A little background, for those readers who haven’t been reading this blog since the inception of written records (somewhere in 2009):
Wadkins Glen, II, III, IV
Just before I head out to Colgate University to install some woodwork, I decided the timing was about right to deal with a niggling issue I have had with the Wadkin table saw since I bought it, namely that the sliding table is not flat: it is bowed upward slightly in the middle. This is unfortunate. For cross-cutting around about the 90˚mark, this is no issue, but for ripping or acutely angled cuts, the bow or bulge upwards at the middle of the sliding table causes stock which I am ripping to be tilted up slightly. It is slight, but it means the rip cut face has the unfortunate characteristic of not being 90˚ to the adjacent face, necessitating additional steps after sawing put things right.
I just want the saw to perform the basics well, and in this case it has been letting me down.
Additionally, the detent positions in the sliding table for the mitre fence are severely worn, and the pivot hole location, which accepts a BSW (Whitworth) 1/2″ x 12TPI pivot bolt, has been heli-coiled -and not very well – so it is not a precise location any longer. As a result, none of the detent positions (there are three: 90˚, 45˚ and 30˚) actually produce accurate mitre cuts. Also, there are too few detent positions in the first place. I would like to have more, if they can be done accurately.
Because of the sliding table bow, the linear bearings for the table cannot be properly set, and thus the table runs with more slop in the middle position, which is exactly where you would want it to have minimal play, since that is where most of the cutting takes place..
Finally, the cast iron support beam for the sliding table is a ‘fugly’ casting which bothers me on an aesthetic level. so I have wanted to have that casting machined clean so it doesn’t look like it was cast by a couple of fellows who had spent a little extra time in the pub at lunch, if you catch my drift.
I’m more concerned generally with how a machine works, than how it looks, but if I have the sliding table off for machining work, then it makes sense to deal with the ugly casting issue at the same time.
In the past I have had machines with non-flat table castings, like my old Oliver 166 BD jointer, and I had a machine shop in E. Hartford CT do the grinding work on the two tables and fence for that machine, however, through that experience, and others, I have found that the shop doesn’t follow the instructions I give them, and haven’t been too friendly when the overcharge for services was their own fault and not mine, so it left me with a slightly bad taste in my mouth.
Also, Blanchard grinding, which is the norm in N. America at least, is not my top choice for correcting a distorted surface: rather, it is single-point planing.
As Wayne Moore put it in his 1970 publication Foundations of Mechanical Accuracy (p. 17) the acknowledged ‘bible’ on the topic,
Single point planing is how many machine surfaces on higher-end european woodworking and metal working machines are created. On the Wadkin, it was used for the table edges at the very least. Typically, on machine tables like those found on jointers and planers, and usually shapers, the planing is done in the direction of stock movement so as to leave the surface with very slight ridges, which allow for easier transport of the wood over the surface.
Shockingly, on some high end machines these days, like a Martin shaper, for example, the table surfaces are just ground whereas they used to be planed. It’s a cost-cutting measure to grind.
Trouble is, finding shops with single point planers for machine table work is tough: these machines are comparatively rare birds. I had, in fact, largely dismissed/abandoned the idea of getting the table planed, and instead had devoted quite a bit of time to the design of a sliding table made from Mic-6 aluminum tooling plate. This idea had its plusses and minuses, the chief-most among the drawbacks was the cost, which would likely be at least half-again what I had paid for the saw.
While humming and hawing over this matter for a good couple of years, I happened to be researching the topic of replacing cast machine ways with linear slide rails, which more than a few people have done with their Bridgeport mills so as to get around the cost and difficulty of machine and scraping everything back to spec. I then found a video of a guy who had his Bridgeport mill’s saddle re-planed, by a guy up in New Hampshire who had a planing mill which he had modernized and reconditioned. Here’s a link to the video that caught my attention. After the intro, which is definitely worth a view, if you jump ahead to the 7:52 mark you can see the planing machine and watch it work. Around the 13:00 minute mark you can see the metal planing most dramatically.
