hand cut dovetaiols

Re: hand cut dovetails

Jim Shaver Oakville, Ont.


>Hi Marty,

That is a good question. I don't use either when I cut DT's, I use 1:7.

I may be wrong but I understand the difference has to do with the strength of the wood fibres in the grain as you approach the end of the tail (the tips break off more easly with a higher ratio in the softer woods?).

I have believed that the soft woods do not have the same fibre strength as do the harder woods, so a lower angle ratio has been the rule. I may be completely off base with this understanding but is one that makes sense to me.

:-)

I have been using 1:7 all along and not really paying attention to the other ratios, it may make more sense depending on scale, such as a drawer side vs a large case application. The drawer may not be as critical to meet this "rule" as say a case assembly that needs the strength.

Any one else have a rule for the differences?

Take care,

Jim

Re: hand cut dovetaiols

Todd from San Jose


>Hi Marty and Jim,

I've gained the impression that it really doesn't make that much of a difference. Maybe in theory it does, but in practice maybe not.

Lonnie Bird, for example, had us make 14 degree dovetails for all types of wood, which is about 4:1. He's been doing this for years without issues. I've gone to this angle as well, because it is very pleasing to my eye. He's a big early american furniture fan, so I would imagine that 14 degrees was used a lot in the 18th century, although I don't know.

Todd

Me too, the middle of the road...

Scott Burr in Ben Lomond, CA


>I use 1:7 for the most part. Also I have never seen a difference between hard and soft woods. Jim Kingshot always used 1:7, just so ya know.

Re: hand cut dovetaiols

Donald Pierson


>I like the looks of the 1 to 4...14 degrees. They are more prone to chipping at the corners but I haven't had any trouble so far...Don in Vancouver,WA,USA

Re: hand cut dovetails

thomd


>I think there are basically 2 reasons for the degree differences in Soft vs. hard wood. While dovetails can be clamped, traditionally they were generally held by the mechanical hold of the joint itself, while the glue dried. Softer woods require more angle to hold the joint because of the ease with which they can be compressed. It's not just the wood that is compressed but the residual waffle pattern from the saw. While it's not good to force the work together there is a little more leeway in softer wood. The higher angle draws that in tightly.

The other factor is that while the lower angle works well enough to hold hardwoods together, it's more difficult to cut, and hit the fit with. It is one of those factors associated with better quality work. The high angle tails are sometimes referred to by the lower standard of the worker doing them like carpenter or joiner. Hey I read it once in FWW.

I don't think the higher angle make for a stronger glued joint, since they reduce the number of tails you could crowd in, and create a short grain condition. I think on purely practical grounds the best reason for high angle is getting a tight fit in soft woods. Of course in glueless construction it's a whole other issue.

I agree with everyone else that these are not hard and fast rules. And the 1/6 vs. 1/8 is just marketing to sell two gages. Or to have two gages at either ends of a spectrum. If you go back to what was happening around 1980, the mark of good workmanship was the low angle variable spacing DT, because of the influence of Krenov's work. That's why we are stuck with a Leigh jig that cuts low angle variable spacing, though fat joints. An adjustable jig like the Leigh is a very good thing but the original push was for variable. I think a lot of folks these days would actually prefer the Leigh if it was steeper. Styles change and we now see more emphasis on cabinet grade dovetails with very small pins, or tails stylized with higher angles. For instance a twisted dovetail with 1/9 angles would look pretty silly, I think.

It can happen that dovetails need to be varied either to suit wood grain problems, were a very short grain would result from a certain angle. An example of that was Simon Watt's step stool where the wood in the sides was aligned with the load path, at an angle with the dovetailed treads. A normal orientation of the tails would have lead to very short grain so they had one vertical side and one steeply angled side. This can arise from grain orientation in the plank.

Sometimes, though this would not happen with the bland woods I normally use, one needs to consider figure and colour in laying out the angles. If one had very prominent striping, as can happen even in Walnut, the result can be awful if it fights the angles on the tails.

The best work, without shouting it from the rooftops in some flashy way, takes all this stuff into account and an experienced eye should be able to determine the intend of the joinery in good work.

1:8 = show, 1:6 = strength,

Andrew F in Australia


>Irrespective of what the wood type is, you use 1:8 for show and 1:6 for strength (carcase dovetails.)

So a drawer side cut in pine is 1:8.

A carcase cut in pine is 1:6

A drawer cut in hardwood is 1:8

A carcase cut in hardwood is 1:6

This is why the commercial router bits are sloped at 1:7

Cheers,

Andrew

See the handtools board for more info - last week there have been a few threads.

Re: 1:8 = show, 1:6 = strength,

Marty


>My Goodness,Thank you all Jim,Scott,Thomd,Todd,Donald and Andrew. I guess I will not spend the$ At LV and try to do the 1:7 myself

Make one yourself Marty...

Scott Burr in Ben Lomond, CA


>Not that hard to do if you use wood or plastic. Brass might not be that difficult either. Make a it a bit long to get the proper slope. then cut it down to a more usable size say 1.5-2 inches.

or just do them by eye so it looks good ;-)

Dan Donaldson

Re: hand cut dovetails

Chris Scholz


>Marty, that seems to be an open research topic.

