As so often, I start with the pull from a post by Martin which inspired the thoughts below:

I think the chap that made and sold these Lord uplifts [for saddelbags] probably compensated for any possible material weakness by using thick plate and rods. The uplift and the support that fits to the saddlebag weigh over 400g, nearly as much as the Tubus Fly steel rack.

You're very likely right. Mr Lord had to work with the materials he could get at the time, the processes he could control. But today we have many more choices.

In such small members of a structure the stresses and strains of the load are mainly reacted in the outer fraction of the material in any support, in other words near the circumference of the rods. It is therefore good engineering to use tubes rather than rods, to apply part of the material mass saved to choosing a thicker tube than the replaced rod, and to making triangulated structures or even to spend some of the weight saving in triangular flitch plates. Proper design in tubes rather than rods will save weight without any other sacrifice except to the expected vintage appearance, which in bicycle gear is often an excuse for incompetent engineering or absent skills. For instance small-diameter thin-wall tubes are a pain to weld without destroying them, and a proper high-strength weld requires curve grinding or laser cutting under computer control.*

A VERSION FOR OUR TIMES
However, if I were to design a modern version of a saddle bag support, it would be cut out of a single flat sheet of quite thin stainless steel for durability and appearance. After cutting out the single shape which will be both the support and the fixing arm, it will be folded, so that the main support will look like an upside down tray and the fixing part like a smaller tray folded upwards from the main tray.

The main part after cutting and folding will look from above like a flat rectangle or square with an X through the middle. All edges will be turned down, except the attachment section, which will be turned upwards only were it forms part of the main platform. That will be the last bend performed, the other edges of the attachment tray being turned down at the same time as those of the main tray.

The attachment tray too can have some of the inside cut away. The attachment tray will have the shape of the rails of the saddle from where they turn down at the back until they turn forward for attachment.

All of that is one piece of metal.

Now you need one or both the offcuts from beside the fixing tray, which you will cut and fold on all four sides to fit vertically inside the fixing tray and also inside the saddle rails, open side forwards so that you can get inside the folded sides, which are now doubled up as in the fixing tray and the fitting tray or trays.

The fitting tray folderovers and the fixing tray folderovers are drilled together and bolted together around the upright part of the saddle rails; no bolt or (horrors!)   self-tapping screws visible or, worse, in contact with your expensive saddle bag.

All interior corners from cutting away metal for weight saving must be rounded with an appropriate large radius; these are the places where cracks will start if they are sharp.

The centre of the X on the main platform and on the fitting tray can be cut to a decorative pattern if you like, as long as you don't take away too much metal.

Both trays and also the bolted inside fixtures can be strengthened by cutting up the offcuts, bending them into angle and spot welding or bolting sections of the angle inside the trays where the flanges meet.

Finally, at last a slightly difficult brace to make. By now both the bearing tray and the fitting tray are independently stiff enough for purpose everywhere that matters, except where they meet in a right angle. To provide a buttress brace on top would steal loading space. So a buttress brace will have to be supplied on the back. This means that at least one vertical edge of the metal in the brace must be at ninety degrees to the surface being braced.

The last thing to make is the nearest to a precision job in the whole assembly. Start with two large L-shaped pieces of metal with letters fat enough on the long arm to make a substantial angle, and on the short arm a lesser width of angle. The long side of the L must be longer than width of the platform, and the short end must be longer than the height of the fitting plate. The excess length in each case should be enough to bend another angle so that the support L can be fitted to the existing fold-downs of visible plates without any screws or rivet visible on the carrying surface or otherwise touching the saddle bag.

The flat bearing angles must all be bent from the inside of the L.

On the long part of the L, the angle will be bent outwards to make this crucial additional support as wide as possible. On the short part of the L, there won't be space to bend the angle outward, so bend it inwards towards the centre of the bike.

Now comes the difficult bit. Where you want to arrive is each L-shaped arm supporting the carrier plate on top of the angle of the long arm with a close fit but without any fixings at this stage. Simultaneously the inside corner of each L shape must be as tight the corner at the bend of the main plate where it becomes the fitting plate as possible. Measure carefully, thrice.

We want at both ends of each L to attach the integral angle we'll form in both the surfaces of each of the four ends without pres-tress to the fold-down we put into the flat load bearing and fitting plates.

Now, you have the measure you want to arrive at, from measuring thrice. From this, subtract twice the thickness of the metal you're using, and again subtract twice the thickness of the metal you're using to allow for bending. Don't worry if you're a bit long or a bit short: you can file or ground the metal for a good fit, or fit a felt or metal washer. Just don't get cocky with your measurement so that adjustment needs to be a lot rather than a little.

Now saw the arms of the L's to length, then mark up for the folder, saw or grind out the corner, and fold the two side over each other. Do the same at the other end of the L, then offer up to the construction and adjust for a tight fit at the folder and along the angle bearing. Do the same for the other side. Glue the corners you've just made into the foldovers of the plate, or rivet or bolt.dd

Whether you bolt or rivet or bond these L braces depends on your chosen material but there will still be no bolt-heads in contact with your saddlebag.

lf built in aluminum or laser-laid in engineering plastic, appropriately thicker sections will be needed.

I'll be happy to provide dimensioned drawings if anyone wants to pay for them.

