How Strong is a Dowel Joint?
The woodworker has many options when it comes to selecting ways of joining solid wood. Some joints are easy to machine but don’t offer much strength. Some are difficult but when made well last a long time. Most joints fall somewhere in the middle. The dowel joint, whether made with the help of an aftermarket jig, dowel center or other hole-location method, is easy to make and offers a lot of flexibility. To find out if the dowel joint could stand up to a beating, I selected two other methods of solid wood joinery and tested them all to failure: the biscuit joint is commonly used for its ease of machining, and mortise and tenon – although more difficult to make, and known traditionally for its great strength – is the other joint in the test. I believe these three commonly used joints will provide a fair cross-section of strength results.
Usually furniture parts join each other at 90°, so I decided to test them at that angle, and chose to mimic the joint commonly found in cabinet door frames. Leaving the door panel out of the mix allowed us to get strength results regarding just the rail to stile joint.
Because dowels are often used in melamine kitchen cabinet construction I did a separate set of tests mimicking the joint used to fasten the gable to the cabinet base. There are two other joints offering up competition in this case as well; biscuits with a butt joint, and a rabbet joint. Both of these joints are commonly used in kitchen cabinet construction.
When finished I found the melamine results so surprising that I did two additional tests regarding dowel joint strength when used with sheet goods.
I used this machine to test the strength of all the different joints
Broken dowel joint
The strongest of the bunch, the dowel joint held together until the gauge read 650 lb.
Broken mortise and tenon joint
Coming in a close second, this joint withstood 500 lb. of pressure.
Broken biscuit joint
The biscuits broke at 325 lb.
Great for cabinets
In the sheet good test, the through dowels ranked the highest.
In the solid wood test ¾” thick clear flat sawn red oak was used. For the sheet good test ⅝” thick melamine coated particle board was used. Each test was performed on five samples and the results were averaged. The testing apparatus was carefully set up so that the tests were as repeatable as possible.
Mortise and tenon joints were created using accepted methods for the home workshop. Mortises were cut using a dedicated shop built jig, a router and a new ⅜” dia. spiral bit. Mortise dimension were ⅜” wide by 2″ long and 1″ deep. The ends were squared up using a regular ⅜” chisel. Tenons were created using a sliding table and a dado set. Tenon thickness was set to create a tenon with a tight fit into the mortise and adjusted with a handplane for a snug, friction fit; i.e., the joint required minor tapping with a mallet to close, but didn’t fall apart if turned upside down. Tenon dimension was set for a ⅛” clearance in both length and depth. Both pieces were jointed and planed to final thickness. Glue was applied to both the mortise and tenon.
The dowel joint was created using the Dowelmax jig. Three 2″ long, ⅜” dia. dowels were used in each dowel joint. Holes were drilled to a depth of 1 ⅛ in both pieces and glue was applied to both the holes and the dowels with a small brush ensuring enough glue to create a slight amount of squeeze-out.
There were two biscuit slots per joint and they were all cut with a new blade. The #20 biscuits were a drop-in fit and glue was applied to both the slot and the biscuit.
The sheet goods test samples were all 5″ wide. Two sets of dowel joint samples were created using three and five 1 ½” long ⅜” dowels, evenly spaced along the test sample. The dowels penetrated 1″ into the case bottom and ½” into the gable piece. Two #20 biscuits were used for the biscuit test samples. In the final test sample the rabbet was machined on a table saw and was cut ¼” deep in the gable piece.
All test joints were glued with Titebond 1 and clamped for 24 hours.
All the solid wood joints failed in a similar manner – the wood around the tenon, biscuits or dowels let go of the mating surface. A small amount of wood fibre was still adhered to the tenon, biscuit or dowel, but for the most part the joint failed at the glue line. This indicates that glue surface area and glue penetration into the surrounding solid wood fibres is a key to joint strength.
Test results show that dowels are the strongest method for creating this type of joint. The dowel joint in solid oak failed at an average of 650 pounds pressure, mortise and tenon joints failed at 500 pounds and biscuits failed at 325 pounds. Variations in the pressure at failure was less than 5 percent.
In the sheet goods test, dowels were far superior to any other method. The dowel joints failed at 240 pounds pressure, the biscuits failed at less than 100 pounds pressure and the failure point for rabbet joint was immeasurably low.
The above results compelled me to complete two further tests, slightly modifying the dowel joint in sheet goods scenario. For the first test I used five dowels on the same size part. Turns out more joints don’t create a stronger joint. The test with three dowels in the 5″ wide piece was stronger than the five dowels by almost 50 percent. The second group of sample joints involved three dowels with full penetration through the gable side. This resulted in a joint almost twice as strong as the “blind” dowel joint with three dowels.
It’s nice to know that dowels can provide a strong option when making furniture or cabinetry. Because a dowel jig assists in machining evenly spaced holes that are square to the surface, and doesn’t allow the drill bit to wander during drilling, they provide a snug fitting joint with a surprising amount of strength. This is particularly true in sheet goods. The ease of joinery and fast construction make dowel joinery extremely attractive to both professional and hobby woodworkers.
I built a Leonardo da Vinci self supporting arched bridge spanning a creek 27′. I used one 1″ wood through dowel at each joint on the 4×6 fir beams. The bridge deck was red cedar post, rails and deck boards. All the wood used was off the property at our boat access off the grid property northwest of Vancouver Canada. The bridge was built in the back yard and then lifted by a large zoom boom (54′ reach) forklift into place. The joints didn’t move.
I am completing the 3D cad drawings and then having an engineer see what the load capacity is on this bridge. The next bridge will be approximately 35′ with small side by side ATV traffic. Should be fun. If you send me your email address I can send photos.
That would be cool to see.