The most recent publication states that for timber construction a ratio of 1:3 is optimal (in fact it is 25% to 35% stronger than a 1:6 dovetail). They also look a the dovetail width (more precisely the base width). Can anybody who reads Japanese tell us what wood they were using? MC appears to be a major factor that influences the strengt of a joint (see abstract of second paper below).

Miyauchi, Kazunari, Masuda, Minoru; Murata, Koji, Analysis of strain distributions of wooden dovetail joints using digital image correlation method, Div. of Forest and Biomaterials Sci., Graduate School of Agriculture, Kyoto Univ.; Zairyo/Journal of the Society of Materials Science, Japan, v 55, n 4, April, 2006, p 367-372 Language: Japanese

ISSN: 0514-5163, Society of Materials Science Japan

Abstract: In Japan, the shapes of traditional wooden joints are based upon carpenters' experiences. However, in terms of strength, it is not revealed whether these shapes are optimum or not. For dovetail joints (ari-tsugi in Japanese), which have been frequently used in traditional timber constructions, an inclination of 1/6 is empirically and often used. In this paper, we conducted the tension tests with dovetail joint specimens of various inclinations and base widths, and examined whether empirical shape is optimum or not. Then strain distributions were calculated by digital image correlation (DIC). The results were as follows: (1) Joints with the inclination of 1/3 were 25-35% stronger than that of 1/6, therefore empirical shape was not optimum. In comparison with the inclination, the effect of the base width to tensile strength was not significant. [...]

Harada, Masaki, Hayashi, Youko; Hayashi, Tomoyuki; Karube, Masahiko; Ohgama, Toshimasa Effect of moisture content of members on mechanical properties of timber joints, Laboratory of Engineered Wood and Joints, Department of Wood Engineering, Forestry and Forest Products Research Institute; Source: Journal of Wood Science, v 51, n 3, June, 2005, p 282-285 ISSN: 1435-0211 Publisher: Springer-Verlag Tokyo

Abstract: To investigate the effect of moisture content (MC) of members on the mechanical properties of timber joints, bending tests of precut joints and shear tests of dowel-type joints were carried out using timbers of Japanese cedar (Cryptomeria japonica D. Don) with three moisture conditions: green, kiln-dried with a MC target of 15%, and over-kiln-dried with a MC target of 5%. For the bending test, timbers were processed with a precut processing machine into "koshikake-ari" (a kind of dovetail joint) and "koshikake- kama" (a kind of mortise and tenon joint).

Re: hand cut dovetails and strength

thomd


>It's only relevant if the tests replicate actual failure mode. That does not sound like the kind of thing I generally cut dovetails for. It's better than the stupid test than take various joined corners and place them under a ram, and destruction test them, those are a joke.

Let's say you are making a timber frame and the dovetails are in various braces or beam members. The joint could be desinged not for max load to failure but max stability. The frames fall down not when they pop apart but when they skew to the point where the posts are no longer in column. That's a contention, but suffice it to say one has to have a specific objective in testing.

With the router bits there are two unusual factors. The slope is related to tuning the joint tightness by agusting the cutter height to open or close the joint. The other factor is that the ost comon jigs and even the leigh on alf blind, are not square shouldered and only provide maximum retention over a very narrow area, so there is a further design consideration that is not relevant.

"Strongest" is not tnecesarilly the best design. The strongest shoes might be casy iron, but I wouldn't wear them. Joinery needs to havea range of factors that determine the ideal from the shop to the 50th year in use and beyond. However even where maximum strngth is concerned one has to deal with the issues of wood strength and glue strength. High angle dovetails actually undermine the continuity of wood fiber, many joints will fail in the wood, so making it easy to simply split off a segment of the tail is not necesarilly a good idea. Also really high angles minimize the number of tail joints that can be cut. On the other hand, high angles have a better chance of survival if glue starts to get dodgy over time. most of the carcase failures I have seen were in the backs of drawers not in the big boxes.

Another major issue are the loads in the piece. Made this dovetailed tiger stripe book case. The top was dovetailed. I cut the joints so the tails are on the top where the sexy angles can't be seen. What one can see, and the joints are exposed, is not that interesting, but it's in the right place for structure. Though structurally the loads aren't all that great. Hmm... It would certainly sell better with the tails on the sides.

Re: hand cut dovetaiols

arw01


>Soft woods crush more therefore they need a steeper angle so there is more wood in the joint to hold.

Alan

Testing Dovetails

Warren in Lancaster, PA


>I wonder how many of those who have "tested" various angles have ever taken apart the dovetail joints and put them back together. This is done routinely in repair work. Tails with larger angles are most apt to split off and make more work.

Re: Testing Dovetails

thomd


>Good point, and that is what they are likely to do otherwise also. Though it can still be a good solution depending on the objective.

myth (I think)

Andrew F in Australia


>Hi Alan,

I think that this is a myth propagated now in the internet age.

