Monday, July 28, 2008

Getting To Know Roubo-Part 1

It's been a couple weeks since I finished the bench, and I've had a couple opportunities to put the vises and workholding to some basic tests. Here's my first observations about the bench. More are sure to follow.

Tail Vise:
I recently reviewed a thin bladed marking knife from Czeck Edge Hand Tool. I did a quick sample joint using hand cut dovetails to test the knife, and in the process the tail vise saw some action. The tail vise is performing just as I had hoped. It works flawlessly. It clamps boards up to 8" wide in the leather-lined jaws with ease, and holds them there without hesitation while cutting dovetails. I love it. There is one issue with the vise, and I was anticipating it. In fact, Scott Landis mentions it in the Workbench Book as the last comment in his write-up of the Powell vise. The problem is with the operation of the screw itself. It's the reverse of what we're used to. This vise is lefty-tighty. And that's the problem. I've only used the vise for a few hours, but in that short time I opened and closed the vise dozens of times. And every time, without fail, I either had to stop and think about which way to turn it, or I flat-out turned it the wrong way. The question is, will I get used to it? Frankly, I don't care, since I don't want to get used to it. The thing is, if it was my only vise, I probably would. But all of the vises, clamps, and handwheels in a shop are all right hand threads. I don't think I'm going to ever train my mind for that one odd movement. So what's the solution? Well, I've already talked to my machinist friend who's looking into Acme screws and nuts, left hand thread of course. So in the next few weeks I'm hoping to have this little issue licked, and then I will have the ultimate tail vise.

Wednesday, July 16, 2008

Bench #3 - Finis

It's finally done. And I couldn't be more pleased with the outcome. The bench, so far, is everything I was envisioning it to be. There are a couple issues here and there, but nothing of concern. No bench is the "perfect" bench. Although I would have to say that Andres Roubo has come mighty close. As I was watching the movie "Master and Commander" earlier this week, I was admiring the construction of the inside of the captain's cabin and thinking about the tools and benches that were used to create these examples of fine woodwork. The cabin is located directly at the rear of the ship, and the windows are all built around a compound curve. This is no Hollywood set though. This tall ship was built as a replica of an 18th century Royal Navy frigate. And in the movie, it was used as the HMS Surprise. After a little Googling I discovered the original HMS Surprise, was actually a French-made boat (the Unite) captured by the British and renamed "Surprise" for the way in which the ship was taken. Andres Roubo wrote his treatise on woodworking around 1770. The Unite was built in 1794. What does all this have to do with this bench? There's a good chance that the original curved windows and panels I was admiring in the movie were built using a bench very similar, if not outright the same as Andre Roubo's bench. I happen to have built Roubo's German Cabinetmaker's bench from plate 279 (perhaps it's my German blood that influenced my decision), but this bench is 100% French at it's heart. I think Roubo would be proud to know that over 2 centuries later, his design is still being used, and enthusiastically at that. Thanks to Scott Landis' "The Workbench Book" for introducing me to the Roubo bench. And thanks also to Chris Schwarz for rekindling my interest. I'm looking forward to many years of work on my new 2008 Roubo German Model.

Sunday, July 13, 2008

Bench #3 - Attaching the top, leg vises

I like a row of dog holes as close to the front of the bench as possible. The bench features 5-3/8" thick legs that are flush with the front of the top. I didn't want to place the dog holes behind the legs, which would mean a reach when working narrow pieces, and a bit of a pain to raise up the dog that falls behind each leg. Skipping over the legs was also not an option, since I wanted a hole every 3"-4". That would mean a lot of extra cranking for pieces that fall around the "leg" hole. So my solution was to place a dog hole directly over the leg itself, and mortise the top of the leg for a dog hole. I then bored a hole cross ways through the leg itself so I could reach in and pop up the dog when necesary. The access hole passes completely through the leg so it's easy to keep clean. I also made a dedicated dog that lives in this hole exclusively. It's a tad shorter so when it bottoms out against the ledge in the hole, there is just the right amount of room under the bottom that I can get a finger in there to raise it up.

The dog holes are canted at 2 degrees, so calculating the offset for drilling the dog hole took a bit of thought since the top of the leg would end up buried in the benchtop. Plus, I had to drill at 2 degrees off plumb to get the hole to meet up properly with the access below.

I squared up the dog hole with a chisel. It felt real nice to chop into end grain, since there is no end grain to contend with. I was a little concerned that my solution might weaken the leg. Well, it does weaken the leg since material is being removed. But I'm not worried about it. The short grain section above the access hole is the fragile spot, and with the rest of that massive shoulder distributing the force transmitted through the top, I don't think there's anything to be concerned with.

I managed to get the base flipped upside down onto the underside of the top to mark the tenons for the top mortises. I marked very accurately with a knife and proceeded to waste most of the material with a plunge router.

On the fourth mortise I noticed I was about 1" short of my target depth. My bit just wasn't long enough anyway. I found a bit that was long enough and it just so happened that it had a pattern bearing. So I squared up the top of the shallow mortise (above), ditched the router fence...

...and took the mortise to final depth. What a great technique!

After I squared up the mortise corners I measured the leg tenons with a dial caliper and made sure they would just slide into their mortises. I also drilled the top for the tenon pegs. I didn't want to drop the base into those mortises and remove them more than once. My careful tenon sizing paid off. The top dropped into the mortises like a glove and the shoulders seated nicely. So I marked for the pegs, and setup four spreader clamps to raise the base off the top.

