From: Jonathan Kirwan
Subject: Re: circuit design help
X-Newsreader: Forte Agent 1.92/32.572
NNTP-Posting-Date: Sat, 19 Oct 2002 07:20:38 GMT
Organization: AT&T Broadband
Date: Sat, 19 Oct 2002 07:20:38 GMT
On Sat, 19 Oct 2002 04:46:46 GMT, chris bergeron wrote:
>I'm not sure if this is the right place for this or not. If it's not,
>please let me know (I'm not a spammer!) and I will go elsewhere. I need
> help designing a circuit. I have a timing design (which is attached).
> I'm willing to pay nicely for help in desiging this circuit. It seems
>like a simple circuit; but as a novice hobbyist in electronics, I'm
>stumped at how to go about it. I've toiled with using a 555/556 and or
>a 7447 (or maybe it's a 7474). Anyway, I obviously don't have the skill
>to design this circuit. I know a decent [small?] amount about
>electronics (transistors, relays, ICs, etc); enough to attempt the
>circuit and fail. Hence my posting here.
>Attached is a [relatively] simple circuit that I'll pay decently for
>help desiging. I've tried, but failed; and as such, I know when to ask
>Thanks to anyone that can provide assistance!
Chris, my own preferred approach to this would be to use a microcontroller. Mostly, because I'm very familiar with coding them up, have lots of tools handy for doing that, and because it's very easy for me to make changes if I discover something else I should have thought of. And the resulting hardware design for something along these lines is relatively more immune to application changes. It also uses very few parts.
You need three outputs, all 0V or 5V. Your input is 0V or 12V. I'm not sure if that input is car-voltage related, but if it is some amount of conditioning may be in order. That can be approached to a degree in software, if you choose that route. But good conditioning will require some hardware.
Assume a micro, for now. You need +5 and gnd, 3 digital outputs, 1 digital input, and 2 analog inputs. That's 8 pins. Turns out, the PIC12F675 might be a good choice here. It's $2.08 in ones from Digikey, $1.29 in 25s. It can use the internal oscillator which is, over the entire temperature range and voltage range, accurate to 5%. That accuracy seems adequate for you, given what specs you do provide. At 5V only, the accuracy should be somewhat better.
It's easy enough to use digital inputs for reading potentiometers. But the 12F675 has analog inputs, so you may as well use them. Saves some external caps, too. So, I'd wire it up thus:
+5 12F675 / /
| ,-----, ,-->\10k '-->\10k
'---|1 8|---gnd | / | /
| | | | | |
A <------------|2 7|--------' gnd | gnd
| | |
B <------------|3 6|------------------'
C <------------|4 5|---,
The values of the potentiometers can be anything from 1k to 100k, I'd guess. The input current of the analog inputs is "to be determined" on the data sheet, but it will be in small microamps, I think. In any case, your application isn't too critical so perhaps even the 100k would be okay. 10k would be just fine.
It doesn't get much simpler, circuit-wise. You may want to use a MOSFET switch and pull-up instead, to convert X. But it's likely the simple resistor will be okay for hobby work.
Of course, there *is* the programming to be done. If you are looking for a no-software solution, I'm sure someone else will post an excellent answer. That solution would have smaller phase delays for X to B, for example. But that doesn't sound all that important to you. It would also not require special development tools, such as a chip programmer. That would be a big plus for many, in fact. But I'm just suggesting an alternative approach for you to consider.