Table of Contents
To adjust for the 1983 gyproc lining of the walls and ceiling, a deeper stool (a carpentry term for the interior window sill - to differentiate it from the exterior window sill) and deeper jamb extensions were added over the top of the originals, which are fortunately still there to be restored.
The original stool had it's nose trimmed off back to the level of the new lining so one of my first tasks will be to add and shape a new nose piece.
On the outside, the verandah panel lining comes down low enough to cross in front of the window opening a little (top left in image 1). This is a detail that is unfinished and it would be nice to fix it.
All of the exterior paint (frame and sash) had crazed long before I got to it. This will be fixed when I strip it all off before repainting.
As usual the window box frame was wedged under the timber lintel but both wedges were loose.
The adaptable weather barrier that I made for refurbishing the window in bedroom two has been installed (image 2).
Observations as the window was disassembled:
Parting beads: both are pine and were cut too short as though the measurement forgot to account for the mitre at the top. One came out complete but the other came out in pieces. Both will be replaced with cedar of the same age as the windows.
Access panels: the right-hand panel and the surrounding jamb are badly damaged from previous removals. A new panel will be made and some of the jamb will be replaced as well. The opposite jamb may also be partially replaced to repair damage.
Sash: the upper sash was effectively painted shut but not nailed this time. One of it's cords had broken, which is not a problem as all will be renewed. Noting the weights: two x 9lbs (lower sash) and two x 8lbs (upper sash).
Glazing: there are no cracked panes and the muntins and glazing fillets look good.
1. Window with extra deep stool and jamb extensions removed, exposing the originals.
2. Weather barrier installed.
The white seal strips were added to keep bats out.
3. Heat stripped stool and jamb extensions.
The original finish was shellac over natural cedar.
4. Damaged weights access panel and surrounding jamb.
This carnage is avoidable.
5. Beads, sash and access panels removed.
Shows the result of painting with sash in situ.
Removing the access panels
Image 4 shows the carnage from previous episodes of removing the access panel. In these windows there is only one access panel each side and the parting bead runs down the middle of it. The panel comes out bottom-first so all that is required after the paint has been cut through to separate all four sides of the panel from the jamb is to drill a hole through the bottom of the beading channel near the bottom of the panel and gently hook it out. I used an Allen key (6mm). The hole will be hidden by the parting bead in normal operation.
Restoring the stool
What was left of the original stool was cut back further just enough to remove the horn at one end and clean up the other end where the horn had broken off. The offcut will not be wasted, it's a nice piece of 35 x 1200 x 12 mm.
To make the new nose piece, two strips were cut from a length of original 1879 cedar skirting board and glued up back-to-back to form a single 40mm x 45mm x 1400mm piece (image 6) that was then dressed all round on the table saw. It was deliberately made several mm oversize for that purpose. Final size will be closer to 36mm x 40mm x 1360mm.
Dados were routed along the meeting faces of the stool and the new nose to take a full-length maple spline before both parts were glued and clamped together (image 7). The upper and lower faces of the nose piece at this stage are a little proud of the corresponding faces on the stool, to allow them to be planed down.
After that the horns were cut to their correct size and the nose was shaped. It's more of an eliptical profile than a full bullnose and after deliberating for a couple of days I ended up doing it by hand using my Stanley #4 plane. I switched to the belt sander on the cross-grain ends after the plane tore a chunk out of the edge (image 8).
6. Making the new nose piece.
7. Fixing the new nose piece to the old stool.
8. Stool horn cut to size and shaped.
9. Sample of original nose profile.
10. Finished nose profile.
Making and fitting the stool apron cove moulding
The original cove moulding wasn't any standard dimensions that I can buy off the shelf locally. It is 40x30mm. I've drawn it up in image 11, as it would be if cut from a minimally sized flat stick. I calculated the angles for the bevels as 37° and 53°.
The bedroom two original is a "plain" moulding, i.e. cut from a solid 40x30mm rectangular section, and the ends are simply shaped to look like they've been mitred.
If I want to make it from matching cedar then I'd prefer to make it from flat stick as a "sprung" moulding to conserve my old cedar stock. I have a length of old architrave that will do.
I would also need to actually mitre the ends.
To cut the 39mm wide x 6mm deep cove I can use my table saw.
11. Cove moulding from a flat stick.
After dressing down the architrave I was left with the nominally 20mm thick profile in image 12 and there was no need to reduce it further just to comply with the drawing. The extra thickness compared to the drawing gives wider contact faces, which is an advantage. The missing corner left over from the architrave profile would have been removed anyway for the cove profile.
Being my first attempt at this process it took a very long time, and it turned out that the cove shaping itself was the last but easiest step. The angles for the flat sides were the difficult part. My table saw can only do 90° or 45° cuts. I used my trimmer router's tiltable base to do the first two 37° faces and then realised it can only tilt in the range +45°/-30° and would not therefore do the 53° complementary angle that I needed. I considered using a 45° bevel bit (tilted 8°) but decided against it. I ran the piece through the table saw as best I could at 90° to the two previously angled faces.
