BLUF or TL;DR: ?could a semicircular laser welding tool head weld one half of a tube-to-tube joint at any angle, or is there a simpler way of automating the fixing of such a joint
1. Paradigm
2. Intent
3. Question
4. My initial thoughts
1. Paradigm
Anyone who has read my posts will know I'm not an engineer; my apologies in advance for any incorrectly used jargon.
While I earnestly believe Barnaby Wainfan's composite-sandwich personal-air-vehicle vision, derived from Jack Northrop's philosophy that if it's in the airflow, it had jolly well better provide lift, represents the future I'd like to see for non-aerobatic GA (among other things), in the present, most of the extant light aircraft, that have profoundly impressed me philosophically, have had tubular steel space frames (specifically @bwainfan 's Facetmobile and @Robin Austin 's VH-SRS).
I'm a scientist by trade, and thus doctrinally inclined to 'believe' in reality defined by statistical likelihood. I find when I consider manufacture, I subscribe to views espoused by Shewart and Deming and other folks who believe that the 'best' quality control happens during the planning (and statistical/actuarial analysis) of how production will occur. At the risk of sounding glib, "make it easy to be excellent". I think that Statistical Methods from the Viewpoint of Quality Control are best applied to systems understood as fully as possible by the user. From the point of view of an individual home-building an aircraft for his or her own education then, they are likely to be 'open-source' systems, so that the individual has as full access to the 'source code' (or, in the case of hardware, the plans) as possible. This is not an absolute of course, as many aircraft plans will, effectively, be purchased by someone agreeing not to copy them without license.
I'm also fascinated by production techniques and technologies that - just like Facetmobile and VH-SRS, through incredible amounts of hard and visionary work - are beginning to show that tremendous capability could end up at the disposal of an adequately motivated and diligent amateur, such as Annin Robotics' AR4 open-source robot arm.
Many years of sterling work have produced software to help plan welds. Alas, I am not aware of any open-source machine-vision projects specifically intended for laser welding automation, and I suspect I would be foolish to hope for open-source plans for a laser-welding tool head, for it is a thing unlikely to be attempted by a motivated amateur due to safety constraints.
2. Intent
I'd love to see tubular metal space frames built to the highest standards available to everyone who can order ready-coped tubes online. Laser welding tools are becoming cheaper, and might further push that into the hands of the motivated amateur.
For a given thickness of metal, formed into a tube of known diameter, and coped to abut - at a known angle - another metal tube of known thickness and diameter, the path in space of the joining surfaces - around the intersect of the tubes' long axes - is known, for it is a function of these aforementioned variables.
Presently they are hand-welded, but this requires skill, and if I have the choice of spending many hours building a skill to use once, or many hours developing a tool which I might hand down to others after me, I am minded to at least consider the latter.
3. Question
What's the simplest way for an amateur to automate, to the greatest degree possible, the permanent joining of these tubes?
4. My initial thoughts
In a fantasy world, I would have a multi-armed clamp that - when clamped to each of the tubes, and with the intersect of the clamps' arms positioned over the intersect of the long-axis of the tubes, would deploy a tool that welds the tubes together. I would hope it could detect the tube diameter from having been clamped to it.
It would of course have to be 'told' the material, the metal thickness, and nature of the coping (in an ideal world from the tool designer's point of view: tube A would be uncoped, tube B would be minimally coped as allows it to abut tube A, tube C would be minimally coped as allows it to abut tubes A and B, and so forth).
The tool would then project an arcuate limb over the abutment it intended to weld, and laser-galvanometers would direct the beam down onto the target.
Now all that might be a bit over-ambitious: It is more realistic to expect we jig the tubes in opposition, and a robot arm +- laser-galvos +- machine-vision-fine-correction maneuver the beam in accordance with our needs.
There are plenty of reasons NOT to do this, but there are also plenty of reasons not to build one's own aeroplane, yet the childhood dream persists, and some hardworking, dilligent and lucky heroes, manifest it in reality. My aim in thinking about this is to find out ?are any of those reasons *good enough* to outsource the task.
I would treasure any thoughts anyone has about this question. No plan survives first encounter with the enemy's main strength, but being practiced in planning and being familiarised with as many variables as possible helps us adapt swiftly. I believe thinking about the current state of play helps us consider how we might best exploit whatever tool appears tomorrow.
Here is a link to real WWII RAF banter for anyone who gets this far and still has the monty python sketch stuck in their ears
(content warning: safe for work, but it is ultimately a recording of fellows at the pointy-end of a kill-chain).
**a note to any moderator: My thanks for your work keeping all this spinning. I beg your pardon; the early drafts were focused on laser-welding tools, and as I evolved the question it became clear that it had to arise from first principles and must thus be a more abstract question that leans toward tool development. I hope discussion will stay narrowly focussed upon tools, but if you feel now or at any time that the tool subforum is not it's place, then I welcome you cataloguing it as you see best.**
1. Paradigm
2. Intent
3. Question
4. My initial thoughts
1. Paradigm
Anyone who has read my posts will know I'm not an engineer; my apologies in advance for any incorrectly used jargon.