*A. In probably 1980, we returned from introducing our child to the grandparents in Australia, and settled in a house with extensive garages on Hayling Island, between Chichester and the Isle of Wight, because I wanted to supervise the shortening of a Bentley chassis I was turning into a sports car. Imagine my surprise when I arrived to meet the primarius of hotrod metal work, Geoff Jago, the first thing I saw as I walked in the door of the metal workshop (there was also a GRP workshop in a separate building) was a computer with on the screen a design by me for hanging Ted Knight's exemplary Jaguar XJ independent rear suspension very neatly and nearly invisibly, with below it some kind of a gas cutter under the control of the computer. As soon as I recovered from my amazement at seeing such sophistication at a hot rod welder, I held out my hand, palm up, as Geoff was trying to shake my hand, and said, 'I don't recollect giving you a license for using my design. How many of these adaptor plates and crossbars have you sold? Where are my royalties?' Not a small tube example, but today a small laser cutter that you can control with your phone costs a couple to a few hundred quid (Geoff's pathfinder setup back then cost in the tens of thousands), and you can buy engineering grades of plastic on a reel to have a laser under computer control build you anything you can program.

B. Innovation also helps. When I designed my small-tube bike, which I decided on aesthetic grounds should have lugs rather than visible welding, I chose 1mm wall stainless aircraft tubing, and made lugs and internal and external butts by layering inside and outside with graduated short sections of the same 1mm wall tubing in different diameters brazed into the ends of pre-cut lengths of the main tube, with the last outside section fully decorated and the other sections decorated at the showing edge, the whole assembly then to be laser cut in three dimensions to within a tenth of a millimetre, and brazed together with other similarly built up sections. The whole thing would have been remarkably light even in stainless, but my limit is neat soldering joints on large valve-era electronics, with tiny brazing being entirely beyond my eyes or my skills. I spent a couple of years talking to bicycle braziers who promised everything as long as it could be built with tubes off their shelves; they were basically not selling innovation in bicycles unless you wanted a bike in wood or bamboo, which I didn't need them for (a 68ft sailing yacht I designed and built of laminated veneer squares and then thrashed in the toughest oceans around the world is still operating year round in both China Seas over half a century later). Eventually a guy who wanted me to give a series of lectures at an art college to advanced students made some calls for me and I had a choice of art school welding instructors, all but one never having braced anything in their lives (and the one who could braze was a lush), at which point a manufacturing jeweler who had one of my designs in her catalogue suggested that her silver brazier could do the job... It was a question I should have asked first: Who will build it?

***
I have to do my post and sleep, so it will probably be tomorrow before I get around to editing this. E&OE.

On the topic of doing it yourself with aluminum rods and bars, it is my experience that the bar and rod stock that I buy at the hardware store is much more flexible than the metal parts that come on manufactured parts. 

Thus, my home made stuff is often weaker with lower weight capacity than commercially made components when using aluminum.

Maybe. I think the chap that made and sold these Lord uplifts probably compensated for any possible material weakness by using thick plate and rods. The uplift and the support that fits to the saddlebag weigh over 400g, nearly as much as the Tubus Fly steel rack.

I took the first attached photo about a decade ago.  Someone made this bracket to support their bikepacking type saddle bag.

George, that support in your first photo consists of the hoop connecting the two side plates of an original Blackburn Lowrider front rack. The connectors are sourced from an original (Mk1) Tubus Cargo rack (rods and adjustable tilt-plate, around 2007), a reflector bracke, and a piece of aluminum 90° angle. I found the same photo back in the day and made one using the same parts!

It worked well enough in concept and was stable, but the aluminum rod hoop was unnecessarily long and began to show signs of stress at the clamp. Also, it wasn't compatible with my rear pannier rack, so it didn't remain in service long. A good try, though, and top marks to the original innovator for repurposing existing parts!

The "handlebar bag" under-bag support on the seatpost is really clever, George.

When Forum member JulianK kindly gave me his old CCL, he included a Karrimor-style uplift bracket that engaged my Brooks saddle loops. It released quickly and easily, but I largely stopped using it out of concern larger loads could apply a lot of torque to the saddle loops, especially on bumpy roads. That started me noodling for something more solid, like the billet Q/R I described above. I really must spend some time out in my metal shop to bring it to reality!

I bought Thorn's extended T-bar uplift back in...2011, I think. It used machine screws and acorn nuts at either end of the cross and the idea was you could leve the CCL's leather straps a bit loose and so ease them over the acorn nuts when you need to quickly de/mount the bag. It worked but wasn't as stable as I'd hoped.

I also tried the nylon boat cleat trick...bolting the cleat through the bag and into the hardwood dowel. Undo the lower seatpost strap, give it a little wiggle, and the cleat's crossbar slipped between the saddle rails, yet was secure otherwise. Unfortunately, it tended to rattle unbearable on rougher roads and the bag wiggle a fair bit as the cleat's nylon crossbar wiggled back-and-forth despite the lower securement.

Best, Dan.

Carradice saddlebag with Lord uplift. ...
...
... On the other hand, this uplift puts the bag a bit higher up and a fair bit further back, giving me more room for my thighs than with a saddlebag attched to saddle loops and seatpost.
...

I did not like my Carradice bags hitting my legs.  I have made a modificaiton on my bikes where I can use the Carradice straps on saddle bag loops or on my Conquest saddles where there are no bag loops, but springs that I can use as a substitute for the saddle bag loops.

I use a threadless stem, shimmed to fit on the seatpost, and a wood dowel to push the bag further to the rear at the lower part of the bag where the strap is.  I also use a longer strap, in this case use a leather toe clip strap.  First photo below.

And since I am showing photos, I have commented before that I put a piece of coroplast inside the bag to give it more structure and better shape when partially empty, second photo.  The coroplast is simply bent into a U shape, not physically attached to the bag.  Thus, if I have one small heavy item in the bag, it does not sag.  I initially tried paper cardboard, but when I was happy with the result I replaced it with coroplast which retains shape better when damp or wet.

The bag shown is a Pendle.

Final photo. Another view of the uplift without the saddlebag.