A second story deck is about to fall. We jacked it up, pour new concrete piers and install bracing.
Skill Level: 3~4 (Intermediate ~ Advanced)
Time Taken: About 8 hours
This article is about structural repairs on a second story deck. Structural repairs should always be performed according to local building codes by qualified professionals. Be sure to obtain the necessary permits before any project. Incorrectly performing structural repairs can lead to structure failure, property damage and possibly causing fatal injuries to people or animals.
This article is for entertainment only. Do not attempt to duplicate any of Gilles’ Outlet techniques, methods or approaches. If you decide to ignore this warning, then you do so at your own risk. In no way will the publishers of this web site be held liable for any injuries or damages, direct or consequential, incurred by any person who attempts to follow the examples shown herein.
This second story deck looked fine from far away but upon more careful inspection, there is something not quite right with its footings. The right footing does not appear to be resting flat on the ground.
While walking on the deck, I noticed that the right side on the picture was significantly lower than the left side, almost 4’’. The whole structure appeared to no longer be level and plumb.
The plan is to replace those pre-cast footings by concrete piers, poured in place. Metal connectors will be precisely located in fresh concrete and posts will be attached to them. Finally, in this area, building code requires second story decks to be braced and this deck is not. This will be a good time to install them.
There were a few plants stored on the deck. They were invited to leave.
Left: Upon closer inspection, the pre-cast concrete pier appeared to dangerously lean on one side. It is just like if it was installed directly on disturbed, non compacted soil and eventually sank. Water erosion may also have contributed to this situation.
Right: An even closer inspection reveals how dire the problem is: the concrete pier moved so much that the metal rod connecting the pier to the post is bent. Critical parts of the connector are badly corroded. The connector could break at any time and send the whole deck crashing down.
The structure needs to be supported during the pier repair. I will need to jack the deck, remove the existing post, perform the repair as the deck is jacked. This is a potentially dangerous procedure.
Left: I pried all nails attaching the left post to the 4×6 beam. This way, I will not be hammering nails when the structured is jacked, reducing chances to de-stabilize it.
Right: I estimated the load on the post to be around 1200lbs. When jacking structures, experience has taught me to spread the load on the ground: I laid a piece of scrap 2×10 close to the left post. I cut a piece of 4×4 to length and connected it to the beam with a metal connector and 10d nails. This will avoid the post from moving as the structure is jacked.
I installed an 8 ton jack under the post. It is overkill but it has a larger base and therefore is more stable.
Left: I pumped the jack. The structured started to move. I made sure I jacked up slowly, listening to the structure. Wood makes cracking noises before breaking and these would be my warning signs to immediately abort the operation.
Right: A few minutes later, all the load was transferred from the old post to the temporary support, through the jack. The old post swung free. I removed it. I was very careful to spend the least amount of time under the jacked structure.
Left: The pre-cast pier did sink in the ground about 3 in deep. This suggested that there was no proper footing under this pier. So I prepared to pour a real footing: a 12×12 x12 concrete pad will be just fine for this application in this region where the frost line is 12’’.
I marked the position of the footing using white spray paint. I started to dig and ….
Right:… I reached pay dirt so to speak. There was in fact an existing footing with the remains of a metal connector in its center. The footing was buried under about 4 in of top soil.
Well, when doing home repairs, nothing ever goes as planned. Time for a change in plan: I will be pouring a new concrete pier and connect it to the existing footing with #5 rebar.
Left: When connecting a new pier to the existing footing, it is typical to drill 1/2’’ in the concrete and drive in #5 pieces of rebar. Sometimes, rebars are also set in epoxy to further enhance the bond.
Right: I will be using Simpson Strong Tie brand epoxy for this.
The manufacturer instructions demand holes are cleaned with compressed air at 80 psi for at least 10 seconds.
Left: Then, holes need to be cleaned with a special brush 4 times. After brushing the holes, the manufacturer demands holes are cleaned one more time with compressed air.
Right: This brand of epoxy is dispensed out of a tube similar to caulking. Pushing the material out mixes the epoxy and dispenses it into the holes through a flexible hose. The manufacturer recommends verifying that the epoxy is properly mixed when dispensing. When properly mixed, it has a light grey color.
Left: I immediately drove #5 rebar in holes. It was a very tight fit, which is a good thing. Overall, this little pier ended up having three pieces of #5 rebar. Rebar was installed so they were all at least 3’’ from any edge of concrete.
Right: I cut a piece of 2ft long quick tube, diameter 8’’. This will contain the concrete for the pier and it will dry in a cylindrical shape.
Left: I need to precisely center the pier under the existing beam. I first made sure that the beam was absolutely level. I then determined the center of the post and drove a nail in the beam. This allows me to attach a plumb bob. Now, I just need to place the center of the quick tube at the location indicated by the plumb bob …
Right: … like this. It is difficult to see on this picture but this is actually the right location.
Left: Once the quick tube was precisely located, I built a set of braces with scrap lumber. This will prevent the tube from moving as the concrete is poured.
Right: I mixed about half a bag of concrete mix and started to pour in the form. The concrete needs to be wet enough but not too wet. The more wet the concrete is, the weaker it will cure.
Left: It took about a minute to fill the tube. I used a margin trowel to form a little dome at the top of the pier. This will promote drainage and lead water to the sides of the pier instead of keeping it in contact with the post. This significantly increases the life span of the wooden post.
Right: I precisely located a Simpson Strong Tie Post Base metal connector in fresh concrete. Again, I used the plumb bob to align the center of the connector and I made sure the connector was set level and plumb on all axis.
Now, the concrete needs to cure for at least 24 hours to make sure it will have enough strength to handle the load. During this period, I made sure the concrete stayed wet. This is called “wet curing”.
Left: With the beam still perfectly plumb and level, I measured the distance between the bottom of the bracket and the beam. Since I am going to use a metal bracket, I removed 1/16’’ to account for the thickness of the bracket and cut a new piece of pressure treated 4×4.
Right: I installed the new post and placed a hot dipped galvanized Simpson Strong Tie Post Cap. Notice that it is not yet fastened.
Left: I slowly lowered the jack to transfer the load back to the new post. Eventually, the jack lowered to its maximum and the structure started to bear only on the new pier / post. I checked the post and the structure for level and plumb one last time. Since everything was right on, I moved on to the next step.
Right: I fastened the post to the base connector using 16d hot dipped galvanized nails. I first installed 4 nails (two on each sides).
Left: I turned my attention to the post cap. I installed three 16d hot dipped galvanized nails on each face, according to the manufacturer instructions.
