Hobbyists and you

I was talking to an engineer at a small company the other day and they were facing a pretty interesting project; I won’t go into specifics but they had a few things which made it a lot more interesting: They had 5 weeks, they were to make ~200, the system had to work reliably, the system was one that one couldn’t source off shelf and the system had to be delivered for sub grand a piece.

Ouch, I really did not envy them in having to deliver that sort of product, I’d love to tackle the problem but not with those sort of restrictions. I started thinking: “Oo I think I saw something like this on x blog” and “this might be a good application for _ sort of technology”, basically I knew somewhere on the internet, someone would have already started solving this problem: almost certainly not a complete solution but ideas and systems to get it started.
So I asked: “Do you ever look at hobbyist projects when you are faced with something like this?” Answer: “Not really, no”. I’m not sure how I would go about this sort of thing half as effectively without having picked up masses of random, and often useless, projects from the web: all made and carefully documented by skilled professionals or kids in their spare time.

Not only do they provide ideas but the solutions are under really good constraints: Most hobbyists don’t have access to massive amounts of fancy fab equipment, they are shooting for low quantity, easy to manufacture. They have basically *no* R&D costs and often not much time. Maybe not a perfect match for your desired end product, but I’m sure you can see where the similarities lie. The problem is that firms don’t document failures and experiments, of course they don’t, there is no reason why they would.

And that is exactly the problem, firms find solutions to problems and discard them for various random reasons, hobbyists too and there is very little communication between the two sides. What I’d like is a world just a little bit more integrated, where smarter designs came about from everyone keeping their eyes and development just a little more open.

–Disclaimer: sorry for the awful title–

Advertisements

Micro-payments

“Traditional media” is (in brief) physical information. We pay for it, not just because we are actually getting matter, but also for the information, however the information has, apparently, started to reduce the amount we actually consume which has been made or printed.

I’m writing this on a computer, probably for a few people to read it on a screen, maybe even on an e-reader or if desired, on paper, I have no idea, but I am not responsible for distributing any of it. In “traditional media”, I would either be paid for this, or I would have no chance of publishing it, as it would have to be printed, but as it is, I do not have to pay anyone, nor do you, my reader to consume my writings.

And I’m quite happy with this system, as I have no desire to force people for money for anything, unlike some (recently The Times newspaper has broken new ground by starting to charge for use of its website) and this has obviously put a lot of people off. We have all got used to paying nothing for information and it seems an affront to have to actually pay for anything. Usually the information we glean from a site is so small that actually we wouldn’t want to pay for it (who knows what kind of thing your search is actually linking to!) but sometimes it yields actual gems of insight.

Possibly what is needed for the future of good quality media and information is a system of paying very small amounts of money (cents/pence) to sites one reads with minimal time delay to the user (possibly built into the browser) while not forcing any particular provider or amount on people.

Elektor (which I have previously criticised for pricing and other bits and pieces) has a pretty neat way of dealing with this: one can buy Elektor credits, for about 70p for 10, which buys you an article, in a lovely pdf, from a previous issue of the magazine. This gives one a cost effective alternative to buying the magazine regularly, without undermining the companies business model.

Of course these are only usable on Elektor, which is a bit of a drawback if one wants to read a well written documents on politics or anything else but I like the mechanics of it…

Open source hardware, software and content all costs something to make, so why not reward the creators with a tiny bit of your hard-earned cash, it might encourage them, or fund some interesting new development…

(See that donate button on the right hand margin? Hit it and donate, go on, And I promise, I’ll write another article for every donation I get, no matter how little an amount…)

The open workshop

I recently purchased the Open bench logic sniffer for debugging various things (I kept finding that whenever I did something, I would think “if only I had more samples on this scope, even though the signals are digital”) and I have to say, I am very impressed with the hardware, as well as the very comprehensive set of PC softwares available for the device. A minor fault I have to say is the lack of live viewing of signals (for testing things really quickly and for testing hookups) however I might well just write a tiny app for that very function. The Bus Pirate tool is one that is very well established now, hitting its 4th revision quite soon and allows very quick, PC based modification and testing of various things (a beautiful list is available here, under chip demonstrations). These tools are beautiful examples of open source hardware, and provide handy bits of diagnosis equipment at low prices (<$50).

The Open bench logic sniffer is based on the Spartan 3E, a pretty low cost FPGA made by xilinx, with 250K gates. Making this kind of hardware out of FPGAs are very popular, even in a commercial context, due to their ability to handle signals in a massively parallel and direct manner, allowing very fast signal paths and therefore sample rates similar to commercial products using more sophisticated hardware. In fact PicoScopeuses Xilinx FPGAs in quite a few of its products (at least according to Elektor’s photos), the scopes are very high performance, however most of the features are placed in software, meaning (I would expect) the scope would get laggy when one is running lots of bits and pieces. Even so, the scopes are very popular. Oscilloscopes are very hard to make, however, due to their required high performance front end, so amateur scopes can hardly be expected to perform with the pro scopes.

However the idea of a set open tools ones workshop or workbench, made professionally but with the software competitive and modifiable for the users good. There are a few tools which pretty much everyone uses, although some of them have higher price form factors which really aren’t far out of the capabilities of open source: PC scopes, bench multimeters, logic analyses, “pin wigglers”, variable loads, supplies, arbitrary waveform generators and simple sensor input (temperature, tachometer). One could also have remote sensors using ZigBee or Bluetooth.

Being as bigger projects dont (really) work in open source (as the hardware costs start to be pretty huge), the chances of filling a full 19 inch rack would be pretty low, and the practicality of doing so would be questionable due to the cost of the mounts. The open bench logic sniffer is about 60 by 100mm (very roughly), meaning a big enclosure would just be wasteful, as well as costly. Most instruments now dont need very large boards if it doesnt have a user interface, and connectors are pretty small. Clearly what is really needed is a metric based rack system, prehaps miniaturized for the new breed of test equipment. I would personally suggest a rack size of around 160mmx160mmx25mm, although prehaps even smaller would still be practical. Power supplies or more complicated systems (computers) could take several units (like 19inch currently). A universal power bus providing low currents for the digital systems would simplify the individual design (5, 3.3 and 12V?) as well as a universal interconnect system with USB 3.0 taking pride of place as the PC interface. If one wanted a stand alone instruments one could screw the power supply and instrument modules together (or put an add-on at the back).

The software that could be created on a computer would allow interconnected equipment for testing and debugging on a new level: peripheral simulation triggered by other events, interconnected triggers, data loggers and configurable displays would completely transform the current level of sophistication. Bring on a massive set of development…

%d bloggers like this: