Dining Table: Trestle Structure

The support trestle is all made out of old Douglas Fir framing lumber that came out of our house, which was built in 1911.  Roof rafters were 2 x 4 and will live on as the stringers and legs, a 6 x 6 porch post will become the feet, and a 2 x 6 joist from the bathroom floor will become the horizontal brace.

The stringers (aka battens) were made and attached several weeks ago, as I flattened the top.

Mmmm.  Lead-based paint.

The big post that became the feet was coated in nasty old paint, had a ton of nails in it, and has been sitting out in the weather for a couple of years.  Even in that condition, I couldn't throw out a big chunk of clear lumber like that.  First I pulled all the nails (one I couldn't remove got driven in deeply with a nail set) and used a thrasher ebay jack plane to clean off the paint and flatten the surface.  The wood was wet from a few rainy days, so it kept the lead dust down.  I cut off a couple of 26" lengths with one of my grandfather's old hand saws, and then ran those through the power planer to clean them up.  After deciding how to orient the parts to hide most of the defects, I used the band saw to rip down the foot blanks to 3-1/4" x 4-1/4" x 26".

Ah, the irony.  I'm not a masochist.  That stock is too long to crosscut on the table saw.

To make the curved shape of the feet, I spent some time designing patterns on hardboard, then used that to draw and cut a couple of test pieces on 2x6 lumber to show my wife and refine the shape.  Once I got to something we were happy with, I used the template to draw the shape on both sides of the foot, and then cut off most of the waste on the band saw.  I used a hand saw and chisels to make the sharp arrises where the flat top transitions to the curve.  A paring chisel took off most of the remaining waste, and then a little quality time with the oscillating belt sander smoothed things out.

Shaped Foot

The legs are glued up from two 26" pieces of 2x4, with a 1/4" strip of walnut between them as an accent.  Final blank size is 25-1/2" x 1-5/8" x 7".

Roughing out mortises at the drill press

I put a 1/2" x 1-1/2" x 6" mortise in the stringers.  I made one with a brace and bit followed by chisel work, but roughed out the second one at the drill press with a forstner bit.  The quality is the same, but the drill press is quicker.  The foot got a larger mortise, 3/4" wide by 1-3/4" deep, because there's a lot more to work with on that part.  Matching tenons were made in the leg blanks, cut close with the table saw and fine-tuned by hand.  I have a Lie-Nielsen rabbet block plane that is my go-to tenon tuner.  It can even be flipped over and used to clean up the shoulders if the wood is nice.

Cleaning up a bottom leg tenon with a L-N rabbet block plane.  Haven't cut the narrow shoulders yet.

The cross brace is just a 2 x 6, planed down to smoothness.  Big 4" x 4-5/8" long tenons were shaped on either end, fitting into 1-1/2" x 4" through mortises in the legs.  Those were carefully chopped from both sides, since the edges would be visible in the finished piece.  After fitting the legs to the cross brace, a tapered 1/2" x 1" (to 3/4") through mortise was made vertically through those for tapered walnut tusks to hold the assembly together.

Drilling for the mortise that will hold the tusk tenon (in background)
that secures the leg to the horizontal brace.

A few more details:  the ends of the horizontal brace's exposed tenons were rounded off at a 5" radius, and 1/4" was removed from the bottom of the feet to leave square pads on the ends.  Most edges were eased or slightly chamfered with a block plane, to prevent splintering and keep the incidence of damaged shins to a minimum.  The stringers and feet were glued onto the legs and that's all the permanent joinery on the base.  Bam.  Done.  Now these parts can get out of the way while I finish the top.

The big question now is, will I finish in time to use this table for Thanksgiving dinner?  My family's not coming up until Saturday, so I've got a couple extra days.

Choosing a water heater

Like everything else in the world, picking a new water heater requires an unreasonable amount of research and nail biting.  Here's a taste, based on a friend of mine's situation.  It should be pretty easy to plug in your own numbers for usage and utility rates to do this for yourself.

Some terms:
Sorry for the units, all you socialist users of the metric system!
BTU: the amount of energy it takes to heat one pound of water 1˚F (aka British Thermal Unit)
1 kWh = 3412.3 BTU  (common unit for selling electricity)
1 therm = 100k BTU  (common unit for selling natural gas)
1 gallon of water weighs 8.35 lbs.
EF: efficiency rating - tells you how much incoming energy makes it into the hot water.  Manufacturers like to bs around with this.  For example, I believe they ignore any electrical consumption made by a gas appliance.

