Results 1 to 17 of 17

Thread: Tuning A Duplexer

  1. #1
    Join Date
    Feb 18, 2012
    Posts
    69
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Default Tuning A Duplexer

    Dear Experts, I have a IFR 2023A Signal Generator & AdvanceTest R3267 Spectrum Analyzer. I believe that I can use them to tune a China Duplexer. Anyone can guide me step by step to tune the duplexer? I want to tune duplexer RX Frequency to 442.2125 and duplexer TX Frequency to 452.2125.


  2. #2
    Join Date
    Feb 04, 2012
    Posts
    2,092
    Thanks
    288
    Thanked 1,036 Times in 464 Posts

    Default

    basicallly you tune one side to give minimum signal on the Tx freq and tune the other side to give minimal signal on the rx freq. many duplexers are marked tx/rx if its a UHF duplexer the TX is usually the LOWER freq and the RX the HIGHER.

  3. #3
    Join Date
    May 23, 2012
    Posts
    801
    Thanks
    475
    Thanked 932 Times in 385 Posts
    Country: United States

  4. #4
    Alpha's Avatar
    Alpha is offline T S - Moderator
    CS Forums $upporter
    Join Date
    Feb 12, 2012
    Location
    Directly above the center of the earth.
    Posts
    2,780
    Thanks
    1,239
    Thanked 1,533 Times in 739 Posts
    Country: Christmas Island

    Default

    That references pass/reject type duplexers If you are talking about one of these pieces of crap:

    duplexer.jpg

    These are reject-only. The setup is to inject signal to the outer connectors, one at a time, plugging up the unused one with a 50 ohm dummy load. Then hook a receiver to the center terminal.

    To tune you put in the opposite signal, i.e. when tuning the RX side, input he TX frequency and tune the 2 or 3 slugs on that side to minimize the signal. Then you do the reverse, you put in the RX frequency to the TX side and then set those for minimum signal. That's it...

  5. #5
    Join Date
    Jun 28, 2012
    Posts
    20
    Thanks
    0
    Thanked 6 Times in 4 Posts

    Default

    OK Now that we have heard from the peanut crowd, here is how you tune a duplexer.

    First most of the small duplexers are a Pass/ Notch type. They will have a Lo side and a Hi side and Antenna port. The sides are where you will place the signal that is to pass through the duplexer.
    442, on the Lo side and 452 on the Hi side.
    Take your signal generator and insert the frequency on the Antenna port. Say the frequency is 442 MHz. Attach your analyzer on the Lo side. Terminate the Hi side with a 50 ohm load.
    Tune the resonators for lowest loss. They are the screw adjustments at the end of each cavity. After you tune for lowest loss, change your frequency to 452 and move your analyzer to the Hi side. Move your load to the Lo side and tune for lowest loss.

    This will set up the cavities to pass each frequency through and notch the other frequency.
    That is for small duplexers that only have cavity adjustments. The notch is controled by the length of the connecting coax to each cavity. On more expensive duplexers there is another adjustment, a variable capacitor near the coax end of each cavity. After tuning the cavity resonance on each side you would adjust the notch with the capacitors, tuning for as deep a notch as posible. Then you will have to retune each cavity for lowest loss. There is a little going back and forth till all the adjustments are made.
    The one measurment you cannot make with this setup is match to 50 ohms for the Hi and Lo ports. That will require a bridge to make that measurment. I use a HP 8753C network analyzer to tune duplexers for Hams around my area, DFW. That setup will display loss, notch, and return loss.
    On small duplexers a split of 5 MHz is about all you will get with a low loss and a High notch.

    On a six can setup you should get a loss of 1.2-1.4 Dbm and a notch of -80-90 DBm. Match to 50 ohms should be lower than 20 DBm. This is at a 5 MHz split. A wider split will do better and a narrow split will do worse.

    Hope this helps

    mm

  6. #6
    Alpha's Avatar
    Alpha is offline T S - Moderator
    CS Forums $upporter
    Join Date
    Feb 12, 2012
    Location
    Directly above the center of the earth.
    Posts
    2,780
    Thanks
    1,239
    Thanked 1,533 Times in 739 Posts
    Country: Christmas Island

    Default

    Peanut crowd? You do realize you have just insulted half the staff here, right? Most of us are electrical engineers.

    He asked about Chinese duplexers, that usually means the cheap reject-only ones.

    Thanks for the information, next time try not to be such an assh*le about it.

