shield continuity with CB-68-LPR

[gavril]

I have an open-ended shielded cable coming from the amplifier of a force sensor. The cable

contains 6 twisted and shield differential channel, 5V, 0V, and AGND. I have screwed the

open-ended side to the CB-68-LPR terminal which is it-self connected to a 6034-Series E

multifunction card by means of the SH-68-68-EP shielded cable. 

sensor------>amplifier------>CB-68-LPR------->6034.

I see that I am picking up lot of interference (50 Hz but also other stuff

like a strong but intermittent  26 Hz signal which I have not yet identified).

I succeeded reducing somewhat 50Hz interference by moving the power line of the PC

but I think that I have another problem with the fact that the shield of the open-ended cable

is connected to nothing. As a matter of fact, it seems that connecting the shielding of the

connecting cable with the connector of the CB-68-LPR helps. 

My question is what is  the best way connecting this shielding to the CB-68-LPR terminal 

or to SH-68-68-EP to have continuity?

I would also appreciate any other suggestions. Regards,

Gabriel

PS: I don't have problem when I connect directly the sensor to the 6034-E.

[JaimeC]

Hi,

Well i dont really know how you ar connecting that amplifier and neither more about that.

The 5V - GND output pins from the card are directly out from the PC and there is no protection for that, so be carefull about connecting any active device to that terminals.

If you are experienciencing 50 Hz noise and you already have a shielded wire the easiest solution would be to implement any kind of filter to a lower frequency.

I dont know how is your input signal, but you can try with a hardware filter or with a programatical one (for example a  low pass to 4Hz) done with labview to improve the quality of your signal.

Regards,

Jaime Cabrera

NI applications engineering Italy/Spain

[gavril]

I have a cable very similar to the SH-68-68-EP but with one open end that connects

the amplifier of my 6DOF force/torque sensor (Nano 17, ATI technology).  As said in my previous

post, I have screwed to wires of the open-ended side to a CB-68-LPR terminal which is it-self

connected to a 6034-Series E multifunction card by means of the SH-68-68-EP.

[F/T sensor] ==> [preamplifier] ==(shielded open-ended cable)==>[CB-68-LPR]==(SH-68-68-EP)==>[6034E]  

I have seen that it is good practive to avoid grounding shield of the cable on both sides to avoid ground loop

currents (NI Tutorial: http://zone.ni.com/devzone/cda/tut/p/id/3344). In this document, Figure 18

shows that the shield should be grounded on the signal source side.

In my case, the amplifier fo the force/torque sensor does not have any alimentation but is, I guess,

powered-on by the 5V-DGND coming from the PC through 6034E cards. I don't see any way

to ground it on the signal source side.

The question is what would be good practice with respect to the shield of the open-ended cable?

Should I try to connect it with the shield of the the SH-68-68-EP? 

If so, how ? (I have the CB-68-LPR screw terminal in the middle)

Thank you.

Message Edited by gavril on 10-23-2006 05:48 AM

[JaimeC]

Hi,

The CB-68-LPR is only a screw terminal (CB: connector block) were you can connect properly the signals, it is only a way to access to the card, so it has the same signals as the card.

If you are supplying the 5V-AIGND voltage from the Card, you can connect the shield of the cable to 0 Volts. If you are not (that means your amplifier-sensor have a different ground) then you should don't do that, because the GND of the card and the GND of the other part (sensor-amplifier) could be different and then you can experience the ground loop. Anyway it is extrange you got better results connecting the shield, because that way you are just making an RSE (reference single ended) measurement, and by the other way you make a differential one which is teorically inmune and rejects common mode voltage.

But anyway, if the 5V-AIGND are the same in the card that on the sensor-amplifier you can connect the shield to the AIGND of the card (or de 68-LPR) without problems.

Hope that helps,

Regards,

Jaime Cabrera

NI Applications Engineering Spain/Italy

[gavril]

Thank you.