Rotorua Linear Repeater  MK2              March 2017 

 

                                                  

 

The original repeater whose photo appears in the original web write up was built over 20 Yrs ago  but was destroyed in 2015 when a psychiatrically disturbed local individual broke into the Branch 33  Rotorua Club Hut , took out all the equipment and burnt it.  The system was temporarily replaced with a recrystalled  Tait 172 E band VHF receiver donated by a local Ham Charlie Tetenburg  ZL1BQJ.  I built a 10.7 to 28MHz up converter and then hung an old yeasu ten to 2m transverter obtained  from ZL1TBG  on the  end . This  got us up and running quickly but I  always wanted to replace it with a better performing unit so I rebuilt the original units in “Warkworth “ boxes   supplied by Ralph ZL1TBG many years ago. The input to the repeater is 144.950 MHz and the  translated output at 144.350 MHz ,The rebuild enabled me to look closely at the concept and optimise the  new design.

Over the previous months I spent many hours scanning through the web reading other folks web sites, jotting down ideas and findings . so i can say there are many parts of the design i have lifted from others and also there is content in here of my own design and experiences  so if there are ideas and concepts you recognise from your publications  thank you for sharing it with me via the w.w.w..

The front end is a low noise BF981 followed by a bandpass filter with an additional tap off after amplification and filtering .This is to enable sharing of the repeater receive antenna system and not have use a 3 db loss inducing rf splitter in the receive feedline , this output is for the HF control receiver on 145.000 MHz .
The mixer is an high level +17  dBm  passive double balanced one with 50 ohm termination on all ports .the IF output passes trough a 50 Ohm diplexer to maintain termination and then into a pair of grounded gate Jfet amp to isolate the mixer from reactive reflections off the 30 Khz crystal filter .

 

The I.F Unit ( the first iteration!! )

The intermediate frequency is at 10.7 Mhz mainly because I have the quartz crystal filters to enable a 10.7 Mhz I.F and I have two that are 30 Khz wide .. I settled on 30 Khz hopefully to enable fidelity for digital modes. Locals are getting into Dstar etc ...
Being a Linear Repeater and able to pass any modulation mode means the i.f has to be AGC controlled .So I settled for an old favourite which I have used before and it is based on the now defunct Plessey range of Integrated circuits made specially for two way radio communications .For SSB/AM they produced a couple of work horses that used a minimum parts count and functioned extremely well . Most work with a 6-9 Volt rail and tens of mA consumption , I used the SL1612 i.f amplifier and its RF controlled AGC chip the SL1623. They do have a better audio derived AGC device for SSB, the SL1621 but since this repeater will be handling FM as well, a carrier derived AGC was more suited compromise .

                 

The input to the i.f amp board is 50 ohms and feeds directly into a mult-pole 30 Khz wide crystal filter correctly matched via a simple L network to achieve an almost flat ripple free band pass ,There is a relatively simple way of ascertaining the unknown  impedence of a crystal filter using a sweep system , two 4k7 variable resistors  and 50 ohm terminations  , the rf output of the sweep analyser  system is terminated at the and of the coax that feeds the filter . the braid connectes to the earth side and the 50 ohm terminated centre conductor connects to the  centre wiper of the 4k7 variable resistor either of the end of the track connects to the crystal filter input  ( do this at the output as well} play around with the variable sestors until you obtain a flat passband response  ,  dont worry about the losses present , measure the value of resistance that gives the flat passband.. that is an approximation of the  filter impedance  with a few random pf across the filter   . Its enough to get you into the balpark  and design an L matching network ! ,thanks to WD4HXG  chuck  for writing this up and putting it on the web!!
 

The output of the filter is matched via a simple Pi network into the first AGC controlled SL1612 amplifier ( with a gain of about 34 dB) , the output of the 1612 feeds straight into a MPSH-10 transistor stage in order to drive the first 10.7 Mhz interstage transformer . Plessey advocate simple capacitive interstage  coupling between I.C’s but I feel this generated  too much broadband noise which would translate to the 2m RF output stage so I limit the rf band noise  via a 10K dampened  tuned 10.7 MHz  transformer .