First, estimate your usage.  For my friend, I figure one shower a day (10 min @ 2.5 gpm @ 105˚F) uses about 20 gal of hot water (the rest of the 25 gallons used is cold water mixed in the shower valve).  1 load of laundry average per day = 4 gal in her high efficiency front loader. Dishes, hand washing, etc. takes another 10 gal.  So, 30 gallons a day.

How much energy is that?  The average incoming cold water is about 50˚F here.  If the tank is set to 120˚F, that's an increase of 70˚F, and that 30 gallons weighs 250.5 lbs.  70 x 250.5 = 17535 BTU per day.

Electric
The most basic electric water heater has an efficiency rating of about 0.90, and putting a thicker layer (3") of foam insulation gets them up to 0.95 pretty quick and cheaply.  Note that the EPA no longer gives Energy Star certifications to any electric tank models except for the heat pump hybrids.  The bottom line is it's too easy to get right close to the theoretical maximum EF of 1.0, so they don't feel like they need to reward anyone for that.  Given how many people and programs base their buying on the Energy Star label, I think that's a mistake, but anyway...  So, I'm looking at a 40 gallon model with a 12 yr warranty at Lowe's for $450.  Nothing fancy.  At EF 0.95, this model will consume 1974 kWh per year.  (17535 BTU / 0.95 = 18458 BTU input needed per day...  18458 / 3412.3 = 5.41 kWh / day...  5.41 x 365 = 1974 kWh / yr).  Our electricity is 0.11 per kWh, so operating cost is estimated at $217 a year.

Natural Gas
A very basic gas water heater is only about EF 0.59, but it's easy to find them up around 0.64.  Energy Star models have to exceed 0.67, and the highest efficiency models get up to 0.70.  I'm looking at a 0.67 Energy Star, 12 yr warranty model at Lowe's for $570.  This model will consume 95.5 therms of natural gas per year.  (17535 / 0.67 = 26172 BTU input per day...  26172 x 365 = 9,552,649 BTU / yr...  divided by 100k = 95.5 therms / yr)  Our natural gas is $1.08 per therm, so operating cost is estimated at $103 a year.

So even though the gas unit costs more, it should pay for itself pretty quickly, and the savings is more if you use more than the fairly minimal amounts in the examples above.  But of course there are other things to think about.  All-new installation of a gas unit can be very expensive.  It needs an exhaust vent (chimney), or for some models they can vent via a plastic pipe out through the wall.  Gas units are dependent on both the electricity (fans, controls, igniter) and gas utilities to operate.  Electric just needs electricity.  Electric units have no moving parts, and are much less complicated.  They also are very cheap to repair (replace an element or thermostat) and the little bit of energy that they do lose ends up in your house.  We live in a coolish climate, so the extra heat is a good thing for eight months of the year.  Gas units dump all their excess energy out the vent, and they also poke another hole in your house's building envelope, requiring installation of a motorized vent damper if you don't want air flowing through the stack when the unit isn't operating.  Electric heaters are much less likely to cause explosions if you happen to spill something flammable near them, and the carbon monoxide hazard is zero.

Dining Table: Battens and Breadboard Ends

The two parts of this table that keep the top flat are breadboard ends and the two battens that also form the top of the trestles.  First I flattened the underside of the top, using a Veritas Bevel-up Jack (blade at 40˚) and my Stanley No. 4 smoother.  Fir is a pain in the ass to plane, in my experience.  High quality tight grained old-growth stuff isn't so bad, but this construction grade crap just loves to tear out, and it tears really deeply.  It's not possible to simply fix it with a scraper because you'd have to dig a crater.  With my No. 4 tuned to the limit of my abilities, I wound up with a couple of really ugly patches.

Evil Tearout

I guess I didn't resort to back-beveling the blade, so my cutting angle was no more than 45˚, but anyway the point I'm getting to is that I decided I needed a new bevel-up smoother.  Going on my old rule of buying as I need (and I need to flatten this big table top), I got the full-size Veritas BU Smoother that shares blades with the other two of their BU planes I've acquired over the past few years.  It's heavy, and it's a real workout to push it with a 50˚ blade in it, but so far it does seem to handle the tough grain reversals better, and it's much quicker to set up than my #4.  The end grain of a couple of knots does some funky things that were handled better by a lower pitch blade, but I'm going to patch those on the top, so I don't care.  I could plane around them anyway.

Rough old framing...

With the underside flat, I started making the battens.  When I remodeled our upstairs bedroom, there was quite a bit of 2 x 4 framing that came out, and I kept nearly all of it, despite the nail holes and mortar and other crap I have to deal with.  Here's why.  Look at that grain.  They were using these trees for construction lumber 100 years ago!  These are rough-cut timbers, measuring pretty close to a full 2" x 4", so I can get real lumber out of them.  I think I've got enough to make all the framing for the trestle if I'm careful and don't screw up.