  7. #7
    Join Date
    Jun 28, 2012
    Posts
    20
    Thanks
    0
    Thanked 6 Times in 4 Posts

    Default

    I thought about that. It was more a swipe the comment about the POC. comment on the Chinese duplexer. I have tuned boxes of the low tier duplexers without real problems. I just got tweeked a little when a member ask for help and gets poor response or comments about his gear. You do the best you can with what you have.
    I did not intend to insult anybody, and I am sorry.
    That being said I enjoy this board and it's great members.

    mm

  8. #8
    Alpha's Avatar
    Alpha is offline T S - Moderator
    CS Forums $upporter
    Join Date
    Feb 12, 2012
    Location
    Directly above the center of the earth.
    Posts
    2,780
    Thanks
    1,239
    Thanked 1,533 Times in 739 Posts
    Country: Christmas Island

    Default

    Apology accepted. I answered his question correctly as I saw it, perhaps briefly, but correctly. The Chinese ones do work, but don't have locking hardware on the slugs and as such detune rather easily, hence my statement. I also was referring to the fact that generally, reject-only duplexers are inferior to pass-reject types as well.

    I was hoping he would come back and tell us what type of duplexer he had so that someone could further elaborate. What you said was technically correct and true, and thank you for your contribution here.
    Last edited by Alpha; Sep 24, 2013 at 10:01 PM.

  9. #9
    Join Date
    Jun 28, 2012
    Posts
    20
    Thanks
    0
    Thanked 6 Times in 4 Posts

    Default

    Your welcome

    I hope drahtlos04 got his duplexer tuned.

    Tuning stuff into 50 ohms from 100KW systems to a Ham antenna on the back of a car is what I do.

    mm

  10. #10
    Join Date
    Feb 18, 2013
    Location
    CLASSIFIED
    Posts
    39
    Thanks
    20
    Thanked 2 Times in 2 Posts
    Country: United States

    Default

    Reiterating a bit of the above data....

    The duplexer shown in Alpha's picture is a typical mobile duplexer. Low power handling capabilities (<50W), no true PASS function, and simple REJECT function. However, due to the design of the overall RF network of cavities, determined by the RG-402 or equivalent semirigid interconnects, there is a high or low favoring characteristic on respective sides. Methods may vary, but my experience is injecting into the Antenna port & analyzing the appropriate high or low ports. Definitely terminate the opposite port with a known GOOD 50 ohm load.

    The ideal tools are a Network Analyzer or a Spectrum Analyzer with Tracking Generator. Failing those, there are ways. First, determine what its present tuning is if possible. Failing that, test for least loss at various frequencies moving slowly up or down. Note the "direction of loss" to plot a curve, i.e. more loss with increasing frequency or less, to determine HIGH or LOW if not clearly marked. Some say "TRANS" and "REC", less meaningful since frequency polarities (i.e. RX HI/TX LO or RXLO/TXHI) can change. A "HIGH" side will plot out (descending in frequency from high to low) a straight line, trailing quickly down to a deep trough (the null), then curl back up to its left & come up say 30dB from the null for a while before eventually leveling out some 10, 20 or more MC downward to match the high side. The "LOW" side will mirror this, i.e. ascending in frequency it will be flat, then dump sharply, have about a 250KC to 600KC null, then climb up slowly. This is NOT indicative of a true pass or even a pseudo-pass/notch or "BPBR" response. A pseudo-pass/notch duplexer will exhibit decent in-band (within 5-8MC) PASS response, shaped like a skirt, with the notch at one side or the other. However, out-of-band (>10MC) will be poor. Additional pure PASS cavities (one per side between TX/RX respectively and the duplexer) are suggested. Anyway, separate discussion/soapbox..

    The idea in tuning is to jockey the minimum possible "Insertion Loss" (how much loss at your desired passing frequency) with the maximum possible "Notch" or "Rejection" null of the undesired frequency. One thing to keep in mind, as you tune the duplexer, a great divergence between the respective sides will cause artifacts in the response of the side you're tuning. Excessive loss, funky results trying to center the notch, etc.; until you "bring it in" on both sides, this will be quite noticeable. Don't panic, it's normal. Most duplexers are designed for between 3MC and 8MC of spacing between TX & RX frequencies. Some do 10MC OK. Too close together or too far apart, they won't tune right without changing their interconnect cabling.

