Levelling rolled toneholes

Here's a shot of what we have to deal with. There are noticeable warps on all the toneholes, but for the purposes of this article I'm going to tackle the low F, E and D.
Why not just show the levelling process on a single tonehole? Well, here's the thing - if you're going to have straight at the tonehole with a file then sure, you can deal with each one in isolation.
However, this means relying on the file to remove metal - which, for the sake of best practice, is something that's best avoided where at all possible. In the case of the Keilwerth it's essential due to the pseudo rolls fitted to the rims - and the same would be true of a traditionally rolled tonehole, such as those you might find on an old Conn.
When you can't or don't want to use a file you have to resort to other methods of levelling the toneholes, and that means manipulating the base of the hole until the rim levels out. It's an effective if rather time-consuming technique, but it can't really be applied to a tonehole in isolation because any manipulation of the body tube is going to have a knock-on effect to the adjacent tonehole(s).
This means that you have to deal with the toneholes in pairs or, for my preference, trios.

Before we get started it's vitally important to clean the rims of the toneholes. Any debris or gunk on them will affect the reading when the flat standards are placed on them. A light wipe over with some 1200 grit carborundum paper followed by a wipe with a spot of cigarette lighter fluid will do the job nicely.
With the rims nice and clean the first order of business is to examine the batch of toneholes with a set of flat standards and note where the peaks and troughs are. If I'm lucky I'll find that they correspond - so that if I'm having to push the bore down to level off a peak on one tonehole, there'll be a similarly positioned peak on the next tonehole up or down the line...which will benefit from the change in the bore.
Unfortunately it doesn't always work out that way, and as you can see here I have the low D tonehole showing a peak that nearly corresponds to the apex of the body tube. To make things clearer I'm going to use clock positions to describe the various peaks and troughs, with the rear of the tonehole (nearest the rod or hinge screw) being the 12 o'clock spot - and I'd describe this peak as being at the 4 o'clock position.
It's usually the case that the peaks and troughs come in pairs on roughly opposite sides, and in this instance there's a similar peak at the 8 o'clock position.

Now look at the 3 o'clock position on the E tonehole. I'm out of luck because there's a trough in it. I'm going to have to lower the body tube at the low D's 8 o'clock position and raise it at the E's 3 o'clock - which means that each operation will have an adverse effect on the adjacent tonehole.
This is precisely why you can't deal with each tonehole in isolation - it's going to be an exercise in give and take as I creep up to a point where the trough on the E and the peak on the D find their equilibrium.
There will be much swearing.

The trick is not to try to aim for precision too early in the game. No doubt other repairers have their favourite way of doing this, but I like to work on reducing the overall size of the warps so that the required adjustments become ever more finer. The smaller the adjustment needed, the less likely it is to have an effect on the adjacent hole. Once the pair (or trio) of holes are at this point, they can effectively be tackled in isolation. However, it's a bit of a constantly moving target because until you get to either end of the body tube there's always going to be an adjacent tonehole - so there's a lot of back and forth going on.

I've decided to start with the low D, and here you can see me tapping down the 4 'o clock peak. I'm using plastic (Delrin) rod and hitting it with a mallet rather than using a mallet directly onto the tonehole rim. The rod provides a far greater degree of control, which is important in terms of regulating the 'spread' of the impact and the alignment with the tonehole wall. Get either of these wrong and you could end up creating more warps than you remove, and you might bend the tonehole wall (which would ruin your day). It's a fairly large rod - as the adjustments become finer I'll use smaller and smaller ones to tap down the peaks.
And when I say tapping I mean just that. It's not too much of a concern on this horn because the tonehole rims are solid, but on a properly rolled tonehole the roll is likely to be hollow...and thus far more susceptible to being damaged by a careless impact. It's all about gentle persuasion, firm but considered coaxing and constant checking.

Here's the result of the first tap, and you can see that the warp has shifted slightly and changed the tonehole's 'table' (its state of evenness). As it changes it also changes the decisions you make as to where next to manipulate the tonehole, and if you're not careful it's easy to find yourself undoing the adjustment you just made because it now looks like the right thing to do.
This is largely where the 'feel' aspect of the job comes into play; there's a certain amount of making it worse before it gets better, so to speak - and part of the skill is in having a plan of action, and knowing when to ditch it when a potentially more promising plan presents itself.

I'm happy with how the first tap turned out, so the next step is to repeat the process at the 8 o'clock spot. The top shot is what it looks like now - and the lower is what it looks like after a tap. There's a huge difference in the table - so much so that the front of the tonehole is now almost level, but note that very slight increase in the length of the trough at the 9 o'clock position. This is pretty much par for the course and I'm none too worried about it because it'll almost certainly change again when I start working on the E tonehole.
But even at this stage we've made a significant improvement to the tonehole. The gaps have gone from being a tad over a thou deep to around about just under a half a thou. If I were to pop the key back on now, I could easily tweak the pad to accommodate a gap that small...and it'd be a fairly reliable seal.

Or it at least it would be if it wasn't for a dirty great trough at the rear of the tonehole.
It's a biggie, and because it's at the rear it's sitting on top of rather a lot of metal. This is going to make it a bit of a challenge, but on the plus side I can be pretty sure that while any changes I make here are bound to have an impact on the table, they're rather less likely to have a knock-on effect to the E tonehole.
I'm going to leave it for now though, because the work needed on the E is bound to affect the D - and I may well come back to this trough later to find that it's shrunk a little (well, I can hope).

