Is the digital I/O on a PCI-6259e connected using a BNC-2110 floating?

[Ibbo]

Hi,

I am using a PCI-6259e with a BNC-2110 (on connector 0) and I've created some digital I/O to control and measure from a circuit running off the digital gnd/5V supply.  I need to connect my circuit to a secondary analogue system which is floating.  I need to tie the digital I/O to the floating system so that the digital 0 and 5V float evenly about a single point in that system.  So my first question is, will the digital I/O normally be floating?  If not, is there a way I can get it to float (removing some jumpers on the board maybe)?  Or failing that is there a way using the BNC 2110 and maybe a few extra components that I can replicate what the BNC box is currently doing but making the power and signals float.

Cheers,

Robert

[Andy.G]

Hey Bob,

Im a little confused as to what you are trying to achieve would it be possible to get a circuit diagram of the setup, just a rough paint job will do. Please include any supplies and grounds. Thanks!

[Ibbo]

I've attached a basic diagram of what I need to achieve, it's a little confusing but it is a difficult thing to explain!

I have a non-NI setup which has a ground connection.  I need to tie the outputs, inputs and +5V of the NI BNC 2110 so that the 5V differential floats but is centred on the ground of the second system (ie. there is 2.5V between the Labview Dig Gnd and the 2nd system ground and there is 2.5V between the Labview 5V rail and the 2nd system ground).

[Andy.G]

Hello Lbbo,

The digital signal being produced must have a reference; therefore it will be 5V with reference to the ground of the NI equipment. However once the 5V is produced and is fed into a circuit which has no ground it can be considered floating. However if you decide to half the voltage using say a potential divider, the current will have to diverted somewhere in order to lower the voltage and therefore will need a ground source.  From this point you would have to use some external circuitry to change that voltage into a floating voltage.

[Buechsenschuetz]

It MIGHT be possible to shift the signal level (or the reference level) by using some tricky resistor circuitry but usually such circuits are not very stable since the voltage shift always depends on the current flowing through the resistor(s).

I think the best method would be a real signal level shift circuit. Since your new reference level is halfway between the 0V and 5V level of the NI card, the signal level in your system would be between -2.5V and +2.5V with reference to your system GND. You could shift the signal level with an operational amplifier, feeding +2.5V into the inverting output. However, to yield negative output levels, you need a split supply voltage. If this is not available in your system, there are circuits (and modules) available splitting a single supply voltage into a positive and negative one.

So very probably it will take a bit more than just connecting some wires (and resistors) to shift the signal level the way you need it.

[Ibbo]

I've been advised that I should use some optocouplers to couple the LabView outputs to the floating 0/5V power supply.  Hopefully that will do the trick.  Thanks for all the suggestions.

[Buechsenschuetz]

From your diagram I see that you want the signal level 2.5V below your reference level. Yes, the easiest way in this case might be using an optocoupler. You will also need a 2.5V power supply. You can connect the (+) output to your system GND and the (-) output via a current-limiting resistor to the emitter of an NPN transistor-output optocoupler. The collector should be tied to your system GND. Thus, the emitter of the output transistor will have close to 0V (with respect to your system GND) when activated and close to -2.5V when not activated.