Genesis Coffee Lab Counter

A couple of years ago, Caitlin discovered a new coffee shop near her house: Genesis Coffee Lab. It was founded and run by a family that had been missionaries in Ethiopia. While there, they became friends with some coffee growers, and “the rest is history”. Caitlin liked their coffee and recommended it to Sue and I. We visited, met the owners, (Jeff and Moriah), and also liked their product. As their business grew, Jeff and Moriah needed a larger place, and contracted a fellow to build a new facility for them. Turned out, the contractor they hired was an old friend of ours, Chuck Morris.

In passing, Jeff had seen some of my woodworking and liked it. He wanted a special counter for their new business. One that would really wow people when they came in, and also have ties to the locale. He told me he wanted it made from local birch, and we discussed design and construction. He asked me if I would build it for him, adding, “I’d be happy to pay you”. I told him that given the design and construction he was interested in, he couldn’t afford to pay me, (or anyone else for that matter), to have it built. It would be at least $10,000. However, I told him that I would do it for “free”, if he bought the materials, which I estimated to be between $1500 and $2000. He agreed. In exchange, I would get free coffee “for life”. 😀 The free coffee was certainly nice, but I would have done it anyway, and told him so. I’m retired; I like woodworking; I liked the idea; and I wanted to see his business succeed. I got started in March of ’21.

The counter was to be made of local birch, 20′ long, 30″ wide, with two returns back to the rear wall, each 8′ long. In one of the returns, there was to be a lift-up door accessing the area behind the counter. The skirt would be tongue-and-groove birch. We discussed adding the business logo carved into the skirt and epoxy-filled, but Jeff decided against that. (I didn’t like that decision. 😉 I think he thought it was too much trouble.)

I knew a local sawyer from whom I had purchased rough-milled lumber before. I asked him if he could supply the needed lumber. He told me that he couldn’t get boards longer than 16′ because the local birch simply wasn’t long enough. I told him that 16-footers would do. I needed seven of them, (six plus a spare), 1.5″ thick, 5.5″ wide, and 16′ long. That would take care of the the long part of the counter top. Fourteen boards 1.5″ thick, 5.5″ wide, and eight feet long would take care of the two 8′ returns. I needed 38 boards that were 1.25″ thick, 5.5″ wide, and longer than 6 feet for the skirt. (Each 6’+ board would be cut into two pieces about 36″ each.) This amounted to approximately 210 board feet of lumber. At $7 per board foot, that came to $1500. (A little more than I expected, but still a fair price.) I asked the sawyer to cut the logs into flitches, (boards kept in the order they were cut from the log) so that I could book-match them later when I assembled the top and skirt. This is no small matter, and part of the reason why the price to have it commercially manufactured would have been so high.

The sawyer did an excellent job, and in a couple of weeks he called me to tell me it was all milled and ready for me to pick up. Being the summer in Alaska, I was able to set all of that lumber out in my driveway while I made room for it in my shop. Unfortunately, there really wasn’t any room in my shop for the 16-footers. Fortunately, there was room in the current Genesis warehouse for them, and I took them over there to dry. The rule of thumb for drying freshly cut lumber outside is “one year per inch of thickness”. You can accelerate that process by drying the wood in a kiln, but the sawyer didn’t have a kiln at that time. (He has since built one.) This lumber would have to dry for at least a year before it could be expected to remain stable after construction. Fortunately, getting the new facility built was taking longer than expected. Furthermore, keeping the wood inside (my shop for the 6 and 8-footers, and the Genesis warehouse for the 16-footers), accelerated the drying process some, and by the time the new facility was “dried in”, the wood was ready to be milled to size.

I moved a table saw, my planer, and all the lumber into the new facility and started preliminary milling all the boards to thickness. These are the 16 and 8-foot boards for the top.
1

2

OY! All of them had warped significantly! However, that was why I had ordered them “extra” thick. It was a good thing I did. Even after milling to final thickness, several were still very warped. I would have to “beat them into submission”.
3

Planing to thickness generated 20+, 50-gallon bags of shavings, and took 10 hours to complete!
4

5

6

Just a couple of hours of planing.
7

8

Once planed to thickness, (not final thickness, but “working” thickness), I had to square the edges. This was accomplished using a power hand-planer first.
9

Followed by “jack” and “jointer” hand-planes.
10

11

12

This just squared the boards four sides. It did NOT “joint” the edges for joining. That would come later when I assembled the top. Never again will I take on hand joining 16′ boards.

