Chris Schwarz's Blog

Thanks Be to Mongo

For this Roubo workbench to work, I’ve got 16 joints that have to come together all at once. There is not an option to glue things up in stages and still guarantee success.

As a result, I tried to give myself some wiggle room. I fit the edge cheeks of the tenons a little looser than normal so they could shift around. I made my tenons’ face cheeks a shade thinner than when I build a typical door, for example. And I cut the through-tenons so that I will have to wedge them.

That said, getting this bench together for a dry-fit required Mongo, the 2-pound English mallet. I tried my other wimpier mallets first. I would have been banging all day with the little guys.

But Mongo made it happen.

Each leg had to be pounded about 20 times to get it to seat. Then I clamped the lower stretchers to the legs to see how all the shoulders looked. The shoulders on the front stretcher look good. The side stretchers each have a gap I need to tend to. The rear stretcher is just going to have to learn to be different. (Like the kid in my high school who had a weird affection for live poultry.)

Next up: Fix the shoulders (if the workbench will come apart), start drawboring the joints and select a glue with a two-day open time.

But I’m about to hit a wall. This weekend I’ve got to build a birdhouse. Yup, you read that right.

– Christopher Schwarz

32 thoughts on “Thanks Be to Mongo

  1. Dave


    Would you seal the bottom of the bench legs with epoxy or some other coating to prevent moisture absorption from a concrete floor?


  2. Brian Ogilvie

    "Mongo just pawn in game of life"

    –Blazing Saddles

    I am going to rename my 3-lb mini-sledge "Mongo"!

  3. Niels Cosman

    Hxtal is a super high purity epoxy that was developed for archival restoration of glass. In my opinion is the best glue on the market for glass- it’s water clear, strong, and versatile… and very very expensive.

    It’s open time and curing are proportional to the ambient temperature in the room. At 70 degrees It’s reasonable to say that the open time is 24-48 hours, meaning that that’s how long you have before you get out the epoxy solvent (nasty nasty stuff). You can also turbo charge he process by heating the glue with heat(lamps) up to 110 degress, reducing the total cure time to well under 36 hours. When mixed, hxtal is super runny (like water), however let sit for 4-6 hours it becomes thicker than molasses. The ultimate bonding strength is not affected at this point allowing you to customize thickness for the task. when it is so thick that that gravity barely has any effect on it- it’s like glass spackle. neet!

    HOWEVER… I’ve only used hxtal glass to glass. Other glues are recommended for glass-to-other bonds. It’s my understanding that is because hxtal is an extremely rigid glue. That inflexibility can cause to problems (self-destruction) between materials with differing coefficients of expansion. Over large surface areas, this is a problem even between different glasses. Also did i mention the price, it’s pretty expensive.

    wow, sorry for rambling- sort of got carried away there.

    The folks at HIS Glassworks distribute hxtal and are fountains of knowledge when it comes to it’s preparation and application. Call them up and ask if how it works with wood, they are very friendly.


  4. adrian

    It took me 45 minutes to assemble my last project, and I used fish glue, which worked well. (Advertised open time of 60 minutes.)

    But when I mentioned the need for a long open time adhesive someone suggested an epoxy called HXTAL. Since the workbench already makes use of epoxy, perhaps you should consider using this product. The information I’ve encountered is a little vague, but it appears to offer an open time of about 24 hours. One set of instructions suggests letting it thicken for "several hours" after mixing before use, for example. It is seven days to full cure.

  5. greglease

    I’m also interested in the concept of drawboring without glue, especially if the through-tenons are set with wedges. I’m working on my first real workbench (all hard maple) based loosely on Garrett Hack’s recent design. In assembling the 2 ends of the stand the cross braces need to be set and then the double through-tenons of the legs set into the feet and tops and wedged. I already planned to drawbore each joint, as much for the look as for stability, but the concept of leaving them glueless would sure make things less frantic in the final assembly of the ends.

  6. Thomas J. Hamernik

    16 joints? Are you counting the combo dovetails and tenons through the top as two joints each leg?

  7. AAAndrew

    This is a serious question, maybe not an intelligent one, but serious.

    What is the possibility of this coming apart just as it is, without drawboring, without glue?

