HOWTO use big screws into end-grain?

[Simon_M]

With big screws do you need a pilot hole all the way? It's into "end-grain" but with a dowel to screw into...

I have a frame that's underneath the top of a "bought" workbench. The frame has a top rail on each side with feet and a lower rail on the floor. They come with each foot attached with "only" two dowels which are glued in. The diagram shows the side of the top rail (100mm deep, that's screwed to the workbench top) and the top of the feet with the "double" dowel joints (10mm x 30mm).

This is repeated for the rail against the floor and the feet are further apart than shown!

I keep thinking that this is a weakness that's waiting to fail and I would like to make it secure first, before it fails. I have added a plywood back to the rear feet, a stretcher between the two front feet and a shelf. This is to prevent racking. The original joints failed (without the additional support) and I don't want it to happen again.

One issue is I can only drill the small pilot drill with a maximum depth of about 100mm so it's more clearance for the screw with no "pilot" into each foot (the dowels and frame are now glued together).

PS I found this in a thread "screwing into end-grain" "Screwing into end grain always used to be a no no but that was with the older traditional style screws. Modern deep thread pozi drive screws will grip fine into end grain but you need very little or no pilot hole". Does it really have strength or is the cross-grain dowel a better solution?

 

[AJB Temple]

This is a lot of complicated questions and hard to visualise without pictures , which is why I suspect you are getting no replies.

 

[woodbloke66]

Modern deep thread pozi drive screws will grip fine into end grain but you need very little or no pilot hole". Does it really have strength or is the cross-grain dowel a better solution?

They don't really and secondly, yes - Rob

 

[Simon_M]

There are a lot of complicated questions and it's hard to visualise without pictures , which is why I suspect you are getting no replies.

I updated the original description to include a diagram of the "problem" area.

Modern deep thread pozi drive screws will grip fine into end grain but you need very little or no pilot hole". Does it really have strength or is the cross-grain dowel a better solution?

They don't really and secondly, yes - Rob

This is the technique (using dowels) and you used it (successfully) on your Assembly Table?

This is how I could strengthen the dowel joint (on LHS is one available option and on the RHS is the "Rob" method):

On LHS I could use a bolt to pull the rail and feet together. There's a 30mm hole drilled into the foot with a Forstner bit and a bolt with nut and washers to pull the joint together.

On RHS I could use a screw into a hardwood dowel with available cross-grain. There's a clearance hole drilled into the rail but (not much) of a pilot hole for the screw in the area around the foot. The 130mm x 6 screw and dowel pull the joint together.

I could just hope that the joint is strong enough without any additional hardware (say prayers), add screws and hope they work in end-grain (and regret it) or adopt one of the two techniques shown. The workbench has two stretchers which are secured using M12 x 120 nut/bolt/washers technique to draw them together and I used this (successfully) when making an additional stretcher for the front legs.

 

[Sideways]

I'll tell you what I did when building a shed recently.
Time will tell whether it's durable.

The sides and ends were build of softwood framing with an OSB skin to support the battens and outer cladding.
I screwed into end grain for all the vertical studs.
In the top and bottom plates (horizontals) I routed a quarter inch deep recess to take the end of each vertical timber stud.
(vertical studs are 3x2" at the end and middle, 2x2" in between)
I clamped everything together dry using the OSB as an assembly surface so the framing was flat and tight.
Predrilled through the top and bottom plates into the ends of every stud to a total hole depth of about 100mm (that's about 60mm into the end grain once I was through the top or bottom timber).
Put a 6" x #6 forgefix cut point torx screw into the end of every stud.
Predrilling and the routed sockets meant that not a single end split.
The length of the screws meant that every one also drove on into virgin timber once it had bottomed out in its hole.
I was happy with the build and the finished shed is very solid.

• I think long and strong screws were essential.
  Predrilling was vital or the 2x2's would have split.
  The routed sockets were overkill but didn't take too long to cut with a simple jig and helped both with keeping square during assembly and again resisting the tendency to split the timber ends.

If the roof and trusses pull the top plates off the walls then I'll let you know

 

[ED65]

PS I found this in a thread "screwing into end-grain" "Screwing into end grain always used to be a no no but that was with the older traditional style screws. Modern deep thread pozi drive screws will grip fine into end grain but you need very little or no pilot hole".

Take that with a pinch of salt.

Actual test data tells us that screw design is far less significant than some sources say, far less than I've read at least going back years, to long before I took up woodworking. In fact screw type seems to make very little difference, there is a surprisingly uniform lesser hold in end grain for all screws – it's a given percentage of their hold in long grain (I can't remember the number offhand, but it's greater than 60%, so not paltry). Well-sized pilot holes are assumed in all cases from what I understand.

...or is the cross-grain dowel a better solution?

Probably always, with a reasonable size of dowel.

I do think that examining your supposition of the weakness of the dowelled connection might be advisable, but if you want to reinforce for peace of mind by all means go ahead. There's rarely a downside to over-engineering something!

