Last Wednesday, I cut a perfect box joint. I did it using a well-known method on my table saw. Unfortunately, I only got to cut this one joint, which was the test joint to make sure the jig was set up correctly before cutting the actual joints of the box I was building. I stopped after that one, because as I finished the last cut, something hit me in the face, and the table saw's throat plate hit me in the pelvis. I shut everything down and walked upstairs, calling out to my wife to let her know that I was okay but bleeding.
I take power tool safety, especially with my table saw, very seriously. That's part of why I called out to my wife - I don't use those machines if she's not around to call out to. That's also why, while I was bleeding from my face, I was, actually, fine. My safety glasses deflected all debris away from my eyes. Several layers of clothing (shop apron, hoodie, heavy-duty cotton pants) took a lot of the force out of the impact of the throat plate.
Once I was patched up (seriously, it was the tiniest of bandages - you all know how easily face wounds bleed, right?), I returned to the shop to investigate. I found a few interesting things:
- The throat plate had a good long gouge in the bottom.
- One of the throat plate leveling screws was way higher than I had adjusted it.
- Half of one dado blade tooth was missing.
- The outer half of the dado blade had rotated until it was touching the teeth of the inner half.
- Both the jig and the work piece clamped to it were completely unharmed.
The last point shouldn't have been a surprise. A common source of table saw accidents is jigs or workpieces binding on the blade. But if this had happened, it would have been the jig or the workpiece that got thrown at me, not the throat plate under them. And in that case, they also would have had the teeth marks - the joint pictured earlier wouldn't have looked so good.
I spent a bit of time skeptical of the blade. Could that tooth have broken spontaneously? It's a Freud blade, which are generally considered good quality. But it's also one of the cheaper Freud blades, and I've certainly given it a workout over the four years I've owned it. If the piece of tooth did fly off suddenly, would that have generated enough force to dislodge the throat plate? This seems far-fetched, honestly.
With fresh eyes, after a few days of cake-baking/hiking/mailbox-post-repairing head clearing, I think I've found the culprit. The summary: I pulled the throat plate into the blade when pulling my jig back to the front of the saw. I hope this makes you ask how in the world that's possible. I also hope that when I tell you it could be operator error that you ask what awful design could allow such an error. The rest of this post is the answers to those questions.
The saw I use is a Harvey Ambassador C200-30. It has been a great saw - very accurate, and generally a pleasure to use. But there is one part about it I have never liked: the throat plates. It came with two: one for rips and crosscuts using normal full-kerf blades, and one for dados. They're a very basic steel plate, a bit under 1/8 inch thick. The table top has a shallow recess for a plate to sit in, with four wide-top screws along the edge to allow you to adjust the height of the plate to match the height of the table.
I've never liked these plates becuase they're not actually flat. They're very close, but they're not completely true. This makes adjusting the leveling bolts complicated. It ends up being a bit of a tradeoff, trying to get as many edges near the height of the table as possible, so that the work piece isn't pushed up, allowed to droop, or able to snag on either the plate or the recess edges. I get close, but never perfect.
What's worse in this setup is that the last couple inches of the rear end of the plate end up unsupported. The adjustment screws aren't at the back edge, they're a good deal forward of it. Since the plate is thinner than the depth of the recess, it's possible to push on the trailing edge of the plate, and lever up the leading edge. If that trailing edge is above the table surface at all, merely pushing a workpiece across it will cause the throat plate to wobble.
The pictures above are a power-off, hand demonstration. With leveling screws adjusted, I can push down on the back of the dado throat plate, then push forward and the plate will slide right toward me.
This tilt and slide shouldn't have happened. I was using a jig that kept everything at table level, so none of it should have touched the plate, let alone put a levering force on it. Unless the rear edge of the plate was higher than the table. I adjusted the plate right before using it - did I leave the trailing edge high? I've adjusted these plates hundreds of times, and sometimes just barely high in the rear is what it takes to keep the front from being high enough to snag. It's possible that's how I set it. It's either that or the screw that was way high after the incident didn't get there from the friction of the ejecting plate alone. Did the back-and-forth of the jig work it that way?
Because the back-and-forth is the last piece of the puzzle. I've made hundreds of cuts with the dado blade, and the dado plate, on this saw. But for almost all of them, I pushed the piece through, then lifted it up to move it around the table and back to the front for the next cut. With this jig, I was pushing it through, then pulling it back across the table. Pulling a loose workpiece backward across a table saw is a huge no-no, because if it touches just about any part of the blade, it will be flung directly at the operator. But a jig, held in place by the miter slots - especially a jig that has been through the saw a number of times to loosen the kerf a bit - is generally safe (see example uses of a cross-cut sled). But that safety is about the jig staying away from the blade - it's assumed that the throat plate never moves.
This is not the first box joint I've done on this table saw. Was I just unlucky this one time? No, I think there's one final bit of safety equipment missing here: the riving knife. The job everyone thinks about for a riving knife is to hold the saw kerf open behind the blade, so that the work piece doesn't pinch the blade and get flung back at the operator. But mine, it turns out, serves an additional purpose: it prevents the throat plate from moving forward.
Or at least, it prevents the normal-kerf blade's plate from moving forward. The dado plate has no opening for the riving knife, so it can't be used. From the usual safety standpoint, that's not a problem. A dado isn't going to pinch a blade, and a 1/8" knife wouldn't stop a 1/4" dado from doing so anyway. In fact, the standard knife would actually prevent a dado from being cut, because dado blades have a smaller diameter than ripping and crosscut blades on this saw, so the knife sticks up too high.
The other box joints I've cut have used my rip blade instead of my dado blade. This has allowed me to use the regular throat plate and the riving knife. So even if the plate had wobbled during that procedure, the knife would have kept me from pulling it forward.
So here's the replay of what I think happened. On my last cut, I pushed the workpiece and jig onto the trailing edge of the throat plate. This levered the front edge of the plate up and out of the recess. When I pulled the jig back, it pulled the throat plate with it. I didn't notice any of the plate's movement, and so it eventually contacted the spinning saw blade. Things happened too fast to properly register the order in which they happened after that. But more or less, the blade started shaving steel off the plate, throwing those shavings my way. Eventually, the blade gained a good enough purchase to accelerate the plate. Between this point and the plate getting past the blade, one of the blade's teeth made violent enough contact to break. The plate hit me low, and either plate shavings or the bit of tooth hit me high.
I think this story has a high chance of being the correct one. And that makes me happy, because I know how to prevent the same issue next time.
The truth is, I don't even use the normal-kerf throat plate unless I'm making an odd angle cut. Most of the cuts I make are with the blade at 90°. For those cuts, I've prepared a custom-made "zero-clearance" plate. This plate is made out of two layers of 1/8 inch hardboard. This gives me two advantages over the steel plate when it comes to mitigating the risk I've described in this post.1,2 The first is that it's flatter than the steel plate. This means that after installation adjustment, there is not as much play on the back end, so it can't raise the front end as high. The second advantage is that the bottom layer is shaped to fit around the lobes into which the leveling screws thread. So, even if the front edge of the throat plate is raised, that lower layer still has purchase on the lobes, preventing the plate from sliding forward.
Preventing this trouble in the future is "just" a matter of making the same sort of throat plate for use with a dado stack.
Do you own a Harvey C200, or does your table saw have a similar design? Do you use the dado plate it came with, or did you make your own? Do you have some other method of keeping it in place, or has this never been a problem for you? Send me a message via email or ActivityPub (links in the menu) to share your thoughts.
Footnotes
1 Okay, a third advantage: a few ounces (about 200g) of hardboard shouldn't hurt nearly as bad as a pound and a half (about 700g) of steel if it should be ejected at me.⤣
2 Fine, a fourth advantage: hardboard shavings won't cut me, and hardboard won't break saw teeth.⤣
Categories: Woodworking
Post Copyright © 2023 Bryan Fink