Before I attached the base permanently I routed and installed two 1-1/4" angle irons for the sliding leg vise, one for each side of the bench. I sometimes work from both sides of the bench, so having the ability to place the sliding leg vise on the back of the bench will be a real bonus for my work. The idea for this comes directly from the Woodworking Magazine Blog. Specifically the sliding vise design by Bill Liebold. Thanks Bill and Chris for the info.

Getting the top married to the base was a turning point. It meant the bench itself was done. All that is left is to install the leg vises and throw on some finish. The bench is 36" high. That's a full 3" shorter than bench #2, and I'm already loving it. I think Schwarz's 34" would be a little too low for me. He must have long arms for being 6'4", since my pinky pretty-much grazes the top of this bench. (he mentions in the book that his grazes the top of his 34" high bench) Either that, or I have short arms for 6'1".

That's the leg with the dog hole access.

The sliding vise took a little mocking-up to get the tongue right and tilting-out action correct.

I couldn't simply cut a tenon onto the end of the sliding portion, since the placement of the dog hole strip forced the angle-iron channel rearwards past the thickness of the slide. Adding a tongue or tenon to the back of the slide was the only option. Here I'm clamping on a test piece to check position and thickness. The trick was to find the right height to allow the slide to rise off the runner at the bottom, yet when lowered into position leave enough of the tongue engaged against the angle iron for strength when the vise in engaged.

After I found that magic number I cut the final piece and mortised it so it would slip around the vise screw's nut.

I didn't glue the tongue in place in case it fails in use (I doubt it), but simply ran four 1/2" bolts through the slide and the tongue piece.

For the smoothest possible action on the main leg vise I overdilled the holes in the chop and leg and mortised in Delrin bushings just larger than the screw diameter. These keep the screw from flopping about and provide a smooth surface to rub against. These don't need to be very thick since the screw is not bearing any of the chop's weight.

My original garter design called for brass. I ditched that idea when I found out what brass is going for today. So I used an almost equally expensive material-ebony. Well, it's actually quite a bit less than brass, but it fills the bill nicely, and I had some hiding away for a couple years, so this seemed like a good use for it.

I taped the two ebony halves to a scrap block and chucked the whole works into the lathe. What better way to make perfect circles?

A few scraper cuts later and it was done.

I popped them off the block and drilled holes for attachment screws.

Then I used a router with a circle cutting jig to define the perimeter of the garter's mortise before drilling the through hole for the screw.

The screw engaged the garter perfectly. The ebony garter with stainless steel cap screws (I tapped the wood for 10-24 threads) also looks pretty nice, I have to say.

The leg vise, with roller supports and delrin bushings works effortlessly. I couldn't be happier. In fact, it works so well, that I can grab the end of the hub and give it a spin to open the vise.

It's looking like Roubo's German bench now.

Thursday, July 10, 2008

Bench #3 -- The Tail Vise

To install the tail vise I first have to fit the front laminate to the end cap by means of a double dovetail. Dovetailing the end of an 8' long, 1-1/2" thick Ash board is not my cup of tea. It's not like I can clamp in a vise and start sawing vertically. So after careful layout with a knife I got close with a jigsaw (yes, a regular old hand-held jigsaw, but a good quality Bosch saw with a nice blade) then pared to the layout lines. If anyone knows a better way to do this, I'm all ears. The tails did turn out pretty good, all things considered. I've never been much of a dovetailer.

I mark out the half-blind pins on the end cap with my Czech Edge marking knife.

I got rid of most the waste with a router. Gee, I won't be able to do this with the new bench's tail vise. That's why I'm keeping bench #2!

I pared to the scribe lines.

They fit, and they are solid. I had to glue in a couple cosmetic "shims" here and there. Good 'enuf for this bench.

I drilled through the end cap for the screw and bolts.

The tail vise hardware. In this entry I posted drawings from Landis' book for David Powell's tail vise design. A fellow woodworker got me in touch with a retired machinist who has a complete machine shop in his garage. He was able to manufacture the parts to my specifications for the vise. The screw and nut I purchased from Woodcraft. The guide rails, plate and angle I had fabricated. As I mentioned in the previous entry, I did away with Powell's crossbar. It's not necessary.

The guide railes are mortised into the underside of the bench. I scribed lines for the mortises.

I cut the mortises with a router equipped with a guide fence.

In this vise, the screw remains stationary as the nut traverses the screw. Contrary to a typical tail vise arrangement where the clamping force is applied against the screw's attachment flange, in Powell's arrangement the clamping force wants to push the screw out of the end cap rather than pulling it in towards the bench (the vise is tightened by turning the screw counterclockwise). Landis doesn't mention this in his book, so judging by the small screws holding the screw flange to the end cap in Landis' drawing, I believe this may have been an overlooked element in Landis' analysis of this vise. In order to counteract these forces, I completely eliminated the screwed flange and decided it would be prudent to bolt the flange to the end cap. This way the clamping forces would be working against the threads of the two 5/16" bolts, instead of the rather weak holding properties of a typical wood screw.

The 5/16" nuts on the inside of the end cap.

The dog block mounts to the angle component of the tail vise. It's rabbeted to cover as much of the iron as possible without interfering with the capacity of the vise.

There is a square hole milled into the angle and plate to allow the dog to pass through.

The view from the top of the bench. This is the slot where the dog block goes. To the right is the first dog hole in the bench top.

The dog block in position, bottom view.

The dog block in place, top view.

The screw alongside the dog block, top view. As the screw turns, the block advances or retracts.

The complete vise from the bottom.

A short video of the tail vise's action.