 The Transformer link couples to the following SL1612 stage and also has a resistive tap off so RF can be removed from the IF amp unit( to a FM demodulator  section )   or if required introduced  from another  front end source  eg 10 m  Front end  ,into the second  SL1612 and processed up to 2m ( the AGC control range is much less via this connection) . The second SL1612 is identical to the first with a similar transformer coupling at the output . Link coupling off the second transformer carries amplified i.f signal to the ’ L’ matching circuit that connects to the second 30 Khz  wide 10.7 MHz crystal filter ..The second crystal filter limits the bandwidth of any noise generated in the 65 db gain i.f  strip to 30 Khz as this .i.f noise will be translated up to the  144.350 MHz repeater output frequency and we do not want to waste RF spectrum on just broadband noise, but also keep potential upmixed broadband noise  away from the  receive frequency of 144.950 MHz

                     

The I.F board was built on double sided fibreglass PC board for maximum screening and to minimize  potential instability problems , when first tested, the board would break into oscillation at 10.7 MHz when ever I removed my hand  from the unit , The offending cause was shown as  a very sharp tuning first IF transformer ,this instability was eliminated by dampening the tuned circuit with a 10K resistor. Me thinks I should have placed ferrite beads on the emitter leads of the MPSH-10 driver transistors  perhaps they are an over kill for use at 10.7 Mhz .  Note that the driver transistors have un bypassed emitter resistors , .and bypassing will provide more gain, (I measured 11 dB ).. if required but you may run into instability problems.  I must experiment to see what this addition will do to overall circuit performance gain and AGC control.   The AGC is carrier derived and a tap from the collector of the final transformer drive transistor capacitively couples 10.7MHz signal into the SL623 . The SL623 also contains a product detector and an AM demodulator but neither are used in this application.    Agc voltage is fed to the appropriate  pin of each of the SL1612’s When adjusted for 2v on the AGC rail with 125mV input the SL623 will hold the  signal to within 4 dB .
The two SL612 stages have a combined potential to provide 140 db of range but this is only in theory…  in practice this will never be achieved due to limiting performance of other components in the receiver

                         

This is the swept response of the I.F at 10.7 . I was a little disconcerted to see the double hump in the pass -band but I suspect its due to the agc  working,  the  gain changes, due to attack and release time constants ?  I will examine this more closely later.     At the moment I am very happy with the performance of the IF , the SL612 supposedly has a N.F of 4 db at 10.7  so it should be relatively quiet in operation.

This is minus 120 dBm into the IF unit,  as you can see this level has a very good s/n !

 

                 

 

                                                                       Plessey If  output

 

 But minus 120 dBm also produces what I will say is a useable s/n  so Im not going to need bags of preamplification in the RF  front end ahead of the I.F !!

 

                        

                                                                     matched crystal filter 

The second 10.7 MHz crystal filter  is “ L” matched to a common gate j fet stage that provides a small amount of gain but whose very good return loss provides an easy 50 ohm impedance with which  to match to,  thus maintaining the flat passband characteristics of the    “ bandwidth  tail  noise limiting” crystal  filter .

 

I should have used 3 of  SL612 !!!

And Now the faux Pas .. It was not until I was designing the 10.7 to 144.350 up mixer did it dawn on me I had made a blunder  with the IF amp . The original 10.7 IF amp in the repeater that was burned by the psychiatrically disturbed one  had 3 SL612 IF amp chips , giving a total stable  I.F  gain of over 100 dB , on double side pc board without extensive shielding (good ol plessey they knew how to make great RF stuff)  .

