10/21-23/08 I went up to the property for 3 days: Tuesday through Thursday.

There was snow on the ground, although not on the roadway, above 3300 feet going over the pass. Looks like it could be an early winter. I arrived at 12:30 and was greeted by Bert and Ernie as usual. I was disappointed to find that the peanut was gone and both mousetraps upstairs were sprung but empty. I searched for a hole but didn't find any. Here we go again.

I had hurt my back and a couple fingers on my left hand wrestling with Qdog. My back hurt enough that I toyed with the idea of skipping going up to the property this week but I decided to go anyway. If I couldn't work, I could rest up there as well as at home.

I started out by building a fire in the wood stove. It was chilly outside and it made it nice and cozy inside with that fire. I set up two big sawhorses on the porch to use to fabricate the ends of the long roof panels. I had already sort of figured out how to do it on the short panels I did last week. Now I was a little better at it. My back didn't hurt too much while I worked, so I had the first panel bent in short order.

Since I was so eager to get the roof fixed, and the weather was so nice, I decided to try to get the first long panel up on the roof. I rigged up a rope and pulley and then lifted the panel up onto the roof. Then I used the rope and pulley to drag the panel up to the ridge over where it needed to go. Installing this panel was exactly like installing the original panels except that it was only 18 feet long instead of 24 or 40 feet like the originals.

I found that working on the roof did not aggravate my back pain at all. In fact, it seemed like the pain went away altogether while I was up there. I think part of that was because when I walked or stood on the ground, I had to hunch over in order to minimize the pain. When I am on the roof, I have to walk that same hunched-over way just to keep from slipping. Another reason might be that when I am on the roof I am so preoccupied thinking about balance, safety, and the job at hand, that I don't have time to think about anything else, like back pain.

Anyway, I snapped that panel down in place in just minutes without much problem. I did slip and fall once, but my rope harness kept me from going very far. One of my knuckles hit something as I fell and started bleeding, but aside from a little blood on my shirt and pants, it wasn't all that bad. At least I was glad I had one of the last three panels in place on the roof.

Charlie the chipmunk came around for peanuts a couple times before I went in for the night.

On Wednesday morning the temperature was 32 degrees when I went out to work. I went up on the roof and screwed down the clips that hold the panel I had put up the day before. Then I fabricated the end of the second long panel and then installed it on the roof with its clips. That left only one panel to go, but the last one was not going to be so easy. Each successive panel overlaps the previous one, but this last one would have to be tucked under a panel that was already installed. I had rehearsed how I was going to do it many times in my head, but now I had to make reality match theory.

I fabricated the end of the third panel on the porch. The noise brought the family of gray jays over to get their peanuts. When I had the panel bent, I went up on the roof to lift up the existing panel which the new one had to go under. Whether I could do this or not would determine the difference between success and failure. I was sure I could do it in theory, but now came the acid test.

I took my tools up on the roof and after a little doing, I found that I could lift the panel up about a half inch which was about what I needed in order to get the new one tucked under. I put a little block of wood under the panel at the very top and another one below where the new panel would end and these blocks held the panel up its entire length. I was super happy to have been able to do that, and went in for the night feeling pretty confident.

On Thursday morning there was an intermittent breeze. It is dangerous, if not impossible, to work with long panels on a roof in the wind. But the roof is pretty well protected, so that even though there is a strong wind blowing the tops of the nearby trees, it is a lot calmer on the roof. Nevertheless, I waited until I thought the wind was going to stop for a while before I started hoisting and installing that last panel.

Since installing a replacement panel is not a common thing to do, (Curt had told me that he had never done it) and since someone else may have to do it someday, I decided to document the process I used, right here in this journal, so that it can be referred to later. I'll describe the entire process beginning with the fabrication of the panel:

How to replace the top part of a panel in the middle of the roof

If more than one panel needs to be replaced, all but the last one are installed the same way they are originally installed, so there is nothing special about them. If only a part of a panel needs to be replaced, then the panels must be installed in sequence from the bottom to the top so that they overlap in a way that will shed rain. The top one is the special case. The lower ones are installed overlapping the ones below, but the top one also has to go under the projecting ridge metal, which makes it a little trickier. So I'll describe the installation of one part-panel at the top of the roof with existing panels on both sides.