While I earnestly believe Barnaby Wainfan's composite-sandwich personal-air-vehicle vision, derived from Jack Northrop's philosophy that if it's in the airflow, it had jolly well better provide lift, represents the future I'd like to see for non-aerobatic GA (among other things), in the present, most of the extant light aircraft, that have profoundly impressed me philosophically, have had tubular steel space frames (specifically @bwainfan 's Facetmobile and @Robin Austin 's VH-SRS).
I'm a scientist by trade, and thus doctrinally inclined to 'believe' in reality defined by statistical likelihood. I find when I consider manufacture, I subscribe to views espoused by Shewart and Deming and other folks who believe that the 'best' quality control happens during the planning (and statistical/actuarial analysis) of how production will occur. At the risk of sounding glib, "make it easy to be excellent". I think that Statistical Methods from the Viewpoint of Quality Control are best applied to systems understood as fully as possible by the user. From the point of view of an individual home-building an aircraft for his or her own education then, they are likely to be 'open-source' systems, so that the individual has as full access to the 'source code' (or, in the case of hardware, the plans) as possible. This is not an absolute of course, as many aircraft plans will, effectively, be purchased by someone agreeing not to copy them without license.
I'm also fascinated by production techniques and technologies that - just like Facetmobile and VH-SRS, through incredible amounts of hard and visionary work - are beginning to show that tremendous capability could end up at the disposal of an adequately motivated and diligent amateur, such as Annin Robotics' AR4 open-source robot arm.
Many years of sterling work have produced software to help plan welds. Alas, I am not aware of any open-source machine-vision projects specifically intended for laser welding automation, and I suspect I would be foolish to hope for open-source plans for a laser-welding tool head, for it is a thing unlikely to be attempted by a motivated amateur due to safety constraints.
2. Intent
I'd love to see tubular metal space frames built to the highest standards available to everyone who can order ready-coped tubes online. Laser welding tools are becoming cheaper, and might further push that into the hands of the motivated amateur.
For a given thickness of metal, formed into a tube of known diameter, and coped to abut - at a known angle - another metal tube of known thickness and diameter, the path in space of the joining surfaces - around the intersect of the tubes' long axes - is known, for it is a function of these aforementioned variables.
Presently they are hand-welded, but this requires skill, and if I have the choice of spending many hours building a skill to use once, or many hours developing a tool which I might hand down to others after me, I am minded to at least consider the latter.
3. Question
What's the simplest way for an amateur to automate, to the greatest degree possible, the permanent joining of these tubes?
4. My initial thoughts
In a fantasy world, I would have a multi-armed clamp that - when clamped to each of the tubes, and with the intersect of the clamps' arms positioned over the intersect of the long-axis of the tubes, would deploy a tool that welds the tubes together. I would hope it could detect the tube diameter from having been clamped to it.
It would of course have to be 'told' the material, the metal thickness, and nature of the coping (in an ideal world from the tool designer's point of view: tube A would be uncoped, tube B would be minimally coped as allows it to abut tube A, tube C would be minimally coped as allows it to abut tubes A and B, and so forth).
The tool would then project an arcuate limb over the abutment it intended to weld, and laser-galvanometers would direct the beam down onto the target.
Now all that might be a bit over-ambitious: It is more realistic to expect we jig the tubes in opposition, and a robot arm +- laser-galvos +- machine-vision-fine-correction maneuver the beam in accordance with our needs.
There are plenty of reasons NOT to do this, but there are also plenty of reasons not to build one's own aeroplane, yet the childhood dream persists, and some hardworking, dilligent and lucky heroes, manifest it in reality. My aim in thinking about this is to find out ?are any of those reasons *good enough* to outsource the task.
I would treasure any thoughts anyone has about this question. No plan survives first encounter with the enemy's main strength, but being practiced in planning and being familiarised with as many variables as possible helps us adapt swiftly. I believe thinking about the current state of play helps us consider how we might best exploit whatever tool appears tomorrow.
Here is a link to real WWII RAF banter for anyone who gets this far and still has the monty python sketch stuck in their ears
(content warning: safe for work, but it is ultimately a recording of fellows at the pointy-end of a kill-chain).**a note to any moderator: My thanks for your work keeping all this spinning. I beg your pardon; the early drafts were focused on laser-welding tools, and as I evolved the question it became clear that it had to arise from first principles and must thus be a more abstract question that leans toward tool development. I hope discussion will stay narrowly focussed upon tools, but if you feel now or at any time that the tool subforum is not it's place, then I welcome you cataloguing it as you see best.**
Last edited:
