10-23-2009 06:18 AM
Hi,
I am using Labview 8.5.1 for passive beamforming with alternative actuation of 5 piezoceramic sensors. I am using the SCXI-1336 to configure the SCXI-1129 high density matrix as 16x16 (2-wire) matrix. Before this I was manually switching the cables which takes quite a bit of time along with lead disconnections. The electronic switching is supposed to take care of this and thus reduce the testing time considerably. I have developed a Labview VI that is supposed to do this with case structures and FOR loops. But I am getting strange and bad results. The signals have non-zero mean and trends in them, and the time delays between the arrivals at the differnet sensors are different from when I was manually switching. I have been working fruitlessly on this for a while, so I would appreciate any help. Thanks in advance.
10-23-2009 09:29 AM
Hi haddu,
Can you describe the strange and bad results in more detail? Perhaps with a picture? If the signals are bad because they have non-zero mean and unexpected trends in them, then this may be due to signal crosstalk (the coupling of a signal to an adjacent path). Try routing just one signal through the switch and terminal block and see if the results are good. Then add just one other signal and note the results.
The signal paths through the matrix won't all be the same exact length/distance. Could that be why you are witnessing the difference in time delays?
Hope this helps!
Chad Erickson
Switch Product Support Engineer
NI - USA
10-23-2009 01:43 PM
Hi,
Thanks for the quick reply. I am attaching a word document with an example of the signals I am getting. I will try your suggestion as soon as possible. Although, for the initial switch connection the signals are good except for the time delays. Things go wrong for the first switch or change in connections. I wonder if that means something.
As for the reason you suggested for the time delays, that could be the case. But it should not be as bad as it is and I guess it should not be a major problem if the same signal paths are used for different data sets, i.e. when I subtract the baseline from the new measurements. Also, the difference in time delays is from those measured by direct connection of the piezos to the PXI.
Thanks again.
10-26-2009 11:10 AM
Hi,
I am switching 4 sensor signals and one actuator signal. By adding the signals one by one as you suggested I have figured out that it is the actuator signal that is routed throught the system to the piezos that is causing the problems. Otherwise, there is no problem. Now, what do I do about the actuation. I need to switch that signal too. Thanks.
Muhammad
10-27-2009 03:44 PM
Hi Muhammad,
Other customers have had success by separating the problemmatic signal from the rest of the signals. Distance helps.
1. What happens if you route the actuator signal through r0 and c0 and the remainder of the signals through the higher rows and columns?
2. Another recommendation is to insert ground references between the actuator signal and sensor signals. Is this possible?
Please write back if there's anything else I can do to help!
Chad Erickson
Switch Product Support Engineer
NI - USA
10-28-2009 06:51 AM
Suggestion 1 makes sense, but in my case that will work only once. There are five steps in my test and in each step the actuator signal goes to a different piezo, therefore, its route will be different. Actually, when I start the test, all the signals look fine because the initial connection is r0_c0, where the actuator signal is coming through c0. Things go heywire after the first switch where other piezos are connected to c0 through the switch matrix. I do not see any way to avoid that unless I can somehow define the connection path withing the switch to keep the actuator signal far away from the sensor signals.
As for suggestion 2, I have to be honest I have no idea what you mean or how that can be achieved. Thanks again.
Muhammad Haroon