Right: I finished nailing the post to the base connector. The two big large holes are for manufacturing purposes and are not to be used to install nail or screws.
I repeated the same procedure for the other post. At this point, I let the concrete piers cure for about 8 days. Because it was very hot and dry, I made sure to spray water on both piers twice a day. I also covered the top of both piers with 6 mil plastic sheeting to reduce water evaporation.
At this point, the deck was now level and plumb. Bracing still needs to be installed.
Left: The bracing starts by cutting four pieces of pressure treated 4×4. Building codes have some precise requirements on those: they must be installed at a 45 degrees at the beam and the post. They also need to be fastened at least 18’’ away from the post top.
Cutting these braces is actually quite tricky: 4×4 posts are never really straight so those two cuts are actually compound cuts. Anyone who makes a mistake cutting these is automatically forgiven. Cut those longer, measure angles precisely and cut them step by step, removing a little bit of material at a time.
Right: I drilled a countersunk hole on each side of every brace.
Left: Braces will be fastened using 1/2’’ lag screws with washer. This hardware needs to be hot dipped galvanized because it will be exposed to elements and pressure treated wood.
Right: The lag screw and washer will sit flat, recessed in the countersunk hole. This is important for the lag screw to be effective when installed on an angle piece like this brace.
Left: I pre-drilled the beam where the new lag screw will be installed. This is necessary with large lag screws or the wood will most certainly split.
Right: I pre-drilled the post as well.
Left: The brace was installed. I fastened lag screws tight with a large 1/2’’ drive ratcheting wrench.
Right: I repeated the operation on the second side of the first post. The bracing is now completed on one side of the deck.
Left: The lag screw sits recessed in the countersunk hole. I expect this connection will look better and better as the wood weathers to a darker color.
You can see the manifestation of the compound cut: the intersection between the brace and the post tips ever so slightly down to the right side.
Right: After a few more days, I peeled away the quick tube. It came off very easily.
Left: After peeling the quick tube, the edge of the footing was slightly sharp. I used a brick hammer to round it up. In insight, I could have done this when the concrete was still wet, by rounding it over with a special masonry tool. Alas, I do not own such a tool. This worked well.
Right: I back filled around the new piers with top soil. I also compacted it with a hand compacter. This will help fight erosion.
The finished project. I do like the look of the new piers in concrete. If I did not want so much concrete to stick of the ground, I could have cast shorter piers. I could also dress the piers build building wooden boxes to hide them.
A 30 years old rusty chimney cleaning trap is removed. Rust protection is applied. A fresh coat of exterior paint is sprayed. The trap is re-installed with mortar.
Skill Level: 2~3 (Basic ~ Moderate)
Time Taken: About 2~3 hours
This 35 years old house is equipped with a wood burning fireplace. In order to facilitate cleaning, some chimneys offer a cleaning trap. In this article, we give an old rusty trap a new life.
The subject of this article. An old, rusty chimney cleaning trap. Rain has started to carry rust away from the metal and bricks under the door are stained with rust marks. The door closes but no longer latches. On high winds days, it swings free and hits either the siding of the house (left) or its metal frame.
When I first noticed this problem, my initial reaction was to replace the trap with a brand new, stainless steel door. I found some cast iron replacement doors online starting at about $40 (including shipping and handling). Stainless steel doors were significantly more expensive. I eventually decided to go with a quick and inexpensive repair instead because:
I can give this door a new life for under $10. This is important in the current economy,
The current door is slightly bigger than 8’’ x 8’’ but smaller than 12’’ x 8’’. I could have enlarged the opening to accommodate for a new door but that would have been more work,
After repair, the door will function as if it was new (i.e. it will latch, no longer rust and stain bricks and even look significantly better),
The repair will not prevent a brand new door to be installed later.
I am in no way linked to that vendor. I do not endorse it nor receive any compensation to mention it here. It appears as a “note to self”, a way to remember where to look at first if I ever decide to replace the door.
Left: I swung the door open. This revealed a large amount of ashes in the chimney pit. This will be addressed later. After inspection, it looked like the door was set in a bead of mortar. I could not find any fastener.
Right: I inserted a pry bar under the metal frame and gently pried away. The entire door assembly came loose very quickly. There was not much holding the door to the building.
Left: The opening after I removed the door. The mortar bead is clearly visible at the bottom and left sides of the opening.
Right: I knew it would take a while to eliminate the rust and paint the door so I taped a piece of 6 mil polyethylene around the opening. This will discourage critters and keep rain away.
It turns out, the wind blew my patch away the same day. Oh well, next time, I’ll find a better cover mechanism.
Left: This detail of one of the hinges of the door best illustrates how much rust was on the door. There is really not a single square inch without rust. There are spots of loose rust.
Right: I had a hunch that removing the door from the frame would make rust neutralization and painting much easier. I used a pair of “channelock pliers” to carefully bend the frame part of the hinge. It took about a minute to separate the door from its flange.
Left: I purchased that can of rust neutralized years ago for another project. I did not see any preemption date so I assumed it was still good. Judging by weight, the can felt about half full.
I carefully read the instructions before doing any work. That specific brand of rust neutralized demands loose dust to be removed by sanding with 80 grit sand paper and requires rust on the metal to bind to it. Instructions clearly warned that the product will fail to perform if there is no rust left. Finally, instructions called for using the product when temperature is above 40F.
In my experience, it never pays to ignore manufacturer’s instructions.
Right: I used a power sander equipped with 80 grit sand paper to remove loose rust. I made sure to wear a respirator.
Left: It was cooler than 40F outside so I move inside, in a well ventilated location where I did not mind the overspray. I sat all parts on a set of ceramic tile scraps and sprayed as per manufacturer instructions (between 8’’ to 10’’ away from the surface, in a continuous movement). I applied two coats, within 20 minutes, as required by the manufacturer.
Right: Once applied, the product needs to dry for at least 24h. After drying, all the visible rust turned black. This is expected and explicitly described in the instructions. At this point, parts can be directly top coated with an oil based paint.
Left: The manufacturer also describes that a smoother finish can be obtained by lightly sanding parts with 400 grit sandpaper. The black part needs to remain for the parts to be protected from further rusting. I sanded as described on the instructions.
Right: Tools for spray painting. From left to right: a respirator capable of blocking some VOC, a spray can handle and a can of exterior paint.
The handle presses the can’s tip and saves your finger for getting numb. I once sprayed a door (about 2 cans) and my index finger got numb for three weeks so I purchased this inexpensive (~$5) handle.