A bit of digging around and I had obtained the contact info for the fellow in NH, named Rees. I spoke with him and his son Fitz, who are both involved in the business. Rees is retired but his favorite machine to run happens to be the planer. I explained my Wadkin situation to him and, a month or two later, he came to my shop with his son to take a look at the Wadkin and inspect.
Rees was initially skeptical that a woodworker would want a woodworking machine set up more precisely, but when he checked over the table on my saw he could confirm that the bow was more than insignificant, and he also said he could deal with repairing the pivot hole and detents for the mitre fence.
So, yesterday, the right moment had arrived, and I pulled the sliding table and its support beam off of the Wadkin:
The saw now without its sliding table looks a bit diminished to be sure:
The angle brackets to the left and right which carry the table were also removed later so they could be inspected at Rees shop. I noticed that the surfaces under those brackets were single-point planed, though the brackets themselves were simply rough-ground.
I was surprised to find that, with a straightedge on the bottom surfaces of the sliding table, where one of the linear rail assemblies mounts, was very flat, while I was expecting it to be concave:
Checking with the straightedge along the (in the above photo) left side mounting position for the rail, however, found a result which was concave, as expected. So the table bow was not uniform, seemingly concentrated towards the sliding table edge which runs along the blade.
Also, with the linear rail assemblies off the table, I found that the assemblies themselves were not straight along their lengths:
Each linear rod carrier was bowed a bit more than 1/16″, the gap totaling more than 1/8″ at the middle:
I’m not sure that bow really matters with those parts, as the mounting bolts will hold the rails in a straight position, but it was curious nonetheless.
With some help from a friend, I got the sliding table and support beam loaded in a rental truck and arrived to Rees’ shop a little after lunch. He lives in a rural area, up a dirt road. The view at the shop door revealed a clean space, a good sign:
He had various pieces of equipment, most of which were quite old, like this Brown and Sharpe universal milling machine:
Here’s a look at the planing mill, with the sliding table laid down for some preliminary inspection:
We flipped the table over and Rees checked the top face along each edge only:
The weird thing was that the inspection did not show the table edges to be as bowed as they had appeared to be in the shop when it was examined using a Starrett 48″ straightedge. Rees was also puzzled, since he had also looked at the table when it was on the machine and had seen the bow. We then checked up the support beam, and found it to be decently flat too along the linear rail mounts.
This was one of those unaccountable moments, kind of like where you hear your car making a strange noise so you take it to a mechanic, whereupon it does not make the strange noise. It is quite possible in this case that the weight of the casting itself was causing the sliding table to un-bow slightly when laid, giving the impression that it is flatter than is actually the case.
Anyhow, Rees will do further inspection and will plane the top and underside of the sliding table, along with a re-do of the mitre gauge mounting and positioning locations. The support table looks like it will not need work, so it will just have the casting edges cleaned up.
Possibly, further inspection will reveal more clearly the location of the table bulge – when it was face down on the planer’s table it definitely was not presenting a dead flat surface, as it spun easily on a high spot. But the good news appears to be that the table rework will not be as extensive as initially imagined, which means the work will take less time, will affect the finished height less, and will cost less as well. So, that’s all good. Rees should be able to turn this around within a week – and he is excited to tackle the work, which in itself is becoming a rarity among those who do the mechanical arts.
Here’s a look at the back of the planer, which dates from 1905 or so:
The front carriage, which you can see has been re-scraped:
I’ve been in several shops where you can see an abundance of old ‘arn (“old iron”); the difference with Rees shop equipment is that it has been scraped and aligned and selectively modernized, but he has held off on the usual tarting up you see with a lot of older machinery, where the parts have been painted all shiny, and brass pieces buffed to a sheen, and new bearings put in, but the important stuff, the alignments and flatness of surfaces, has been left out. It’s like cleaning up an old engine and repainting it, buffing the valve cover, etc., but not actually rebuilding the internals properly. There’s a lot of ‘restored’ equipment out there like that – because it is the low hanging fruit, so to speak, anybody can do it, and gives the most ‘wow’ factor to the average person looking at it.