Thus previously  I had a designed and well thought out system with  about 15.db of headroom, from broadband IF noise output and the start of full on agc ,This gave me a 15 db range in the transmitted RF output  from an input sigal of less than –127dbm to  -112 dBm where the agc now is fully on and controlling the receiver  gain . I tried to close the agc range into the upmixer with another outboard I.F amp with its own extra AGC but ran into a mass of trouble with both IF interacting  ,in the end I gave up and redesigned the receiver I.F board ..I needed over 100 db so I had a small bucket of Siemens BF960 UHF dual gate mosfets and pressed them into action .Three stages will give me a total of  104 db  I look up the net and found a lot of IF amps use almost direct broadband coupling from a RF choke in the drain how ever I decided to  use bandwidth limiting 10.7 MHz interstage IF transformers  on the input ,to couple each stage and the output .and design my own RF derived AGC to control them .

I am trying to minimise the amount of broadband IF noise generated in the  Amp so as the Upmixed RF wont be drowned in the signal generated by the IF contribution and also cut down on the amount of unnecessary broadband  rf output.. After much reading on the net and QST magazine articles , it was decided to use red LEDs as reverse bias generators in the source of each BF960 stage ,This trick enables you to apply more AGC control voltage than simple source  resistor bias and achieve  even greater AGC range  ..something I lacked in the previous amp .

                       

                                                                    Second iteration of the 10.7 IF  amp

 

 So I built up a standard three stage IF amp along the usual mosfet Amp designs on double sided PC board  and used manual control of the IF to see how tame the Amp was  ..well it was not !  It was a reliable oscillator at around 10.7 MHz at high gain setting.  Some of the IF tuning was quite peaky and sharp indicating some sort of regeneration was attempting to take place . Stagger tuning helped  stability  but I did not want to use that as a mechanism to exert control .  I then set about biasing all the RF points as much as practical ..RF chokes in the B+ leads to each stage the odd 10k resistor across tuned circuits  t o dampen things down and it still looked like it was ready to explode on the spec an screen, tuning was still very peaky , The RF was till getting from the outpuit to the input !100 db is  a lot of gain in a small area !! it wasn’t until in an act of final desperation did  try close shielding of the printed circuit side of the board with some thin shaped brass shim sheet soldered to the ground plane topside that everything settled down and the amp became useful . I then made up a simple RF generated AGC control circuit  with a good load resistor on the AGC line and we were in business .The  Agc could be made to kick in very early in the piece when the RF input was still quite low level and thus I had obtained my small dynamic range of potential RF output from 100 mW up  to almost  10w at 144.350 MHz.   The next thing was to redesign the L match to connect the two  30 Khz BW I.F filters to the input and output of the  10.7 MHz I.F

                        

                                                            10.7 MHz   if filter response 

My only concern with these old wideband crystal filters is the overall stop band is not that good compared to  narrower, more modern multipole  filters, These  have a better stop band  and shape factor, Using these wideband filters    can potentially make the system more susceptible to out of band nearby interference , but we are lucky here in Rotorua we are a long way from high power signals .

                 

                                                               Linear Repeater  Front End

 

The line up is as follows.      RF    input matching configured for minimum noise ( under 1 dB, hopefully 0.7! )  into the BF981 dual gate Mosfet ,  this is a design published on the net by VK2TAB ,  I have changed the output to a simple link coupled tuned circuit.   The BF981 provides a measured  23 dB gain , more than enough …The output of the BF981 feeds a three section series  filter with a loss of 3 dB and  good stop band performance to eliminate the image frequency .

( The image frequency must be suppressed by at least 16 db or the presence of the image frequency will increase the noise figure of the rf amp by 3 dB .)

                   

 I decided to use a high level +17 dBm  passive double balanced mixer for good intermod performance. ( and because I had some  too !)The mixer was terminated as is recommended in 50 ohm matching,  to guarantee this , 6 db attenuators where placed  on the rf and local oscillator ports  . The I.F port was terminated with a 50 ohm grounded gate j310 in push pull amp,  this feeds the 30Khz wide crystal filters via a “L” matching circuit. The jfet amp provides a 50 ohm buffer between the High level mixer and reactive 30 KHz crystal filter via its matching network . 