Next came the skirt boards. Before they were cut to length, (34″), each board had to be planed to 3/4″ thickness – the proper thickness for cutting the tongue-and-groove profiles in the edges.

Here they all are after planing to thickness and grading.
13

Once the skirt boards were planed to final thickness, (no need to edge them, as that would happen when the edges were profiled), they were returned to my shop where they were cut to final length, 34″.
14

From there, back to the warehouse to be set up for matching.
15

The “white” boards you see in the above picture are the center of the front of the skirt. If you look closely, you can see that the boards match in pattern from the center out.
16

17

Once matched left and right, each board was numbered and its side noted, starting at the center and moving toward the end. For example, “7R” meant the 7th board on the right side. That’s the tape you see on each board. Once that was finished, it was back to my shop to shape the edges – cut the tongues and grooves.

In process. (The bottom ends of the boards.)
18

All of the pieces for the front of the skirt duly annotated and the face noted with the arrow direction.
19

The “o” marked the skirt boards for the returns, and the “-” were the spares. There was no “ordering” of the return skirt pieces.
20

You’ll note in the above numbering that there are no “1L”, “2L”, “1R”, or “2R” boards. You’ll also note that the four center boards are completely “white”; no brown heartwood. Recall that I wrote that I didn’t like the decision to eliminate putting the Genesis Labs logo on the skirt. I found a fellow that, for a very reasonable price, would rout the logo into the wood. I would then fill the recess with black epoxy. So… That’s why boards 1R&L and 2R&L were separated out from the rest.
Here they have been jointed, and are being glued up to create the center panel. You can see the numbers on the blue tape.
21

There was a bit of trepidation on my part in “doing the epoxy”, as I had never done that before. I had some scrap pieces of planking that were the right size, so I had the fellow doing the CNC, (Computer Numerically Controlled), routing do a practice piece. He could get the cutting details worked out and I could get the epoxying method worked out, without jeopardizing the actual IRREPLACABLE, (I thought), pieces. The cutting went great.

One of the issues with filling wooden cavities with colored epoxy, is that the colorant ‘bleeds’ into the grain of the wood. I knew about this from reading about doing the process, and decided to try sealing the grain with shellac. Shellac is a great grain sealant for staining wood, so I thought it would be for this process too. Nope. Didn’t work at all.
22

You can see how the black dye bled into the wood grain. No sanding that out. It runs as deep as the cavity is. Whew! REALLY glad I tried a practice piece.
23

Here’s the starting point. I was being stingy with the epoxy because it was $150 per gallon!
24

The dye was “India” ink.
25

I formed a little hot glue dam to constrain the epoxy to a small test area.
26

The tiny little bubbles that form as the epoxy starts to hardens.
27

Blowing the surface with a heat gun or a torch, (I tried both), bursts all the little bubbles leaving a relatively smooth surface.
28

After fiddling with the various elements of the process, I was sufficiently confident to work on the “real” piece. Oh yeah… Since the shellac didn’t work as a sealant, I went with what the manufacturer suggested: heat the epoxy AND the wood prior to application, and apply a thin CLEAR coat of epoxy to seal the grain before adding the colored epoxy.

I got the birch panel back from the “CNC Guy” in a few days. It was even better than the practice piece. (In part because the practice panel was pine, a softwood. The birch, a hardwood, cuts cleaner.)
Here’s how that piece turned out.
29

After sealing with the clear coat.
30

31

32

33

34

As you can see, I added some eyes. Those are glass, “competition grade”, African lion taxidermy eyes. I thought they really ‘livened’ the logo. However, after looking at both versions – with and without eyes – Jeff and Moriah decided they didn’t want the eyes. Not a problem. I’d simply refill the eye sockets with epoxy. But it was a problem. When I drilled the holes for the eyes, the drill bit cut new openings in the grain of the wood. The dye bled into the grain. Unusable. Jeff and I discussed the options. Too late to even go back to the eye version. One option was to turn the panel over and recut the logo on the other side. The “issue” was that the panel wasn’t all “white” on the opposite side. There was some heartwood on the pieces. The second option was to start anew with four new pieces. I chose to do both, and see which one looked best. As it turned out, the opposite side, with the small amount of heartwood, actually looked better to both Jeff and I than the “all white” side did. Here’s the final version.
35

36

37

38

We like this better because it doesn’t look “inserted” into the skirt. In other words, the “all white” panel looked “square” and “inserted”. This side has mostly sapwood, but transitions less abruptly to the adjacent pieces because of the small amount of heartwood on the outer two boards (2L and 2R).

