Learning simplified versions [was Re: [cctalk] Re: Omnibus to USB

I doubt that anybody knows the entire vocabulary of any living human language. In a sense it isn't. I guess it's a matter of degree... For example, let's conisder 'buildign a computer' You might start from various ores dug out of the groudn and refine your own materials. You might start with metals, etc and make your own electronci components. Copper oxide transsitors, albeit poor performers, can certainyl eb made at home. So can point-contact germanium transsitors if you have sutiable germanium crystals. I know of nobody who has made a computer startign there, though. You might start from simple electronic components, discrete transistors only. Peophe have certainly done that. You would prboably start by designing and building flip-flop and gate circuits that could easily be reproduced, and use those as the basis for your computer design You might start with SSI and MSI ICs (TTL) You might use an LSI procesosr (a microprocesosr) You night bolt a ready-made moterhboard into a case, plug in the PSU cables, the hard disk, some memory modules, etc. To me, the last is not 'building a computer', it's 'assembling a computer'. All the others do count as building one to me. But if you claimed ot have 'made a processor', I don;t think that soldering a Z80 onot a bit of stripboard is wqutie ewhat is normally meant by that. To 'make a processor' you need to start with something simpler than the processor. And I guess that's how Ifeel about 'buidling a radio' by connecing an RF/detector moduel to an audio amplifier module. The parts are too high level. The other thing is tghat TTL devices are somethign that are/were gernally used. There are excellent databooks for the giving intenral scehatmics for the simpler devices. But the FM radio module you find in an electronics kit is not something you will come across elsewhere. It's probably possible to do things like this, but there are no clues as to how to go about it, and the modules are not that well docuemtned. A true hacker can peroblaby do it, but the kits are not aimed at such people. It ias _amazing_ how few prople truely understnad the 555 timer IC. IMHO there is a differnece between learnign and using (yes,this might partly contradict my idea of not learnign things you have ot unlearn, it's more that ymay have to learn thigns you don;'t actually use). There is probalby nothign wrong with using an op-amp as a black box. Until, that is, you tun into problems fro mthe input current, or the slight offset, or... Or you end up using many mode components than you have to becuase you don;'t realise you cna use a particualr IC in a different way. And aexample of that. The 7474 Dtype flip-flop has acrtive-low direct set and dirct clear inputs. They work like an SR flip-flop. And the behaviour if both are asserted at the same time is docuemtned. Both ! and Q/ go high at the same time. This condition is not latched, if either set or rest goes high and the other remains low, the flip-fliop does what you'd expect. What this means is that if you tie reset low (say, then Q/ is high. But Q will be the inverse of the set input. It has rather more of a porpaction delay than a 7409, but I once used this when I have half a 7474 unused, not other gates fee, and I needed ot invert a slow-ish signal. A lot of people would have fitted anotehr IC. I beleive it is impossible to be totally consistent. And to understand why you reson the way you do. Part of it is 'gut feeling; To me, here, it's the idea that at least when learning you should use thigns considerablly simpler than the device you are making And that you should, if at all possible, learn with real components that you will come across later. I also feel that if you are nto careful, just plugging modules together (which includes things like Arduino boards), although economically sensible in that it cuts development tiem and production coses, can end up with an over-complex and less reliable device. If yopu are elarning you should elaern how to start from first principles. -tony