What’s your approach to methodology?

RobertP said

Sadly, affordable 3d printers haven’t achieved enough reliability for me to use them without becoming an expert. Also, YouTube is chock full of poorly presented information.

I have had this experience too. I bought (a couple of years ago now, discussed it here I think) an amazingly reliable, small extremely cheap machine (an Ender 2 Pro) which has made me a lot of very nice highly precise prints, and all I did to customise it for ages was replace the bed springs with stronger ones and set up OctoPrint. I wanted to just print, and have as little expertise in printers as possible, and that worked out incredibly well for a while with PLA, especially eSun's PLA-ST.

But ultimately I wanted to safely print PETG at the higher end of its temperature scale, and that kicks off a series of upgrades: really need a bimetallic heatbreak to stop the PTFE tube heating up and giving off Death Smells, which in turn requires shorter retractions so really a direct drive extruder. Did the bed levelling sensor at the same time, during the upgrade learned a bit too much about gantry wobble and correcting it, discovered I'd need to add a reverse bowden tube, learned all about split-sleeve wiring tubes, replacing fans, discovered that in fact filament dryers really do help, etc.

So now I have expertise that was not always that fun to acquire. In fact rather stressful, if not particularly expensive. Though that expertise does give me confidence and opens up more machine choices. But now the ability to print hotter PETG, lower-temperature nylons and softer flexibles, etc.

I think it's probably increasingly possible to avoid most of this stuff with affordable off-the-shelf printers for a lot longer, but probably not forever. But then we get back to the acquiring of more expertise than one originally anticipated in CAD software. Because once you can design anything, you start getting clever. 

(FreeCAD reached version 1.0 yesterday as it goes. Still rather a geek lifestyle choice though)

Edited by Unfocussed Mike

Unfocussed Mike said

(FreeCAD reached version 1.0 yesterday as it goes. Still rather a geek lifestyle choice though)

I can’t draw for toffee but I do have a maths degree so it’s OpenScad all the way for me.

RobertP said

Unfocussed Mike said

(FreeCAD reached version 1.0 yesterday as it goes. Still rather a geek lifestyle choice though)

I can’t draw for toffee but I do have a maths degree so it’s OpenScad all the way for me.

(It's a tangent, pun not intended, but it's my thread so ehhh ;-) 

I can't draw either but FreeCAD is not a direct-manipulation CAD package, so it doesn't tax me. (You can draw freehand on the grid in the Draft workbench but I never have!)

FreeCAD and OpenSCAD are actually pretty conceptually similar, it's just one is code-first and the other is not. (Though you can script it). In FreeCAD you can do all the SCAD-type CSG stuff (cylinders, cones, cubes, boolean operations), or you can add constrained sketches (equivalent to 2D shapes in OpenSCAD) and then extrude/revolve/sweep them along a path, cut them out of other shapes etc.

But constrained sketches are a lot simpler than technical drawings or freehand drawings. You just rough out your shape as vertexes and edges, then you specify lengths/distances, tangency, angles, expressions, equality constraints, etc., to lock them down. Simpler and IMO richer than specifying 2D shapes in OpenSCAD because you're only concerned with the facts of the shape, you don't have to think in a co-ordinate space.

(You can open OpenSCAD files in FreeCAD. That way you can get a STEP output. Except for, er, hull() and minkowski(), and maybe one other operation that can only result in a mesh, I can't remember offhand.) 

I have a CS degree (therefore I'm shit at maths, almost by definition!) so I love the idea of code-CAD. I did my first designs in OpenSCAD, and was prepared to learn various libraries like BOSL and get into the geometry. But it doesn't have a two-way CAD solver, only a one-way geometry renderer, which does eventually get in the way.

CadQuery/Build123D and Replicad get a bit closer because the CAD kernel isn't so limited. You can do all the SCAD type maths but then also do operations directly on the resulting vertexes/edges/faces of a previous operation, which is enormously simpler. And you can use other Python (or in Replicad, JS) code.

But the way that FreeCAD is fully parametric means you're not stuck editing big bushy sketches to make changes. You draw simple sketches, and set up designs where changes to a measurement or parameter neatly cascade through the design's 'tree'. I don't think I've used a B-spline yet, and I'm getting into some geometrically quite complex things, like parametric slot helicoids, where the whole thing is driven by a single profile for the slot edge and a few configuration parameters that update the design, including even the number of slots.

Which does circle back to the methodology thing. In the case I mentioned above, I feel the way the person I was trying to help was sort of dodging the methodology meant that it was really hard to help him avail himself of those tools at the end of the process. One has to be thinking in the right methodology throughout.

Edited by Unfocussed Mike

And speaking of methodology and looping back to my thread about coffee stuff the other week...

I tend to be as un-methodical as possible about coffee because I think a lot of "Aeropress lore" is woo-woo. Not that much of it is reproducible, quite a lot of it is Shoreditch top-knot wishful thinking.

But I bought that grinder I mentioned, which is so much more consistent than the previous and this evening it finally dawned on me that what I felt I have been observing over the years -- that darker roast beans are easier/quicker to grind, but counterintuitively harder to press -- is a real, reproducible, consistent thing.

So I googled, and yeah, science. Beans that are roasted for longer are more brittle, smash up easier in the grinder to make finer grinds, so they are harder to press. And that means I can adjust my brew for the beans I am using and improve its sweetness. Which is good, because these darker beans are decaff and it needs the help.

And now my brain is like: ooh, learn more coffee methodology and I am like: NO, BRAIN. STEP AWAY FROM THE POTENTIAL HOBBY. DO NOT RUIN THE SIMPLICITY.

Edited by Unfocussed Mike

Unfocussed Mike said

And speaking of methodology and looping back to my thread about coffee stuff the other week...

I tend to be as un-methodical as possible about coffee because I think a lot of "Aeropress lore" is woo-woo. Not that much of it is reproducible, quite a lot of it is Shoreditch top-knot wishful thinking.

But I bought that grinder I mentioned, which is so much more consistent than the previous and this evening it finally dawned on me that what I felt I have been observing over the years -- that darker roast beans are easier/quicker to grind, but counterintuitively harder to press -- is a real, reproducible, consistent thing.

So I googled, and yeah, science. Beans that are roasted for longer are more brittle, smash up easier in the grinder to make finer grinds, so they are harder to press. And that means I can adjust my brew for the beans I am using and improve its sweetness. Which is good, because these darker beans are decaff and it needs the help.

And now my brain is like: ooh, learn more coffee methodology and I am like: NO, BRAIN. STEP AWAY FROM THE POTENTIAL HOBBY. DO NOT RUIN THE SIMPLICITY.

Edited by Unfocussed Mike

A lot of coffee lore in general is utter woo-woo but ultimately there is a strong underpinning of chemistry and chemical engineering.

I went down that rabbit hole a decade or so ago. It's an interesting hobby but not ultimately that rewarding - pick a blend/roast, finesse your process (whichever one you choose - I'm an espresso guy) to extract the "best" flavour from those beans and then hand it to someone else with a different palette and they'll most likely go 'meh' :-D

I make espresso and espresso based drinks for myself and non-coffee people, which they seem to like but I'd never consider myself a coffee guru.

The Ghost said

 lot of coffee lore in general is utter woo-woo but ultimately there is a strong underpinning of chemistry and chemical engineering.

I went down that rabbit hole a decade or so ago. It's an interesting hobby but not ultimately that rewarding - pick a blend/roast, finesse your process (whichever one you choose - I'm an espresso guy) to extract the "best" flavour from those beans and then hand it to someone else with a different palette and they'll most likely go 'meh' :-D

Well this is more or less where I am at -- like, I know there's more to coffee flavour science than there is to wine-tasting, but I also know that the returns really must diminish so fast in terms of technique with an Aeropress, because it really doesn't have much surface to optimise. People are imagining things that can be optimised and then imagining they have solved them. You can often spot this by the way the jargon becomes more aspirational and imprecise.*

So I look upon it as something that shouldn't need much optimising, because it'll spoil the sort of simple joy of how well the thing works already. I've already got that dark decaff tasting much better with simple changes.

But it's a very interesting lens through which to observe other areas of my life where I might have let methodology and technique weigh down the fun of hobbies in the past. I was sat playing my odd lap-steel guitar (weissenborn type) the other night and realised that the coffee grinder thing had shown me that I should probably just focus my efforts on the one main area of my playing technique that needs to find some more balance (more or less literally), and leave other aspects aside for a while.

Unfortunately it's not much help with my day job; sacrificing precision for feel is no help there.

* Photographers do the same thing. Consider the way the phrase "tack sharp" often seems to mean "I am proud of this expensive thing I bought and I believe it has imparted a quality whose actual quantifiable benefits I will not be measuring empirically"

Edited by Unfocussed Mike