NOISE

**June of 2009 important update**

It is never too late for learning they say... Indeed, after a few members from the Perseus newsgroup alerted me about the strange behaviour found on my wav files (available for download thanks to Ken Baird in Scotland), I was left intrigued. Actually, when the preamp was engaged, it was contributing in lowering the noise floor while the signal was enhanced by a large margin. I was really convinced that the RPA-1 was doing just an exceptional job despite an animated debate on this issue over the forum. However, with all that chat around this theme, I began to have serious doubts about my setup as it was.

Then, the key message (#4966) that hit me came from Bill W4ZV in response to the following (beginning of quoted text):

>>What *is* interesting is that the noise drops while the signals increases, like magic....A >>"normal" amplifier can not see any difference between wanted signals and un-wanted ones!

>I suspect the problem here is common mode noise being fed to the Perseus but being >rejected at the RPA-1. Knowing how fanatical W8JI (RPA-1's designer) is about common mode >noise, I suspect he has an isolation transformer on the input to the RPA-1 which is shunting >the noise. I can't say for sure because there is no published schematic and the photo on DX >Engineering's website is too small, but that's my guess. (end of quoted text).

Yes, common mode noise rejected by the RPA-1... Just presuming of course. I then decided to follow Jürgen Bartels’ advice to fight the noise source from my system using a feedline choke in order to get back to a "normal" behaviour. It seemed to be a fast and easy way to achieve that.

This is the choke

I used for

preliminary testing. I made it according to Jürgen’s design using a FT140-77 core and I've wound over it as many turn as possible of  RG174, each end  terminated with a BNC connector

While I was locating the test choke in shack hoping to find where it could make a difference, I found that inserting the choke between the preamp and the phaser made a huge change on the noise floor. Actually, it dropped a lot. I also found that engaging the preamp made both noise floor and signal to rise. It wasn't "recovering" the signal as it was before. Bingo I said to me. Now I'm getting a proper s/n ratio from my system. Nevertheless, the noise was still a little high. I knew that I had to pursue in that direction with further testing.


The next morning, I was horrified to find that even with the choke in place, the bad behaviour was back again so I began to move the cables around the choke and the problem  went on and off. It was definitely something around there. First, I thought of a faulty PL259 socket on the MFJ phaser chassis but after it was replaced, the problem was still there. Now, only a 30cm piece of cable that was used to connect the RPA-1 to the MFJ-1026 was left to be responsible.


Indeed, THAT was the problem. Replacing the cable, the s/n ratio came back to normal.

An apparently harmless cable causing all that trouble! There must have been a fault in my coax preparation or a hidden imperfection when the cable was manufactured. One thing for sure is that I couldn't find anything unusual after a visual inspection over both connector areas.

Now, with this major problem fixed, the time has come to "really" play with chokes. Noise floor was still a little high so there was room for improvements. I decided to start from feedpoint and later compare from the shack. So, I brought my laptop and Perseus outside straight at my feedpoint. I connected the receiver with a 1m cable to the matching transformer. This way, I had no long feedline involved in the measurement. I used a 470 ohm resistor to ground  instead of the regular 950ft wire. The two Perseus shots bellow are explaining the effect of the choke.  

The choke mounted in a waterproof plastic cabinet It literally connects through the feedline close to feedpoint. One leg of the matching transformer coax side has to be grounded for the choke to be efficient. Connectors are simple F female chassis mount. Hot glue has been used to tie up the choke to the box.

This without the choke inserted, noise floor @ 895 KHz  was at -110dBm. Note higher noise peaks around 200 and 600 KHz. This noise was probably generated from the laptop and picked up by the short feedline

This with the choke inserted showing how the noise peaks were cut besides bringing the noise floor down to -114dBm @ 895 KHz

This shot has been taken with the Perseus in shack using the faulty cable. Antenna switchers (in shack) were in the "off" position leaving a 75 ohms dummy load for the receiver through both preamp and phaser that were in their bypass position. A total of  ±50cm cable run. The noise bellow 750 KHz was generated by the external HDD switching power supply. Note that @ 930 KHz, you can read -113dBm. Not shown here but with the faulty cable fixed, the power supply noise disappeared and the noise floor was kept near -118dBm from 100 to 1700 KHz. Engaging the RPA-1 would have reduced that noise I presume, by blocking common noise as stated by W4ZV. A test I didn't made.

So, -114dBm with a dummy load and receiver directly at feedpoint. Now, see bellow with choke and the beverage connected. At 895 KHz, s-meter (not shown) was showing -94dBm hence it makes me assume that the antenna and/or location are/is indeed a little bit noisy.

Now, what I  have left in order to get the noise lower is moving the feedpoint farther which is out of question. Don't expect me to make another trench in the forest and get the feedline buried. However, I still have that MFJ-1026 phaser along with at least a second Beverage which can be useful for getting rid of the extra noise. Watch this short video showing how a probable nearby light dimmer is interfering over the longwave range. For the purpose, I tuned a reliable TA LW: Morocco 171. Then watch and listen what is happening with phaser engaged. Impressive isn't it?


Initially, this phaser was used to null out adjacent or very close MW stations helping out revealing weaker TA . Frankly, I rarely thought in using it for cancelling noise. My only concern is during a Perseus wideband recording. I doubt that over both LW-MW bands, it could give a flat and stable result. Some stations could be altered and other noise sources could remain present. Anyhow, this should probably be the best to be done for now.


I really want to thanks Jürgen Bartels for his great availability in this quest and also Leif Asbrink SM5BSZ for taking time analyzing my on/off preamp file (provided by Ken Baird) with his legendary strictness and kindness.


I've been triggered after numerous comments on the Perseus newsgroup about the "miraculous" effect of the RPA-1 in cutting the noise while giving back the signal strength I should already have had. As I've stated, fixing this problem brought back to life my K9AY and also revealed the usefulness of my MFJ phaser (modified) down to 100 KHz in which I had given up in the past.


choke_demo.wmv

***2011 update***

At the time the chokes were installed, I went using a single large common ground connection for both Beverages, the matching transformers, feedpoint’s remote control gear and chokes. Since I’ve been plagued with some buzz noise especially in the LF-VLF range after the above article was written, I decided to use a dedicated ground rod for each of the feedpoint component, five in total. Even

if they aren’t much widely spaced between each others, the final

Click on the image  to play in-shack choke demo

result was positive nevertheless and the noise went down significantly. It also made a noticeable change in the Beverage directivity which is always good news. Encouraged by the results, I decided to add two more chokes in-shack to see what happen. Fortunately, the result was positive again. I made the same attempt in the past but for some reasons, I didn’t succeeded. Maybe I was just looking for results on MW where the difference is far less noticeable, I don’t recall exactly. See the short video showing the difference once the in-shack choke inserted over

a portion of the VLF band.