 

                                            

                                                                   50 Ohm Jfet amps

 

                                  

                                      Three section post Rf amp  bandpass filter

 

The rf output of the three section bandpass filter also  has a tap off , which will be brought to the the outside of  the receiver as a  female BNC,  for connection of the 145 Mhz command   receiver ( icom 2a)   for remote controlled HF  receiver ICR75. That way we do not have to use an external antenna signal splitter to share the repeater receive antenna  for the 145 Mhz  control ic2A receiver  and suffer the additional 3.6 dB loss and thus an increase in noise figure of the repeater receiver. And also the 145 MHz control signal will just sneak through our 144.950 receive filters !

                                    

                                                                   145 Mhz  control frequency response of  repeater filters

                                   

                                                                          The RF Mixer

The mixer is a high level +17 dBm minicircuits unit with low noise and high third order intercept, ( provided the mixer is terminated in 50 ohms at all ports)  To achieve this, the mixer is terminated with 6db pads made with resistors ( although you can purchase smd pads !) in the RF input port and L.O. port ,the I.F output is terminated in 50 ohms via the frequency selective  diplexer circuit that passes the 10.7 MHz to the  IF amplifier and diverts other mixing  products to the 50 Ohm termination thus keeping the mixer  happy!

                         

                                                                                           simple 50  diplexer

                                 

.                                                                       10.7 Mhz diplexer  response 

 The +23 dBm L.O signal is passed through a simple 3 pole low pass filter to clean the 134 Mhz injection frequency then through a 6dB pad before it enters the   E300 Double Balanced  Mixer .

                                  
 The output of the mixer passes through the diplexer to a dual J310 Jfet grounded gate broad band amplifier with a gain of a few dB, but  its main use is to isolate the  mixer from the  reactive components generated by the out of band frequencies reflected  from the crystal filter ..The 50 Ohm output of the jfet mixer connects to  the “L” matching network that connects to  the first 30 KHz  wide crystal filter and gives it a flat pass band .

                                           

                                                                      Rx Local Osc unit       + 23 dBm output

 

The whole front end has a total gain  @ 144.950 MHz input of ……   –70 dBm INPUT converted    down  to 10.7 MHz gives  –61.2 dBm OUTPUT  that    equals 8.8 db gain   ..  at the testing stage the only signal generator I had to drive the local osc port and thus  test the front end ,  that is capable of +18 dBm is the HP8920 test set , but in actual fact I need more than +18 because of the loss in the 6db termination pad and the low pass filter ,Think I’ll need at least +23 if the hot carrier dbm is to perform as designed .
The passive hot carrier dbm mixer used in this receiver has  a +17 dbm osc  requirement if it is going to perform as per the manufacturers  design , One has to be a little careful to make sure the rf output of the oscillator multiplier chain is clean and free of spurii and noise as this will reflect on the rf performance of the 2m  front end , the 44.75 MHz  oscillator is a standard collpits based one, common to a lot of crystal based RF  oscillator designs . The third harmonic at 134.25 MHz is selected by a lightly biased High level Local osc injection multipler stage with dual tuned circuit output , producing approx +2dbm at the link coupling that feeds a J310 in grounded gate configuration .The drain rf  output is passed through two lightly coupled tuned circuits to achieve +8.5 dBm as a drive level to a single 2n4427 class C stage which produces   +23 dBm into 50 ohms .       

                

 

                                                                             Circuit of receiver  L.O rf chain ………………

 


 This RF feeds the L.O port of the mixer via a pi configured caur  low pass filter to attenuate higher  harmonics . 

  I  decided to place voltage control on the osc to maintain stability and used a three terminal 10v low drop out regulator to achieve this.. Be wary of the effects of voltage regulators especially LDO types , they have high feedback gain . and  they can be prone to instability despite using the recommended capacitive bypass on the in and the out  .This was a problem I had with my Rx osc chain causing this spectral output on the transmitted signal

                           

                                   YIKES   !!    This was a spectrum plot of the 144.350 MHz output .

