26-12-2012, 01:01 PM
FM constant impedance combiners. Theory and tuning
1FM constant impedance.docx (Size: 934.91 KB / Downloads: 34)
INTRODUCTION
FM combiners are an integral and crucial part of high power FM broadcasting. This document will cover combiner signal flow, tuning procedures, and intermod calculations for balanced FM combiner systems. Also included is a list of tests to perform during acceptance testing or following combiner retuning.
Many new combiners come with a manual detailing theory, tuning, and response curves. Read it to become more familiar with the particular model in use. RTFM!
Robert Surette at Shively Labs1 has written a paper that explains the operation of FM combiners and includes signal flow diagrams. It is a good starting point to gain a better understanding of the history, components, and operation of combining systems. A pdf version is available. Jampro has an article that simplifies RF combiner theory.
http://www.jamprouploads/tech_docs_pdf/c...theory.pdf
Objectives
- Install FM combining systems to minimize signal distortion and attenuation. Ensure optimum performance for all stations.
- Perform acceptance testing on new systems as supplied prior to installation and operation.
- Recognize symptoms of problems. Perform basic troubleshooting.
- Understand tuning techniques.
Purpose of combining systems
Due to (or to prevent) tower congestion, a broadband antenna and single feedline can be used to serve multiple FM stations. This may result in revenue generation by supplying antenna capacity to other broadcasters. Advantages of balanced combiners
Bandpass filters provide good intermod reduction Emergency (wideband) port available Easily expanded Disadvantages Can be physically large, especially for high power and narrow spacing. Footprint can sometimes be more than four times that of the transmitter!
Adapter kit
High quality adapters from Type-N to EIA flanged and unflanged line sections will allow measurements on coaxial lines and patch panels. Bent, dented, or home made adapters have questionable characteristics and should be avoided. Final results will only be as good as the poorest adapter.
Type-N Adapters and Connector care
Never twist Type-N connectors during installation. Only the outer threaded shell of the connector should rotate. The RF contacts should slide in and out only. Twisting can score the RF contacts. This could damage the chassis connectors, adapters, or the 50 ohm calibration kit. Vector network analyzers must be calibrated each time they are used. Their calibration is only as good as the reference calibration kit. Damage the connectors and test results will suffer.
Test cables
In order to make trustworthy and repeatable measurements high quality cables are needed. They should be properly labeled and kept in a protective case when not being used. I have seen miscellaneous cables that have center pin protruding too far or not properly centered. Faulty center pins can cause damage by spreading open the center pin socket of the female Type-N connectors. Always check the center pin prior to use. Other cables have had boxes stacked on top or have been walked on. This can deform the shape of the cable resulting in impedance mismatches along its length.
Testing
Manufacturer supplied test data is often incomplete and inconsistent even between modules. There is no defined series of tests or response curves to be taken upon receipt of a new combiner during acceptance testing or following frequency change. Prior to accepting new combiners or following retuning or relocation a full test should be performed to ensure the combiner is properly tuned and suitable for use. A thorough set of response curves could be useful later if addition of another station is being planned. It also provides a good reference point for troubleshooting or if the combiner must be tuned in the future. Along with the tests listed later, replicate the test data supplied by the manufacturer for confirmation.
Signal flow
- RF appears at the tuned input port. It is split by the input 3dB coupler and flows through the pair
of bandpass filters. The output 3dB coupler recombines the signal at the output port.
- Other signals appear at the wideband port. Signals are split by the output 3dB coupler and appear
at the bandpass filters. All signals are rejected and returned to the 3dB coupler where they
recombine at the output port.
- The output port of one module can connect to the wideband port of the next. A long chain can be
created, limited only by floor space and the power rating of antenna system components.
3dB Quadrature hybrid couplers
- When a signal appears at one port it will appear at –3dB at two output ports with one signal
lagging by 90 . The fourth port will experience isolation typically in excess of 30dB from the
input.
- When two signals of equal amplitude and 90 phase shift appear at adjacent ports, they will
combine and appear at the output port. The fourth port should receive very little energy, with
typical isolation in excess of 30dB.
- Crossover vs adjacent / non-crossover couplers. Most 3dB couplers in use are crossover types.
Kathrein combiners often use adjacent couplers. An ohmmeter or study of circuit configuration
can be used to confirm each type.
Tuning
- Prior to tuning take measurements to confirm operation. This may assist with troubleshooting.
- If necessary, remove input and output 3dB couplers to gain access directly to filters.
- Confirm the 3dB coupler power division, insertion loss, bandwidth, 90 phase shift between ports.
- RF tuning is a compromise. Minimum insertion loss requires wide bandwidth and may lead to poor isolation. The goal is to optimize all parameters to create overall acceptable performance.
- Adjust cavity tuning. Typically raising the tuning rods will increase resonant frequency. Monitor S11 (return loss) and S21 (insertion loss) while making adjustments. Slowly move each cavity towards the new frequency. Once all cavities are rough tuned they must be carefully peaked for return loss (26dB minimum), bandwidth, and minimum insertion loss. There will be compromises
between all three.
- If the 3dB couplers can not be removed, tune the assembled combiner by monitoring S11 and S21 simultaneously. Remove the dump load and wideband port terminations to display true response of the cavities. Occasionally put the terminations in place as a quick check.
- Bandwidth becomes narrower as number of poles increases. It is also affected by cavity coupling. There is a tradeoff between bandwidth and insertion loss. Tighter coupling will reduce insertion loss and provide wider bandwidth. Loose coupling will provide better isolation by reducing bandwidth, but will also increase insertion loss. Adjacent channel spacing and number of poles determine requirements. Tuning response too wide for the intended installation will not provide adequate isolation. Too narrow will result in increased insertion loss and heat generation.
Group delay
Group delay will reduce stereo separation. It should be less than 350nS at 200kHz referenced to the carrier frequency to ensure 30dB stereo separation. Performance can be improved by pre correction in the exciter or by a passive group delay correction module inserted between the exciter and IPA. A group delay anti curve is created to cancel the combiner delay curve. External correction will increase the overall delay, but equalize it across the FM channel. Measured group delay of 4-pole combiner with cross coupling. Group delay at 200kHz is 941ns and 671ns. Correction is needed to maintain acceptable stereo separation. Measured group delay of external group delay corrector. The correction module usually installs between the exciter and PA, but higher power models can also be installed at the input to the combiner module. This is the response curve for the blue Shively corrector pictured on the next page. This anti-curve is slightly excessive and overcompensates for the combiner. Overall corrected delay is within 220nS.