Using notch filters for Rx and TX and extending the range of a notch duplexer (Draft)
I have a six cavity notch duplexer for 70 cm. At the original frequencies, outside the amateur band, the RX and TX notch responses are a mirror of the other, as notch filter responses are asymmetrical, compared to symetrical for band pass..
The mirroring of the responses are achieved by adding a quarter wave line to each cavity in the TX half. Being a quarter wave length, it changes with frequency when the duplexer was tuned about 20 MHz lower from its original frequencies. This was enough to stop the mirroring, making the RX and TX curves much the same; high losses for TX.
The desired TX shape can be restored by adding about 10 mm to the quarter wave lines on the TX side. The length was calculated as the difference in quater wave between the two TX frequencies, in this case about 8 mm. For this duplexer, it was quite easy as the join to the cavity was soldered. The cable was lengthened by adding a 10 mm piece of coax with the cover and shield removed. The join is wrapped with insulation tape. The shield was made from a strip of copper shim wrapped around the cable and soldered to the other cables shield.
The duplexer RX cans on the right, TX cans on the left.
The RX cans showing a simple T join on the coupling loop.
The TX cans showing the extended quarter wave link between the T and the coupling loop.
The normal RX response with the low-loss on the low frequency side. The T is connected directly to the cavity.
Sunday, 16 October 2016
Tuesday, 4 October 2016
DVB-T TX Filter: 23 cm filter using 70 cm duplexer
I came across a commercial UHF bandpass cavity filter that was as long as a VHF cavity. I removed a coupling to see inside. The probe was nearly 500 mm long, rather than the 170 mm in most UHF filters. They were using the probe as a three quarter length, rather than the typical quarter wave length. I checked the cavity at VHF and, sure enough, there was another peak.
The obvious question then was whether the 70 cm duplexer filter would have another bandpass at three times the frequency, about 23 cm? It does!
With careful tuning it should be possible to use the ubiquitous 70 cm duplexer at 23 cm.
Raw responseWith the filter set to 7 MHz at 446.5 MHz, I had scanned three times that frequency and a very wide span. Sure enough, there was a passband, although at a different frequency and three times wider.
The frequency of these cavity tuners are determined by a fixed length probe (tube) and a capacitive screw adjustment. The relationship between the probe and the adjustment screw is probably affected by frequency.
The bandwidth is three times greater as the upper and lower frequencies are multiplied by three.
TuningWith a bit of tuning, an acceptable 7 MHz filter at 1290 MHz can be achieved. The losses across the channel are a little high but there is usually reserve TX drive. The notches are about -50 dB, which should be adequate as a TX filter.
DiscussionThe general idea of using the filter at three times its normal frequency works. However, the duplexer needs adjust to achieve the desired frequency, bandwidth and notch depth.
I am happy to just measure the performance at the moment. I have a Darko 23 cm amplifier but it is not attached to a heatsink.
It may be better to modify a common 900 MHz duplexer by reducing the length of the tubes, but it is a fair amount of work to pull them apart.
A notch duplexer can be used at three times its nominal frequency as the probes act as three quarter rather than one quarter "antennas".
Monday, 3 October 2016
Duplexer DVB-T TX filter: Technical details
Bandpass is addition of two notches
Three cavities RX to ANT
Reverseded notch TX to ANT
Sum of two notches with six cavities RX to TX
Two left hand couplings are conventional, the two right hand couplings, with plastic spacers are unusual, give reverse notch (not sure how).
SWR of Duplexer filter7 MHz on 70 cm, not great but probably affected by tuning, possibly by coupling design.
Coupling for reversed notch needs further investigation