First, some terminology. The panels are 16 inches wide and run straight down the roof from the ridge to the eave. Each panel has a standing rib on the left side (of course they could be installed the other way) and a raw edge on the right side. When installed, the raw edge is held fast to the roof by clips that are screwed down to the roof and that hook over the top of the raw edge. The standing rib is shaped like an upside-down 'V' and it snaps over the top of the clips and the raw edge of the panel next to it.

The raw edge is about an inch of the edge of the panel that is bent up at 90 degrees. It is not flat, but has a couple creases in it that engage the standing rib when it snaps over it.

Two panels, being the same size, cannot overlap nicely, because the top one won't fit inside the lower one. In order to fabricate the end of the panel so that the top one will fit inside, either the lower one would have to be widened, or the top one has to be narrowed. Since the bottom one is already on the roof, and the top one is the one to be installed, it is easier to narrow the top one because the work can be done on the ground instead of up on the roof. That was part of the wise advice I got from Curt Pritchard.

To narrow the end of the panel, I bent the side with the raw edge rather than the standing rib. First, I flattened the last 9 inches of the raw edge (the panel was going to overlap 8 inches so I added an inch for comfort) by clamping it between two steel bars using vise-grips. The bars were 8 inches long (they should have been 9), a quarter of an inch thick, and an inch and a half wide. I used two vise-grips to squeeze the creases out of the raw edge by alternately tightening up on the vise-grips and working them up from the end of the panel until the metal was flat. Now the last 9 inches of the raw edge was still sticking up at 90 degrees, but now was flat and it stuck up a little higher because the creases had been taken out.

Next, I put one of the steel bars flat on the panel, near the end of the raw edge but with the edge of the bar in from the raw edge about a quarter of an inch. Then I put the second bar under the panel and directly under the first bar and clamped the ends of the two bars together with a vise-grip that went over the end of the panel. The vise-grip jaws could reach in about an inch to grip the bars.

On the other end of the bars, I used a big C-clamp that could reach over the raw edge and clamp the other ends of the bars together tightly. I turned the C-clamp as far as it would go up against the long end of the panel so it would be out of the way.

Using a small hammer, I tapped the corner of the panel, where it bent up to form the raw edge, from underneath, working the hammer across the 9-inch span back and forth. The hammer blows bent the metal up to essentially move the bend in a quarter of an inch. The new bend was formed around the edge of the top metal bar, and as that bend was formed, the hammer gradually took out the original 90 degree bend and flattened it. When the clamps and bars were removed, I used a pliers to straighten up the newly configured raw edge so that it was nice and straight.

That part of the raw edge was now nearly a half inch too high, so I used a tin snips to cut a half inch off trying to make as straight a line as possible that went in to the un-bent part of the original raw edge. The entire raw edge of the panel, including the 8 or 9 inches that I had modified, needs to get tucked under the existing panel, so the straighter it is the better.

Next, I turned the panel over on the sawhorses so that the standing rib and the raw edge were underneath and the underside of the panel was on top. Then, with the tin snips, I cut a slit down the top of the outer edge of the standing rib for 9 inches, and then cut loose the metal strip. This removed about a quarter inch strip that originally went inside the standing rib and engages the raw edge when the panel is snapped in place. Since the bottom 9 inches of the standing rib will overlap another standing rib, it won't fit over unless that strip of metal is removed.

Finally, with the panel still in the upside-down position, and with the broad top of the saw horse directly under the end of the standing rib, I placed one of the steel bars inside the last 9 inches of the standing rib, and beat down on it with a hammer. This action widened the top of the standing rib, and opened it up. Then, holding the bar down tightly, I squeezed the standing rib back together again. What this did was to widen the top of the standing rib so that when it goes over the top of the standing rib under it, it will fit better and bottom out. With that done, the panel was fabricated and ready to install on the roof.

Before describing the installation of the panel, it will help to describe how the panels fit into the ridge. At the very top of the roof, there is a Ridge Cap which looks in cross section like a very shallow upside-down 'V', each side of the 'V' being about 6 inches wide.. The Ridge Cap runs the entire length of the roof ridge. Since the panels have standing ribs of an inch and a half sticking up, the Ridge Cap necessarily must be up off the roof an inch and a half so that the panels can go under it. To hold the Ridge Cap up, and to tie the whole thing together, there are two C-channels running the length of the ridge, one on either side. In cross section, the C-channel looks like the letter 'C' except that it is made of three straight lines instead of a curved line. The open side of the 'C' faces down the roof. The bottom flange is flat on the roof, the 1 1/2" side sticks up at 90 degrees, and the top flange goes over the top of the panels and it fastens to the ridge cap.

