Glazing bar mistake

[Jacob]

Just been making a peculiar window sash - 2 panes high, 6 panes wide. It's for an internal borrowed light and unlike anything I've made before.
In view of the width I thought the short vertical bars should go through - to keep the top and bottom rails straight and tied together.
This means cutting the one horizontal bar into 6, with stub tenons at each end.
Sounds good so far - but what I found was that a tiny error on the horizontals was accumulative and causing the stiles to bow out slightly.
This meant time wasted carefully checking and easing all twelve ends, and a couple of dry runs.
I normally get it right first time with perhaps just having to ease one or two joints, so why a problem here?
Answer is: say each piece at each end was 1mm too long, that'd accumulate to 12mm error at the stiles.
But if the horizontal had gone through and the verticals cut instead there'd only be a 4mm error top to bottom, and 2mm end to end.
Furthermore - the rails were not moulded, so the joints could all have been eased in seconds by dropping a saw cut down the gap.
Moral of the story - a multi-paned window then let the long bars go through and only cut the short ones.
Except of course you can't do this with a vertically sliding sash as the the verticals must hold the top and bottom rails together.

 

[ColeyS1]

Wang a sash clamp on and just crush em up- it'll be our secret [WINKING FACE]I would have cut them the same way you did.
Cutting in infills between stair spindles always use to be the one that tested my patience. I've finally got the hang of that now [SMILING FACE WITH OPEN MOUTH]

Coley

Sent from my SM-G900F using Tapatalk

 

[Jacob]

Wang a sash clamp on and just crush em up- ....

That's what I normally do but it wasn't enough this time!
I take your point about the spindles; 0.5 error is precision for woodwork, but 20 little 0.5mm errors gives one big 10mm error.

 

[AndyT]

Interesting point, and I defer to your experience - I've not made any windows with glazing bars, so my hindsight is purely theoretical.

Could the problem have been avoided if you had first cut the long horizontal bar to the overall length (ie matched against the overall inside measurement of the surrounding frame) then marked the shoulders for the six individual parts, then cut the long bar into six pieces? The width of the five saw kerfs would have provided clearance where the ends of the tenons meet in the shared mortices and the shoulder marks would have defined the length of each component.

 

[Jacob]

..
Could the problem have been avoided if you had first cut the long horizontal bar to the overall length (ie matched against the overall inside measurement of the surrounding frame) then marked the shoulders for the six individual parts, then cut the long bar into six pieces?

yes thats how you do it, not marking from the frame (amateur bad practice!) but from the rod, including marks for the all the details.

he width of the five saw kerfs would have provided clearance where the ends of the tenons meet in the shared mortices and the shoulder marks would have defined the length of each component.

Yes, thats how you do it. The tenons don't meet, they stay a saw kerf apart - it's the shoulders which determine the fit.
But cutting by hand involves error. A natural tendency is to play safe and tend to be on or over the line, never under.. If over you can trim back a touch if necessary. This means the errors are all plus rather than minus.
Better in this case would be to saw close to the line so the errors would be more random and balance out - sometimes a gnats under, sometimes over, sometimes spot on.

 

[AndyT]

Ah yes, I see now. You did write "cutting the the one horizontal bar into 6" - but there are still ten shoulder cuts, each of which can be over or under length.

It all helps my admiration for the original craftsmen of the 18th and 19th centuries who made windows by the thousand, day after day.

 

[Jacob]

Ah yes, I see now. You did write "cutting the the one horizontal bar into 6" - but there are still ten shoulder cuts, each of which can be over or under length.

It all helps my admiration for the original craftsmen of the 18th and 19th centuries who made windows by the thousand, day after day.

Thousand is the key word. Mine would have been better second time! By 1000 I'd have been doing them perfectly in my sleep. :shock:
Repetition is a key feature of real craft work - endless copying and refinement of the process. Often overlooked by those who see bespoke one-offs as the ideal. In the real world you haven't cracked it until you've done a few - the first (insert number) are unprofitable prototypes.

 

[RobinBHM]

Accumulation errors on glazing bars are a bit of a pain. Usually its on pattern 10 or 12 Georgian doors, which have the same issue, all the verticals are in between.

If they are too tight you get bulging stiles / rails but too short and you get gaps on the shoulders / scribes.

Cranking up the sash clamp could result in a row of very wonky bars!

most of our work is double glazed so we don't often do real bars apart from slimlite and then the problem is slimlite only has a 2 or 3mm deduction so its important to get the rebates all the same.