...Is tight-grained Douglas Fir

I drew out a nice cyma and made a hardboard pattern.  I then traced that out on the pieces and shaped them to the line with bandsaw, spindle sander, and some hand tools to clean up the details.  I attached them to the underside of the table top with metal figure 8 connectors to allow for movement.

Final shaping of the battens

The breadboard ends are cherry, cut from 5/4 stock, which was barely thick enough.  I flattened the ends of the top with a jointer plane run across the grain, so it would join nicely with the flat breadboard pieces.

5/4 Cherry breadboard stock

To make the cherry end pieces I trued up the faces and one long edge of a piece of 5/4 x 39" x 4" stock, then I used a dado stack blade on the tablesaw to cut a 1/2" x 3/8" wide groove in the edge.

I used a router with a 3/4" straight bit and clamped the cherry end board as a straightedge to cut a long rabbet 1-1/2" wide x ~5/16" deep down the entire length of the end on each side, leaving a single huge tenon 3/8" thick.

With dividers, I marked out six 2" wide sections to leave full length and cut the material between to leave a 1/2" tongue, using a coping saw and sharp chisel.  Then I marked out the location of the 1-1/2" mortises from the tenons on the top and made those with a 3/8" auger bit followed by some chisel work.  The depth gauge in the lower left of the photo is very handy when checking mortises.  That one is a General 444 that I found at a garage sale for a couple bucks.  I usually set it to my tenon length plus 1/32" or so and slide it along as a go/no-go gauge.  A small high spot somewhere in the bottom of a mortise can be a real hassle.

Test fitting and tuning the tenons with a rabbet block plane got the ends in place.  I tend to make my mortises, cut my tenons a hair fat, and tune the joint by adjusting them if necessary.  I'm going to radius the ends, but not until final assembly.

A couple of suggestions:  take a small piece of scrap and put the same dado as in the end piece.  You can see the one I made from a chunk of 2x6 in one of the photos.  Then you can go along and test that dado against all the individual parts of the top until you get the same fit everywhere.  Trying to test fit the whole thing at once doesn't really work because you can't tell where it's tight for sure.  Leaving the stock over length is very very handy when it comes time to knock the end off during final fitting - otherwise there's no place to apply the mallet to remove it and you're kind of stuck.

How to use Dividers

Dividers were a mystery to me for a while, but they are a low-tech, elegant way to accomplish very accurate work.  I love stuff like that.  Here are some different scenarios.

The fundamental task is to divide a length into an arbitrary number of sub-lengths.  Let's say you have a board about a foot wide that you want to make nine evenly spaced holes in for a coat rack or game or something.  Draw a centerline, and set the dividers for about 1/10th of the distance across the board.  Just eyeball it if you want.  Then start at one end of the line and walk the dividers along.  When you get to the far end, whatever error you made in the setting will mean over or undershooting the end of the board.  Say you're about your thumb's width over.  Narrow the dividers' setting by about 1/10 of your thumb's width, and step off the line again.  It might take two or three tries, but it's pretty easy to get extremely close to perfect, at which point you step down the line again, but this time press as you go to make a clear pinhole mark at each center point.  Note that at no point in this do you need to measure anything.  The line you're dividing can be any length, and at any angle, although it does need to be a straight line.

Once you've done it a few times, adjusting the dividers to close in on your desired value isn't fiddly.  It is trial and error, but you can home in on a very accurate measurement within three or four tries.  So don't get stuck in the trap that dividers are some kind of old-school imprecision thing.  Let go of your fear, Luke

A common use comes up in laying out dovetails.  This can be used for pins or tails (whatever you do first) but here's my tails-first approach.  Make a mark a half-pin width in from each end of the board  Set you dividers to about what you want a pin plus a tail width to be, and step off down the joint line from one of the half pins.  Do a few trials until the last step goes off the board and lands where the far half-pin would end if it were a full pin.  It doesn't need to be super exact.  Then step off that distance from both ends, pushing in the points to mark out the joint.

Another use is finding the center between two points on a line.  This is the simplest case of division (into two).  Set the dividers for what looks like the center by eye.  Take one step from either end.  The mid-point of those two new marks is the center.  Widen or narrow the divider setting by half the distance between those points (by eye) and try again.  When the points land in the same spot, you have the center.