    Putting your sig gen into the analyzer's Antenna or input port, get a rough idea where your "zero" (base line) is. Start by generating 0dBm to make the math easy, or -20dBm, something high. Next, having recorded this, hook up the Sig Gen to the duplexer ANT port. Spectrum Analyzer on one device port, 50ohm termination on the other. Walk the tuning up or down to achieve a null at the desired reject frequency. Keep the screws pretty close to the same number of turns in or out for the respective change in tuning on each side, i.e. turn each 3 or 4 turns in sequence from inner to outer. Test above and below approx 5MC, note the loss, one should be significantly different than the other. Get the idea which side is HIGH or LOW if not so-marked. Swap sides, repeat the process, nulling the frequency you need to reject - NOT to PASS. Then check your pass path. You may need to touch up the other side, so swap frequencies and repeat. Do this at least 3 times. When you have an acceptable insertion loss on both sides for the PASSing frequency, check your NOTCH again on both sides. Likely will need touching up. Get the best notch you can, and strive for least PASS loss but favor more notch. Without a tracking generator or network analyzer to visually display this, it won't be easy, but it's not impossible either. When you're feeling the duplexer out, wide spans are fine. As you walk it in, zoom in the span & change reference level to see the finer details & go down lower. Honestly though, it might be hard to really set the nulls "dead nuts" since most Spectrum Analyzers aren't terribly sensitive.

    Using your channels:

    TX: 452.2125
    RX: 442.2125

    Tune the "HIGH" side to reject 442.2125. Tune the "LOW" side to reject 452.2125. The duplexer mirrors itself, one side rejecting the energy from your transmitter in its attempt to enter your receiver while sharing the antenna. The other rejects (attenuates) your receive frequency within the transmitter's path to the antenna and your receiver. Thus any noise at or near your receiver's frequency as generated by the transmitter is further attenuated before it can affect your receiver's performance.

    NEVER tune a duplexer under substantial transmitter power. Even I fudge this rule a little by making the occasional "trim" adjustment, but DRASTIC changes in tuning WILL cause arcing inside and damage your cavities. Minute changes, while I discourage the practice overall, can be made at lower power levels (<5W). But it's just Not a Good Idea (tm). If you want to really tweak on it and feel brave, get out your power meter and use at MAXIMUM 500mW (1/2W). Tune your notches for the least amount of power and check your insertion losses at the pass frequencies as before. This is only AFTER you've brought the duplexer into close tuning with the Sig Gen and Spectrum Analyzer. Hook the transmitter to the duplexer ANT port, wattmeter with a 50ohm load on the other side to the respective HIGH or LOW port, terminate the other with another 50 ohm load, and give it a go.

    Some pictures or screenshots would be great, but I really am up too late..perhaps a revision to the post later. Good luck! The Commieplexers do work, although I prefer not to buy them myself. I support Made in USA. (or up in Canada der eh!) A true station duplexer is always preferable, but for certain applications or in a pinch, mobileplexers/flatpacks can do a job in low RF environments. Anyway, time to crash...so I won't on the road tomorrow..


    *** no warranties, insurances, blessings of the gods or hell & damnation guaranteed, provided or expressed ***

  11. #11
    PRC148 No Longer Registered

    Default

    Quote Originally Posted by Microwave Mike View Post
    Your welcome

    I hope drahtlos04 got his duplexer tuned.

    Tuning stuff into 50 ohms from 100KW systems to a Ham antenna on the back of a car is what I do.

    mm

    WUT .... ?

  12. #12
    Join Date
    Jan 19, 2013
    Location
    In Your Network
    Posts
    2,755
    Thanks
    3,172
    Thanked 2,011 Times in 1,014 Posts
    Country: Holy See

    Default

    Why would you run a 100kw system at 50 ohms? -just sayin'.
    "God as my witness" - Jeremy Dewitte - Felon

  13. #13
    PRC148 No Longer Registered

    Default

    Quote Originally Posted by com501 View Post
    Why would you run a 100kw system at 50 ohms? -just sayin'.
    Actually, I have no idea what you were talking about. 100kw ham antenna what??? I'm confused. What's going on in here?

  14. #14
    Join Date
    May 23, 2012
    Posts
    801
    Thanks
    475
    Thanked 932 Times in 385 Posts
    Country: United States

    Default

    I think he was trying to state that he works on a broad range of equipment and tuning it properly. From 100kW systems to simple ham antennas on the backs of cars.

