Multi-pin collecting device

[mentler2]

I'm looking for some sort of device that has about 60+ pins. The pins would sense current, and the device would output some other marker on a single channel. Voltage would work nicely. So, say a current comes in on pin 1, the device would output 1V. A current comes in on pin 2, it would output 2V, etc. Does anyone know if something like that exists?

 

[IRstuff]

So, pin 60 would output 60V? That's a bit impractical, given that most systems would want digital data anyway, so why generate analog from digital data only to have to digitize it back to digital?

TTFN
faq731-376

 

[mentler2]

You're right, 60V is impractical - I was only using it as an example. Digital output is fine, as long as a unique binary number can be assigned to each pin.

 

[VEBill]

Sounds like a resistor network (assuming it's one at a time).

 

[VEBill]

...but 60+ pins might be impractical.

 

[LiteYear]

As VE1BLL says, a simple resistor network might suffice, provided:

* The input current magnitude is fixed and known.
* Only one input is active at once

Something like the attachment.

http://files.engineering.com/getfil...b81d-9404f80cfc81&file=analog_multiplexer.png

If you want a more flexible input, I'd be looking at a digital solution with a multiplexer on the input. A counter would cycle through each of the multiplexer inputs and a latch would output the counter state when current is detected on a multiplexer input.

 

[RajeevCell]

A priority encoder?

 

[IRstuff]

You have given woefully indequate information to accurately assess your problem:
> Are more that one input active at any time?
> How long are they active?
> How long are they inactive?
> What level of current?
> Is it really a current, or is it a voltage?

TTFN
faq731-376

 

[mentler2]

Sorry for the poor information. I was still trying to figure out specifics regarding how the circuit would work when I started the thread.

The inputs are all active, and the signal that I want converted would be a half-second period of inactivity, from some percentage of the inputs. The percentage would change with every signal, though if the inputs are ordered sequentially, the system could work by identifying the highest deactivated input, and then outputting the binary number assigned to that input. The circuit isn't finalized yet, so the signal could be a voltage or current, though I'm currently leaning towards current.

After speaking to a friend, I think that 64 NAND gates and 2 8-to-3 digital encoders might work. Once I study the data sheets I should be able to hammer out a few more specifics.

Sorry again for the poorly presented problem. My experience with circuit analysis is limited to a year of university classes analyzing existing circuits, not designing them. If anyone has better ideas, or even criticisms of the current plan, I'd love to hear them. And if anyone is curious about any other specifics regarding the circuit I'll try my best to answer them.

Thanks!

 

[IRstuff]

Nope, you're going to need way more than that. You need enough gates to encode your 60 inputs into hexidecimal or BCD, ala:

http://www.utm.edu/staff/leeb/DM74185.pdf

which requires ten chips for your application. Since each digit has a unique binary representation, the resultant 10-digit BCD number fully describes all the inputs that are high and all the inputs that are low at any given instant.

TTFN
faq731-376