That more or less defeats the purpose of the idea. Almost the entire machine is cast from melted-down scrap metal and then (variously) machined, bootstrapping from the first tool, the lathe, which is hand-finished.
Does it presume you have like, scrapers and a surface plate and so on?
I can't see someone making an accurate lathe or mill without an expensive suite of metrology tooling (or going full whitworth three plates method and lapping like forever).
It's hand scraping all the way. For initial standards, the author recommends just going with the straightest thing you have in your shop, then just scrape away until things mate well.
The books are light on the theory of precision and accuracy, but I would be surprised if most readers of Gingery books didn't also have a PDF of 'Foundations of Mechanical Accuracy' somewhere :). These days nothing about this process is practical, it's all basically an exercise in bootstrapping.
Duckduckgo for 'gingery lathe' if you're interested in seeing results, people have been building these for decades.
Float glass is pretty flat, but I'm not sure regular glass is temperature-stable enough for the task (and of course, it's fragile). Temperature-stable glass (e.g., Pyrex) usually requires a lengthy annealing stage.
Granite surface plates are pretty popular, though. Maybe glass would be okay for light duty.
Speaking of glass, amateur telescope makers can grind mirrors accurate to within 1/4 wavelength of light, using nothing but abrasive powders and patience, but they start out with a pre-cast and annealed Pyrex blank.
I do use a piece of float glass on my 3D printer bed, and it works well enough. Maybe I'll see if I can figure out how much it distorts when heated (not enough to cause a noticeable problem with the relatively low-precision stuff I've been making thus far).
Some people do use borosilicate glass (like Pyrex) for this, but I've been satisfied so far with a piece of regular float glass window stock I had cut to size at Lowes.
There are glass compositions that are very thermally stable, this is a very well researched area because so many pieces of the science puzzle depend on it.
No, but it can get you a pretty good starting point to make something that is really flat. For instance, three pieces of float glass alternated using successively finer grains will get you insane levels accuracy in a relatively short time. To the point where you'd need a laser interferometer to detect where you still have high spots.
You know, I've never really thought about it, but I would think that grinding two same-material surfaces together with an abrasive compound would result in flatness if the pressure is even.
With two surfaces you can still have some bias (convexity/concavity) even though they appear to be relatively flat, you need three surfaces for absolute flatness. See the Whithworth Method described earlier in this topic.
He uses a piece of glass on plywood, with a ground file for a scraper. He's only scraping aluminum.
I don't think Gingery wrote these books as a way to demonstrate bootstrapping. When Gingery wrote these books there were no Chinese machines available, and hobby lathes (say the Southbend 9" or the myford 7") were way out of reach for a lot of people. He wrote the books to let people get to a machine with what they had. Everyone would be able to find a piece of glass and a file.
I built this lathe, but with much heavier sections and out of ZA12 rather than aluminum. I did a pretty good job, I think, but it is still inferior in every way to a Chinese 7x12. Zero regrets that I did it though, even with those better machines available.
I don't think too many Chinese factories are scraping in their ways on their mini mills or lathes.
And also lots of old machines that likely have lost their flaking and true trueness a while ago. Still being run by old operators that understand their limitations.
You can make accurate parts on an 'inaccurate' conventional machine.
The slightly nicer ones have scraped mating surfaces. The very cheap ones will at a minimum be surface ground very well, and have some oil scrapes.
> You can make accurate parts on an 'inaccurate' conventional machine.
To a point, sure. At some stage, it doesn't matter how much of an old timer you are and how much knowledge of your machines' limitations you have - you just need to make a part the machine cannot.
You can basically always rescrape a machine, give it new gibs, and be back to 100%.
IIRC, wood chisel and 3 plates. Prussian blue to evaluate the thickness. Portland cement, a hairdryer and crucible. I made tongs with strap iron and a drill. hairdryer on the charcoal to melt the aluminum.
I started with sand casting aluminum cans. I realized I was woefully unprepared to deal with molten metal, if something went wrong. I wasn't willing to iterate at the bottom, and buying the next level of tech seemed like cheating. Still have the books though.
Surface plates are actually one of the easiest tools to bootstrap, using the three-surfaces method and lapping with fine grit. The only three surfaces that can all slide laterally against one another are three perfectly flat surfaces.
Here's a video (four parts) of a guy doing it. He uses modern tools in the video, but really that's just about saving time (and showing off). You can simply subsitute (a whole lot) more time spent lapping for the lathe part in his video.
