Now to the details I have used.
 The first device was used on my 12 ele home built NBS and a 28 MHz dipole . I shall not go into the details of sizes etc. - there are many software programs, articles etc. and even the original  article to use.
 The radiating device was constructed for the 2 metre antenna as follows.
 Radiator was 3/8 inch diameter aluminium tube. The Clemens arm was  made from 3/16 inch diameter brass tube. The outer of the feeder (URM 76) was attached to the radiator by means of 2 solder tags to the outer sheath. The inner was looped round to the brass arm and soldered. The end of the coax and joint to the arm was sealed with a silicone sealer.
 The capacitor was constructed by inserting the inner from a length of URM 76 ( about 6") into the brass tube and fixing it by means of  a solder tag to the radiator. Again the ends sealed by silicone sealer. The stand-offs were made from bits from the spares box - they were made so that the distance could be adjustable from about 1" to 3" spacing.
 Next came the tuning . I made the tail of URM 76 about 6 feet long - this is were I put the SWR meter (Bird) and tuned up. I used a low power source (TS700S  - 10 watts) .The aerial was mounted on a pair of wooden steps pointing skyward with the reflector just off ground. The frequency of lowest SWR was found fairly  quickly and adjusted by altering the radiator length (the length of the radiator isn't critical) Then the spacing / length of capacitor was varied and the effects noted. Gradually  was able to get the SWR down to 1:1 from 144.1 to 144.6 - only slightly rising out side of this. As a separate point of interest as I had used 1/10 diameter stainless steel elements when it rained the resonant point moved down by about ½ a MHz and the SWR rose to about 1.3:1 - not much but noticeable.
The ends were then sealed and the aerial erected - just a very slight drift down in tuning was observed - I decided to give it a go. This aerial performed excellently - I worked some 50 Countries on 144 in all sorts of openings over the next 8 years and no problem at all. The aerial was taken down  and put up at my present QTH were it did a further 4 years service. It was replaced by a 17 ele NBS built to the same specification and that lasted for a further 5 years. It only failed when the Crows that my wife's chickens attracted damaged the coax capacitor by cleaning their beaks on the arm!!! - this also occurred on the 5 ele NBS for 50 MHz a bit later - the lesson I learnt was to mount the Clemens arm UNDER the radiator !!!
 I also built a 10 metre dipole at this new QTH. The radiator was 1/2" aluminium with a 3/8" diameter arm - this time I used URM67 as the feeder and the capacitor (the inner fitted nicely into the 3/8 tube). This aerial is still in use today - tattered but working still after 10 years.
 My first real 6 metre aerial was the 5 ele NBS . I put all I had learnt from building the 2 meter aerial  into this aerial . The radiator was ½ " aluminium tube and the arm again was 3/8" aluminium tube. The feeder was a 12' length of URM 67 and the ends were sealed with silicone sealer. Again this aerial performed excellently - my record of achievements with this aerial speaks for itself.
 About 3 years ago the crows got to this one also so I decided  to build another aerial - this time a 6 ele NBS. As is was Roger G4HBA was also looking at aerials and we had many hours of designing and trying to choose which aerial to use - in the end we came to the conclusion that, for 6 metres and given the large size of aerials the best , REPRODUCIBLE aerial was the NBS design. We toyed with 2 times 4 ele, 2 times 5 ele and even 2 times 6 ele yagis - in the end it was decided that a single 6 ele would be firstly an improvement over our previous aerials and also so easy to produce and construct. We also had a good source, down near Wimbourne, of Aluminium at a good price.
 We built the 2 aerials at roughly the same time - however they were different - mine had the elements mounted above the boom on  metal clamps and Rogers were through the boom. There also was a slight difference in the way the Clemens was made - this point took up many happy hours of debate, argument and counter argument and come the end it was felt that either way would (should) produce the same usable results.
 I built my Clemens with the capacitor on the feeder end of the arm with the far end grounded to the radiator - Roger built his with the capacitor  at the far end from the feeder.
 On my first aerials I used solder tags bolted with stainless steel fixings to ground the outer feeder covering to the radiator or arm - on the last one stainless steel circlips were used with the outer of URM 67 as the grounding  link - this does seem to be satisfactory - both aerials have been up a couple of years now and as can be seem by our achievements are performing most satisfactory.
 Lesson's learnt were :-
 Clemens arm to be under the radiator - as above - stops large birds damaging it.
 As the effects of rain can be noticed I suggest that the radiator be made from 3/4" diameter tube with 3/82 diameter tube for the arm.
 Using silicone sealer has only a little effect on the tuning of the arm - self amalgamating tape DOES affect the tuning which has to be readjusted as the tape is installed.
 It may be that a grounding of the feeder at the centre point of the aerial is useful - I did not - this is as is shown in the 2 RSGB articles.
 I was not in any position to have equipment to conduct any scientific tests as to  effectiveness of the aerial or properties as to the degree of  lack of imbalance ( amount of RF on outside of feeder) although I was able to check the polar pattern against that predicted by either the Ham Radio article or software programs which matched very well and was also observed in practise on air.
 My final comments are that I thouroghly recommend the NBS design for all used on 6 meters  for Home builders - it IS easily reproducible and I also recommend the Clemens as the matching arrangement - it is also very easy to build for Home builders and reproducible - I do not understand why it has not come into prominence before for the 28 - 144 MHz bands.
 It is quite conceivable that if you have only to cover a length from the shack to the aerial of around 80' or so that it can be done in one length of URM67 with no joints - solid all the way to the arm!! - Happy building!!