So to install a panel, the top of it needs to be squeezed into the C-channel, the hard part, of course, is the standing rib which just barely fits into the C-channel. The rest of the panel has plenty of clearance.

So, with that out of the way, the installation of the panel is described next.

Before bringing the panel up on the roof, the standing rib of the existing panel on the roof needed to be lifted up in order to get the raw edge of the new panel under it. I had done that the previous day, but for continuity of this description, I'll describe how I did it in more detail here.

I used a small nail-puller (or maybe it's called a wrecking bar) that is a stiff steel strap about 10 inches long with a right angle bend of about an inch at one end. Both ends have a nail-pulling type slot in them. The tool is normally used by driving one end or the other of the tool under the head of a nail using a hammer, and then either by hand, or again using the hammer, using the leverage of the tool, pulling the nail out. My plan was to use a hammer, if need be, to drive the bent end of the tool under the standing rib of the panel and lift it up off the roof.

Since there was no raw edge under that standing rib, there shouldn't be much holding it down. The only thing holding it down would be the clips, but they were spaced a few feet apart, and they didn't engage the standing rib directly. I started at the very top of the panel, up under the C-chanel, and found that with the nail-puller, I could easily get the tool under the standing rib and lift up on it. As I lifted, the standing rib ran into the top of the C-channel and kept me from lifting it more than about a quarter of an inch. By forcing the tool, and bending the C-channel up a little, I got it lifted about a half inch and was able to stick a little half-inch block of wood under it to hold it there.

I moved down the panel and lifted it up in several places and it came up easily, and even was held up off the roof a little by that block of wood at the top.

I went down to the bottom of the space for the new panel to lift it there. From there, the panel continued down the roof for the entire 40 foot length and also from there, the standing rib was engaged in the raw edge of the panel directly below the one to be installed. This made it a lot harder to lift the edge.

I had to use the hammer to drive the nail puller under the standing rib, and for a while I was unsuccessful. By trying a different spot, I realized that I had been trying to lift it right at a clip. By trying a foot or so away from the clip, I was able to force the tool under the rib and by prying up on it with considerable force, got it to pop up and disengage from the raw edge of the panel underneath. I forced a 3/4" piece of wood under the standing rib which then held it up. Looking at the panel, I could see that the standing rib was up off the tarpaper for its entire length. This would allow me to slip the new panel underneath. The next step was to bring the panel up onto the roof.

I had two big anchor ropes that went over the ridge and over the entire building. The ropes were tied to trees on the side of the small roof to anchor them. Since I didn't need to get on the small roof, the other ends of the ropes didn't need to be tied off. I used one of the ropes I'll call the safety rope to anchor me, and the other, the anchor rope, to anchor a tool bucket, the panels, and whatever else I needed to hold.

I attached a pulley to the anchor rope up at the ridge and ran a long pulling rope through the pulley so that both ends of the pulling rope reached the ladder. I used this to pull and hold the panel up on the roof.

To grip the panel, I used a fabric strap with a buckle and cinched the strap around the panel a foot or so above the middle. I put a chunk of cardboard between the strap and the raw edge of the panel to protect the strap.

I stood the panel up on end up against the roof next to the ladder. Then, I rigged myself up and climbed up. I tied one end of the pulling rope to the end of the strap and then lifted the panel up high enough so that it lay down flat on the roof. Holding it in place with one hand, I used the other hand to take up the slack from the other end of the pulling rope. When the rope was taut, I tied it to the ladder. The pulling rope then kept the panel from slipping back off the roof.

Standing on the ladder at the edge of the roof, I fastened my harness to the safety rope. I use a form of a tautline hitch to tie my harness to the safety rope which allows me to slide the knot up and down the rope and yet have it snug up if I pull on it from the harness. Once I was hitched up, I got up onto the roof and slid my knot up the safety rope so I could lean back on my harness and free my hands.