Left: I read all instructions on the can of paint and respected it to the letter. They call for a 12’’ spraying distance, two or three thin coats in 20 minutes and 24h drying time at least. I sprayed and left the part dry overnight.
Right: Meanwhile, I went outside and concentrated on cleaning the ash pit. By judging at the amount of ash, it appears that it was never cleaned in the 35 years this house has been standing for.
I first started to clean with a shop vacuum cleaner but wood ashes were so small, they clogged the filter almost immediately. After cleaning the filter three times in less than 3 minutes, I decided to scoop most of the ashes and then finish up with the vacuum cleaner. It took about 35 minutes.
Left: After all was said and done, about two 35 gal trash bags were filled with ashes, partially burnt pieces of papers, burnt poultry bones …, yuck! I will later put those bags in a heavy duty contractor trash bag (3 mil thick at least) so they do not burst open easily. I my community, ashes can be disposed in the garbage.
Right: After the door dried, I re-assembled it. I also took a minute to adjust the latch: a tongue of metal fitting in an opening of the frame. This fixed the non latching problem. As I did this, I made a few minor dents in the paint on the hinges.
Nevertheless, the door looked almost new. A few dust particles got caught in the paint (I should have been more careful containing dust from another project I worked on as the door was drying) but I was not concerned about those tiny imperfections. From where it is installed, no one will see them.
Left: Re-attaching the door to the masonry opening first requires the removal of all loose mortar. I used a brick setter’s hammer to gently hammer out any loose material. I also removed all dut, ashes and other debris. This will help the new mortar to bind to the existing structure.
Right: A bag of “SAKRETE – Type S High Strength Mortar Mix”. It is designed for brick settings, among other things. I purchased that bag a while ago and kept the leftovers (a few pounds) in a dry location. I scooped some of the mix in a 1 gal buck.
Left: I always mix concrete by adding a little bit of water a time. Too much water weakens the concrete. In fact, the drier the mix, the stronger it will be. There are of course exceptions to this rule.
Right: After a few minutes of mixing, the concrete reached the texture of a stiff cake icing.
Left: A tool specifically designed for setting bricks. This narrow trowel makes it possible to pack mortar in between bricks. I purchased it as part of a “masonry kit” at a discount tools shop years ago. I have never used it before and it still has the manufacturer’s sticker on the handle.
Right: I used a water sprayer to wet all surfaces. This promotes good bond between the old mortar and the new mortar. It also removes all traces of dust. All masons I know never skip this step.
Left: The trowel is ideal to fill this joint. It is sometimes a little tricky to load the mortar on such a small trowel so …
Right: … the best way to repair a joint is to load a margin trowel with mortar and use the joint trowel to pack it in the joint. This operation is sometimes called “tuckpointing”.
Left: The joint filled. In theory, after filling the joint, masons make one last pass with another thin trowel shaped as a half moon. This creates a small recess in the mortar. I do not own this special trowel so I achieved the same result by wetting my finger and using it to smooth the joint as one would smooth caulk. Mortar attacks skin so I only do this for tiny joints and immediately wash my hands with clean running water.
Right: I once again applied water to all surfaces and packed mortar all around the opening. This will hold the door in place.
Left: The opening with mortar, ready to receive the door.
Right: I immediately pressed the door in place. It took a little effort because the mortar created some friction. After setting it there was about 1/16’’ gap between the flange and the wall. The door actually sat that way for 35 years so I decided not to correct this small issue.
Left: Details of the bottom of the door: I troweled the mortar smooth and made sure there was a little bevel leading to the inside. This will avoid breaking the edge of the mortar when cleaning the pit.
Right: At the top, I filled the minor gap between the flange and bricks with mortar. I also smoothed the joint as explained before. The only thing left now is to close the door and wait for the mortar to cure.
What an Improvement! It is difficult for me to cost this repair. I already had all materials and tools required with the exception of the black can of spray paint. It costed me $4.99 at a local hardware store and I only used half of the can.
Margin Trowel, Brick joint trowel.
Spray Can Handle
1 Gal bucket
About half a can of “Loctite Extend Rust Neutralizer”
About half a can of Rust-Oleum Exterior Black Spray Paint
80 Grit / 400 Grit Sand Paper
About 2 lb of “SAKRETE – Type S High Strength Mortar Mix”
A GE Profile dishwasher is repaired. We identify the problem, disassemble the door, clean and lubricate the sequence switch. A full "normal" cycle is run to spread the lubricant on all parts.
Skill Level: 2~3 (Basic ~ Moderate)
Time Taken: About 1 hour
During a long washing cycle (called "Pot-Scrubber" by the manufacturer), the dishwasher abruptly stopped and started beeping. The "Normal" indicator was also continuously blinking. Pressing any button including the "Stop/Reset" button did not help. This suggested to me that the dishwasher detected a serious problem and decided to stop operating.
I carefully unlatch the door. Luckily, no water came gushing on the floor.
Left: I immediately attempted to locate the dishwasher model number – it is usually located on a side of the door or on one of the sides of the dishwasher’s recess for the door. I found it at the top left, on the door’s opening. It is difficult to read but the model number is GSD4030Z05WW. Knowing the model number allowed me to find an exploded view of the dishwasher’s door on the Internet at: http://genet.geappliances.com/IPCNet/Dispatcher?REQUEST=IPCNETGETPDF&file=00000000/00000600/00000622.p01.pdf. This greatly helped me fix the issue.
Right: I inspected the interior of the dishwasher: soap was released from the locking cup but it seems that it did not yet see any water.
This told me that the first pat of the "Pot-Scrubber" cycle (wash with fresh water only) completed properly and the dishwasher failed right at the beginning of the second part of the cycle (wash with soap).
I went to the electrical panel, located the breaker for the dishwasher and turned it off. I also verified that the dishwasher was no longer energized.
Left: I used a Phillips screwdriver to remove the screws holding the door cover to the door. I resisted the urge to use a power tool here. I have seen many cases where old and rusty screws break and I did not want to deal with two problems at once.
I made sure to set all screws aside, in the exact order they were removed. This will make sure there is no confusion on which fastener goes where when I re-assemble the door.
Right: After removing about 6 screws, the door cover swung loose. I removed the door and inspected it.
There were signs of soap leaks as well as some minor signs of water leak at the bottom. I cleaned all spots very carefully.
Left: I turned my attention to the door itself. It also did have traces of soap and water. I cleaned those too. Also visible on the door are switch panel (left) and the heart of a dishwasher: the sequence switch (white component on the right).
