Post

More people should build stuff with vacuum tubes. They’re easier to understand on a conceptual level than semiconductors, and it’s absolutely worth learning the limitations of things from the past at a fundamental level, plus there are still millions of them in circulation

Similarly, it was much easier to get my ham license than I thought it was, more people should do it! But be careful about sharing your call sign unless you’ve registered with a PO Box, because the FCC database is fully public, including your address and phone number

Vintage computing, while rewarding, can be hazardous; not only do you have to deal with leaky batteries and the occasional exploding tantalum capacitor, but mice and rats often make a real mess of things and especially in the American Southwest, this carries the risk of hantavirus

If you want a really good crash course in analog engineering, start building analog synthesizer parts; you will gain a much more visceral understanding of how transistors really work by making a log converter (for 1V/octave) with a BJT than any amount of book exercises

Also, this is not an opinion so much as a fact that is not communicated very well in basic electrical engineering books and courses, but: BJTs are voltage-controlled, not current-controlled, and (here’s the opinion part) teaching them initially as current-controlled does a lot of lasting damage

Speaking of tubes, it’s worth learning some of the military and related part number equivalents for popular tubes; a lot of better and more durable versions of, say, the 12AT7 are out there for a fraction of the price because they don’t have “12AT7” stamped on them and the audiophiles aren’t looking

Medium Dave Lillywhite@cursed.monster

More people should build stuff with vacuum tubes. They’re easier to understand on a conceptual level than semiconductors, and it’s absolutely worth learning the limitations of things from the past at a fundamental level, plus there are still millions of them in circulation

Learning what to read in datasheets is an art and takes practice. Op amps are especially tricky, because we like to treat them as ideal devices, but even nominally similar ones (e.g. the TL072 vs the 5532) can have huge differences in bandwidth, input impedance at different frequencies, etc.

Radar is incredibly exciting because there are like twelve different ways of trying to interpret the pulses that come back from the pulses you sent out, all with different merits, but the jobs that involve figuring it out do not pay as much as they should, frankly

Radio stuff is regarded as wizardry, and sometimes it still feels like that (especially when you get into at atmospheric effects), but it’s rewarding to learn about electromagnetic waves and how they interact with the world around us, plus you can call Elon a poser when he talks about Maxwell’s Laws

3 reposts50 likes

Circuit board autorouters are seldom useful; they require massive amounts of attention and tweaking to get just okay results and in the end you spend as much time or more than you would have doing it properly by hand, which is part of what biases me against a lot of generative AI, which is similar

For those who got excited about vacuum tubes but don’t know where to start: go find a copy of the ARRL handbook from the ‘50s to the early ‘70s (the latter being great to see the transitional transistor circuits), they’ll set you up right

For everyone who really, REALLY wants to get into the weeds of how transistors work, The Physics of Semiconductors (Sze et al) is a great one, newly revised ($$$) but older versions are available on ThriftBooks and similar for good prices from time to time

I’ve always wondered what prevents autorouters from being good. (I say this with genuine curiosity. I’m not nearly hubristic enough to think that means it’s easy to *make* them good)

Same thing that prevents generative AIs from being good; they are “throw spaghetti at the wall until it gets approximately right machines” that can’t actually understand what they are doing, so they are only as effective as the constraints you give them and are unable to learn from mistakes

Though I also have to point out that using encryption on amateur bands is a big no-no (with some very small exceptions for craft control), so encrypting a Baofeng is only a good idea on other licensed bands like DMR