Without relay, but without Arduin or Fighting in the sandbox

Forgive me readers, but this is a replica of the replica. It all began with an interesting article , where the author introduced us to his ATMega8 based development. Obviously, this stone is redundant for the task that was implemented on it, and the opponent paid attention to it .

But now I have zasverbilo ... um, in the palms, and I could not help but try to insert my five kopecks. I understand that my real note is too big for a comment, but I don’t draw on a full article, but I’ll try.
')
Retelling the content of the previous series makes no sense, who cares - I gave the links. Let's go straight to the point.

The scheme on the relay, proposed by a respected author, caused me some semblance of a toothache. Perhaps because in my youth I had to deal with decadic-step and coordinate ATS, after which I hated the relay for the rest of my life. Yes, relay circuits are easy, yes, they will work even in a nuclear explosion, but ...

The first but - reliability

What do you think are the most unreliable components of electronic circuits? Perhaps I am mistaken, but in my experience these are mechanical contacts and electrolytic capacitors. Contacts burn, oxidize, capacitors dry. And here the whole circuit consists of contacts and capacitors. In addition, in this circuit, polar capacitors are connected in parallel to the windings of the relay, and as a result, when the EMF is broken, self-induction closes through what? - correctly - through a capacitor, which in this case will work with such a poor diode. And that turns out to be connected in reverse polarity. I do not think that it will benefit the so not very reliable component.

The second but the cost

How much does one relay cost? Looking for "Ali", I did not find anything decent cheaper than one American dollar. Okay, even if I do not know how to search, take the price of 50 cents. Four relays (after all, in the original scheme, four channels) - already two bucks. It turns out only one relay will cost more than the entire bundle of options on the microcontroller, including the microcontroller itself. And if in that scheme there is still more extra details ...

Third but extensibility

But tell me: what will we do with this scheme if we need to introduce additional functionality into the product? No, I do not ask to play military marches, but at least squeak a buzzer after the expiration of the time that is given to the player to answer?

Voooot ...

Okay, that was a saying. And now to the point. See the diagram. I quote: "soldering was faster than drawing."



One shareware chip, three resistors, three capacitors. Piezopischalka. And that's all. It is possible to feed this economy from three elements AA. Without any stabilizers. At rest, it consumes less microampere (power down mode) - you can refuse the switch - another hateful contact.

The AVR AT90S1200 controller is not because it’s good for something, I just have one polvedra left after one of the ancient projects, and since it’s been out of production for a long time now I’m not going to stick it anywhere. Only in some home-made. Therefore, shareware. But, as the author of the first article correctly noted, anyone will do here.

Now about the program. Since there is no RAM in this controller (from the word in general), and the hardware stack is only three-level, there are no high-level languages. Only Assembler, only hardcore ...

Put the button on the interrupt also does not work. There is only one external interrupt, and there are six connoisseurs at the table. Well, we will poll in a loop. And to interview all at the same time, so that there is no advantage to one player. The polling cycle time at a clock frequency of 1 MHz is 6 μs, that is, we catch the difference between the presses of 6 μs. If two players press the button at the same time, then all of a sudden, you never know, two LEDs will be illuminated. All honest. Although, I think the probability of such an event will be very low. But still, if the accuracy of determining the leader in 6µs is not enough - you can connect quartz, flash the fusion and start the controller at a frequency of 12 MHz. In this case, the polling rate will increase 12 times. Your KO

Contact bounce does not affect anything, since we react to the very first level drop seen. This is followed by the processing of pressing, and what is already going on there on the button - we do not care.



I agree, different buttons with different time and bounce character can give someone an advantage in a microsecond or two, but, in my opinion, this is catching fleas.


Thanks for attention. If something is wrong - write in a personal, correct.