: Splitters, Attenuators, Filters, Diplexers, Other Signal Gear
2011-11-27, 04:01 PM
Yes, the frequency response is somewhat narrow band, but the common mode rejection of a half-wave balun is probably more narrow band than a well-constructed voltage balun for the VHF/UHF band.
I think you're right about that. On the other hand, at low frequencies the differential voltage across the parallel line becomes shorted by the low impedance of the halfwave line, which is no longer a half wavelength. That's why I can't use a halfwave coaxial balun on my UHF antenna - I also want to receive VHF signals with it.
2011-11-27, 07:35 PM
I have several push-on ferrite baluns with series capacitors in the balanced lines. I always figured they were there to help reject AM signals (I have one tuner that had intermod on FM from a local AM signal). These indoor baluns would have long twinlead feedlines attached. I wonder if those capacitors are there to lower the common-mode response to nearby lightning strikes.
2011-11-28, 11:43 AM
...push-on ferrite baluns with series capacitors in the balanced lines. I always figured they were there to help reject AM signals ... I wonder if those capacitors are there to lower the common-mode response to nearby lightning strikes.
I don't know. It does seem plausible.
For some time, I have appreciated your FM & related articles.
You discuss several concepts and principles that apply to much of RF & communications engineering.
2011-11-28, 12:27 PM
This morning I found a paper that consolidated a number of measurements of lightning specta from several sources (Google "lightning spectrum" and click on the first result). When I saw that lightning power at 10 Hz is 100 dB greater that at UHF, I became a believer in the importance of low-frequency common-mode balun rejection (or some other method of protecting a receiver input from low-frequency power). Ferrite baluns at least have a transformer winding across their output, which tends to short low-frequency energy. Below its operating range, a halfwave coaxial balun passes common-mode energy with no attenuation all the way down to DC. Thanks for the heads-up on this.
2011-11-28, 03:25 PM
I think there's TOO much emphasis on LIGHTNING, overlooking STATIC ELECTRICITY
buildup that occurs nearly every day as wind blows across the antenna elements
(like shuffling your feet across a carpet). This DC voltage can easily exceed the
breakdown voltage of a Preamp's input capacitor (which it SHOULD have). And
will be much worse when there is nearby electrical activity:
A conventional transformer Balun provides DC (and AC) isolation between the input
and the output.....an additional layer of protection NOT provided by a Coax Balun....
2011-11-28, 03:48 PM
The capacitively coupled input circuits I've looked at all use a resistor to drain any static from the antenna terminal. I can't recall ever seeing a floating input capacitor. But if your receiver has one, just add a 1meg resistor from the input terminal to ground.
There's no reason for a preamp to use an input capacitor if its input circuit is biased at ground. Example:
2011-11-28, 04:15 PM
CM7777 & HDP-269 use C-L-C Pi-Networks on their inputs, so no DC discharge path.
CM0265DSB Spartan (300-ohm input) has a DC discharge path thru the input Balun
on both ports, followed by a C-L-C Pi-Network (VHF inserts an additional filter).
2011-11-29, 01:35 PM
Last night, I gathered all the combiners that I had at this location, planning to measure isolation and excess loss.
After some effort, discovered that a TruSpec FAM-3 (3 dB attenuator) is somehow faulty.
This particular attenuator causes poor isolation to be seen between the two output ports, when it is used along with a 75 Ohm terminator to terminate the input port of combiner/splitter. This probably means that it does not exhibit a 75 Ohm resistive impedance.
I am fairly sure this same attenuator was used with the SD-3700 measurements, maybe at the output port of the SD-3700, between the SD-3700 and short coax plus another attenuator connecting the RF power meter.
Bottom Line: The attenuator could have affected the test results presented:
I will try to remove the previous chart, until its accuracy can be verified.
2011-11-29, 11:59 PM
Someone was asking about where to get attenuators. I just bought 2 at the source today, so yes they do still stock the 6db dc-pass attenuator.
2011-12-01, 12:17 PM
The graph of Winegard SD-3700 loss, originally located in a previous post (http://www.digitalhome.ca/forum/showpost.php?p=1355409&postcount=1382), is shown below.