While the logo panel was being dealt with, the rest of the counter was being constructed. Jeff had a “welder friend”, Justin, that would fabricate the frame for the counter. The main structural components would be made of 2×2 square, steel tubing. To those would be 1.5″ steel angle-iron. The skirt would be screwed to the angleiron. The top would have a substrate of 3/4″ MDF (medium density fiberboard), which would be bolted to the 2×2 square tubing frame. The birch boards, once glued together into panels, would be glued and screwed to the MDF substrate.

Justin welding up the frame.
39

40

The counter prior to the MDF being attached.
41

42

43

44

The MDF pieces were joined with “biscuits” (https://www.woodmagazine.com/woodworking-tips/techniques/joinery/biscuit-basics?mode=step_by_step).

Next came the real challenge of fabricating this counter: Jointing the 16-foot long boards to create the top itself. That was a bear. I will never again hand-joint 16′ boards! I wouldn’t have this time, but I really had no choice, and as my friend Alain would say, “You made a rod for your own back.” I should note, that I discussed shortening the original 20′ length of the counter to 16′. My rationale was that there was no need for 16′ boards if they had to have joints where the extra 4′ would be added, thereby breaking up the “book-matched” nature of the top anyway. In other words, if he really wanted the top to be book-matched, (all the effort to get the boards cut into flitches, etc.), then adding the 4′ extensions was out. Jeff agreed.

The process started as you saw in the pictures above of using the power hand-planer to square the faces. (Pictures 9,11, & 12) Unfortunately, that doesn’t remove “crook” or “twist”. (The 18″ floor planer that made all the shavings took care of the “cup”.) “Twist” would have to be dealt with by brute force, but “crook” had to go away before the boards could be joined. I purchased a 16′ piece of 1.5″ angle-aluminum. (You can see it on the floor in front of the counter in pictures 41 and 42 above.) That piece of angle-aluminum would provide the straight edge against which I could run a trim bit in a router. That would get me ‘close’, but the final fitting had to be done with hand planing.

The older I get, and the more skilled I get with hand tools, the less I like to use power tools. Reality is, regardless of skill, if you screw up with a power tool, (your fault or the machine’s), the odds are very high that the piece will be ruined. If you screw up with a hand tool, the odds are very high that you will recognize the error long before any permanent damage is done, and very low that the piece will be ruined. Here’s what can happen VERY FAST with a power tool:
45

46

Lotsa sailor talk around that event… Which indicated that it was time to quit and have some home-brew.
47

Here’s what the joints looked like before hand jointing the boards.
48

49

50

That’s the level of crook I was facing.

I needed some straight “standards” that I could use as a guide. I purchased two 16’x6″ pieces of painted, laminated, ‘trim’. You can see them in pictures 48,49,& 50. Here’s how they were used.
51

Jointing two boards that are adjacent in the top, ensures that minor errors in holding the plane perpendicular to the face of the boards is compensated on the “other” board. In other words, whatever angular error is present on one board due to tilting the plane ever-so-slightly, is compensated by the complementing angle on the other board because they are cut simultaneously. Jointing the edges pairs-at-a-time ensures a tight joint. Of course you need a plane with a blade wide enough to span the combined thicknesses of the two boards.

It is always best to plane “downhill”, or with the grain ends flowing away from the cut, as opposed to the blade cutting “into” the grain (AKA “uphill”). Every time a knot is encountered, the grain changes direction. Because of the frequent knots in the birch, the grain changed directions frequently.
51

52

So, if we number the boards 1 to 6, from the “outside” edge of the counter to the “inside” edge of the counter, I started by jointing boards 1 and 2 at their common edges. Next, I joint boards 2 and 3 at their common edges. Then 3 & 4, then 4 & 5, then 5 & 6. So while there are in reality only 3 pairs of boards, there are 5 pairs-of-edges to joint. Each 16 feet long. That’s just the front of the counter. There were also the two returns.