 

At first I thought the instability was generated in the 2m Linear amplifier  output stages and I couldn’t understand why the spurii were  regularly spaced and remarkably frequency stable and  no amount of tuning would change or tame the beast. So I connected a pick up loop to the spec an  and started waving it overtop of the upmixer  module and local osc circuitry , I turned off the 10.7 MHz drive so there was no influence from that source  and low and behold the same interference was in the upmixer  local Oscillator output. , so I retuned it and there was no change.         

On the Spec An it looked like an AM signal not FM!.. then I put a scope probe on the 10v rail and low and behold this is what I saw !! a 4 volt sawtooth on the 10v rail ..no wonder !!amplitude modulating the local oscillator chain

                

 !  Fixed by soldering a 10uF tantalum directly between the output pin and ground of the low drop out three pin regulator.( I had already connected the manufacturer suggested decoupling capacitors)

                            

                                              Finished Receiver for 144.950 MHz  with 30 KHz if  filters 

 

The RIGHT hand section is the high level mixer local oscillator with +23 dBm Rf output at 134.250 MHz

. The middle section is the BF981 RF amplifer with three section bandpass filter , high level mixer, post mixer diplexer and dual J310 jfets  50 Ohm post mixer amp in grounded gate configuration..

The LEFT  hand section is the 3 BF960 mosfet based IF amp using  leds as constant voltage  in the source and required carrier derived AGC

 

             

 

                                                                       The RF  Up Mixer

After much internet scrutiny it was decided to use an SBL-1 DBM as the mixing element in this section . With the sbl-1, rf losses are low and it  has  good strong signal handling ability, good balance to minimise L.O feed through and  an easy obtainable  level of RF requirement for the local oscillator at +7 dBm with few active  components  . I chose the port that gave the greatest isolation for the 134.850 Mhz local oscillator as  this enables me to relax the amount of 144.350 selectivity . I did try the parallel tuned circuit  in the grounded gate  JFet amp at the wanted rejection frequency  133.650  but couldnt get to perform  ( works well at 10's of MHz ) . so I went back to the simple amp configuration.
  The up mixer  linear amplifier chain consists of three stages of amplification .
The first a common gate Jfet  amplifier (J310) with a 50 ohm input impedance to match the output of the sbl-1 DBM , through 2 tuned circuits to enhance 144.350 Mhz and reject the L.O up mixer frequency leaked from the SBL-1.

  This feeds a BF960  dual gate mosfet through more lightly coupled tuned circuits, for good selectivity, gain and linearity.  The output of this stage  feeds  a 2N4427 Bipolar linear stage   for an output of up to  +18 dBm at 144.350Mhz . The RF output hopefully will drive a Toshiba SAV-8 Linear 2m power brick to achieve a nominal 17 W peak RF output .

 

     Modification to upmixer RF driver                               

                                                                                                        