    There is so much tension built into and working on the whole structure. Each piece is holding the other pieces in place because of tension from somewhere else. Think of the amount of force it took to get everything in place, and then think that the same amount of force, or more, will be required to work it all loose, applied in very specific amounts at very specific locations in a fairly specific order.

    I ask this in all seriousness because when I built mine, I didn’t glue the long stretchers, nor did I do anything to the blind mortises into the bench top but get the tenons in. There is so much mass and natural friction that I can’t imagine what I’d have to do to the bench to really make it even begin to work loose.

    I wonder if some of it is because of the shorter size. Mine is only 5′ long and looks similar in scale to yours. Without the great lengths of bench top, I suspect the racking tension and other forces working on the base and base-to-top connections is less.

    Just a thought. Perhaps you’ll try a glue-less bench next time. 🙂 (ducking and running)


  8. james

    John & Peter make a good point, why glue at all? Is there any advantage to gluing VS drawbore M&T joints?

  9. Keith

    Here is a different way to seat the legs. Start them in their mortises then turn the bench right side up. Lift the assembly by the bench top and drop it so it’s own weight drives the joint home. Should only take a couple of lifts per side.

  10. Christopher Schwarz

    Don’t know about the stability yet. Seems fine now. Time will tell. I have some things up my sleeve.

  11. Christopher Schwarz


    That might work. It might not. Once you start adding more and more joinery surfaces, you increase what some people call the "dynamic tension" of a piece. Lots of parts interacting with one another, making the tolerances tight and tighter.

    Because you want the leg-to-top joint perfect (it’s the strongest int he bench) you want those joints dead-on perfect. Once you make that decision, you really have no choice as to how to assemble.

    Plus, this isn’t a machine-made project. Every joint is individually fit, so fitting the top to an assembled base would be a difficult task. Nothing in this bench is perfectly square.

    Hope this help explain my approach.

  12. Thomas J. Hamernik

    How is the stability? Do I recall correctly that you had wondered about it since the bench top is narrower than what you build typically?

  13. adam

    Amateur question: If I had to glue up something like this, i’d do it in two steps: 1) glue up the legs to their stretchers, clamp them together and then just tap the tops of the legs an inch or so into the bench top to make sure they were oriented reasonably well with the mortises there. 2) after the legs were dry, I’d then glue/pound/drawbore the legs into the top.

    Seems like it’d save me a bit of urgency doing it in phases but i’m a tyro so perhaps missing something.

    By the way, I think the whole thing is lovely. the contrast of the pine vs old cherry is really striking, and the joints into the table are really cool. it’s going to be anice piece of furniture for the shop!

  14. Tommy

    Is there any clue as to how this bench would have been assembled in the pre-clamp days? Such as drawbores or wedged tenons on any of the old drawings?

  15. Shawn G

    Since she’s had the practice why don’t you just have Meagan jump up and down on it to seat the top?

  16. John Walkowiak


    Forgot to add, if you put a couple of 2×4’s directly on the floor, then put the benchtop on them, seating the legs will be much easier. As you have it the picture, the saw horses and rubber matt are absorbing much of your mallet blows. And, being lower, you can use a huge Commander, like were used to drive in the pins on post and beam barns and like Roy is using on the end of his shows.

  17. John Walkowiak

    I don’t really think you need glue. About 15 years ago I built an assembly table, 36"x66". I used doug fir 4×4’s for the legs, and the stretchers are construction 2×4’s, all were cleaned up a bit. I used drawbored M&T joints, 1-1/2" pin for each tenon. I drilled thru the legs and left the pointed ends protrude on the inside. I used no glue as it would have to be taken apart to get it out of the shop. It has been in constant use, being pushed around the shop to make room and it is as solid as when it was built. As long as there is tension on the shoulders, which the drawbored pins provide, the the joint will remain tight and the glue is really doing nothing.
    Live on the edge – go glueless!

  18. Chris F

    Mongo seems pretty lightweight. Wouldn’t this be the job for "Luigi" the 10lb "persuader"?

  19. Eric R

    Well, if there is another Whopper earthquake, anywhere near you, I bet that sucker won’t even tip over!!!

    And a two day open time sounds about right for that bad boy!

  20. Ethan

    You can almost see the juxtaposition of tension and depression build in Lambo with the onset of "middle-bench syndrome" as Minibo nears delivery, er, completion.

Comments are closed.