Right: Close-up of the sequence switch (also known as "Control Module"). This is what controls every single aspect of washing cycles in a dishwasher. Because the dishwasher stopped abruptly in the middle of a cycle, I immediately suspected this component.
Left: I disconnected the two electrical connectors and …
Right: … removed the two screws holding the module in place. The screws came out very easily and the module swung free. I picked it up and moved it to a well lit, flat area.
Left: Close up of the sequence switch seen from its back. The gray wheel at the top controls the locking soap cup. It appears to have just released it (the wheel has a notch – as the wheel turns, the notch will push a lever releasing the soap and the notch’s location matches the position of the lever). This is consistent with what I already observed.
Right: I inspected the module carefully and noticed that the top cover is held in place with plastic pins. I used a small flat screwdriver to release them. This must be done very gently to avoid breaking any of the pins. Once freed, the cover was easily removed.
Left: The internals of he sequence switch. From left to right: the small motor feed by red and black wires causes the white and black gears to turn. The black gear itself causes the large gray wheel to turn at a very slow speed. As the gray wheel turns, various plastic pins on it causes switches to open or close. These switches command various parts of the dishwasher: pump to remove dirty water, water valve to admit fresh water, heating element …
I tried to gently move the gray wheel. It was stuck and did not want to move. At this point, I knew I was right to suspect the sequence switch.
Right: There are three things which can fail in such a module: the motor can be burnt, the various mechanic parts may be damaged or not longer moving as designed and switches may not provide a good electrical contact any longer.
I decided to verify them all. I carefully removed the first bank of switches and inspected it. The spark created by the switch opening when power is cut caused the contact point to accumulate sooth I removed the first bank of switches …
Left: … and cleaned up its contact points with a strip of 400 grit sand paper. It needs to be a very fine grit to not damage the contact points. I performed that operation on the bottom contact points as well as on the second bank of switches.
Right: I applied a lubricant (in this case WD-40) to all moving parts, including the two gears, the axis of the wheel … I also applied some of the lubricant on every single pins on the gray wheel. This is designed to facilitate the movement of the wheel as switches ride it.
Finally, I disconnected the motor and measured its internal impedance. The reading slowly grew to "1.". This suggested that the motor was still working. This is not a definitive test but I had a hunch that the problem was not at the motor so I took a calculated risk.
I put the switch back together, re-installed it and put the door cover back on the dishwasher. I powered the dishwasher and ran a "Pot-Scrubber" cycle. At the time the second part of the cycle started, the same problem happened again. I decided to run the shorter "Normal" cycle once. It worked.
Over the course of a week, I ran the "Normal" cycle twice more. This had the effect of spreading the lubricant on all moving parts. I decided to run a "Pot-Scrubber" cycle and voila!, it worked flawlessly.
While removing a recessed medicine cabinet, a dangerous junction box is discovered. The dangerous box is removed, a new box is installed and re-wired.
Skill Level: 2~3 (Basic ~ Moderate)
Time Taken: About 1 hour
While remodeling houses, it is not uncommon to find work which was not properly done. Today’s article describes gross violations of the National Electric Code (NEC) resulting in a fire hazard. The junction box is replaced and re-wired to comply with appropriate building codes.
After pulling the medicine cabinet, I used my digital camera to take a picture of what was in the stud bay, just out of curiosity. Well, I was not disappointed. There was a metal junction box lurking in the wall. On this picture, you can see at least the following NEC violations:
314.29 – Junction Boxes shall be accessible without damaging the building finish
–>This box is concealed behind drywall.
312.5 – Use of proper fittings to ensure wires are secured properly
–> No clamps are used to secure wires.
314.16 – Volume of junction boxes
–> This box is way to small for 8 14/2 wires. This is a fire hazard.
Left: I turned the power off at the main panel and verified that all wires were no longer energized. Remember that you sometimes must trip more than one breaker (up to six in fact) to completely de-energize the whole house. Always double check that the wires you intend to work on are no longer energized.
Right: I used a pry bar to loosen the top nail of the box. It was much more difficult that it looks like. I had to be gentle enough not to damage wires, yet apply enough force to remove the nail.
Eventually, the top nail gave way and I was able to pry the bottom nail. The box was freed from the framing.
Left: This is a 34 cu (cubic inches) remodel junction box. It can contain up to 17 conductor equivalent 14 gauge wires. NEC article 314.16 describe how to calculate minimal volume of a box.
In this situation, there are 8 14/2 wires in the box, no device and no internal clamp. So we have 2 (hot + neutral) * 8 (there are 8 NM wires in the box) + 1 (all ground wires in a box count for one conductor equivalent). The box needs to accommodate 17 conductor equivalent. For 14 gauge wires, a conductor equivalent requires 2 cubic inches. The box must be at least 2 * 17 = 34 cubic inches.
Right: I decided the location of the new box and marked the opening with a pencil.
Left: I progressively cut the drywall for the new box. This revealed more of the dangerous J-Box. It is obviously very crowded.
Right: Another detail of the J-Box. It is difficult to see on the picture, but there is another NEC violation:
314.4 – All metal boxes shall be grounded
–> The box itself was not grounded.
Left: I finished cutting the hole with a drywall saw.
Right: Detail of wires entering the box. They are not clamped as required by code but the jacket of the top NM wire has been slightly damaged by the sharp edge of the box. This is a fire hazard. If the metal box cuts the insulation and reaches a wire, this can cause a spark and start a fire.
At this point, I marked all wires and made sure I recorded all connections on a piece of paper. I unwired all connections. Wires were not even twisted properly. This was another fire hazard.
I pulled wires in the new box and secured the remodel box to the wall. Plastic boxes like these have built-in clamps which secure wires at the box.
Damaged wires were replaced. It was now a matter of making the connections according to the notes I wrote in the previous step.
Making a connection is easy: twist wires with a Lineman’s pliers so both wires are twisted tight against each other and cap the twist with an appropriately sized nut (I used yellow nuts which handles two 14 gauge wires according to the manufacturer).
After finishing the wiring, I carefully pushed all wires in the box. I tired to arrange the as neatly as I could to facilitate identification of circuits. Later on, the box will be capped with a 2-gang blank wall plate to conceal wires yet allow for easy access if needed.
This work was clearly not performed by a licensed electrician, not permitted and not inspected. This "do it yourself specials" exposed occupants of this house to fire hazards. When doing any kind of work on any building, always consult your local building department, professional(s) and respect all local and national building codes.