The above graph was removed from the original post, per comments in this post (http://www.digitalhome.ca/forum/showpost.php?p=1356681&postcount=1408).
The graph results are labeled as having a known equipment anomaly.
Frequency responses were re-run as originally described, with substitution of attenuators used in test setup.Results are plotted below:
The average of each test: - The first four (with known equipment anomaly)
- The recent four measurements
Shown together on this graph:
In the recent measurement, an increase of attenuation at FM and VHF is observed.
In lower portion of the UHF band, improvement is noted, but increased loss is evident in the higher portion.
The cause of the FAM-3 attenuator to appear anomalous is unknown. The device does not represent a very close match to 75 Ohms. Several other attenuators were tested for 75-Ohm match. As one might expect, there are some variances between devices and between manufacturers.
Since, in real-world applications: the impedance of various devices and transmission lines are rarely perfect matches.
It could be that either measurement, the first or the current, represent something akin to what might be seen in an actual installation.
2011-12-01, 01:27 PM
tripelo, did you take into account the suggestion that the SD-3700 is designed as a combiner and will likely be a very bad as a splitter? Most of the losses could be in the splitter half of your setup.
2011-12-01, 01:53 PM
...Are we sure that ...the SD-3700 works as well as a splitter, as it does a combiner? That symmetry is a fundamental assumption your testing was based on.
... did you take into account the suggestion that the SD-3700 is designed as a combiner and will likely be a very bad as a splitter? ...
That the SD-3700 could not function as a splitter was not considered.
Combiners and Splitters have much in common. A combiner may have greater precision requirements than a splitter, especially with respect to phase. For example, a splitter may not require that the two divided signals precisely track in phase. A combiner requires some degree of phase alignment for two signals to combine without loss (addition of vectors).
For a combiner to not be capable of performing well as a splitter (when reversed), would mean the device is non-reciprocal.
Non-reciprocal devices in RF are commonly associated with special use of ferrites and wave fields, such as circulators (http://en.wikipedia.org/wiki/Circulator) or isolators (http://en.wikipedia.org/wiki/Isolator_%28microwave%29). At the moment, my thoughts are:
If an SD-3700 can combine efficiently, then it can split efficiently.
Of course, my thoughts are subject to change.
2011-12-01, 03:02 PM
In post #1390, 300ohm provided a photo of SD3700 interior....some Baluns and
a Ferrite Transformer...all of which are bi-directional. But someone should run an
insertion loss test when used as a Splitter to verify the currents go the way they "should".
He also cites a Patent, which is found in post #1316:
2011-12-01, 04:46 PM
Thanks for redoing the combiner measurements. I think the losses are way too high for combining two UHF-TV antennas.
You can match 150 ohms (paralleled 300-ohm antennas) to 75 ohms with a quarter wavelength of 106-ohm line. A 5" length of air-dielectric line yields an SWR of 1.00 midband, 1.25 at 470 MHz, and 1.22 at 698 MHz. The mismatch and conductor losses are negligible.
Two #12 or #14 wires with their surfaces 1/32" apart exhibit an impedance near 106 ohms. However, it's not clear how to maintain that spacing without using periodic spacers, which will alter the impedance and the electrical length of the line. The effects can be accounted for, but you'd have to use a specific material in exact amounts.
Air-dielectric coax constructed with an inside diameter of 0.375" and a center conductor of #14 wire has an impedance of 106 ohms. I think a 5" length could be constructed with end supports only (or possibly a single thin center spacer, which should have negligible effect). Place a couple ferrite sleeves over the tube at the 150-ohm end to inhibit outside surface current. Or parallel another quarterwave tube and connect the two at the 75-ohm end.
A chassis-mount female F-connector is slightly smaller than 0.375" and cannot be directly threaded into the tube. However, one I tried fit snugly with good electrical contact in a 3/8"-ID tube when I shimmed the threads with aluminum foil. This would have to be epoxied in place. This arrangement provides both center-conductor support and the feedline connector. At the other end I think I would try to use a rectangle of plastic so that incoming 300-ohm twinleads 180 degrees apart could be tie-wrapped to it to avoid broken wires. This piece could cover the end of the tube with a hole for the center conductor. I'm not sure how I would keep it in place. I would probably use a hose clamp to make electrical contact with the tube. Perhaps the clamp could also secure the end piece. This end of the tube would have to be carefully weatherproofed.
