Dennis J

Hollow grinding is a quick way to establish a low or high angle bevel on a chisel, gouge or plane blade and it makes sharpening easier, especially with very hard carbon steel or powdered metal blades like the PMV11 ones. I have an 8 inch aluminium oxide dry grinder with a tool rest to make initial hollow bevels which runs at 1725 r.p.m. I have to be very careful using it with thin plane blades to avoid overheating. I also have an 8 inch water drip grinder which I use freehand, no rest, to establish a primary bevel on chisels and gouges which need a lot of attention. I simply hold them horizontally to the wheel which results in a flat bevel surface. In the case of gouges that hollow grind disappears after a few sharpenings. When it becomes harder to resharpen without increasing the honing angle I go to a very coarse water stone to grind back the whole bevel and a little hollow grinding using the wet grinder can speed up that operation. I only use a honing guide to sharpen plane blades on steel plates charged with diamond paste after a very careful hollow grind using a tool rest on the dry grinder. After repeated sharpening of course the hollow grind will disappear. And using steel plates with diamond paste works well on narrow bevels but not wide ones so it becomes necessary to go back to the grinder. With a good set of water stones and a little help from a low speed grinder sharpening chisels and gouges is not hard if you use a routine that works. There are plenty of videos showing the Japanese using these without powered grinders. However, I think that they don't approach sharpening plane and chisel blades in terms of primary and sharpening bevels. The reason for this is that they use thick laminated blades with soft iron backing. So they usually regrind the whole bevel starting with coarse water stones. Once some of the soft iron backing is ground back the thin high carbon steel at the edge can be sharpened with finer stones.

I watched a video recently featuring a successful older maker having trouble gouging a maple back. It appeared to me that the main problem was a blunt gouge. Perhaps even experienced makers don't realise how important it is. A slightly domed, low angle single bevel of about 20 degrees is needed and it surprises me how many makers have to use brute force to do the job. If the bevel is domed it is possible to use a light mallet smoothly without the danger of it digging in too far.

It has only 5 inches of sole forward of the blade opening. So it is fairly awkward trying to plane a 15 inch back or top length of wood with the plane locked in the bench vise. And using such a long plane is also awkward with the wood clamped in the vise. The sort of plane pictured in Woodland's post would be the best option.

I've tried a 22 inch bevel-up jointer held in a bench vise and it is not long enough to comfortably use. I can see how a longer wooden plane might work, but finding a good one would be a problem. Making one would be the best option. But using one would not necessarily produce a flat surface. I use a No. 4 Veritas bevel-up along with a smaller block plane which I use to level any high spots if necessary. I don't take for granted that the No. 4 will automatically produce a perfectly flat surface either. I carefully check for flatness with a straight edge after each pass with the No. 4 until it is perfect. If it is not I go back to the smaller block before a finishing stroke with the No. 4.

I know that is the case but it can be a bit of a guessing game. Establishing exact profiles at the usual cross arch locations any wood removal between them can be completed quite smoothly using that principle. It is very easy to take too much away around the edges without templates, especially at the centre bout. And going in too far, especially at the lower bout, can result in very thinly graduated plates there. There are plenty of examples of that which I think is a carving problem, not deliberate.

What looks good is right and that looks about right. Trying to categorise the curvature is pointless. I make templates having fixed heights and inflection points using French curves to draw a compound curve on to anodised aluminium and then cut and file them to a smooth compound curve. But those inflection points have be located pretty accurately on the horizontal and vertical axis at each cross arch location. The upper and lower bout arches are fairly straightforward because the inflection points have to be at edge height level and about the right distance from the plate edge. The middle bout arch at the narrowest cross arch position is critical. It needs to be just about as full as practical. So a template there should be pretty important. A good arch profile there should be close to circular.

Andreas, by absolutely correct I mean that arching plans I have posted previously comply with known violin geometry such as the proportionality of the four circles method. I make arcs from the upper bout to the lower bout using double-string-length generated lines. This fixes inflection points, at edge level, for the upper and lower bout. The arcs drawn between those two points locate the inflection point positions at each of the usual cross-arch positions, specifically distance from the centreline and proximity to the plate edge. From there inflection point heights at each corner position and centre bout one can be calculated. That information can be used to make accurate arching templates. I don't know what the sonic implications are for all possible, practical arching variations. My guess would be that arching height would be more significant than most other aspects of arching. All I know is that using templates makes arching very straightforward and predictable. One aspect of making templates like this is that the inflection points of the top plate can be the same as the back plate, on the horizontal axis that is. A different back long arch with a lower height for the back will result in a different arching but the inflection points for both front and back will be the same. It would be interesting to investigate if that is the case for early makers' instruments. Probably not easily assessed.