DISCLAIMER: Procedures demonstrated in this article may or may not meet code in your area. Consider this article for entertainment purposes only. I cannot be held responsible for anything that may happen when trying to duplicate anything shown here. Before attempting to perform any work on any building, always consult a licensed professional.
A furnace filter is replaced. During this process, it becomes obvious that the filter is poorly installed. An air return filter grille is added to correct the problem.
Skill Level: 2 (Basic)
Time Taken: About 25 minutes
Forced Air Furnaces have a blower which takes air from the living space, warms it and pushes it through ducts to all rooms. It is important to make sure air blown in the system is free of dust, particles, allergens because:
You are breathing this air – dust, particles, allergens … can cause lung irritations or other respiratory diseases,
When dust in the air comes in contact of impellers in the blower, they tend to stick to it. This builds up a coat of dust on internal elements of the blower and causes the motor of the blower to have to work harder to pump air. This can lead to premature stress of the motor.
In this article, the filter is poorly held in the furnace and cannot effectively perform its function. We clean the dust accumulated on the blower’s motor and provide a better way to hold the filter.
When I tried to change the filter on the furnace, I could not find a "filter box" (a drawer like device where furnace filters are held). I had a hunch that the filter would be inside the furnace.
Just to be sure, I shut down the power at the furnace.
Left: I unscrewed the cover of the furnace and sat it aside.
Right: I inspected the furnace and noticed that the blower compartment was located at the bottom, hidden behind a protective panel. The panel was held by a few sheet metal screws. I removed them using a simple screwdriver.
Left: There was the filter – held by duct tape. Actually, the glue on the duct tape dried and the filter was pretty much floating freely at the bottom of the blower compartment. This is bad because when the blower starts, the flow of air causes the filter to be sucked tight at the bottom of the cylinder (the blower fan). This leaves large gaps around the filter where air can get through and bypass filtration.
Right: There were signs of dust and particles collecting on the blower. They are clearly visible on the back of the blower’s electric motor.
Left: I pulled the filter out. It is very dirty.
Right: Using a 5HP shop vacuum cleaner, I cleaned the blower as well as I could. I could not access the impellers in the fan assembly so I did not clean these. I was able to touch a few impellers with my finger and observed that some fine dust already started to collect on them.
I could have removed the complete blower assembly, pulled the motor’s arbor and cleaned the fan perfectly. This would have taken at least a few hours and I did not have this kind of time.
However, it was now obvious I had to find a better way to hold the filter in place. I put the furnace back together and turned the power back on.
Left: All forced air furnaces have an air return grille: this is essentially where air is sucked in the system. This simple grille hides the duct work and prevents items to be sucked in by mistake.
In this installation, the air return grille is easily accessible, making it the perfect candidate to hold an air filter.
Right: This is a new Air Return Filter Grille purchased online. It is an air return grille which can hold a standard 1” thick filter. It offers a removable face which allows for easy filter replacement.
When ordering air return grilles (with filter or not), the width of the duct opening comes first and then the height. In my case, the duct opening is 14” wide by 20” tall so I ordered a 14 x 20. It is easy to forget this and order the wrong part.
Left: Using my impact driver, I removed the four screws holding the old grille in position.
Right: I slid the new grille in position and secured it to the framing at four corners using 2” coarse thread drywall screws.
Left: In order to prevent unfiltered air to enter the system, I applied metal foil tape to seal any gaps between the edges of the air return grille and the duct. I made sure all the duct work to the furnace was properly sealed as well.
While it is impractical to make ductwork perfectly air tight, sealing with metal foil tape is, in practice, pretty effective.
Right: The standard 14×20 1” furnace filter slides in the opening. The removable grille cover gets installed above and secured to the frame.
The brown mark at the top left of the filter is a paper tear I caused when I took the filter out of its packaging. It does not affect its effectiveness.
A freeze proof faucet (also known as "sillcock", "bib hose" or "spigot") is an exterior faucet specifically designed to prevent water from freezing at or close to the faucet during winter freezing time. This is achieved by the two following unique design aspects:
The water is actually shut as far away as 20 inches from the handle. This ensures that water stays in the building, far away from the freezing conditions,
The faucet drains itself when the water is shut off to ensure no water remains accessible to freezing conditions.
Now, the automatic drainage can only happen if when the hose is disconnected from the faucet. This is why it is critical to always disconnect the hose from the faucet at the beginning of the winter. Many people did not respect this simple rule and saw their freeze proof faucet … freeze. This is clearly written on every single faucet in home centers.
In this article, we replace an old faucet with a new one.
The existing faucet. It is a multi-turn valve with a rubber washer and has previously shown signs of leaks. Moreover, it does not have an integrated backflow preventer (sometimes called "vacuum breaker").
In North America, we expect to get clean water when we turn the faucet. There is a rare but dangerous phenomenon called "back-siphoning". It happens when the following conditions are met:
A faucet (or garden hose is turned on)
The faucet or hose is submerged in water that has "left the system" and is therefore considered dirty. For instance a shower head left submerged in a filled bath tub,
There is a sudden loss of pressure in the water supply system, perhaps because the water was shut off for maintenance.
These conditions can cause the water to be sucked into the incoming water pipes, contaminating the system. Most building code require a backflow preventer to be attached to all garden hoses. Some cities require the whole house be protected against backflow at the meter level.
There are in-line vacuum breakers for spigots but it is easy to forget to use them. Besides, the wall behind this faucet is open so replacing the faucet is a faster operation.
Left: A replacement freeze proof faucet. I purchased it for about $20 at a local plumbing supplier. It is 10” long which means that the water is cut 10 inches away from the handle. The faucet is actually located close to the threaded piece on the far left of the picture.
Right: The faucet offers a two in one connection: threaded or sweat soldering. The heat of the torch can melt the rubber gasket in the faucet so a warning on the faucet body reminds installers to remove the cartridge before soldering.
Left: The old faucet seen from the inside of the house. It is an 6” freeze proof faucet which means that the water is shut off 6 inches away from the handle.
The new 10” faucet will provide a little more protection.
I went inside and shut off the water at the main valve.
Right: From outside, I removed the two screws holding the faucet to the building.
Left: Using a pipe cutter, I cut the existing faucet as short as I could. The idea is to leave enough copper pipe so I can adjust to the size of the faucet without having to cut and solder an extension.
The pipe cutter is easy: You set it up on the pipe, make one turn, tighten the screw a little, make another turn and so on until the pipe is cut.
