Author Topic: multiplexer problem?  (Read 14681 times)

0 Members and 1 Guest are viewing this topic.

Offline TechGuy

  • Regular Contributor
  • *
  • Posts: 79
Re: multiplexer problem?
« Reply #25 on: October 13, 2010, 09:59:33 pm »
The distance would not be that far, typically less than 100' and no more than say 200'. It would be 18ga solid wiring, and I have a max of 3 transmit wires (2 would be best if that is workable). What I need is as follows. I have 8 switched output channels that are either 24v or 0. I need to take these 8 IO pins and transmit their status across 2-3 wires as described above. This will be read and the ouput replicated on the other end. Keeping the cost down is important. The output side of this will be used to switch 24v relays.

Hope this helps to clarify..

Thanks!

With RS-485 two signal wires would provide bi-directional communications. RS-485 to a MCU with a built in UART is extremely simple. If you look at the four channel relay board I posted earlier, it uses 4 connections Vin (Pwr), Gnd, and two signal lines for RS-485. I used a 8 pin connection so I could run an extra 4 lines, so I could have two separate runs daisy chained. The advantage of RS-485, is that you can daisy chain them so if you want connect more than one node to the same wire run. Perhaps you need to install two units, one at 200 feet and another at 300 feet. You just connect the units in serial. Soure -200 feet- Node A -- 100 feet -- Node B.. You can go on much futher with RS-485, up to about a mile, but you have to drop the baud rate to 2400 bps at long distance runs. The biggest issue, you'll have with long runs is the voltage drop across the power lines.

This webpage:
http://www.csgnetwork.com/voltagedropcalc.html

will estimate the voltage drop across a cable length. I put in 200 feet with 24V with 18 ga, with a 1 amp load and the voltage dropped to 20.4V. That may or may not be an issue depending on if your end use device can tolerate a 4 volt voltage drop.

A better option would be to go with a 48VDC supply and using a point of use DC to DC convert like the LM2576HV to convert the input voltage to 24V at the end-point to ensure you have the correct output voltage. Plus with the higher voltage you can go with a smaller gauge wire. Perhaps even using Cat-5 cable using two or three pair in parallel to deliver power and the one pair for signalling.





 

Offline TechGuy

  • Regular Contributor
  • *
  • Posts: 79
Re: multiplexer problem?
« Reply #26 on: October 13, 2010, 10:04:30 pm »

A colleague of mine has chosen not to use 1-Wire chips because they are expensive and sometimes unavailable from his preferred distributor. He uses Atmel ATTiny microcontrollers instead. He developed a protocol similar to those used in IR remote control units. Since the communication is in one direction only (simplex), it is well suited to low power galvanically isolated systems (only one optocoupler needed, <100 uA). Another possibility is to use some kind of digital current loop (simplex or half duplex). Yet another idea: at least in theory, a signal (voltage) can be superimposed on the existing 24V power, so you'd only need 2 wires!

The problem with 1-wire is distance and noise. 1 wire serial I/O is affected by noise since it acts like an antenna picking up noise. A Differential serial signal is prefferable for long distance connections.
 

Offline Frant

  • Regular Contributor
  • *
  • Posts: 54
Re: multiplexer problem?
« Reply #27 on: October 13, 2010, 10:57:18 pm »

A colleague of mine has chosen not to use 1-Wire chips because they are expensive and sometimes unavailable from his preferred distributor. He uses Atmel ATTiny microcontrollers instead. He developed a protocol similar to those used in IR remote control units. Since the communication is in one direction only (simplex), it is well suited to low power galvanically isolated systems (only one optocoupler needed, <100 uA). Another possibility is to use some kind of digital current loop (simplex or half duplex). Yet another idea: at least in theory, a signal (voltage) can be superimposed on the existing 24V power, so you'd only need 2 wires!

The problem with 1-wire is distance and noise. 1 wire serial I/O is affected by noise since it acts like an antenna picking up noise. A Differential serial signal is prefferable for long distance connections.

I agree. However, if I understood correctly, in this case only 3 wires are available and that includes the power. Therefore, I don't see any possibility of using differential signaling. Well, it might be possible to use the common mode of the differential signal instead of a dedicated power conductor (as in Power over Ethernet), but I don't think it would be practical since it would lead to a more complex design. Current loops have proven to be very robust because they use current signals instead of voltage signals. They are widely used in industrial environments and for long distance connections. The 2-wire 4 mA to 20 mA analog current loop is still the dominant industrial standard for field instruments.
« Last Edit: October 13, 2010, 11:04:37 pm by Frant »
 

Offline brettddglasTopic starter

  • Newbie
  • Posts: 9
Re: multiplexer problem?
« Reply #28 on: October 13, 2010, 11:41:05 pm »
That would be correct, there are 3 wires total. The way it is structured now is that it only switches 1 device on and off. You have a +24v, 1 common (for powering the lighted panel) and 1 switched channel. The switched channel switches either a relay or contactor which is 24v. Now we are adding devices and need more wires. That is what this is about. I cant imagine the contactor or relay needing more than 1A. It has to be inexpensive as running new wire is "inexpensive" but ugly, ill-protected, and sometimes not doable because of where the machinery is located relative to where it is being controlled from. So what we are stuck with is 3-18ga solid cu wires that need to control up to 8 24v/1A loads. Upgrading to better 2 core mic wire or whatever would render this idea useless as we need to use the wire in place or I would just pull 18/8 wire and be done.
 

Offline JohnS_AZ

  • Frequent Contributor
  • **
  • Posts: 499
  • Country: us
    • About.me
Re: multiplexer problem?
« Reply #29 on: October 14, 2010, 04:03:10 am »
Well the first thing you should do is go slap the snot out of whoever built the thing with bell wire in the first place.  :)
I'm either at my bench, here, or on PokerStars.
 

Offline brettddglasTopic starter

  • Newbie
  • Posts: 9
Re: multiplexer problem?
« Reply #30 on: October 14, 2010, 05:00:33 pm »
yeah, would love to but these are 30+ year old equipment and they figued that they will never need more than one control channel for this. Kiinda like we will never ever need more than 512k of ram or no one will ever fill a 10mb drive.  :P

Funny where technology takes us.




 

Offline tycz

  • Regular Contributor
  • *
  • Posts: 99
Re: multiplexer problem?
« Reply #31 on: October 15, 2010, 12:29:11 am »
I am trying to come up with the easiest way to do simple 24v dc switching over 7 channels using only 2-3 max for transmission wires. Would a multiplexer/demultiplexer do what I am asking? Is there a simple way of doing this?  ???

Thanks in advance for any help

Can't you buy a device like this? I've got one sitting on the shelf here, though it looks pretty old.

8 inputs, 8 outputs, serial connection... The person who gave it to me called it a multiplexer.
 

Offline brettddglasTopic starter

  • Newbie
  • Posts: 9
Re: multiplexer problem?
« Reply #32 on: October 19, 2010, 06:16:13 pm »
that is what I thought it was suppose to be. mux/demux would take the inputs, convert them into a signal then output them via serial where it would be picked up by the demux and then reversed. I would be surprised if this didnt exist in a simple chip of some sort. Seems there would be a lot of uses for this. well.. a lot of applications with one use I guess would be more correct.

So am I wrong on the mux/demux thing? seems there should be an easy cheasy way of doing this. Is this what the DS2408 is all about?

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf