Elliot’s back from vacation, and Dan stepped into the virtual podcast studio with him to uncover all the hacks he missed while hiking in Italy. There was a lot to miss, what with a smart meter getting snuffed by a Flipper Zero — or was it? How about a half-gigapixel camera built out of an old scanner, or a sonar-aimed turret gun? We also looked at a couple of projects that did things the hard way, like a TV test pattern generator that was clearly a labor of love, and an all-transistor HP frequency counter. More plastic welding? Hey, a fix is a fix! Plus, we’ll dive into why all those Alexas are just gathering dust, and look at the really, REALLY hard problems involved in restoring shredded documents.
Check out the links below if you want to follow along, and as always, tell us what you think about this episode in the comments!
Episode 223 Show Notes:
News:
What’s that Sound?
If you think you know what this week’s sound was, enter here!
Interesting Hacks of the Week:
- Using FreeCAD To Replace OEM Parts
- Flipper Zero “Smoking” A Smart Meter Is A Bad Look For Hardware Hackers
- A 489 Megapixel Camera For Not A Lot
- Do Not Attempt Disassembly: Analog Wizardry In A 1960s Counter
- Arduino-Powered Missile System Uses Ultrasound To Aim
- Good Vibrations: Giving The HC-SR04 A Brain Transplant
- Measuring Air Flow With Ultrasonic Sensors
- El Cheapo Phased-Array Sonar
- Octosonar Is 8X Better Than Monosonar
- Dual Sensor Echo Locator Gives High Accuracy At Low Cost
- I forgot one of the best phased-array explainers of all time: Bend It Like (Sonar) Beacon With A Phased Array
- Recreating An Analog TV Test Pattern
Quick Hacks:
- Elliot’s Picks:
- Dan’s Picks:
Not seeing it on Google Podcasts.
Episode 222 is latest one on there.
Thanks! I don’t see any of them over at Google Podcasts: https://podcasts.google.com/search/hackaday WTF?
Will have to dig into this.
Yes. Searching for the term “Hackaday” or “Hackaday Podcasts” returned only results from other podcasts that referenced Hackaday.
I could only see the list of existing episodes by clicking through my subscription entry in the Android app. At first, I thought y’all must have pulled out of that platform.
I did get the notification and downloaded the episode successfully. I guess you kicked it in the right place.
The thing about CAD/CAM is that you kinda want to model your parts the way they’re supposed to be manufactured.
If your part is made out of thin sheet, you don’t want to start modeling it by simply projecting intersecting sketches, but instead you start out by defining a sheet and then add cuts and folds. If your part is to be machined out of a billet, then you start with a box of material and subtract your sketch. If the part is to be 3D printed or cast, then you have a little more freedom in how you define the shape. In any case, your workflow should reflect the part that is being made, which makes it easier to spot errors, like creating a pocket which can’t physically be milled out, or a fold that is impossible to bend. That’s why the different workspaces exist.
However, compartmentalizing the program into separate workspaces, especially in FreeCAD but also in other tools like CATIA, is that each one has their own toolsets, each follows their own logic, and some tools that could be used crosswise (e.g. why not mill something and then fold the rest?) may not exist or are implemented differently with different features available within a workspace. In the worst case, one thing named such and so means a completely different thing in another workspace. In the end, it’s not a cohesive unit but simply multiple simpler or half-implemented CAD tools mashed in one – which mostly reflects how a program like FreeCAD came to be in the first place. It offers both too much of everything, and too little where it counts, like an old greasy van with random crusty tools thrown in the back – the 13 mm spanner is always lost.
What you want instead is a nice organized toolbox with tools that are not like that weird bearing puller that you made once in a hurry out of piece of angle iron, that kinda sorta does the job while chewing up all your sockets.
“not a cohesive unit but simply multiple simpler or half-implemented CAD tools mashed in one” I think that’s basically the gist of Thomas Sanlanderer’s complaints as well. (Although I think “half implemented” is less true now than in the past. The workspaces that matter to me are all pretty solid.)
I think all of the workspaces play (reasonably) nicely together, but having so many of them is confusing. The “part” and “part design” workbenches are confusing at first — which do you use?
But then “part design” is much better tailored to subtractive design, and “part” to additive, just as you said, so maybe merging them isn’t reasonable? I’m not going to “create a torus” if it’s going to be milled out later. That’s just asking for trouble.
As it is, I could live entirely happily with a FreeCAD that had only “part design”, “sketch”, and “path” for CAM. But it doesn’t bother me to see the rest of them either.
The so-called topological naming problem is the big elephant in the FreeCAD room, still. It’s a bummer to have to adapt your design process to what amounts to a software deficiency. (It’s not a show-stopper, but a bummer.) My guess is that they’ll hammer that down in the next year or so, though. That, plus a good assembly package, will make a huge difference — I’ll kick the tires when they do, and let you know if it’s worth a second look.
>so maybe merging them isn’t reasonable?
My point was that they should be. They share enough similarities, and there’s no reason you can’t both add and subtract for the same part: cast and mill, print and mill, mill and then weld in a tab… negative extrusion is the same as milling or drilling.
>I’m not going to “create a torus” if it’s going to be milled out later.
Sure, but that’s a special case, and then sometimes you just have to make it that way.
Also, there’s powerful stuff you can do more easily if you have the same tools available in a single same workspace.
For example, I’ve done folded designs where the basic structure is bent out of sheet, but then I need a hole that goes right through the entire part to fit a metal bar that will be added later. I simply model the metal bar as a solid cylinder and subtract it from the folded part as a last step.
That’s a “virtual operation” that doesn’t match any real production step, but it reflects back to the first step where the sheet is defined. When the part is unfolded, the holes appear in the right places in the flat sheet as if they had been planned ahead. If the two operations were done in different workspaces that can’t talk to each other, it’s a no can do.
However, you have a good point in tailoring the workspace to definite needs.
Much like folded designs have their special considerations that model how the sheet will actually bend given the material you use, 3D printed parts have to mind things like overhang, cast parts should mind their draft angles, so they may as well be handled as special cases within their own workspaces. Part and Part Design would be cluttered by having these special features, yet all the other workspaces would benefit having their tools in addition to the special stuff.
That points towards a tree structure, where your root design tools are the additive and subtractive methods, and then you branch out into the folded, printed, cast, etc. methods, not as separate workspaces but different “modes” of the same workspace.
Dan, regarding the Alexa voice assistant, I don’t know, being able to make wheat thins and mayonnaise appear at your door by just saying the words has some value. Even if it’s a little weird.