The finished moulding is not perfect but will do (image 13).
12. Architrave partially trimmed to size.
13. Finished cove moulding.
The angle and offset for the cove-shaping table-saw fence setup (image 14) was derived by running test pieces over until the correct offset, width and depth were achieved. I used the sanding pad as a push pad/gripper block for safety.
14. Cove shaping setup.
Image 15 shows the finished apron fixed to the stool. I used brads and no glue in case I need to separate it again in future. The cove needs a little more sanding to remove more of the kerf marks from the table saw blade.
Without thinking I ran the moulding to 30mm from the ends of the horns whereas the original would have gone closer to where the radius of the nose starts, maintaining the same setback as along the front of the stool. However, looking at image 16, I prefer the accidental version because the transition from the inward cove to outward nose curvature is tangential and looks "right" to me.
15. Cove apron attached and mitred
16. Temporarily back in place, the finished stool with apron.
At least one panel (image 17) needs to be remade and the surrounding frame restored to remove the gouging from previous maintenance. Rather than use an epoxy filler I want to try the more traditional method of rabbetting around the opening and laying in cedar strips. This is another skill-building opportunity.
First, I removed the complete window frame from the wall - a simple matter of removing the wedges that were holding it in place - to make it easier to work on and to see what
17. Utterly disfigured weights access.
the opening in the wall looks like without it. It looks very neat after cleaning away the dust and debri (image 18).
There has been some significant shrinkage perpendicular to the grain in the timber lintels and nailing blocks. The blocks are also a little loose and will be packed or wedged tight again, probably by driving a few screws into the gap at the top of the block.
In the workshop the frame will get a thorough going over (image 19).
18. Bare window opening.
19. Window frame removed to the "workshop".
The RH access panel itself was missing the top lap entirely after the parting bead was removed (the piece missing out of the top was past damage only weakly glued back in place). So the panel was cut right back to the depth of the bead channel to remove all damage to the visible face. Then, using small scrap pieces from my 1879 cedar stock, I cut and glued two strips to what remained (image 20). The gap separating the two strips forms the bead channel. The ends of the top and bottom laps are bevelled 10° to match the opening. The critical faces of the new parts were left proud to allow them to be dressed down to fit nicely (images 21).
20. Access panel reconstruction.
I routed out the damaged edges of the jamb around the opening, but rather than fill with a single piece then cut most of it out, I've laid in six individually cut pieces to fit together around the opening either side of the bead channel (images 21).
21. Restoration of the RH weights access panel and jamb.
I wanted to use offcuts to conserve my 1879 cedar stock but separate pieces also made it much easier to do the bevels at top and bottom of the opening. I ensured that all inlay pieces stood proud of the jamb so that they could be dressed flush.
An exterior grade glue (cross-linked PVA) was used because of the potential for getting wet.
The LH access panel and jamb were visibly significantly less damaged, however, the panel had been broken and glued back together (image 22), and so I decided to make a new panel entirely, a single piece of cedar.
As for the jamb inlay rabbets, not as deep this time (images 23).
22. Crack through the LH access panel.
23. Restoration of the LH weights access panel and jamb.
Frame and jamb extensions
The inlays and the panel were all in place and planed together whereas the RH restore was done in two stages.
Repaired the major areas of gouging and broken edges along the rest of the left and right parting bead channels also using inlayed cedar pieces (images 24).
Undercoat and first oil gloss coat completed on the trim components (image 25) and the frame. The oil paint had been stored in a cold location and was a little stiff to apply although I didn't realise why at first, but luckily the work area and components were warm, which soon relaxed the paint enough to produce a good finish.
24. Bead channel inlays.
25. Stool and jamb extensions painted.
The sash are in good condition, including the glazing where little or no work is required. They will be heat stripped, sanded and painted as for previous windows, and new hardware fitted. The meeting rails are about 2mm different in level and a timber strip had been attached to the outer face of the lower sash lower rail for draft prevention and will be removed.
After heat stripping the paint off both sides of the rails and stiles I noticed that all the rails were slightly bowed (all convex on the exterior side). Laying a straight edge along the concave side the gap was about 2mm at the centre. I doubt it was from the heat stripping, but how can it be fixed?
Fixing bowed rails - part one
One woodworking web page describes a method of soaking the timber on the concave side to restore the original shape. Another method also soaks the timber but on the convex side while the timber is clamped straight, relying on what is called "compression setting" to permanently fix the shape. I'm going to try the first technique.
I did not want to submerge the sash, preferring to selectively soak just the rails that were bowed.
I started with the lower sash bottom rail, folding a towel to width and length and soaking it before laying it along the concave side of the rail (image 26). After 12 hours I lifted the still soaking (thanks to cold weather) towel off and was able to scrape some residual coating off which had turned to mush, then with a scalpel I cut light incisions along the rail with the grain about 5mm apart to aid penetration because I wasn't sure if there was still an impenetrable coating there.