The Linear repeater was bought down for servicing as I detected via received signal strength .It had dropped its RF output by some 6 db's or so when stimulated by a signal sent from my signal generator into to my 2m collinear antenna , the change in received signal Level was seen on the spectrum analyser .
I went up to the site and swaped it out leaving the old lashed up repeater as a replacement . When I put the repeater o the bench and fired it up I measured just over 2 watts RF out subsequently this was traced to a burned out capacitor in the RF lowpass filter at the output . But while testing and looking on the analyser via a 30 dB power attenuator I noticed instabilities on the rf upmixer driver circuitry , after much fiddling and bypassing and component substitution I decided it would be easier to just redesign the upmixer driver ..perhaps the instability had been there all along ?? I also noticed the high level of 133.650 MHz mixer L.O rf output in the output , a little higher than I would have liked due to it being in the top end of the aviation band .
The next thing to do was work out the amount of gain I will need between the “no signal” 10.7 I.F level -53 dBm and the amount of signal power (+13 dBm ) I will need to drive the 10 W rf output brick (RA08H1317M)to full power . This was approx 66 dB ! This time I wanted to use minicircuits brand MMIC in the chain instead of discret devices ( only because I had some) I did play around with a BFR96 as a pre driver for the power brick but had nothing but trouble with instability in the GHz range so avoid this device like the plague .
This time I took a different approach ,I would split up the total required gain into two chunks, place some at 10.7 MHz and less at 2m, because the 2 m will be the harder gain block to stabilize.My final line up consisted 10.7 into a Mar-1 (17 dB) into SBL-1 mixer then into a 2m bandpass filter Mar-8 (30dB) 2m bandpass filter , Era-5 (18dB) then into a Mav-11 (12dB) this gives a total of 76 dB ,In theory this will take the no signal I.F output to +22 dBm ,9db more than enough(+13) to drive the power brick to 10W , however remember there will be 6 db loss in the SBL-1 and semi large losses in the bandpass filters , and the rf level of the 10.7 I.F will rise to -27 dbm at full AGC . The 10.7 IF level from the output of the receiver can be set with a pot so the Max 2m RF (+40dBm) out can be adjusted the max I.F output of the 144.950 MHz receiver.    

                                

                                                                  new up mixer driver board  installed .

The measured gain of the final upmixer RF driver board was 68dB ie -46dBm in at 10.7 MHz gives +22.8 dBm at 144.350 MHz ( mute just opens... as mute circuit is in the transmitter unit ) -26dBm in gives 39.2 dBm out at 2m ( slight loss of output in the low pass filter) the apparent drop in Rf output is also due the top of the response curve of the power brick .Note to minimize this “power drop” in the total amplifier chain, is to pick mimics whose rf output compression point is well above the maximum signal expected to pass through the devices mimmics with low compression points at the beginning and highest at the end .The pc board was made with double sided copper , with the earth ground plane on top and used to earth components .

                                   

                       

                           Upmixer  Board  linearity test....... 1 Khz  @ 80%  modulation at +13 dBm ( im happy with this)

      Mute    circuit                                              

All Repeaters require a mute circuit to gate the transmitter on after receipt of a signal, The last repeater used a CA3075FM  FM detector with a noise mute circuit,  this switched a relay which powered up the linear amplifier stages as they are not required until a signal appears ,thus saving power in our solar powered site.

 The rf oscillators are left running to minimise off-on  frequency drift. I tried all types of mute circuits but none were stable tame and consistent so I looked at alternatives. Some I will investigate later  when time permits . ( dual pll correlation mute ??)

With previous experience many years ago with Plessey RF stuff I decided to try a plessey SL6600 dual conversion  PLL based fm detector IC of which I had several .
 I have always had good results with their active components, such a pity they have gone the way of many good RF  semiconductor companies. I etched up a small board , populated it and tested the performance of the chip at 10.7 MHz and I was impressed . ..I somehow expected it would perform well  and the mute action was very decisive even under the  high level of broadband noise that came in along with the wanted signal from the 30Khz wide 10.7 MHz signal into pin18 ..the crystal is 10.6 Mhz  as the pll runs from 50 Khz up to 455 Khz .

(  The plessy notes give a simple capacitor value  formular for PLL VCO frequency  but I found I could "select on test"  with a variable  and get better weak signal performance ) 

                    

 

I can have reliable muting at –127 dBm !!if I want   . This I.C rocks !  I also took the audio out of the fm demodulation port in case dtmf control function is required at the repeater site.
. The Squelch output was taken through two transistor stages so the mute polarity shift would drive a small relay. I  finally settled on a –124 dBm mute open signal as this appears on the border of useful readability of a nbfm signal picked up on the HP8920A test set .