Right: The pipe separated and a little bit of water drained. I put a small bucket to collect the water.
Left: I pulled the faucet from outside. It came without a fight.
Right: Using an adjustable wrench, I removed the packing nut of the faucet and extracted the cartridge, as required by the manufacturer.
Left: The disassembled faucet. The body of the faucet is at the top. The cartridge is the brass part at the bottom.
Right: I slid the faucet’s body in the opening. I could insert it into the cut pipe but as expected, the pipe was too long and the faucet stayed proud of the siding.
Left: I measured the distance between the siding and the faucet’s flange. It turned out to be about 1/2”.
Right: Back inside, I pulled the faucet as much as I could and marked where the pipe needs to be cut. It is not easy to see on this picture but you can actually see the bottom of the fitting in the faucet and it is possible to estimate where the cut needs to be done by eye.
After marking, a quick check with a tape measure showed that I would remove about 5/8” of the pipe, which felt about right.
Left: I cut on the mark with the pipe cutter.
Right: This pipe cutter as an integrated reaming tool. It is designed to cleanup the inside of the pipe after cutting. The burr left by the cutter is minimal but any obstruction in the pipe can reduce the water flow and create a lot of noise when water flows.
Left: I removed the reamer from the cutter and reamed the end of the pipe by hand. It took a while but this step is important.
Right: I cleaned the end of the pipe with a piece of waterproof 100 grit sandpaper designed for this application. This cleans the copper and allows the solder to bind to the copper properly.
After this operation, the copper should be very shinny. I removed any dust and sandpaper grain with a damp rag. The pipe needs to be perfectly clean for the solder to be successful.
Left: I used a round wire brush to clean the inside of the fitting on the faucet body.
Right: I then applied tinning flux on the outside of the pipe …
Left: …. and inside the fitting.
I could have applied flux later but my experience told me that the flux helps lubricate the joint a little bit and facilitates inserting the pipe into the faucet’s fitting.
Right: I pulled the faucet out and applied two thick beads of 100% silicone caulk to the flange of the faucet.
Left: I positioned the faucet in the opening, inserted it into the pipe in the wall and secured it with two 1 5/8” exterior screws.
Right: I added a more caulk at the bottom of the faucet and smoothed it with my finger. This should achieve a water tight seal.
It is not the most perfect bead of caulk I ever did but this area can’t be seen so effectiveness matters more than aesthetic.
Left: Turn the torch on and start warming up the fitting. Make sure you do not directly put the flame of the torch on the join you will be soldering: the sooth of the torch will contaminate the joint and it will be difficult to achieve a water tight seal.
Right: After about 15 to 30 seconds, move the torch away and apply a little bit of lead free solder to the fitting. If the solder melts like water, quickly put about 3/4” of solder into the fitting. Solder will be sucked in by capillarity and will seal the joint.
If it does not melt and becomes liquid, move the solder away and apply some more heat.
It is critical to not apply heat directly to the solder: it will melt too quickly and will not allow you to produce a water tight joint.
Left: As the joint is still hot, I take a slightly damp rag and clean up the excess solder at the bottom of the pipe.
You have to be quick and gentle during this operation of you will remove solder from the joint and it will leak.
At this point, I left the fitting alone for a few minutes to allow it to cool down. I then cleaned all residues of flux with a damp rag. Flux residues can corrode pipes so it is a best practice to always clean flux up.
Right: I applied a generous bead of expanding insulating foam from the inside of the wall. Washington energy code mandates that all opening be sealed properly to reduce air infiltration.
I am using Great Stuff Pro Gaps and Cracks with gun applicator. It is a overkill here (it is a residential type V fireblock which is not required in this application) but I had a can that was going to expire so I finished it up here.
Left: I turned the water on for a few seconds to flush the pipe from all flux residues.
I then threaded the faucet stem back in. This is a delicate operation: the cartridge needs to engage in its housing and go all the way in.
Right: I tightened the packing nut by hand and then made one more turn with the wrench.
I turned the water back on and checked for any leaks.
Left: Alert readers have noticed that the pipe is located close to the edge of a stud. Plumbing codes call for nailing plates. This will prevent nails or screws to go and puncture the pipe.
Right: It is matter of positioning the plate and nailing it home, like that.
Cedar panels are glued and raised on a router table. Posts are cut to length and decorated on a router table. All parts are sealed and assembled.
Skill Level: 3~4 (Intermediate ~ Advanced)
Time Taken: About 6 hours
There are many designs for cedar planters and I have already showcased a simpler version in a previous article. It was meant to hold the soil directly and has served me well. Today, I am going to show a more sophisticated design.
Unlike its older brother, this planter is designed to hold a pot. It is entirely made out of solid clear cedar and features an adjustable shelf. This allows the planter to host various kind of pots.
Pieces of the planter before of the assembly:
Four raised panels: these are the sides,
Four posts: these act as corners (only 3 visible on the picture)
Eight supporting strips: panels rest in groves machined in those (only 7 are visible)
All pieces received a coat of Thompson Water Seal before assembly. This picture was taken immediately after application which explains why the wood looks darker. Thompson Water Seal is a clear product and I decided not to apply any other finish: cedar is beautiful as is.
This is the first time I used that product and while I have been satisfied with it, it did not always goo smoothly. Suffice it to say that it is critical to read and respect all instructions exactly. More specifically, be sure to respect the temperature, drying time and maintain the wood dry as the product dries or you will ruin your pieces.
But, let’s not get ahead of ourselves.
Left: The construction starts by cutting four 18 1/2” long pieces out of 2×2 clear cedar stock. This wood is already surfaced on four sides and is designed to build deck ballusters, among other things. These will ultimately become the posts installed at the four corners of the planter.
Left: The top of each post gets decorated with a narrow and shallow groove made at the table saw. I have installed a stop block (the piece of scrap wood clamped to the fence on the bottom right). This ensures a consistent location of the groove on all posts.
It is also critical to safety: to prevent kickback when crosscutting on a table saw, the wood must never be in contact with the fence.
Right: The blade was raised about 1/2” above the table and each post was ran through using the miter gauge. It is necessary to groove each post on for faces.
I decided to leave the blade guard for safety. It was mildly awkward to run pieces through.
Left: I moved to the router table and installed a chamfer bit. A chamfer will be cut at the top and bottom of every four posts …
Right: … like that (example at the top of a post). The chamfer is a little less than 1/2”. I just kept moving the fence away from the bit until I was satisfied with the depth of the chamfer.
You can also see the result of the previous grooving operation on the table saw.