The OP's original question was about the difference between a full and scooped arching. I've pointed out that the difference is not much when you look at arching profiles. The simple fact is that there is only a small amount of variation possible although it can be visually significant. I'm not suggesting that an aspiring maker try to construct a full scale geometric plan like I have to design arching templates although it would give them the ability to design any sort of arching. It involves drafting tools like a large compass, long ruler and large sheet of paper. It's not something that can be done on a PC. Things commonly used by early makers to design violins by the way. Templates have to be based on inflection points consistent with a particular outline shape and specified inflection point positions. And a long arch pattern is the starting point which is needed to fix heights at cross-arch template positions. The long arch is not tied to any violin geometry that I can see but the cross-arch ones are. The position of the inflection points (distance from the plate edge) at the upper and lower bout is delineated by the same geometry that defines the general convergence of the strings, fingerboard and neck of the violin. That results in the distance of those points from the plate edge being greater at the lower bout than the upper bout. No point in making templates which don't fit that sort of criteria.

I'm surprised that the critics of what I post can't point out where my approach to designing arching layouts which can be used to make templates is wrong. If they understood the geometry behind what I do they would be able to put on paper what I have done. I'm sure early makers would have known about what I have laid out.

What I see in those drawings is a fairly high arch for the top. The convex part of the centre bout arch is fairly narrow with a wide recurve as are the upper and lower bout arches. So the corner bout arches subsequently have large concave elements. But overall it looks to me to be geometrically what I would expect. Large radiuses like this have to become almost flat to merge into the perimeter scoop. The central convex element is even narrower for the lower back height and seems to match the top geometry.

Having spent a huge amount of time over the past few years I know the underling geometry I use to design arching profiles is absolutely correct. It is quite flexible as far as recurve is concerned but there is a limit as to how far in from the edge the inflection points can be at the upper and lower bouts. If they are too far in from the edge there it is not possible to create a smoothly integrated transition from the upper convex curve into the edge scoop. Because the inflection points at the upper and lower bouts have to be at or near edge height (to blend into the scoop at the upper and lower block areas) there is a limit as to how far in from the edge an inflection point can be. If anyone doubts that try drawing a smooth, convex arch profile with a very wide recurve, especially the lower bout one.

It does, at least with smoothly contoured arching profiles, particularly the upper and lower bout ones. And higher arching there can accentuate that. The centre bout arch needs to be about as full as practical. In the photo I have drawn in pencil what is a possible at the upper and lower bouts. It is no more than about 1 mm, and most of that is achieved by making the scoop a little shallower. That involves moving the inflection points closer to the plate edge. And that in itself will have little impact on the corner cross-arch inflection point locations. Inflection point positions on each arch are the blue dots. Those corner locations have by far the largest recurve components.

Diamond plates are useful for some things but are a waste of money for sharpening. I've found finer grade diamond plates always have irregular surfaces which leave deep scratches, particularly noticeable when flat lapping. The only decent diamond plate I have is made in Japan, and that is past its best. The pic shows what I use to sharpen gouges and chisels. Two Shaptons, 5000 one broken, and Kings along with slipstones which I use regularly, followed by a loose cotton buffing wheel. I've just reground the bevel and sharpened the number 5 sweep gouge. It is bevelled at 15 deg. and can be used at a low angle. The reason it stands up by itself is because it has one bevel only, no rounding at the cutting edge. I've added a pic showing how I have skimmed the bevel on my wet aluminium oxide wheel. It is almost flat, stopping short of the edge in red. I'll progress to my coarse stone getting as close to the edge as I can without breaking through. I'll then use a 1000 Shapton to refine the whole bevel to a slightly domed shape and finish with a 4000 King to raise a bur on the incannel surface following with the cotton buff.

The pics might show what I'm talking about more clearly. The third shows the gouge standing at about 15 deg. with the bevel close to vertical, tilting a few degrees. All of the flat edge rests on the board. If I tilted it back further as LCF says the sides of the gouge would be ground back further. The first shows it from the in cannel side, the fourth from the bevel side. And the second shows a slightly radiused edge profile. I'm regrinding the bevel and have used the grinding wheel to speed things up. I also use a felt marker pen to check on where I'm going with it. My priority is to maintain the bevel angle being careful to not break through the cutting edge until final sharpening with fine stones. At present the edge is no more than about .1 mm thick. But it will take quite a bit more work to raise a burr on the incannel surface.

I thought this post was about sharpening. The fact is sharpening gouges can be tricky. My method of getting the edge straight before anything else especially where a gouge might be have a wavy edge, rounded over corners or uneven bevel is the best way to start. Holding the gouge bevel vertical to the honing plate is a perfectly normal way to set the edge and tilting it over will result in a radiused edge. I don't think there would be any difference performance wise between a moderately radiused or straight edge. Some of mine are slightly radiused. Both cases are a slice through a cylinder type of geometry.