 Dimensions are as follows:-

144 MHz   length of Arm = 8.2"  (.1 wavelength)
   Spacing Arm to Radiator = .875" - 1.5"
    Arm diameter = 3/16" (brass)
   Radiator diameter  = 1/2"
   Length of capacitor - start with 5" and cut to suite - should end up around 4"

  50 MHz length of Arm = 23.5" (2 times .05 wavelength)
  Spacing Arm to Radiator = 2.52 - 4.5"
  Arm diameter = 3/8" tube (copper or Aluminium)
  Radiator diameter =.5" (I now recommend 3/4")
  Length of Capacitor = start with 12" and cut to suite - should be around 8"

Using a copper Arm can be advantageous in that by using the outer of URM67 or URM 76 soldered joints of great strength can be made - stainless steels circlips are still needed to connect to the aluminium. Thanks to Roger the 2 variable standoffs are now made from threaded nylon shafts and nuts and fibreglass tube cut to provide the required spacing.
 It can be a bit time consuming to do this but the rewards are very great - I have had much satisfaction that for less than £30 I have built an aerial that has performed so well for so long.  However one thing should be borne in mind - all aerials should be regularly inspected and I did not expect any of the aerials to last more that 3 to 5 years - I have been very surprised.
Here is Rogers account also.
Addendum 15.8.2000.
Firstly I should pint out that the Clemens has been used by others on 50 Mhz - is was most interesteing to browse through past issues of 6 meter magazines - I found that G3MY had in fact used it too good effect back in 1985!!!

I have built another 5 ele NBS - for 70 Mhz. -  here are the photos of the Clemens construction

Here are the pics of the nearly finished arrangement - all can be clearly seen. All that requires is to put on another piece of cover for the coax (6" pieces of UR67 covering) and to seal the ends with amalgamating tape.
Dimensions are as follows.
Radiator diameter - .5"
Tapping points - at 5.885" (.05 wavelength)
Length of Capacitor tube - 5.5" (end nearest N connector is sealed off.)
Length of matching arm - 5"
length of coax - 32.5"
Spacing of support arm to radiator is 53 mm
Diameter of capacitor and matching arm tube - .5"(12 mm) - I in fact used 2 tubes - a 10 mm inserted inside of a 12 mm - of course a thick wall 12 mm is the best (B&Q have a good selection of this tube) Remember to make sure the Radiator tubing is thick walled - this is the weakest point of the design - the Jubilee clips do tend to crush the thinner walled tubing!!
Diameter of support arm .375" (10 mm) I have used aluminium tube here - I did use plastic tube but it did flex a bit - I observed no difference with either in adjustment.
All tubing is sealed with Silicone sealant.
Increase in capacitance raises the frequency.