Dining Table Design and Top

Well, I got the top glued up, finally.  I sliced the base of my thumb up pretty good while disassembling a handplane, so there was a week of nothing accomplished while that healed.  This is by far the biggest panel glue-up I've ever attempted.  Ten pieces were jointed by hand and glued up one at a time into a 7' x 3' panel.  I used four Rockler 3/4" pipe clamps (with cauls) and a biscuit every 12" to help with alignment.  The biscuits probably could have been further apart but I was cautious.  One mistake I made was gluing the two halves together separately (into 7' x 18" sections) and then trying to glue that joint down the middle last of all.  Hoisting one of the 50 lb halves up to test the joint while I planed it to match was no fun.

Top all glued up. This is the underside - the top has slightly fewer defects.

My wife thinks it looks good, but I think it looks like a bunch of 2x4's made into some kind of fancy picnic table.  I'm planning on inlaying some dutchmen to hide a couple of knots and other ugly bits, and it will have breadboard ends, so I might be happier with it in that context.

There's around 3/16" of cup across one end, and 1/8" at the other.  Cumulative errors of all the jointed surfaces show that I made a systematic error while jointing, since the whole top curves the same way.  Or maybe it was the way I clamped it during glue up.  I can flex it flat pretty easily with a couple of clamps and a piece of lumber, and I know a top this size will move some on it's own even if I planed it perfectly flat now in it's unrestrained state.  I'm not entirely sure how to proceed though.  How flat does it need to be before I put the breadboard ends on and screw stringers underneath to get it the rest of the way?

Here's the plan.  I'm going to plane the bottom side with my #4 smoother and jack.  That won't straighten it fully, just level the joints and smooth the surface.  Then I'll install the breadboard ends and attach the stringers with figure-eight connectors, as it will be when the table is assembled.  Thus, all the physical flattening devices will be on, and I can use my long planes to get the top properly flat so it looks good.  As long as I don't feel like things are super stressed out when I assemble it, I think this will be ok.  And like I keep telling myself - worst case, I blew $40 worth of framing lumber.

The rest of the table plan is still a little fluid, but will look something like the Sketchup rendering below.  I picked up a 5/4 cherry board to make the breadboard ends and butterfly patches for a couple of spots on top, and I'm going to use walnut for a few small bits like the breadboard pins and the wedge that will hold the trestle together.

A fairly standard trestle table design

El Cheapo Dining Table

I'm making us a new dining table, out of Douglas Fir.  I started with 2 x 10 framing lumber from Home Depot...  I found a couple in my outdoor wood pile that had been temporary supports during one of our remodeling projects.  After two years outside and a trip through the planer they looked too nice to cut up for the fascia boards I had intended them for.  So I went to the big orange store and picked out some more decent pieces of green, soaking wet framing lumber, and let them dry outside for awhile, then brought them to the basement for a couple of months.  Then I ran them all through the planer, bringing them down to 1-1/8" thick.  Today I ripped the pithy centers out of them, leaving strips of quartersawn grain ranging from 3" to 5" wide.  Now I get to joint all the edges in preparation for glue-up.  I don't have a powered jointer, so I'll use the tablesaw to even up the worst of it and then fine tune with my new Veritas jointer plane.

Ten planks ready for jointing

Design?  I don't have a design yet.  Well, it will be a trestle table, and the top will be about 7' x 3'.

Panoramic Head - prototype

A friend recently sent me a link to a panoramic photo of Mars, stitched together from a bunch of images from the Curiosity rover that landed last week.  It was as close to I'm likely to get to the experience of really seeing the place, and that got me interested in making panoramic images of my own, of cool places that I have actually been to.

There is free software to stitch images together, but it works a little better if you take all the photos by pivoting the camera around a particular point in space, usually called the nodal point or entrance pupil.  That point is generally near the center of the camera's lens, so a regular tripod doesn't quite cut it because it rotates around the tripod socket under the camera body.  It's also typical to take photos with the camera in portrait orientation (sideways) to capture the vertical dimension as widely as possible.

Enter the panoramic head.  Yeah, sure, you can buy them, but that's not how I roll.  I made a bunch of sketches and here's the first iteration of a working design.  The goal is to have two axes of rotation that intersect at the nodal point of the lens.  This prototype is nothing fancy, just some scraps of plywood and some nuts and bolts from the hardware store.  I tacked it together last night and took a quick set of pano shots of our living room.  They stitched together nearly perfectly, much better than a similar set of shots I took with the tripod alone.  So yay.  Not so yay are some of the details.  The base plate needs to be bigger, and the elevation arm can't swing down to 90˚ to take a shot straight up because of the way the camera is attached to it.  I also didn't take time to rout the groove in the elevation arm that will allow adjustment for different nodal point positions.

Canon 50D with 10-22 zoom on my plywood panoramic head.
The idea is to allow pivoting about the nodal point (white) in two axes: azimuth in red, and elevation in blue.  The green rotational axis is fixed.

Round two of the design will be a lot better.