A Popular Communicator Article from 2018
Several stimulating discussions around the Saturday morning club breakfast table have taken place recently in connection with our use of bandpass filters, diplexers and triplexers. This article is designed to remove some of the mystery surrounding these devices, which we use both at the OTC and at Field Day. Although the discussion relates to HF devices, the same general principles apply to VHF and UHF.
Introduction
As propagation conditions change throughout the day, week and year-to-year, HF stations need to have the flexibility to change to those bands which are open. Typically 20 m is open during the daytime hours with 160, 80 and 40 m opening up in the evening and nighttime hours. In years when sunspot activity is greater, 15 and 10 m also open up during the day. Currently we are near the sunspot low with the result that DX contacts are a challenge at any time of the day with only the low bands consistently productive for DX.
SARC’s first exposure to these electronic devices was ca 2015 when we acquired a set of bandpass filters and triplexer for use with our 10-15-20 m TH7 beam antenna. This was successful and allowed us to have the one antenna on a high tower serve multiple transmitters without significant mutual interference.
Then a couple of years ago at Field Day, we began using an off-centre fed long wire for 40 and 80 m. During the late evening hours these two bands were the only game in town, so the antenna was in demand by two stations simultaneously. Again, a triplexer and bandpass filters allowed this to happen. Alas, one of the devices failed at the critical time.
In 2017, we acquired an identical set of the devices described above for use at the OTC, where we have a tri-band beam for 10-15-20 m plus an OCF dipole for 40 and 80 m. Once again, the 160-80-40 triplexer failed during use.
This could not continue as failures in these devices place expensive radios in danger of serious front-end damage (i.e. smoke) due to strong other-band signals not being adequately blocked. It was time for serious reflection about our physical setup.
A Review of Some Basics
An inductor connected to a capacitor will have a unique frequency at which the pair resonates, called the resonant frequency. At exact resonance, the inductive reactance equals the capacitive reactance expressed as XL = XC and the impedance will either be very low or very high depending on their parallel or series configuration. The effect of resistance in any practical circuit does not change the resonant frequency but it does affect the sharpness (or Q) of the tuning.
In other words, an inductor in series with a capacitance has a low impedance at its resonant frequency, but the same pair connected in parallel exhibits a high impedance to the flow of current. These properties are the basis of many types of radio circuits, used most notably for tuning purposes. They can also be deployed in various combinations as RF filters and in power supply filters to change pulsating DC to “pure” DC.
A low pass filter will pass low frequencies and block high frequencies. A high pass filter does the opposite. Bandpass and bandstop filters allow a band of frequencies to pass or be blocked, respectively. The figures above show the generalized frequency response of the 4 basic filter types.
Below are some simple examples of L-C circuits used in practice for the various kinds of filter devices. The presence of R in the circuits represents loads but otherwise does not affect the general type of filter and can be ignored for the sake of this discussion.
Intuitively, it is not difficult to determine which type of filter it is by examination of the circuit, if you think of the way L and C respond to low and high frequencies, whether in isolation, in series or in parallel when presented with a range of different frequencies.
More complicated circuits have been devised that improve the performance of these basic circuits and make them more useful. A study of such devices will bring forth variations named for the engineers who studied their properties, such as Butterworth, Chebyshev, Cauer and Bessel. More complicated circuits are not within the scope of this introductory article, but a comprehensive discussion can be found in any ARRL Handbook.
The complexity of a filter circuit is described in terms of its “order”, a measure of the number of L and C elements. Here, for example, is a 4th order high pass filter:
Practical Devices
A diplexer allows two transmitters to feed one antenna or, conversely, two antennas to serve one transmitter (don’t confuse a diplexer with a duplexer, which is a different animal). A diplexer simply consists of a low pass filter and a high pass filter operating in parallel, with the cutoff of each somewhere between the two operating frequencies. With an HF unit used to separate 40 m (~7.0-7.3 MHz) from 80 m (~3.5-4.0 MHz), the cutoff frequency typically would be 5 MHz.
A diplexer may be able to discriminate 80m from 40m signals by 20-40 dB. While 20 dB represents a power suppression of the unwanted signal by a factor of 102 it is insufficient to protect the radio.
That is why an HF diplexer is seldom used by itself. A bandpass filter in series with the diplexer might suppress the unwanted frequency an additional 40-60 dB depending on its design. So the diplexer and bandpass filter, operating together, would typically suppress the adjacent band signal by a total of 60-100 dB or a factor of 106-1010.
If a triplexer rather than a diplexer, is desired to facilitate a third band, the problem becomes more complex. The “middle” frequency would necessarily have to be a bandpass filter.
One problem is that the size of components for diplexers and triplexers for 160, 80 and 40m bands will be large. This size factor and associated high cost generally make high power diplexers, triplexers and bandpass filters quite costly.
Our Devices
Our Dunestar triplexers appear to be rather simple filter circuitry. Why do these units fail repeatedly, even with the radios operating at 100 watts? It can only be inadequate current or voltage ratings on the components or excessive SWR, or both. This would suggest that the antennas connected to the triplexer should be close to resonant at the desired frequencies. Operating at extreme ends of the band, especially under Field Day conditions when time does not always permit “tweaking” of their length, height or configuration may produce unacceptably high SWR.
Here is the lesson we have learned: carefully research the characteristics of the diplexer or triplexer you are considering for purchase. Not only are the band isolation and insertion loss important, but the need to have conservative voltage and current ratings on components is critical. Then do not deploy these devices on antennas where a near resonant condition cannot be achieved.
Conclusion
We will probably replace both our Dunestar 160-80-40 triplexers with more robust devices to ensure another failure does not happen. Units available from VE6AM (www.va6am.com) and DX Engineering (dxengineering.com) and 4O3A (www.4o3a.com/products/high-power-filters/combiner/) are under consideration to meet this need. [In the end we went with VE6AM's product, which has given excellent service]
More good reading can also be found at:
https://static.dxengineering.com/global/images/technicalarticles/lbs-pb-tp500_sn.pdf.
~ John VA7XB
No comments:
Post a Comment