  To generate a small “tail” on the repeater I placed a 1000uF electrolytic across the transmit  relay coil . This serves two functions, it reduces the indecision  “ mute chatter” on very weak signals  and places a small tail on the signals such that it may hold the repeater open in SSB contacts between voice peaks.

The relay contacts were paralleled up to increase current capability  and they were used to switch +12v to the transmitter linear amplifier stages . Im sure a switching power mosfet would have been a good substitute ,but I didn’t have any .. In the interest of dc power saving in our solar powered site I decided to use instead of the 17W Toshiba power brick , a  more modern 135-175 MHz linear  Power amp module ,The $21  two stage mosfet power module RA08H1317M. 20 mW in gives over 8 W out , typically 10 w into 50 ohms 

                                                     

                   . The unit is biased from class C to class AB1 by increasing the gate bias on the active devices  . A simple resistive divider  from a stable 5 volt supply to give 3.4 V  at 1mA is all that is required . The small RF power amp is attached to a very large heatsink to keep the electronics as heat stress free as possible , thus  cope with long overs and also the local club net on Wed evenings at 8 pm.. The RF output of the power brick is filtered by a five pole filter  to suppress  the 2nd harmonic by 30    dBm   and the third to greater than  46    dBm . .to keep the radio inspectors from the door!!.
 In the end I removed this from the tail of the power amp and placed It on its own pc board very near to the RF output socket .

                         

                                                           RF output  filter swept response  


                                                     

                                                        The Finished transmit  module  before replacing upmixer  driver board

The coax from the transmitter and receiver to the relevant coaxial  bandpass band reject  filters  is 6mm  corrugated 50 hardline which has absolutely no leakage as the shield is a solid but somewhat flexible copper tube.

 The 10 W transmitter at 144.350 MHz  is going to generate “noise “ on the receive frequency of 144.950 MHz at a level of 100’s of uV at  its RF spectrum  output and you are relying on the High q  filters  to strip this 144.950 MHz noise from the receive frequency so it cant mask any weak signals on the receive input .

So it is very important that no unwanted “144.950 MHz noise” leaks out of the transmit feeder connecting to the filters  and  leaks into the receive coax or receive antenna, thus  appearing on top of the weak uV  wanted signal .
 You must use good quality dual screened coax from at least the transmitter and receiver to their respective filters to minimize this potential problem and  it is suggested you use dual screened coax also  for your coaxial filter quarter wave filter  interconnections as well ( thanks to john  zl4rf  for the coax)  . Read up all about this screening  aspect on Repeater builder .com on the ‘web.

 In our Branch33 repeater, we use separate vertically stacked  2* 5/8 co linear antennas about 3m apart and which  have a measured isolation of about 40 dB . These Hustler copies  give a low angle of radiation and a gain of approx 6 dBd each ,easy to match ,tune , are robust and low maintenance. 

A test was conducted where I fed 0.0 dB at 144.1 MHz from the N9310 Agilent signal generator  into the antenna at my workshop in town, the antenna is a hustler copy of a  2* 5/8 in phase vertical collinear ( identical to the ones used at the linear site,)  and,  is  about 6m above ground .  I took my Agilent E4007B spectrum analyser to the linear site and plugged into the receive and transmit antennas . . I measured the received signal of –100.8 dBm on the lower antenna and  -98.7 dBm on the upper antenna …this I guess this less the =12 db gain of the antennas and minus a few db for coax feed line loss , ??90 db??  represents path loss from my workshop antenna  ??  I  did bring up a handheld telescopic dipole  too and  when held vertical read –114 dBm on the analyser  though it was at head height varied between –110 and  –114 …an interesting exercise .. reflections  perhaps ??

 

When the repeater was installed I measured the minimum amount of signal to open the mute over the air  using a signal from the signal generator in the workshop.   The old repeater  based on the tait 172 E band AM radio opened at –18 dBm into the antenna at the workshop.. The new repeater mute opened at  a setting of  -25 dBm    approx 6 - 7 db improvement !!