British trade tradition from about the late 1700's has been to cut show dovetails at about 1:8 slope, and strength/carcase/framing dovetails at 1:6 slope. This is what I was trained as well.

I may be wrong and am regularly.

Following references to common US books show that:

Carcase dovetails cut about 1:6 (10 degrees) Frid p85

Hayward: 3/8":3" decorative hardwood (1:8)

5/8":3" softwood (1:4.5)

Walton: 1:8 hardwood (Australian book)

1:4 softwood

Joyce: 1:5 max - 1:6 Carcase, 1:7 general show work, 1:8 fine show

quality work. (p171)

There appear to be two schools of thought - Joyce corresponds to my trade training and British tradition, looks as though two other reputable books go the hardwood and softwood route having different bevel angles.

I'd stick with 1:6 carcase and 1:8 show dovetails - just letting you know there is another school of thought.

Cheers,

Andrew

Reformatting text so it's clearer

Andrew F in Australia


>Following references to common US books show that:

Frid p85:

Carcase dovetails cut about 1:6 (10 degrees)

Hayward:

3/8":3" decorative hardwood (1:8)

5/8":3" softwood (1:4.5)

Walton: (Australian book)

1:8 hardwood

1:4 softwood

1:6 average

Joyce: (p171)

1:5 max - 1:6 Carcase,

1:7 general show work,

1:8 fine show

quality work.

I'd stick with 1:6 carcase and 1:8 show dovetails - just letting you know there is another school of thought.

Re: myth (I think)

arw01


>Personally I have no data to back up anything. Rob Cosman told his in class that was the reason for the differences. His marker that is sold by lie nielsen is 1:7 and he uses it for aspen sides.

Alan

Get Twisted!

Donald Pierson


>I doubt that the choice of handcut dovetails is based on strength of the joint. Do dovetails made with jigs and router bits qualify as handcut? As a newbie to woodworking about a year ago I decided to try use dovetails on a shelf I wanted to make. After looking a jigs...more $ than I wanted to spend...tried my hand at marking and ran into my limited skills and lack of patience, I wrote a Windows program where I enter the design parameters and it prints out a pattern that I glue onto the boards. Works fine for me. Somewhere along the line I found out there was something called Twisted Dovetails. Ordinary dovetails have mechanical interlocking along one board and rely on glue to hold the tail board to the pins boards. Twisted dovetails have mechanical interlocking along both boards. I wrote another windows program that generates pattern for twisted dovetail joints. This program is more complicated due the many design options. I have been writing a web program do the the twisted dovetail patterns. It is my first web program which means a lot of learning on my part and someday I will get around to finishing it up. It works on my computer but there a lot of checking and details to finish. Don in Vancouver,WA,USA

No one has mentioned thickness!

David Charlesworth


>I find the thickness of timber being used has a great impact on the visual appearance of slope angle.

I like 1:8 for drawer fronts, single lap dovetailed, with not too thin a lap to avoid ghosting through. I like drawer fronts to be 7/8 thick.

For the backs/sides, where I use 3/8 or even 5/16 thick hardwood, (classic english work is often done in quartered oak), I use 1:6.

If I draw 1:8 the dovetails look almost parallel because of the thinness of the material, almost like a finger joint.

It has become clear to me over the years, that the slope which pleases your eye best is the one to go for. Strength issues are almost completely irrelevant in the world of modern glues. The wood will shear along the grain before the glue fails if all else is well done.

"Rules" are best treated with suspicion/scepticism as there will always be exceptions!

David Charlesworth

PS I was pleased with the drawer making and fitting articles in my second book Pages 26-41

Do you mean this one? *LINK*

Scott Burr in Ben Lomond, CA


>This was posted a week or two ago...


DT Generator

Re: Do you mean this one?

Donald Pierson


>Yes, like the blocklayer site. My DT program provides for each board having a different thickness, number of pins/tails, board leaning from vertical. Here is a pic of a bookend/box I did that has 1,2,3,and 4 pins/tails on each corner and the front leans at 75 degrees and the end at 60 degrees. My TDT programs has a nine parameters including number of pins/tails, sequence of the pins/tails...ptptp or pptpp for a five fingered joint for example...and "proudness" of the joint...Don in Vancouver,WA,USA

That Rocks!

Scott Burr in Ben Lomond, CA


>Donald, I for one would love to give your program a try! Please keep me (us) posted.

Very nice bookend/box combo!

Re: That Rocks!

Donald Pierson


>Scott,

I am all for sharing the programs I write and I think putting a software program on a website is the future. Distributing programs via cd's or downloading is yesteryear and blocklayer is a good example. I am having to do a lot of learning to come up to speed on the latest software techniques and dealing with the difficulties of writing a website program. Writing software is down the priorty list that is topped by taking my dog Ling to the dog park followed by working on my latest wooden gift I am making for a daughter or grandchild. Writing sofware is fun but so is making sawdust. I have a lot more to learn about woodworking than I do about software...Don in Vancouver,WA,USA