Left: A decorative stopped chamfer was added on the external edge of all four posts.
There is a special procedure to make a stopped chamfer. First write marks on the work piece to indicate where the chamfer will star and end.
Right: With the router running, make sure the left side of the post rests on the table and hold the left side above the bit, without touching it. Position the mark on the work piece on the center of the bit.
Now, slowly plunge the work piece into the bit.
Left: When the piece is flat on the table, push it through as you normally would until the center of the bit reaches the end mark.
Pull the right end of the work piece up, making sure the left side is in contact with the table.
It is critical to make sure that at least one side of the work piece is on the table otherwise, the bit could catch the wood and send it flying in the shop.
Stopped operations on the router are dangerous and you can get hurt badly. Always consult a professional.
Right: I cut eight supporting strips out of 1×2 cedar. One side will require two pieces: one to support the panel at the bottom and one top hold the panel at the top.
Left: Each piece was chamfered on three edges using the same chamfer bit. The last edge is left square because it will not show.
Right: I cut a grove on every single supporting strip. This groove will receive the panel and hold it in place. The groove was cut using a straight 1/4” bit. 1/4” bits are usually more fragile because they are thin so it took three passes to machine the groove.
It needs to be deep enough for the panel to be held and allow for wood movement. I decided that 5/8 was deep enough in this case.
Left: Using the same 1/4” straight bit, I machined two grooves on each post. These will receive the raised panels.
Left: All pieces were sanded smooth and paired together so the grain of the wood would match top and bottom supporting strips.
At this point, the wood has not yet received any finish or sealant. However, assembly steps can already be outlined.
Right: A panel slides into two supporting strips: one at the top and one at the bottom.
Left: The right and left sides of the panel slide into a grove on the right and left posts. The supporting strips rest on posts and will be later fastened to them ….
Right: … like that. Each side is completed after driving the four pockets screws home.
Left: After assembling each panel, I drilled a set of holes to hold shelf pins. This allows the mobile shelve to be moved up and down to accommodate various pot sizes.
I drilled free hand and some holes were not exactly at 45 degrees. This is clearly not ideal.
The best tool for this job is a drill press: the work piece can be maintained tight on the table and drilled precisely. Unfortunately, I do not own one.
Right: A shelf pin installed in one of the holes.
The imperfection in the drilling did not have any consequence. I guess I was off but not off enough to have a noticeable impact on the stability of the shelf.
Left: I turned my attention to the shelf itself. It is essentially a few strips of rough saw cedar fencing stock cut to length and screwed onto two pieces of 1×2 cedar. I did not put any glue but it would not hurt.
If you glue this, make you you use a waterproof glue: watering plants will definitely pour water on the shelf.
This design is rather primitive but the shelf is not going to be visible so I can save some labor and materials here.
Right: I drove a couple of screws to hold the shelf together. I used the same pan head screws (aka screws suitable for pocket screws) because I had them handy and they were of the adequate size.
Left: The shelf face up. It looks reasonable and it is fairly sturdy.
Right: A view of the planter after assembly, looking down at the shelf. The shelf rests on four shelf pins.
It is often necessary to remove ceramic tiles without breaking them. For instance:
You want to match an existing tile job and need to bring a sample to the store,
You need to gain access to a shower faucet concealed behind a tiled wall.
Removing tiles without breaking them is only part of the job. If you plan on re-using them, you will also have to clean up the back of the tile. That means removing old mortar or adhesive. Finally, removing tiles usually damages the substrate and you’ll have to patch it up before you can tile over it again.
I have about 70% success with the following technique (for 100 tiles removed, I’ll break at most 30). The key is to progress slowly and refrain from using brute force.
In this article, I’ll show how to remove two tiles, both set with mortar directly over drywall.
This blog sometimes illustrates dangerous tasks and removing tiles is one of them. Be sure to wear the necessary safety gear. That includes but is not limited to: gloves, safety glasses … I cannot be held responsible for anything which may or may not happens to you or anyone as a result of reading and / or applying procedures described in this blog . Be safe. Always consult a professional.
Left: The first tile to remove. It is a ceramic tile which was ripped in half to serve as a baseboard. This is a common practice.
With a utility knife, I cut through the paint, caulk, mortar and grout all around the tile. I used a dull blade: no need to waste brand new, sharp blades here. They get dull in seconds.
Right: There was a thin line of grout at the bottom so I cut through it. I wanted to remove most of the grout / mortar (about 1/4” deep in this case).
It is important to remove as much grout as possible before attempting to remove a tile. Failure to do so will greatly increase your chance of breaking the tile. Do not rush.
Left: After freeing the tile from all around it sides, I gently inserted a stiff putty knife between the tile and the substrate. I gently tapped the blue handle with a hammer to help the blade go in further.
I repeated that operation on the other side of the tile (right on the picture). Again, it is critical to be gentle with the hammer.
Right: I inserted the flat side of a mini pry bar at the top of the tile and gently pushed it in. The tile came without a fight.
The tile separated from it substrate. As you can see, part of the drywall came off. There is even a screw visible.
The substrate will have to be patched before tiles can be re-applied. In this case, I’ll just fill the void with drywall joint compound.
If the tile is to be reused, its back needs to be cleaned. I usually take tiles outside, lay them on a piece of carpet face to clean up. I then spray them with water to control dust and use a belt sander with a 40 grit belt to grind the mortar off. It usually takes a minute per tile and it can get a little messy.
Now, let’s remove a small tile set on the wall. It is part of a decorative border around a corner shower unit.
A critical tool for this task: a grout saw. This is essentially a small piece of steel onto which diamond powder was deposited. It makes removing grout between tiles much easier, especially when grout lines are at least 1/8” wide.
This saw was purchased at my local Harbor Freight store for about $5. So far, I have used it to remove about 4 linear ft of group and it is still working fine. I can feel it is not as durable as more expensive saws but it has served me well for a very reasonable price.
Left: I used the saw to cut the grout on the left and bottom side of the tile. the saw is pressed against the grout and a back and forth movement performs the cutting, like a regular wood saw.
Right: The tile after the grout was cut. I went as deep as I could. In this case, this was the substrate (drywall).
Cutting the grout will reduce the risks to break adjacent tiles during the removal.
Left: Using a dull utility knife, I cut through the caulk and paint on the left side of the tile. The tile is no only held by mortar under it.
Right: I inserted the flat side of a mini pry bar under the left side of the tile and gently pushed it in.
The tip of the pry bar dug a little bit into the drywall. This is OK.