  15. #15
    Join Date
    Feb 18, 2012
    Posts
    69
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Default

    Quote Originally Posted by Alpha View Post
    That references pass/reject type duplexers If you are talking about one of these pieces of crap:

    duplexer.jpg

    These are reject-only. The setup is to inject signal to the outer connectors, one at a time, plugging up the unused one with a 50 ohm dummy load. Then hook a receiver to the center terminal.

    To tune you put in the opposite signal, i.e. when tuning the RX side, input he TX frequency and tune the 2 or 3 slugs on that side to minimize the signal. Then you do the reverse, you put in the RX frequency to the TX side and then set those for minimum signal. That's it...
    YUP... This one.
    China made mini duplexer...

  16. #16
    Join Date
    Feb 18, 2012
    Posts
    69
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Default

    Quote Originally Posted by Microwave Mike View Post
    OK Now that we have heard from the peanut crowd, here is how you tune a duplexer.

    First most of the small duplexers are a Pass/ Notch type. They will have a Lo side and a Hi side and Antenna port. The sides are where you will place the signal that is to pass through the duplexer.
    442, on the Lo side and 452 on the Hi side.
    Take your signal generator and insert the frequency on the Antenna port. Say the frequency is 442 MHz. Attach your analyzer on the Lo side. Terminate the Hi side with a 50 ohm load.
    Tune the resonators for lowest loss. They are the screw adjustments at the end of each cavity. After you tune for lowest loss, change your frequency to 452 and move your analyzer to the Hi side. Move your load to the Lo side and tune for lowest loss.

    This will set up the cavities to pass each frequency through and notch the other frequency.
    That is for small duplexers that only have cavity adjustments. The notch is controled by the length of the connecting coax to each cavity. On more expensive duplexers there is another adjustment, a variable capacitor near the coax end of each cavity. After tuning the cavity resonance on each side you would adjust the notch with the capacitors, tuning for as deep a notch as posible. Then you will have to retune each cavity for lowest loss. There is a little going back and forth till all the adjustments are made.
    The one measurment you cannot make with this setup is match to 50 ohms for the Hi and Lo ports. That will require a bridge to make that measurment. I use a HP 8753C network analyzer to tune duplexers for Hams around my area, DFW. That setup will display loss, notch, and return loss.
    On small duplexers a split of 5 MHz is about all you will get with a low loss and a High notch.

    On a six can setup you should get a loss of 1.2-1.4 Dbm and a notch of -80-90 DBm. Match to 50 ohms should be lower than 20 DBm. This is at a 5 MHz split. A wider split will do better and a narrow split will do worse.

    Hope this helps

    mm

    I tried to follow the instructions.
    Successful tuned the duplexer, I think. :-)

    But got problem, when we test it, If nearby the duplexer, it work fine.
    But if miles away, people cannot hear my audio via the repeater.

    Any idea?

    My equipment, IFR 2023A Signal Generator & AdvanceTest R3267 Spectrum Analyzer really can do duplexer tuning?

  17. #17
    Join Date
    Feb 18, 2012
    Posts
    69
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Default

    Quote Originally Posted by grandnational View Post
    Reiterating a bit of the above data....

    The duplexer shown in Alpha's picture is a typical mobile duplexer. Low power handling capabilities (<50W), no true PASS function, and simple REJECT function. However, due to the design of the overall RF network of cavities, determined by the RG-402 or equivalent semirigid interconnects, there is a high or low favoring characteristic on respective sides. Methods may vary, but my experience is injecting into the Antenna port & analyzing the appropriate high or low ports. Definitely terminate the opposite port with a known GOOD 50 ohm load.

    The ideal tools are a Network Analyzer or a Spectrum Analyzer with Tracking Generator. Failing those, there are ways. First, determine what its present tuning is if possible. Failing that, test for least loss at various frequencies moving slowly up or down. Note the "direction of loss" to plot a curve, i.e. more loss with increasing frequency or less, to determine HIGH or LOW if not clearly marked. Some say "TRANS" and "REC", less meaningful since frequency polarities (i.e. RX HI/TX LO or RXLO/TXHI) can change. A "HIGH" side will plot out (descending in frequency from high to low) a straight line, trailing quickly down to a deep trough (the null), then curl back up to its left & come up say 30dB from the null for a while before eventually leveling out some 10, 20 or more MC downward to match the high side. The "LOW" side will mirror this, i.e. ascending in frequency it will be flat, then dump sharply, have about a 250KC to 600KC null, then climb up slowly. This is NOT indicative of a true pass or even a pseudo-pass/notch or "BPBR" response. A pseudo-pass/notch duplexer will exhibit decent in-band (within 5-8MC) PASS response, shaped like a skirt, with the notch at one side or the other. However, out-of-band (>10MC) will be poor. Additional pure PASS cavities (one per side between TX/RX respectively and the duplexer) are suggested. Anyway, separate discussion/soapbox..

