Saturday, February 8, 2014

Low Pass filter for MF/LF (630 meter and 2200 meter) reception

About a month ago I fired up the SDR-14 (a wide-bandwidth software-defined "receiver") to "listen" to some longwave signals from some (relatively) high-power stations back east.  These stations had obtained FCC Part 5 authorization to transmit on or about 74, 137 and around 470 kHz (yes, kiloHertz!) but it also included FCC Part 15 operations between 160 and 190 kHz, the so-called "LowFER" band.  With the Part 5 operators typically running several hundred watts of RF into their antennas, the low frequencies (for 74 and 137 kHz, at least - 470 kHz is a bit more manageable) meant that they were radiating mere watts - if they were lucky.  Since others in the continental U.S. and a few in Europe were receiving those signals I decided to dust off some of my LF/VLF receive gear and see if I could "hear" them.

I put "listen" and "hear" in quotes as most of these transmissions have been using very slow CW ("QRSS") and/or some very slow digital modes (OPERA, WSJT and WOLF) to transmit their signals. Using these techniques, audio from a receiver would be piped into a computer and detected/decoded at signal/noise levels far below those at which they could be detected by the human ear.

Antenna problems:

Many years ago (in 1986 or 1987) I bought an LF Engineering LF-400B E-field active whip antenna.  While not necessarily a top-of-the-line performer compared to some active antennas these days that use some rather "interesting" circuits to obtain good dynamic range and bandwidth, this whip's claim to fame is that it has a half-decent low pass filter built into it and is able to handle fairly strong, nearby AM broadcast stations without wilting and causing intermodulation distortion.  In my use of this antenna over the years I have had little cause to complain about its performance on that regard.

At my present QTH this antenna has been on the roof for about 15 years and it had worked every time I'd powered it up but on this day, the first time that I'd powered it up in a few months, I heard nothing other than an elevated noise floor and observed the inability to hear all but the strongest signals such as the powerhouse VLF stations run by the U.S. Navy in the 20-30 kHz range in Washington State and WWVB on 60 kHz in Fort Collins, Colorado.

Braving the ice on my roof I retrieved the antenna and opened it up to see what was wrong.  Other than some obvious exposure to moisture at some point - probably due to condensation that had not caused any electrolytic damage since the unit was not left powered up at all times - it looked pretty good.  Touching the gate of the FET on the front end brought a roar of noise but touching anything on the input filter past the second inductor resulted in practically no change.

Removing the three 10 mH inductors in the front end filtering, I put them on test equipment.  The first one in line with the antenna was open, the middle one had much higher than expected (and varying) DC resistance and terrible Q while the one closest to the FET seemed to be OK.  Inspecting them under a magnifier I noticed that on each of them, the epoxy potting the winding and the core seemed to have shrunk away from the plastic, outer casing on all of the inductors, as well as around the leads as they entered the coil.  I can only guess that the thermal cycling of the antenna from well below freezing to hot summer days in the sun, on the roof - along with moisture - must have gradually infiltrated the coils' potting material and broken them down.

Rummaging around I didn't find find an exact match, but only some 27 mH inductors from the same manufacturer and product line (same color, size, etc.) as the originals so I put those in, instead, tested the antenna indoors, re-sealed it, and then put it back on the roof.  Turning on the receiver I was greeted with very strong signals, some 40dB or so stronger than they had been before!  What I also noticed was that I was now experiencing some intermodulation distortion from some of the local AM broadcast stations that I'd not noticed the last time I'd used the antenna.

Simulating the antenna's front end filter using LTSpice I was somewhat surprised to notice that simply replacing the 10mH inductors with 27mH inductors resulted in a worse low-pass response than the original:  I had sort of expected that more than doubling the inductance would have just dropped the low-pass cut-off frequency.  This filter response degradation, allowing the AM broadcast stations to get through better, possibly explained my problems with intermod.

Not sure if it was the SDR-14 or the antenna I threw together the bandpass filter depicted schematically below to place on the output, in front of the receiver:

Figure 1:
Schematic diagram of the (approx.) 500 kHz low-pass filter that could be used for reception at "600 meters."
This filter is intended to be sourced/terminated at 50 ohms.
This filter is bilateral - that is, the input and output are interchangeable.
Click on the image for a larger version.

Initially consulting the low-pass filter tables in an ARRL Amateur Radio Handbook and rescaling the values for the desired frequency, I entered the filter in LTSpice and juggled standard inductor and standard capacitor values that I had in my parts collection until I found a design that was a reasonable performer using more-or-less standard components.  The design of the filter itself is an "Elliptical" filter that includes "notches" to more-quickly achieve a low-frequency cut-off using fewer sections than might be achieved with a "standard" filter such as a Butterworth or Chebychev - this, at the expense of a bit of ripple and ultimate rejection at higher frequencies.