Anyway, I would definitely try to build something like this before I sacrificed a couple dB to a lossy commercial combiner, patent notwithstanding.
2011-12-02, 04:24 AM
FYI: antennahacks used the same test configuration to test two back-to-back
(75-ohm) RF Splitter/Combiners as was used to test the (300-ohm) SD-3700,
except for the pair of different type cables interconnecting the Splitter/Combiners:
Be sure to divide the TOTAL Loss displayed on the Spectrum Analyzer by TWO.
Bottom Line: NONE of them are what I would call "low loss" in UHF Band,
although VHF isn't TOO bad.....Wilkinson Stripline devices are much better:
See the back-to-back test configuration at the bottom of this webpage:
So one is used as a Splitter (generating two "nearly identical" signals as from
two different antennas) and the other is used as a Combiner....hence losses
are due to internal resistive losses plus internal amplitude and phase
mis-match losses....which would be IN ADDITION to antenna mis-match.
It's amazing that anyone gets antenna combining to ever work at all.....
2011-12-02, 05:31 AM
Some LOW LOSS, possibly Wilkinson Stripline Couplers:
Telewave (San Jose, CA) stocks PARTIAL UHF BAND (Ch14-20 or Ch14-34)
2/3/4-port splitter/combiners (ANTPD24, ANTPD34, ANTPD44).
Call to verify whether Wilkinson and ask about other freq. ranges:
Telewave's entire line here...note UHF band gap...and freq. coverage diffs:
Kathrein/Scala made PD2-LLN, PD3-LLN and PD4-LLN low loss splitter/combiners, but
they only operate on UHF Ch14-19...perhaps they have another suitable product???
Triax (in U.K.) makes full-band UHF combiners for 2 and 4 antennas (see pg36):
You'll have to email them to find a distributor (prob. in U.K.) and find
out what the REAL loss is through their box...their specs are confusing:
I have not seen confirmation whether they are truly Wilkinson Stripline couplers...
2011-12-02, 08:05 AM
Splitting and combining are generally reciprocal functions of the same device, same as diplexers.
2011-12-02, 10:19 AM
Here's another way to combine two 300-ohm UHF-TV antennas with low loss. Tie the 300-ohm feeders in parallel at the balanced terminals of a halfwave coaxial balun. The optimum cable impedance for this kind of balun is half of the balanced-side impedance, so 75-ohm RG-6 is perfect. The output impedance is 37.5 ohms. Match this to 75 ohms with a quarter wavelength of 50-ohm cable. Total loss should be no more than 0.15 dB from 470 to 698 MHz. Using a velocity factor of 0.83 (the exact value depends on the specific cable), the halfwave line would be 8-3/8" long and the quarterwave line 4-3/16".
Although not absolutely necessary, I would place two Fair-Rite 2661665702 ferrite sleeves ($1.14 each at Mouser) over the 50-ohm cable at the 150-ohm end to inhibit current on the outside of the coax shield. Weatherproof everything.
I thought I had read that mixing subscription satellite, OR FTA, with OTA is ok, and doesnt present any performance issues in any of the transmissions. Is this correct ? (because I also remember reading that some satellite frequencies will cause interference problems on OTA)
Also, from what I read, a simple diplexer (satellite/OTA) is used to combine the leads to a single downlead, if this is correct, are there any other requirements for the splitter, or its placement position?
2012-01-02, 02:37 PM
You certainly can mix satellite or FTA, with OTA on the same coax, but there are certain issues. The first one is that the diplexers do provide a certain amount of loss to the signal, so check the specs of the ones you buy. Secondly the diplexer won't pass power to your OTA pre-amp.
What you might want to do is have separate leads from your antenna and dish to the distribution panel (where you put the pre-amp power injector), and then combine the signals using a diplexer for distribution through your home (and then use additional diplexers to separate the signals at each TV for your satellite receivers and OTA tuners).
It is cable TV that can't be mixed with OTA.