I guess I like Rees approach better – I think he has the priorities right – where the machines he has work really well, and as intended from the factory, but he couldn’t give a damn about the paint or the buffing of hand wheels, etc.. All the same, I will probably attend to the paint on my saw at some point, but it won’t be this round. I look forward to getting it’s geometry back to where it (hopefully) was when the machine left the factory sometime around 1970.
I’ll follow up with anther post in the near future. Thanks for dropping by the Carpentry Way.
Post II in this thread can be found here.
Chris,
I do remember that post! How does a cast surface get that far out of flat? Was it dropped, or just internal stresses working themselves out over the years? I’m glad you’re getting it planed- that looks like a beast of a machine in Ree’s shop! I’d love to see it in action.
Joshua, thanks for the comment.
Ah, indeed, how is it that some machine tables become warped, bowed, twisted, etc.? This is the third used machine i have bought which has issues like these, and funny enough, in all three cases the seller made no mention of the issue. Possibly they didn’t notice mind you.
i’ve done a fair bit of reading on this topic. There is a common story you will come across which goes like this:
“In the old days, companies used to season their castings for long periods of time before they machined them. Companies no longer do this, and as a result, you find more problems with tables that do not stay flat.”
It’s not quite true, at least not in the way things are associated. I’m sure it is true that, in cost cutting efforts, some companies shorten the time in which castings are left to ‘season’ (sitting outside in the yard to rust, basically). However, the length of time a casting sits in a yard rusting has little to do with whether the castings as such remains stable after machining. Companies season some types of iron castings (it depends upon the type of cast iron; for example engine crankshafts, which are cast iron, are machined shortly after they have been cast) not for reasons of long term stability, but because the rusted surfaces prove to improve machinability. in other words, the process of seasoning the casting is largely for the benefit of the producer, not the consumer.
Castings move out of their ideal form for one simple reason, and it is the same reason that some pieces of wood will keep moving every time you cut them: there are residual stresses in the casting. These stresses are unevenly distributed and will resolve themselves over time, and that time frame can take many years.
How those stresses arise however is an area in which one finds outcomes are not purely random. Indeed, the factors which affect how/where stresses accrue in a casting boils down to the casting design (a casting with thick and thin areas, for instance, will tend to develop uneven stresses as the thin areas cool and shrink faster than the thick ones), and how material is apportioned in the casting, and to the process by which the casting is cooled. If a casting cools too quickly -and here’s where cost-cutting measures can again affect the finished product adversely – then it will tend to have higher internal stresses.
If you look at old cast wheels used on line shafts, etc., you will see a design where the spokes are ‘S’-shaped or curvy. this was not done to make them look pretty, but to deal with the fact that the mass or iron around the hub, and around the rim, are going to cool more slowly and shrink to a greater degree due to their size, than the thinner spokes. The spoke shape helps them accommodate the shrinking stresses by elongating as the hub and rim cool.
At the end of the day, many manufacturers do not care if their machine castings eventually warp into non-flatness many years after the product has been sold. In fact, it would be considered an advantage by the philosophy of pre-planned obsolescence.
So far, in my experiences with machinery, the only ones I’ve owned which have not had any issues with their castings have been made in Germany or Japan. They are doing some thing right in that regard for sure.
As for my saw, the main table is quite flat, and by all accounts, Wadkin machines tend to have stable castings. That said, I know of two other Wadkin PP saws with the optional long sliding table which have the same issue as mine does, so I suspect that casting design might be the culprit. It would be hard to prove though, and the original company is long gone, so one is left with suppositions and not much else.
The table looks to have a relatively even thickness. Not the thick/thin differences that are the bane of metal casting.
In engineering school we had to study the the iron/carbon phase diagram exensively. It is unavoidable that significant stresses arise in casting. The cooling speed is also important to control the solidification process and basically which form the carbon takes in the cast iron.
It wasn’t immediately clear from the pictures, but you might be able to find the remains of the sprue and the risers on the underside of the table (at least that’s where I would expect them). If you can find the sprue, that will tell you where the material solidified last.
For the fastest filling, I would expect the sprue in the centre with risers on the four corners or on the ends of the ribs. But that could very well induce some bow into the table if the centre is the last part to cool.