Look at the twist in those boards!
53

And this is what is necessary to “beat them into submission”.
54

55

56

These are for the 8-foot return.
57

58

59

Once they were jointed and glued together, the miters where the returns and the front join, had to be cut. There was no way to fit those pieces, 5 feet long, 8 feet long, and 16 feet long, onto any power saw I owned, so they were cut by hand. In the following picture you can see the 8′ return up against the wall on the left, the 16′ front on the floor on the left, and the 5′ return with uncut mitre just peaking out above the substrate on the counter to the right of center.
60

The rough, first cut with the circular saw. Notice that even with “beating into submission” during glue-up, a significant bow remained. More beating. (Bolted from below through the MDF substrate, and that’s not the end of the ‘beating’. The whole top has to be flattened. More on that later.)
61

Finishing with the Dozuki.
62

After doing one preliminary circular saw cut and finishing with a hand saw, I just decided to start with the hand saw.
63

Once the mitres were cut, the returns were joined to the front.
64

65

66

67

These mitres, and their resulting joints, were further complicated by the fact that the returns were 24″ wide, and the front was 30″ wide. That meant that 1) The angle wasn’t 45 degrees, and 2) that the inside edges of the front and return wouldn’t meet at the inside corner of the joint. Not only did that complicate cutting the mitres, it complicated getting the edging, and the insert in the joint correct. More on that later. (Note groove routed along the mitre joint to accept the contrasting wood insert.)
66

There was of course no way those hand-cut mitres were going to be “like it grew there”. Planing the end-grain on a “weird” angle and matching two joints, (one for each end of the front), was out of the question. In order to dress up the joint I routed a groove 3/8″ wide by 1/4″ deep, along the joint. This groove would accept an insert of contrasting wood.

The next task was flattening the entire top. Again, since there were no planers available to me that could handle 24″-wide boards, let alone 30″-wide ones, the flattening would have to be done “the hard way”, and the “really hard way.” The hard way was building a cradle and rail system for a router and routing the entire 30 feet of counter top. That worked fine right up to the inside corners. There was no way to “go around the corner” with the cradle. That meant that I had to flatten the corners (24″x30″ square), with a hand plane. This was the “really hard way”.

Here are some pictures of the process of rough flattening the counter top. (The final flattening would be accomplished using a random orbital sander.)
This is a little “roadside repair” of the power switch on the router. Grrr…
68

69

70

Here, 71 & 72, you can see the amount of wood being removed on the middle of the front.
71

72

Here is how thick it was on one of the corners. That had to be removed with a hand plane.
73

74

What it looked like after routing. Clearly, this is not “finished”, but it is “flat”.
75

76

77

78

Sanding to finished “flat” came next. In the below picture, you can see that the machine marks from the router have been sanded out. What you are seeing is the final sanding. Because the grit of the sandpaper is so fine (320), pencil marks are put on the surface so you can see where you have sanded. Also, you’ll note that at this stage, the grooves have not yet been routed in the mitres.
79

Once the whole surface was finish sanded, it was time for the edging. After a couple of stutter steps, it was decided because this was a working, commercial counter, a more substantial edging was called for. I decided to inlet angled aluminum. In the final analysis, I think it looks great!
80

In the following pictures, the first three preliminary coats of the finish – Wipe-on Polyurethane – have been applied.
81

82

83

84

85

86

87

There is no “final product” picture yet. I applied all the polyurethane in the evenings, so that; 1) other activities in the building would have finished for the day, 2) there wouldn’t be any more dust generated, and 3) the finish would have over night to dry undisturbed. For some strange reason, unlike other nights, the final coat got A LOT of significant dust particles on it. While these would wear away quickly due to the commercial use, its current condition is unacceptable to me. I have purchased a buffer and some buffing compound, and when they arrive, I’ll buff all of those ‘nibs’ out. When that is completed, I’ll post “final” pictures.