On to the E then - and the big question is, is it a trough at 3 o'clock or a peak at 5 (and 1) o'clock? I have to make a judgement call on it, and by examining the table all the way round I reckon that I should treat it as a trough.
We need a different tool for this job because we want to raise the body tube rather than knock it down. You could use a dent bar down the bore of the horn, but it's a bit crude and it doesn't really lend itself to quick adjustments and inspections - so I'm going to use a burnisher and a spreader. The burnisher is used as a lever while the spreader provides a fulcrum point for it and spreads the load out over the opposite side of the tonehole. This helps to minimise any unwanted deformation of the tonehole - though it's impossible to eliminate it altogether (unless you reach for a file).
I'm not sure whether you can buy such tools, but they're quite easy to make. At its simplest the burnisher/lever is just a bit of round bar that's been slightly flattened at the ends, bent over at one end and then polished to a smooth finish (this is very important). Any old steel will do - you could even use brass at a pinch - and it's useful to have a range of sizes.
The spreaders are also easy to make...if you have a lathe. Delrin's the best choice of material, but any stiff plastic will do. You can even use wood...but you might find you need to make them rather thicker to prevent them splitting. There's an inner diameter that sits inside the tonehole, which prevents the spreader from being pushed off the rim, and an outer diameter that sits over the rim...and then the whole thing is sawn off at the 3/4 point and a rounded notch is ground dead centre (the hole in the centre of the large spreader is just what was already in the bit of plastic offcut I made the thing from). Again, you'll need a range of sizes but there's no requirement for them to be a snug fit - and if you don't have a lathe you could always try glueing a couple of bits of shaped plywood together.

Here's the tool in use on the E tonehole. I've positioned the lever dead centre of the trough, and because it's quite a wide one I've pulled the lever out so that the flatter part of the end sits beneath it. The spreader is pushed back against the opposite wall of the tonehole, and you can see that the downward force from the lever is going to be spread over at least half the circumference of the rim.
Just as with the tapping down, the force applied is quite gentle. You can perhaps see from my grip on the lever that I'm looking to tease the trough up rather than yank at it.
Make no mistake, no matter how gentle you are the spreader will still have an effect on the table - but it'll be a great deal less than the effect of the lever on the trough.

Here's the result of my levering, and I reckon it looks like I made the right call this time. The trough has all but gone and the width of the peak has increased so that the table now presents a shallow trough at the 7 o'clock spot...some of which will be down to the action of the spreader at the 9 o'clock position (i.e. directly opposite the portion of the tonehole that was being lifted). I'm pleased with how this turned out, and as the tonehole looks pretty good from 8 o'clock round to 12 I think it'll pay to lift this trough out.

Well, so far so good - the front of the tonehole looks pretty even, but have a close look at the 3 o'clock position. The trough I just pulled out has deepened again, so I'll need to return to it and give it another tweak...which'll probably throw the front of the tonehole out again. But not by as much as before, and that's the point. Little by little I'll even the anomalies out. There's still some business at the rear of the hole to deal with, but I reckon it's good enough to allow me to go back to the low D and see what I can do about bringing it closer to spec. I'm also interested in seeing what effect the work on the E tonehole has had on it.

The sax gods have smiled down upon me, because all I can see is a slight extension of the shallow trough that was there before I started work on the E. I'll deal with that in a moment, after I've sorted out that trough at the rear because it's deep enough to affect the whole table when it's pulled up - but from here on in the adjustments get smaller and smaller.
With the E and D close to level I can think about finishing up on the D and moving along to F to make a start on the coarse adjustments. This is a tricky one because not only does it have toneholes either side of it, it has one below it (the side F#). Nasty.
And I wouldn't mind betting you're thinking "Wow, that wasn't so hard! What's all the fuss about?". Here's the thing - what you see above is the entrée...the starter... the amuse bouche. Getting out the major peaks and troughs is pretty easy - it's dealing with the finer levelling that's really punishing.
Essentially it's the same process as described above, though the adjustments are smaller and follow in rapid succession - and there's a lot of tail chasing as the table gets closer to level. There are also an increasing number of judgement calls to be made - and you can't expect all of them to go as planned.

And here's the result of all that pushing and pulling. I've left the F in its original state for comparison, but the E and the D are now even across the table. For me the result is the warm glow of a job well done, along with a certain amount of relief that it all went reasonably to plan...though once I start working on that F hole, the E is going to be thrown out of whack - and I'm not much looking forward to tackling the G tonehole...because it's always a tricky little bugger.
As you can see, it's a vast improvement over the opening shot of this article - and when I've finished the job the result for the player will be a noticeable boost in the depth of tone across the range, increased response and stability and a great deal less finger pressure required in order to get the low notes to speak. And it'll stay that way for far longer when the pads start to harden up with age.
If you were able to look closely enough at the toneholes you'd see one or two faint spots of light. If you could be bothered to measure the gaps you'd probably find they were just a few tenths of a thou. This is as close as it's realistically possible to get with what is, in reality, quite a crude method of levelling - but it's well within the long term tolerance of a pad seat. If you wanted to go the extra few yards you'd have to 'dress' the tonehole by running a very fine grade slipstone over them (or a flat disc with 1200 grit carborundum glued to it). If I were fitting new and firm pads to this horn I'd probably go for it, but as I'll be using the existing pads (which aren't all that firm) I really don't feel it would be worth the bother.

It all takes a heck of a lot of time and demands, shall we say, a certain degree of experience...and a lot of care. All of which ends up on the invoice as a rather unimpressive "Level off x toneholes".