Then, by pulling on the pulling rope going up to the pulley with one hand, I could slide the panel up the roof. I used the other hand to lift the end of the panel as it went so that it wouldn't scrape the roof. Then I would slide the knot connecting me to my safety rope so I could move up the roof and get into position to pull the panel up further. In this manner, I pulled the panel all the way up to the ridge.

I slid the bottom of the panel over so that the panel was situated right over the tarpaper space where it was supposed to go. Then I fastened a C-clamp to the top of the panel about three inches from the raw edge and used a small tether rope to tie the C-clamp to the anchor rope with a tautline hitch. Then I released the rope in the pulley and the panel slid down a few inches until the tether and the C-clamp took over the job of holding the panel up. When I was sure this connection was secure, I untied the rope from the strap and removed the strap from around the panel and let the strap and the cardboard slide down off the roof.

The next job, which is the key to the whole thing, was to get the raw edge of the new panel tucked under the standing rib of the existing panel on the right that was now up off the tarpaper a little bit. To do this, the new panel needs to stand on edge at 90 degrees so that the raw edge is lying flat on the roof. I lifted the panel up in this position, and while holding it with one hand, I used the other hand to push the raw edge under the standing rib wherever I could. The idea was to get that raw edge under the standing rib for the entire length of the panel and then gradually tip it back down so it would lie flat on the roof while the raw edge worked its way up under the standing rib. Not as easy as it sounds.

I had made several lengths of 1x2 about 16 inches long to hold the panel in place by wedging the stick between the panel and the clips or panel on the other side. The idea was to have these sticks hold parts of the panel in place that I successfully got tucked underneath while I worked to tuck more sections under. It worked pretty well, starting at the top and working down, but when I got to the bottom where I had modified the raw edge, it got real tricky. Since the edge was no longer straight, but had that quarter inch jog in it, I had trouble getting it forced under the standing rib. When I did, then part of the upper portion would pop back out, and most of my sticks slid off the roof.

After fighting with the panel for quite a while, I realized that I needed some sort of tool to pull the bottom of the standing rib out so the raw edge could go under. I got down off the roof and made a tool. I used the same piece of steel that I used for the tool to spread open the kerf on the phone wire conduit on 7/30/08. I bent a hook on the end that could grab the bottom of the standing rib and I bent a shoulder in it that would allow me to use a short 1x2 as a lever to pry the bottom of the standing rib out using the top of it as a fulcrum.

I took the tool back up on the roof and tried it out. It worked perfectly so after using it to tease the raw edge under the standing rib in places where it was getting stuck, I was able to get the entire raw edge tucked in. Once the entire raw edge was underneath, I started pushing the panel over to lay it down. It would only go a little way, so I put a 1x2 against the bottom of the panel and hit it with a hammer to try to force the panel uphill. This caused the raw edge to move up a little and let the panel down a little. By pushing and pounding in this way, I finally got the panel to lie down flat. What a huge relief.

The panel was almost in place, but it was about six inches too low. It couldn't be up where it was supposed to be while it was standing up because of the top of the C-channel. But now that it was flat, it could be moved up. Again using the 1x2s and a big hammer, I pounded the panel up from the bottom. I had to go up to the ridge a couple times during the process to get the panel to ride up over the bottom flange of the C-channel and again to tuck the standing rib under the top flange of the C-channel when it got up that far. After that, I simply pounded it all the way up. In the process I completely destroyed all the 1x2s I had up there and had to go back down to the ground to get a short 2x4 to finish the job.

With the panel up where it was supposed to be, it was a simple matter to use a short board and a hammer, working down the length of the panel, pounding on the standing rib until it snapped in place. Then the same process was used on the other side of the panel to snap down the standing rib that I had lifted up.

There were a few more details, like making a mess with the Vulkem I had put in the overlap of the panels. Some of this was exposed when I drove the panel up and I got my knees in it and then it got all over the tools and other parts of the roof. Finally, I screwed down the top of the panel and the job was done. All the damage done the previous winter had been repaired and I was greatly relieved. I still need to put more screws in all the rest of the panels to make sure the same thing doesn't happen again, but for now I was very happy with what I had done.

I put all the tools away, except I left the two ropes in place over the building, had lunch, and left for home at 3:15.