Left: With the help of a hammer, I gently tapped the pry bar under the tile. Again, it must be gentle or the tile will break.
Under normal circumstances, I would hold the pry bar with my left hand but I was taking the picture.
Right: The tile popped off. It is difficult to see on this picture but the drywall will slightly damaged: the pry bar created a low spot as I was prying.
Again, the substrate will have to be fixed and the back of the tile cleaned before reuse.
While preparing a door frame for casing, we notice the drywall is not flush with the door frame.
The drywall is shaved to allow the casing to sit flat on the wall and on the frame.
Skill Level: 2 (Basic)
Time Taken: About 20 minutes
You have already seen me installing door casing here. It is usually an easy task when the drywall is flush to the frame. Sometimes, this is not the case. This short article describes one common technique used to correct the problem.
The starting point is pictures above: a piece of scrap casing is held on the drywall and a gap of about 1/4” appears between the casing and the window frame.
The tool for this project: a single blade tool
Painters use this tool to scrape paint of glass windows, or from ceramic tiles, among other things.
This tool can be purchased at home centers for around $3. I once purchased a box of 100 blades on sale at my local Harbor Freight for about $3.
I am usually not a big fan of HF tools but I figured that $0.03 a blade was a bargain even if they last 5 times less than more expensive blades. I must admit that so far, those blades have met and exceeded my expectations: they worked at least as well as more expensive, brand name blades.
Left: It is easy to see that the drywall is sticking out w.r.t. the door frame.
Using the single blade tool, I shaved a little bit of drywall at a time …
Right: … like this. It is important to remove only enough drywall for the casing to rest flush with the wall and ensure the shaved region is still concealed under the casing.
I find it is best to cut a little bit at a time and use a piece of scrap casing to test. If the casing still does not rest flat, it is easy to see where material needs to be removed. It is a iterative process.
Left: A section of the wall where drywall was cut. It looks bad but everything will be concealed behind the casing.
Right: A piece of scrap casing now sits flat on the drywall / frame.
I continued all around the door, on both side. The door is now ready for casing. You can see this article here.
As you shave drywall, it is not uncommon to uncover finish nails, drywall nails or drywall screws.
Finish nails: I like to remove them with a pair of carpenter pliers. Finish nails are leftover from a previous casing installation and if you pound them in, they will eventually make their way out, possibly pushing the casing out.
Drywall nails / Drywall Screws: I pound them / drive them in. Drywall nails hold the drywall to the framing. It is best to pound them in because removing them would weaken the bond to studs.
Almost all building contractors and auto mechanics have been exposed to broken screws. This article explains how to extract the broken screw without damaging the piece it threads into. While the example of a wood screw broken in dimensional lumber is taken, techniques described below can be used to remove virtually any piece of hardware threaded into another material.
The broken screw of interest. The head fell off when I tried to remove the hinge from a door jamb. As far as broken screws go, it is fairly bad because:
The screw must get out otherwise the hinge will not be strong enough to hold the door,
It is not practical to replace the whole jamb,
A large part of the broken screw is still inside (I could judge by the head of this wood screw that it was probably 8 x 2 screw) so I can’t just use a smaller screw and ignore the problem,
The broken piece is not visible from outside and it is about 3/4” deep inside the material,
The door jam must not be damaged in any way during the extraction.
In general, extracting a broken screw is done following those steps:
Look for solutions involving discarding the piece into which the screw broke. It usually does not make economical sense to discard the work piece but every once in a while; it is actually an option,
If the end of the broken screw is visible and protrude outside of the material by at least 1/16” or more, use a hack saw or a rotary cutting tool (aka "Dremel") to cut a groove in the broken screw shank. A flat screwdriver can then be fitted into the groove and the screw can be removed,
Consider chipping material around the screw hole to expose the broken screw shank and apply technique described in 2,
If everything else failed, a special tool called a "screw extractor" can be used. This is the purpose of this article.
Left: Tools for extracting the screw. From left to right: a 5/64” drill bit, a #1 screw extractor and a "T-Handle".
Right: The screw extractor has a square drive meant to be inserted into the T-Handle.
The chuck is tighten until it holds the screw extractor tight.
Left: The business end of the screw extractor. It is basically a spiral with threads going counterclockwise. Regular screws or drill bits have threads going clockwise.
Driving screws clockwise will cause them to penetrate the material. The screw extractor needs to be turned counterclockwise in order for its threads to grab.
Right: A regular screw (top) compared to the screw extractor.
This difference makes the screw extractor capable of extracting broken screws quite effectively.
Screw Extractors are used as follows:
First, a small hole is drilled into the broken shank of the screw,
The screw extractor is then threaded into the hole previously drilled and turned counterclockwise (aka from right to left),
Since driving the extractor counterclockwise causes the extractor to bite into the screw, it is now possible to drive the broken screw out using the screw extractor.
Screw Extractors come in several size and are commonly referred to by a number. For instance, this article features screw extractor #1, the thinnest of all common extractors. Screw extractors are available at most home centers and well supplied hardware stores. Choose the screw extractor which best matches the size of the screw to remove.
Every extractor calls for a specific hole size. The #1 extractor is usually used with a 5/64” drill bit as shown in this article.
Left: I drilled a 5/64” hole into the broken shank of the screw. I tried to drill in the exact alignment of the shank but this was difficult because I could not see the shank.
After about 45 seconds, I had drilled a hole which I estimated to be about 1/8” deep.
Right: I threaded the screw extractor into the hole and turned counterclockwise (from right to left).
I felt the tool engaging into the screw and starting to drive it out as I turned.
Left: After about three full turns, the screw appeared outside of the wood. It came fairly easily. I continued turning the extractor until the whole screw was completely out.
Right: The extracted broken screw still attached to the extractor.
The more you turn to remove the screw, the more the extractor tightens its grip into the hole.
I after only 5 turns, I had to hold the broken screw with a pair of pliers to remove it from the extractor.
A close up of the hole drilled into the broken screw. It is about 20% smaller than the shank of the screw.
NOTE: It is important to manually drive the screw extractor instead of using a power tool like a drill or an impact wrench. Using a power tool could cause the screw to break again or worse, the screw remover itself to break.
Hand driving the tool allows the operator to feel if the broken screw is correctly engaged and to make the necessary corrections as required.
It is sometimes necessary to extend the size of the T-Handle with a piece of metal pipe in order to increase the torque provided to the screw extractor. In this case, proceed very carefully to avoid breaking the screw and / or the screw extractor.