I've rather struggled with this one. There were only four of them, and they were all built by one builder. How difficult can it be? Well, one source of confusion was that I had 4 drawings, one Barry weight diagram, two GWR weight diagrams and a distorted photo of a drawing by Trefor L. Jones, whose work is generally excellent, but I think may have been struggling with some of the same issues. They were contradictions all over the place. I also had few photos, and all of those were front 3/4 view, so particularly muddy in the tender region. 
So lets go through some of the issues, and the choices I made. 
These locomotives were built by Sharp Stewarts for the Swedish and Norwegian Railway, and the Barry acquired them. They were from two different batches, and the first ones were acquired unused, but the second two were older and had seen some service in Norway. The first two were Barry 35/6, GWR 1387/8, and the second two Barry 92/3, GWR 1389/90.
The first sketch is intended to represent 35 and 36 from around 1902 when they received the tender weatherboards.  At this stage the locomotives seem to have been mostly used for heavy local coal trains in the Barry area.

In 1909 however the Barry decided that the second two should haul main line coal traffic, and they were modified with new boilers, and new cabs, and the tenders given increased water capacity. All the references state the increased water capacity was from adding a well tank between the frames, but I think in photos I see the tank above the frames as having been extended to the end of the frames, and so I've chosen to draw that. I don't have anything that gives me any clues about the well tank. Another puzzle is the cutout in the tender frames. Both the later GWR weight diagram (B) and Jones draw a cutout coming nearer to the top of the frames, but I don't see that in photographs, so I've chosen to ignore that. Another feature drawn in GWR diag B and Jones is coal fenders on the tender, but there's no photographic evidence for these and RCTS states they weren't fitted so I've chosen to omit them. 

On the locomotives there is variation in sanding arrangements. According to the photos at least 1390 lost the big sandbox alongside the firebox and had it replaced by one in the cab, so I've attempted to reproduce that.
So the second drawing is intended to be representative of the second two in their GWR days, but the only external modifications that are GWR are the safety valve cover and chimney. Other differences from the first drawing were made in Barry days. The first two retained the round side window cab and tender weatherboard into GWR days, although at least one of them acquired a GWR safety valve cover.
Other issues - tender brakes were especially contradictory, and the result is little better than a guess. I'm also getting footplate height variations between drawings, so I'm not as confident as I might be about some of the detail and proportions in that sort of area. 
But for what its worth, this is my first pass at this interesting and unusual class, but modelers especially should note all my caveats. The NRM have an appreciable collection of detail drawings from the D as well as the weight diagrams. I can't possibly justify purchasing copies, but the prospective modeler might want to consider a trip to York to see if they provide more useful information.