# Pre multiplexer stage/circuit

Discussion in 'General Electronics Chat' started by mcgyvr, Sep 22, 2011.

1. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
4,798
979
go easy still learning..
I'm going to attempt to use a multiplexer connected to my arduino to detect the presence of voltage at the output of multiple channels (up to 20) of a separate device for a testing procedure (basically continuity and looking for shorts). The voltages at the outputs of this device could be either 5VDC or 12VDC or 24VDC or 48VDC depending on which version the device is under test. But all my multiplexer needs is either high(5V) or low (0V).
What circuit could I put at the output of each of these channels to provide a high 5V signal if either 5,12,24, or 48VDC is present and then low 0V signal when no voltage is present?

I'm basically going to sweep through these channels 1 at a time and #1 make sure voltage is present at each channel when it should be and #2 make sure that when channel "x" is on that no other channels are on too which would indicate a short between 2 channels.

2. ### SgtWookie Expert

Jul 17, 2007
22,183
1,729
Here's one way to do it with inexpensive parts:

R3 should really be 12k, 1/4 Watt. 10k will get a bit too toasty with 48v in.

A similar result can be obtained using a quad 339 comparator. That way, the R3 equivalent can be increased significantly. You could also use a voltage divider instead.

File size:
107.4 KB
Views:
26
• ###### Detect 4.5v to 50v.png
File size:
115.1 KB
Views:
26
Last edited: Sep 22, 2011
3. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
4,798
979
Sgt,
I really appreciate your help... but
I'm sorry I missed one very important part (well I didn't think it wouldn't be part of the solution for some reason).. I need to verify that the ground of that channel is present too. So not just the +Vin but also the -Vin is there. (Each channel has 2 terminals the + and the -) I need to verify if someone missed attaching the gnd too.

See my initial thought (before the wide voltage range requirement) was simply to power the coil of a form c relay with each channel and get 5V from the contacts if the relay was energized and 0 if not. That would essentially verify that a each channel was a complete circuit + and -. If not the relay wouldn't energize.

Attached is a very basic representation of my device I'm testing.
I need to verify that each of the Ch- are connected to -RTN hence the need to also verify the gnd is present

Right now the test is literally to take a multimeter and connect it to CH1+ and CH1- and if voltage is present when that switch is pressed go to the next channel.

• ###### DUT.PNG
File size:
29.5 KB
Views:
13
Last edited: Sep 22, 2011
4. ### SgtWookie Expert

Jul 17, 2007
22,183
1,729
Well, that pretty much blows things out of the water. :/

That's quite a large input voltage range, and large percent of change; from ~10% to 100%.

Is there any chance that these connectors could be hooked up backwards? Or maybe pins mis-wired so that you have a "hot" line where a ground should be and vice-versa?

5. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
4,798
979
No chance of backwards connection luckily.
How about if I narrowed it down to only 24 and 48V.. I only do a few of the 5 and 12 so its really not worth spending any time on it if its an issue.

6. ### SgtWookie Expert

Jul 17, 2007
22,183
1,729
Here's another angle:

I've just assumed you have isolated supplies on the tester vs the equipment. It doesn't matter if they are common or not with this one.

Vin is your 0v-48v supply. R2 is just there because SPICE doesn't like it when there is no path to ground.

D1, R1, Q1, and R4 are a constant-current supply. It's not really very constant, but it's good enough for this. When the voltage across V2 gets above ~3v or so, current ramps up to ~5mA, and by the time V2 reaches 48v, current is ~10.5mA. The transistor will be dissipating ~330mW by then, but that's well within the 625mA limit of a TO-93. You don't have to use the 2N2905, but you will need to use a PNP transistor rated for Vceo >=55v and Ic > 10mA. A PN2907A will work, as will an MPSA56. 2SA1145's will work as well; Mouser stocks them:
http://mx.mouser.com/ProductDetail/...GAEpiMZZMtvtNzZ3W%2bLwLX0qU/JmPIAaonjPlLIkwo=

U1 is your typical optocoupler. You don't have to use that exact opto; adjust components to suit whatever you choose - but 4N25's are everywere.

You may need to adjust R1 if your optos' transfer ratio isn't so good. 1.2k will give you about 4.1mA current to get 5v out.

So, if Vin is <~3v or the return is open, Out will be 0v.

• ###### Wide range opto.png
File size:
90.7 KB
Views:
26
mcgyvr likes this.
7. ### mcgyvr Thread Starter AAC Fanatic!

Oct 15, 2009
4,798
979
Excellent/simple idea.. I'll give that one a try and heck its even got LED's to give them something to watch while they are just sitting there now.
And yes my supplies are isolated.

Thanks..I really appreciate the help.

8. ### SgtWookie Expert

Jul 17, 2007
22,183
1,729
The LED isn't there for lighting; it's being used as a cheap low-voltage Zener. The LED will barely come on at 5v, and will only have ~3mA current with 48v in. It won't be terribly accurate as a Zener, but it won't have to be.

That's one of the old Vf=1.7v with 20mA current. I suppose I should give you an update for an LED with a Vf of 2.1v.

If you use an LED with a Vf of 2.1v instead, increase R3 from 120 to ~170 Ohms.