26. Straightening a bowed sash rail by soaking.
Moisture seen in the tenon (image 27) confirms penetration.
It's hard to see from images 28 and 29 but after three days the rail is straighter - confirmed with the straight edge. Clamping has not been used, the sash is simply laying flat. This result suggests the bow is caused by unusually high and uneven longitudinal shrinkage as the wood dried, and there are several reasons why it can occur, all related to conditions of the tree and the specific location within from which the piece was sourced.
27. Moisture visible in the tenon.
28. Start of soaking - bowing evident.
29. After soaking three days - straighter.
30. Lower sash, bottom rail mid point - almost no gap after soaking/PVA complete. But will it stay that way?
Once the timber has recovered its shape and while the timber is still wet to aid permeation the method calls for saturation with a diluted PVA solution to fix the shape permanently.
Because I don't think soaking the towel in diluted PVA will work all that well I will instead apply a thick even coat of full strength PVA glue to the rail leaving a border to keep it centralised. Then the towel soaked in plain water will be laid along the rail as before, covering the PVA. This will dilute the PVA while hopefully keeping the PVA from curing as it permeates the timber.
Unfortunately after 24 hours drying out the rail has reverted to its original bow. The problem has to be in the way I've applied the PVA step to fix the shape.
31. Top sash, top rail bow.
The bowing is much easier to see in image 31 of the top sash. The concavity at the mid point of the rail can be seen in image 64, which is 2 to 3mm.
After 24 hours of saturation the rail has straightened. Now I'm brushing on diluted PVA frequently to keep it wet, and I've clamped the ends to a straight backing mainly to see if that helps as the rail dries out.
32. Top sash, top rail gap to straight edge.
Fixing bowed rails - part two
The method above did not work. The bow returned as the rails dried out, so on to a more established method - a hot water bath.
Again, this may not work because softwood is generally less ameanable to bending and the setup I'll be using is suboptimal, but here goes.
The setup is a simple trough made from a sheet of copper, from the demolished main house gravity hotwater tank, sitting across the woodstove heater in the workshop. Because this use is temporary and copper is very expensive these days I wanted to avoid cutting the sheet, so it was simply bent whole over the narrow side of an 8 x 2 inch plank to form the trough then the ends were plugged using 4 x 2 inch pine blocks wrapped in soft foam and clamped watertight.
33. Crude wood bending setup
34. End plug
35. Leveling arrangement
With the sash standing with the bowed rail submersed this setup got to 55 degrees after several hours. Starting with less water initially and topping up with boiled water it was able to hold at 68 degrees.
The upper sash top rail was first. After two hours at 55 degrees it was clamped and held overnight bent in the opposite direction by a few millimetres. On release it is almost straight but it remains to be seen if this holds as it dries out completely.
The lower sash bottom rail had more of a bow in it. After an hour at 68 degrees it was clamped overnight (images 36 and 37). It is not fully straight, although it may be enough to finish by planing. Again, it remains to be seen if it holds.
36. Lower sash bottom rail clamped in line with muntins to minimise stress
37. Reverse bend is just a few mm
38. This may be the best I can do
The bottom of the bottom rail on the lower sash was sanded flat, retaining the angle of course, so that it would sit on the sill with minimal gap. The hot water bath described above had softened the putty fillets along that rail quite nicely such that they came loose and needed replacing with new putty. Then I filled, sanded and painted both sash and all the beads with oil based paint. This took some time as the oil paint was particular slow drying in the cold humid middle of winter, even though it was done indoors.
Finally, I checked both sash for loose fit in the frame to ensure the meeting rails lined up correctly. I'd learnt this from the bedroom two window which required some significant reworking after I'd started the reinstallation.
39. Both sash finished
Reassembly followed the same process as previous windows. The frame was placed back in the opening and wedged solidly. Then the weights and sash cords were prepared and attached through the pulleys, first to the upper sash, which was then fitted in place. The weights pocket covers and the parting beads were then installed as a press fit (no nails). The sash lifters were attached to the lower sash bottom rail before that sash was fitted in the frame, then the stop beads were installed. The window was finally completed with the fitting of the lock lever.
New trim has been fitted and only some exterior grouting remains to be done. There was no original trim on this window due to the room having been lined and so this is a first look at the new trim that will be fitted to all windows. It's 90 x 18 mm, which is thinner but a little wider than the original trim that was too damaged on some windows and most doors to replace, and so I've chosen to renew it all for aesthetic consistency.
Cox Bros Crookwell
As with the previously refurbished lounge and bedroom two windows, I was expecting the supplier to have labeled the bundle of window components as destined for the contractor. I was disappointed when it wasn't to be found on the back of one of the jambs, but it was there, this time, on the underside of the stool.
I also found it written on the box frame, which would have been delivered as a complete assembly.
42. The now familiar labeling with the name of the building contractor.