The actual signal level to open the repeater  mute on the bench was –124 dBm  so given the path loss of approx 90.0 to  100 dB   and the signal level into the workshop antenna of –25 dBm   gives 95 ( midway ) plus 25  equals approx  –120 dBm at the  receiver,?? Im happy with  that !!

When on the bench  I set the receive osc and the transmit osc as close as I could to exactly 600 Khz split to minimise R.I.T use   on SSB    contacts .  The HP8920A test set has a useful  receive  offset frequency readout  function, on received carriers ,so I calibrated the HP8920A and the N9310  by using a GPS disciplined 10 MHz external  reference  oscillator to feed both instruments . The test of the signal generator output frequency locked to the same reference as the HP8920 on receive mode showed 000 kHz when both instruments were tuned to 144.350 MHz . . I connected the HP8920A on receive, to a small 2m antenna outside and the N9130 to the tall collinear  and sent a signal at 144.950 Mhz,  the test set showed an off set of 101 Hz on the translated signal from the repeater…   so far drift  has not been too bad.

 

DC Power supply .

I have used extensive use of three terminal regulators in the various modules in both the receiver and transmitter,   that minimises the changes in voltage of the 12v supply chain affecting biasing of stages, oscillation frequencies  rf level of drivers   etc, in all it makes for a  more stable system. However  being a solar site our battery voltage can potentially be from 12 to almost 15 volts  and I had a small problem with the Mitsubishi RF power module in that is was recommended from maximum longevity  not to exceed a supply  input 12.6 Volts .Standard regulator circuits require at least 2-3 volts drop over the pass transistor be they three terminal or bipolar discrete to perform well  ..This is too much voltage drop and I had to explore alternatives   A lot of manufacturers produce low drop put three terminal regulators , they are much better than the standard integrated regulators but a lot of them have up to 1-2 volt drop at 3 amps or more  some of course are lower than that it depends what you can get hold of ..I even tried a switch mode regulator but they still require a voltage drop !!

So search the net I did and found a very good article by Serge Patrishin, he is an Audiophile and designed a very good low drop out regulator based on a tl431 adjustable Zener  controlling the gate voltage  of a series fed power mosfet  , The voltage drop across the mosfet is tens of millvolts at 2 Amps !! . just what I needed to give the repeater a constant voltage. Look here … http://myelectrons.com/about/

So simple but performs so well  and subsequent looking on the net,  showed lots of others copying and modifying Serges work …including Me !! for the circuit to function you must generate a .  23 v positive  supply to bias the mosfet  gate and open the d-s conduction (I  used a 555 in voltage double configuration)  the tl431 monitors the regulated  output voltage and adjusts  voltage by robing the gate bias to control conduction hence regulate  and it does it very well,  just don’t forget to connect the tl431 like a zener diode ie anode to negative and kathode to positive or it will not work.. believe me I know from bitter experience.

                           

So far the repeater has been working well and I am satisfied with its performance . You may well ask why build a linear repeater ?  Well it is an RF  challenge to start with , something to get the brain going .. The 20 yr old  previous version, which was burned by the psychiatrically disturbed one, was 16 Khz wide,  this version is 30 Khz wide, hopefully for better digital  rf fidelity .
 NBFM contacts through a linear repeater sound so much clearer than your standard FM repeater as there is no Audio processing to distort and modify voice  . I can check deviation levels of remote transmitters over the air . I can send calibrated deviated signal out over the air as well and know they will arrive basicaly unaltered .. .We have been data/digital capable here in Rotorua for over 20  years !!  we can take, p25, dstar, dmr , AM, SSB , . it has multimode capability  . We don’t need a pc at the site generating all sorts of digital noise to interfere with our low background noise HF receiving site !  .That computer processing can all  be done at home at the users residence.

 Hopefully multimode capability will stimulate more activity on the bands to everyone’s benefit

 

                           73,s  Mike ZL1BTB

                                    

 

 

                                                   back to home page