heat produced by arcing) and this accelerates the failure process. This broken equipment was located 2.2 miles east from my QTH. There is still a smaller issue in another location but they promised to search for it soon. This RFI source I am able to hear with my 120 MHz AM handheld radio, which can determine the location more exactly. Interestingly, it can only be heard weakly on the 10 m band at my QTH, mainly because of the higher natural noise floor on the lower bands. It is approximately located 2 miles northeast from me. I used my 10 m yagi to get the direction for the search (signal maximum & side null). It has the same signature as the 160 m noise (120 Hz + multiples and goes quiet when it rains).

Now my day time noise floor sits neat at S3 on 160 m. It turns out that the grandfather of the electric company technology chief is a ham and that probably helped.

Excellent service from Craighead Electric!

Update: The smaller RFI source has disappeared on its own. The hunt will continue if it comes back.

My Acom 2000A amplifier started complaining “bias too low at power on -56 V” when turning it on. I pulled out the tubes (GU74B / 4CX800A) and did some simple resistance checks, nothing stood out. Then based on the advice from Acom (fast and good customer support) and a local ham, I decided to put in another pair that was known to work. Well, the amplifier came back on without any issues. Here is the amplifier with the new tubes put in:

And these are the old tubes (of which at least one is bad):

Now I need to finish plate current calibration and I should be all set to go again! If you have to change your tubes, drop Acom an email and ask for instructions. They have a very nice writeup on how to do the procedure. Even though Array Solutions is cooling down their operations (retirement), Jay can still do tube testing. However, there is a bit conflicting information on “gettering” this type of old tubes. On the internet most people say that one must do this for old Russian GU74b’s whereas Acom technical support said that it is not needed. Go figure… Just to be safe, it is probably a good idea to getter them before using.

Except then leads of one of the temperature sensors snapped. This has LM335GE label on it. Now need to replace that. According to Acom, LM335Z will work. I will try to put flexible leads on them in order to avoid breaking them again.

Update: New LM335Z is in and the amplifier seems happy. Also the “new old” pair of tubes look produce full power output.

In order to better catch 6m DX conditions, I wrote a small program to alert me about the band openings (called ealert). I have WSJT-X Improved running 24/7 with filters set to DX only, so it will broadcast only those decodes (see WSJT-X network settings). The program then catches those UDP multicast messages, parses them, and forwards them either to a given email address or XMPP address (using Gajim). The latter works better for me since it does not litter my email box. It has a simple GUI that allows turning the alerts on/off. I have XMPP client (Conversations app) on my cell phone that catches the alerts.

Even when I am not at the radios, I can use my cell phone to run VNC remote desktop to my station and run FT8 QSOs. This combined with the alert system should improve my presence on 6m whenever the band opens up.

Installed Morgan M-303 lightning protection to all coax cables:

It is not fully weatherproof so decided to install the first one inside the shack:

with grounding outside the shack:

The other two units are outside and just in case I wrapped everything with electrical tape:

I always disconnect all cables during lightning stroms but this is just to protect against those random strikes…

One interesting feature of this unit is that it fully blocks DC. This is important during rain/snow/sleet precipitation when static electricity can develop at the antennas and appear at the coax cable. I experienced this last winter when my K4 suddenly started flashing overload!

Note: This unit can handle full legal power on 6m band unlike most other surge protectors (e.g., polyphaser).

Update: If you have antenna that has elevated SWR (greater than 2), the DC block circuitry in the unit can heat up and cause gradual increase in SWR when using full power (1500W) on digital modes like FT8. There is a beefier unit in preparation, model 305, but that does not support 6m like the 303. It is possible to bypass the DC block in 303 to avoid this issue. Or even better, tune the antenna properly…

My NW/SE BOG suddently went dead. It had 30 dB less signal than on the NE/SW BOG. After inspection, I found the obvious reason:

and then this one was the “final straw”:

Added a fresh piece of cable there and all is well again. However, I definitely need to get better/thicker cable for my BOGs so that mice can’t chew them through.

Update: There are apparently wire coatings that are either rodent proof or have some chemical added that they taste bad. Need to look into this.

Update 2: I installed conduit in the area where the critters were feasting:

Update 3: The mice got falsely blamed for this – it was the rabbits! It is now wabbit season….

Due to humid summers and “cold” winters in Arkansas, I need to keep an eye on how the temperature and humidity are in my shack. So, I decided to use Adafruit SHT4x Trinkey M0 USB half-key device for this purpose. It costs about $10. The GTK+4 code (Linux) that I wrote can be found from github. This is what the device looks like:

Since it is USB A half-key, I used electrical tape to secure it to the USB extension cable. Here is what the simple GTK+4 widget looks like:

I will probably add graphing to it at some point but this will enable me to monitor my shack from remote locations. Too bad that there is no USB device that can get rid of insects! One had crawled inside the LCD display of my amplifier and now I have to figure how to get it out 🙁

Well, here is the graphing added to it: