Patrick,
Going back to you earlier question about the length of the belt, remember that you will need it to be adjustable to two lengths, for wearing with and without armour.
To determine the length to the 'full width' part of the belt I would recommend the following. Put on your tunic (and here I mean a proper Roman style tunic, which is almost square and hangs unbelted to mid calf) and hitch it up in the normal way over a waist tie so that it hangs above the knees. This will ensure that your upper body will have the right amount of bulk around it. Make sure you wear nothing underneath other than an undertunic. Having done this, pass a measuring tape around your waist, tightening it until it 'hugs' you but not tightly. Take off around half an inch (to allow space to do it up), as well as the length of the buckle and this should give you the minimum length for your belt. To this you would need to add a narrow 'tongue', which should be part of the same piece of leather, rather than a separate piece sewn on.
This 'minimum' length will be the right length of belt for you to wear when unarmoured and in 'tunic order'. It is the leather in this piece which can be decorated with plates. The belt will need to be longer though to fit around your armour. This extra length can be found by making sure the tongue is long enough to allow for the extra length required.
Before determining the length of the tongue, put your armour on over the tunic and measure again, this time bringing the tape quite tightly around your waist, as you will need the belt to hug your armour tightly (especially if you wear segmentata). Don't forget that the tongue will need to pass through the buckle by two or three inches at least to do the belt up so make sure you allow enough length for this. I would suggest punching holes half an inch apart along the tongue to allow for a good fit.
Having thus made your basic leather belt, you could then decide to stitch along its edges with linen, following the example of the apron strap published by Lindenschmidt.
Now to the glamorous part everyone thinks of first: the metal fittings. As you will know, most re-enactors wear belts completely covered with plates. This may not be correct however. Sets of belt mounts which have been found together never number more than a handful, and the largest cohesive set found, that from the Velson burial, which was almost certainly complete, was comprised of only seven plates (four plain and three hinged). Added to that are a pair of funerary stelae (which are pictured in the article on aprons Kevin linked to) showing belts with three to five mounts of different designs. These are supported by the mounts found with a sword from Vindonissa, which comprised five plates of different designs (winged thunderbolt; square wolf and twins; square emperor and cornucopia; circular chasing animals; and acanthus leaves [too damaged to determine the shape of the plate]). The Vindonissa plates were found in a way which suggested the belt had been wound around the sword.
http://i17.photobucket.com/albums/b73/Crispvs/belt.jpg
What are we to make then, of the belts depicted on 'standing soldier' type stelae from the Rhineland, which appear to be completely covered with plates? The answer may be given by a combination of the complete Velson set and a literary reference to trouble makers using sharp knives to cut soldiers' belts. These miscreants would certainly not have been able to approach from the front so they probably approached from behind to cut the belts, suggesting that the portion of the belt which passed around a soldier's back may well have been left free of plates. It is exactly that portion of the belt which is not shown on the Rhineland 'standing soldiers', as they are all shown from the front. Most people assume that what can be seen at the front must continue right around, but as I hope I have demonstrated, this is an assumption too far.
This then, means that you have two plating styles to choose from: either a few differing plates spread out along the belt, or the front of the belt covered with plates of more or less the same type meeting edge to edge.
Because of considerations about which side of the front of your body you want the buckle to be on, for the edge to edge plated type belt, the plates might be on the buckle end of the belt or they could be on the other end, depending on what part of the belt was going to pass across the front of your body.
Next, the hinged mounts. These could be of three types. Firstly, if you are using the narrower type 'A' plates, the hinges should be cast into the ends of normal plates. If you are using type 'B' plates, then the hinge can either be cut into the rolled and of a plate (which is commonly found on surviving plates) or a hinged tab inserted under the end of a plate (also sometimes found with the originals). Buckles can be of either copper-alloy or bone. At this stage you need to consider whether you are going to have a pugio or not. Despite what some people will tell you, these are not compulsory and we have evidence to show that not all soldiers carried them. There are very few decent reproductions on the general market unfortunately, the present Deepeeka 'plain pugio' being about the only one, and even then it only represents a plain sheath, which only comprise a tiny minority of those so far found. If you want anything approaching a realistic reproduction of something more typical you will be looking at commission work from craftsmen such as Hans Binsfield, Erik Koenig or Holgar Ratsdorf, which will set you back a goodly sum of money. Therefore, especially if your budget is limited, you might want to consider not getting a pugio. If you do decide to get one, set your frog plates close enough together on the belt for the frog buttons to touch the upper suspension rings of the pugio sheath. Do not even consider using short straps to suspend your pugio as you will see many re-enactors doing. Those people are simply copying what they see other people doing and the reality is that the short straps belong to the same fantasy world as leather wrist 'bracers' and fixed crests. The evidence we have shows the upper suspension rings of pugio sheaths in line with the belt itself, touching and sometimes even overlapping the frog buttons. The best guess is that they were tightly attached to the frogs with leather thonging. Make sure the frog plates are close enough together to allow this. Do not make use of the lower suspension rings of a pugio sheath. All the evidence tells us that these were simply a memory of earlier times and that they were NOT used.
Lastly, the apron. If you have read Mike Bishop's article on these (linked to above, as stated) you will be aware of the different options for how these can be done. There is also another option though - that of no apron at all. Although this may seem strange to people used to seeing Roman belts with aprons, the Velson belt showed no evidence of having had an apron and depictions of soldiers on column bases from Mainz dating to the Flavian period also show belts without aprons, as do the later Adamklissi metopes.
Finally, to the matter of making plates yourself, as you suggested in your first post. Despite the fact that most 'reproductions' of figural type 'B' plates are cast, the vast majority of the originals which have been found have been pressed out of thin copper-alloy sheet and then often further decorated with repousee and/or punching (it is quite easy to get hold of embossed/pressed type 'B' plates with concentric circles). If you have worked as a pattern maker, I am sure you will have the ability to make formers yourself and will have a better idea of how to make these things than many people. You may find some inspiration in the original belt plate stamp from Colchester Sheepen, which is depicted on the page of illustrations by Mike Bishop which Jurjen linked to.
Regarding tinning and contrasting colours, I believe there is evidence for some plates being only partially tinned, allowing raised details to be highlighted in brass against a tinned 'field'.
I hope this helps a bit with your project.
Crispvs