The filter itself was built "dead bug" on a scrap piece of copper-clad circuit board material and then frequency-swept using a function generator with a 50 ohm output, an oscilloscope in parallel with a 50 ohm load and also with a homebrew signal level meter based on an AD8307 logarithmic amplifier chip that also presents a 50 ohm load.

Based on the two methods of measuring the filter attenuation (the 'scope and the meter - both of which actually agreed!) I found that measured attenuation was reasonably close to what had been predicted, achieving at least 50 dB above about 685 kHz:  Not too bad for just a few minutes of number crunching, component tolerances, and the rather mediocre performance of some of these small chokes!

Figure 2:
The completed 500 kHz low-pass filter.  Plastic capacitors should be used, but if you use ceramic units be certain that
they are NPO (C0G) types!  The 22uH chokes were small, molded devices while I happened to have
a different (solenoid) style for the 27uH choke.
Click on the image for a larger version.

Placing the filter in series with the receiver I noticed.... No change in the amount of interference.

My guess is that the "temporary" inductors in the active antenna have compromised the low-pass filter performance of the active antenna enough that its amplifier is being driven to distortion - either that, or one of the transistors or diodes has somehow degraded:  Some new inductors of the same style as the old are on my "running" list of parts next time I place an order.

When I first installed the LF-400B E-field whip antenna antenna at my present QTH I heard intermodulation products from several local AM broadcast stations, but soon discovered that it was occurring due to nonlinear effects in the final amplifier stage of a low-power FCC Part 15 ("MedFER") beacon transmitter only a few feet away.  Disconnecting its antenna made the problem go away.
That wasn't the case, this time!

What is this filter good for?

If I were to use an antenna such as a long-wire or a wide-band active whip antenna that doesn't have built-in filtering, the aggressive roll-off of the AM broadcast band offered by this filter would keep these strong signals from clobbering the receiver - a common problem with many amateur-band transceivers and receivers that include coverage of this frequency range!

In some parts of the world there currently are amateur allocations around 137 kHz and/or in the 400-520 kHz area:  An amateur allocation in the 470-480 kHz range (the so-called "630 Meter" band) in the U.S. is being considered and a filter such as this would be necessary for many existing communications receivers!

As designed, this filter is not suitable for transmitting - at least at anything more than a few 10's of milliwatts - mostly owing to the inherent lossiness of the small, molded inductors and the fact that its cutoff frequency is a bit low, possibly including frequencies of interest:  On receive the 3-6dB loss would hardly be noticed at these frequencies but would be prohibitive for a high-level transmitter stage!

I'll keep this filter around:  It hardly cost me anything to make and it may come in handy if we ever do get a "630 meter" amateur band!  (I may even put it in a box.)

* * *

Did I ever hear the signals from back east once I got my antenna working?

Yes, actually:   While the intermod is annoying, it's not too crippling.  I got a pretty good signal from a station (WG2XRS/4) on 74.3211 kHz in New York state - a distance of approximately 1560 miles (2900 km).  I also received a number of stations around 137 kHz from both Canada and the U.S.

Follow-up on the LF-400B:

I finally did get around to replacing the 10 mH inductors on the input of my LF-400B active E-field whip.

While the original manufacture of inductors were no longer available from Mouser, I did get some Fastron 07MFH-103F-50 units (from Mouser) which were the same size - although lacking what appeared to have been the thin ABS or PVC exterior case if the original and were about a millimeter smaller in diameter.

Upon replacement of the (temporary) 27 mH inductors with these 10 mH inductors, the intermodulation problems went away and the LF-400B antenna is once again working as it should!

Update on U.S. Amateur band allocations at 630 and 2200 meters:

As of mid-October, 2017 the first U.S. Amateur Radio Operators received permission to operate on the 630 and 2200 meter bands.  The easiest activity to detect are the WSPR transmissions occurring around 475.7 kHz (dial frequency of 474.2 kHz, USB) using the "WSJT-X" program by K1JT (go to THIS web page - link).

If you plan to transmit on either of these bands you will need to register with the UTC (Utility Technologies Council) using THIS ONLINE FORM - link:  If, within 30 days you don't hear from them - or get a notification of rejection - you may operate according to the FCC rules specific to these bands.


This page stolen from

No comments:

Post a Comment


While I DO appreciate comments, those comments that are just vehicles to other web sites without substantial content in their own right WILL NOT be posted!

If you include a link in your comment that simply points to advertisements or a commercial web page, it WILL be rejected as SPAM!