Hi Graham,
In the rexresearch link (Chet kindly gave above) the document lacks two photos at the very bottom and also lacks useful-sounding correspondence between Al Francoeur and Ben (k4zep).
I found the missing part, see the top part of this page:
https://www.novakcorp.com/energy/experiments/wesley2.htm Though you may be already aware of that piece of information.
I would suggest using a flat (and thin) ferrite rod positioned in the neutral line. ( the line like shown in Figure 2 drawing in Gary's Canadian patent 10239)
Such flat ferrite rods has relatively recently come to be used in induction cooker coils, see 1st attached picture taken randomly from ebay.
Here is a seller
https://www.ebay.com/itm/135111122358 and I think this size would be good for experimenting: L*W*H 50*10*2.5mm (sold in a quantity of 10)
The specifications of the ferrite material is given, they seem reasonable. The output coil would be wound onto this flat ferrite rod, of course and the mechanical 'fixing' of the rod
should be such that it could flexibly and easily be moveable in front of and beyond the neutral line. The back and forth movement for the rod could be solved by a cam-like mechanical drive and spring(s), this not so easy.
Because two magnetic poles are needed to create the neutral line, and horseshoe magnets are relatively expensive, a "straight horseshoe" magnet could be built like
picture 2 shows below. Perhaps two Neo 10mm cube magnets stacked at each end of a 10 mm wide soft iron keeper would serve well.
I agree using rectangular pole pieces and cores versus say cylinder ones. In this video a cylinder ferrite rod is used for testing the poles change:
https://www.dailymotion.com/video/xw7f1f I think the small distance experienced by the guy was due to the cylinder shaped ferrite rod (facing surface areas count).
Edit: the back and forth movement can be applied for the magnets + keeper assembly of course and the ferrite pieace with the output coil on it would be stationary.
Gyula