 

[Jake]

There are some very aggressively threaded structural screws these days which I would not hesitate to use in end grain in any realistic non-structural scenario (most of them are designed to be used diagonally rather than horizontally into 'end grain' sections for actual structural use, but that's holding together houses not benches).

 

[Robbo3]

Sometimes even a cross dowel isn't enough.
I tried that on a gatepost (so not end grain but poor quality wood) to try & prevent the screws holding a hinge mounting pin from pulling out. Ended up using carriage bolts.

 

[dzj]

I'm just guessing here, but judging by the joinery used, your bench has rather thin legs
to start with. Not really adequate for everyday power/ handtool use.
One way around this is to give them a bit of mass by gluing some 18mm stock on both sides of the legs and rails.
This way, you'd also get a sturdy mortise & tenon joint connecting the members.

 

[Simon_M]

Thanks everyone for the comments. I have implemented the LHS solution (see above).

The workbench was replaced under warranty by Axminster so I already know that the glued double dowel joint is not effective at preventing racking and ultimately for these type of joints to have already failed on my old workbench, so it's time to act...

Modern deep thread pozi drive screws will grip fine into end grain but you need very little or no pilot hole". Does it really have strength or is the cross-grain dowel a better solution?

They don't really and secondly, yes - Rob

I tend to agree that the screw will just pull out of the end-grain so doesn’t add too much and using the dowel will greatly increase the strength.

Sometimes even a cross dowel isn't enough.
I tried that on a gatepost (so not end grain but poor quality wood) to try & prevent the screws holding a hinge mounting pin from pulling out. Ended up using carriage bolts.

This prompted me to go with my “LHS solution”, so I’m using an M12 x 120mm bolt with washers and a nut to bridge the rail and feet.

I drilled a 30mm hole drilled into the foot with a Forstner bit and also a 25mm recess from above and a connecting them with a 12mm hole.

With the bolt, washers and nut pulling it together it looks like this:

NB the hardware is bigger than it looks e.g. M12 x 120 bolt and nuts - each rail and foot is 100mm deep/wide with "only" 2x (10mm wide x 30mm long) dowels to connect the rail with the feet with wood glue. When the worktop is screwed to the top rail, all the twisting (racking) forces are transmitted through to the legs via "just" the two dowels in each foot.

The nut and bolt help bring/hold these joint together and because the pair of dowels are unlikely to shear off the joint is now much stronger. The other "racking" solution is a 1/2" plywood back and a plywood shelf above the horizontal stretchers and a third stretcher added below at the front (also using an M12 x 120 bolt at each end).

So overall I have added bracing (especially across the diagonals) in the vertical (back) and horizontal (shelf) planes and closed off the remaining lower front edge (additional stretcher) so it's now a "braced" and "closed" cube. When the top is screwed to the upper rails, I may also bridge/screw the worktop to the back panel.

In the future I may add a concrete building block (B&Q 15kg for £1.50) under each shelf end (to lower the top-heavy centre of gravity because the 95kg workbench has 3/4 of the weight in the worktop) and add to the front legs to make them L shaped (stability) to bring the front level with the worktop and add a useful row of vertical dog holes for support (e.g. planing a cabinet door).

 

[Simon_M]

I'm just guessing here, but judging by the joinery used, your bench has rather thin legs

You are right about the legs. The dimensions of the legs are 100x40mm for the two stretchers and 80x40mm for the top/bottom rails and legs. Here’s a picture to illustrate the “problem”, the total weight may be 95kg but 3/4 of this is in the worktop:

https://www.axminster.co.uk/axminster-p ... eal-717727

I asked the question about racking (Artisan) because one option was to buy the cupboard as a way of stiffening the structure. I didn’t go down this route because I may simply replace the whole leg setup with something more substantial. I also didn’t think upgrading to the next in the “series” was going to improve things as some other feedback was “they are all like this”.

The answer given about the racking was that the cupboard doesn’t stop it. The cabinet looks better than it probably is and being made of chipboard it doesn’t provide much rigidity judging by the answer given. One issue with adding the cabinet, is that the stretchers are replaced by the cabinet and they are put back below the cabinet. So their effect is diminished. One simple “fix” may have been to add two lower stretchers (I added one with also a back and shelf).

If I posted a question about building a workbench like this it would likely get a response that I should rethink the legs. If I proposed using dowels to join rails to legs it would not be viewed positively. The dowels are the weak link here and as well as the top rail to legs, the bottom rails are also (now) held together with M6 bolts like the pictures for M12. I wasn’t expect the lower rails to be a problem however adding the nut/bolt/washers like the addition to the top rail, was easy to add.

So why is it made like this? To reduce the volume of material used and provide an “acceptable” solution? One clue to how it “works” is to look at the stretchers. They use M12 bolts to pull the legs together - something I copied when attaching rails to legs. It is effective especially if the bolts are tightened sufficiently. Adding a lower stretcher using the same design certainly helps too.

It’s taken a while, but the worktop will soon be added and I will be able to comment on the old (unmodified) vs new (replacement with additions).