    The idea in tuning is to jockey the minimum possible "Insertion Loss" (how much loss at your desired passing frequency) with the maximum possible "Notch" or "Rejection" null of the undesired frequency. One thing to keep in mind, as you tune the duplexer, a great divergence between the respective sides will cause artifacts in the response of the side you're tuning. Excessive loss, funky results trying to center the notch, etc.; until you "bring it in" on both sides, this will be quite noticeable. Don't panic, it's normal. Most duplexers are designed for between 3MC and 8MC of spacing between TX & RX frequencies. Some do 10MC OK. Too close together or too far apart, they won't tune right without changing their interconnect cabling.

    Putting your sig gen into the analyzer's Antenna or input port, get a rough idea where your "zero" (base line) is. Start by generating 0dBm to make the math easy, or -20dBm, something high. Next, having recorded this, hook up the Sig Gen to the duplexer ANT port. Spectrum Analyzer on one device port, 50ohm termination on the other. Walk the tuning up or down to achieve a null at the desired reject frequency. Keep the screws pretty close to the same number of turns in or out for the respective change in tuning on each side, i.e. turn each 3 or 4 turns in sequence from inner to outer. Test above and below approx 5MC, note the loss, one should be significantly different than the other. Get the idea which side is HIGH or LOW if not so-marked. Swap sides, repeat the process, nulling the frequency you need to reject - NOT to PASS. Then check your pass path. You may need to touch up the other side, so swap frequencies and repeat. Do this at least 3 times. When you have an acceptable insertion loss on both sides for the PASSing frequency, check your NOTCH again on both sides. Likely will need touching up. Get the best notch you can, and strive for least PASS loss but favor more notch. Without a tracking generator or network analyzer to visually display this, it won't be easy, but it's not impossible either. When you're feeling the duplexer out, wide spans are fine. As you walk it in, zoom in the span & change reference level to see the finer details & go down lower. Honestly though, it might be hard to really set the nulls "dead nuts" since most Spectrum Analyzers aren't terribly sensitive.

    Using your channels:

    TX: 452.2125
    RX: 442.2125

    Tune the "HIGH" side to reject 442.2125. Tune the "LOW" side to reject 452.2125. The duplexer mirrors itself, one side rejecting the energy from your transmitter in its attempt to enter your receiver while sharing the antenna. The other rejects (attenuates) your receive frequency within the transmitter's path to the antenna and your receiver. Thus any noise at or near your receiver's frequency as generated by the transmitter is further attenuated before it can affect your receiver's performance.

    NEVER tune a duplexer under substantial transmitter power. Even I fudge this rule a little by making the occasional "trim" adjustment, but DRASTIC changes in tuning WILL cause arcing inside and damage your cavities. Minute changes, while I discourage the practice overall, can be made at lower power levels (<5W). But it's just Not a Good Idea (tm). If you want to really tweak on it and feel brave, get out your power meter and use at MAXIMUM 500mW (1/2W). Tune your notches for the least amount of power and check your insertion losses at the pass frequencies as before. This is only AFTER you've brought the duplexer into close tuning with the Sig Gen and Spectrum Analyzer. Hook the transmitter to the duplexer ANT port, wattmeter with a 50ohm load on the other side to the respective HIGH or LOW port, terminate the other with another 50 ohm load, and give it a go.

    Some pictures or screenshots would be great, but I really am up too late..perhaps a revision to the post later. Good luck! The Commieplexers do work, although I prefer not to buy them myself. I support Made in USA. (or up in Canada der eh!) A true station duplexer is always preferable, but for certain applications or in a pinch, mobileplexers/flatpacks can do a job in low RF environments. Anyway, time to crash...so I won't on the road tomorrow..


    *** no warranties, insurances, blessings of the gods or hell & damnation guaranteed, provided or expressed ***

    Thank you for the detailed explaination..
    I get clearer information now.

    will give a try.