For parts where stress deformation is unwelcome, you’d expect the part to be heat-treated to relieve stress. Maybe that has been skimped on?
Roland,
good to hear from you. There certainly are many unknowns in regards to how the part was cast, and treated afterwards, then machined, etc. I can’t speak to those, as I have no idea what Wadkin’s practices were, and the company is long gone as is likely anyone who worked there at that time on that machine.
It is a curious thing that the longer table seems to have been more prone to bowing, over some period of time, as compared to the standard length sliding table, when one would not expect any significant difference in the arrangement of the material in the castings. At this point, based upon when Rees is telling me as he works his way along, there are no grossly warped or twisted areas, but it is more a matter of a little bow here, a little twist there, which add up in a cumulative manner to produce a table which is more significant;y out of whack.
Awesome post, Chris. Very interesting to see his machine shop, which is very well kept.
I am also of the opinion that the working surfaces are the important part of a piece of machinery or an engine. Not that I mind a nice paint job, or polished brass parts, but they’re no doubt second in line…..unless they’re a sickly yellow, then they should be attended to :).
Yes, Brian, you and I are of the same mind on this. I also don’t mind the blingy surface stuff, but I have looked through dozens of threads where people are refurbishing/rebuilding a woodworking machine, and I have yet to see much in the way of attention paid to dealing with the working surfaces, beyond the odd table regrind. It’s still helpful to strip a machine down and clean it up, put in new bearings, fix oil seals and clogged grease lines, etc., but I don’t see anyone really going beyond that. And it should also be noted that in some cases the machines were not all that accurately made coming out of the factory. Wadkin, for example, did not scrape any surfaces, so they only took stuff so far, and some would argue that it doesn’t need to be much better with woodworking machines anyhow. I asked Rees if he would teach me about scraping cast iron, which he agreed to do, and I plan to practice on some of the Wadkin parts, like the mitre fence, angled support blocks, and, in Phase II, the sliding surfaces of the trunnion.
Certainly, I think my telltale sign recently is the need to build wooden fixturing to overcome shortcomings of the machine. If this is due to a lack of accuracy in the machine itself or a lack of sturdiness then it should be further attended to. We seem forever in pursuit of that ‘machine precision’ that seems exceptionally difficult to acquire.
Well, Brian,
it seems to me we are either chasing precision in the machine/tool we use, or we are chasing it afterward in the wood, due to the sub-par result we obtained with the machine/tool. I’m tending to prefer the former solution, not the latter.
Hi Chris
At first I was thinking that perhaps part of the problem could be the level of the machine relative to the floor, but I discarded that idea since the tablesaw has sort of a squarish footprint. Metal lathes are fairly sensitive when it comes to being levels correctly, but then they are also longer.
If the two angle brackets are a little bit off, then I guess that the table could bend whenever you tighten the bolts. A way to check this might me to mount a dial gauge on another part of the saw, so that it is touching the middle of the “bowed” table. Then when you tighten the bolts you might be able to see if there is any movement.
But I guess that Rees has checked that already.
A properly scraped surface of cast iron looks SO cool in my opinion. It is like a hallmark of precision and good engineering. I once knew and old engineer who really knew how to scrape. He made fabulous looking model steam engines, and those surfaces where just amazing to look at.
Brgds
Jonas
Jonas,
thanks for the comment and your considerations as to my Wadkin saw.
We checked the angle blocks when i was at Rees shop, and they were excellent, thought they are simply rough ground. I plan to scrape the upper surfaces, as well as the lower ones on the beam, as an exercise (not so much out of need). It will make the assembly a lot easier to slide in and out when changing blade,s so it will be a plus for sure.
So far about 0.023″ has been removed from the top, and a bit more on the bottom surfaces of the sliding table. There was about 0.005″ of warp in the last foot of the vertical bearing face of the underside, so that has been corrected. The support beam has a slight amount of warp in the vertical plane, after more careful inspection. Little by little, all the various distortions in the castings which conspire to make the sliding table work less than perfectly are being carefully ironed out.