01-22-2009 01:40 PM
I have a few questions about the behavior of my PXI-2503 switch module. When I close a relay on my switch I hear the predicted click. As such, it leads me to think that something within could wear out over time... which I would like to mitigate if possible. Can someone please help with the following questions about the attached VI?
1. If I send a close relay request and that relay is already closed, does the switch perform the action anyway? If so, does adding logic to determine the current switch state and only switch if neccessary prolong the life of the switch module?
2. How confident can I be that this construct is properly switching? It seems wicked fast. Do I need to build in some sort of settle time?
Solved! Go to Solution.
01-23-2009 05:15 PM
Great questions! First off, you are right in assuming that because you are using mechanical relays, over time, they will wear out. The life of the relays can be found in the NI PXI-2503 Specifications, on the first page. It should be noted though, that though it lists the "relay life", this simply refers to the average amount of time before the first switch has a path resistance greater than 1Ohm (so the actual life could be longer). We also have replacement kits for if/when individual relays die.
As for your questions:
1. In the VI you are using, the relay will not open to close again if they are already closed and you tell them to close. Ideally, you don't want to unnecessarily switch, but unless you have redunant switching built into your code, you shouldn't need special logic.
2. The debounce VI that you have in step 3 of your program creates a pause to allow for all the created paths to settle. They switch wicked fast, but they also settle wicked fast.
You seem pretty concerned about your relay life though, which makes me curious...how much switching are you going to be doing? Is this a program that will be running for extended periods of time (days, weeks, or months without stopping)? If you are going to be doing that much switching, you might want to consider using a switch that doesn't have mechanical parts, like a FET switch.
Applications EngineerNational Instruments
01-26-2009 02:51 PM
The most exercise I could ever imagine would be once every second for 12hrs straight, two days a week.
As a side note, originally I purchased PXI-2501s. I generate AO signals with a PXI-6259 and routing them via switches, and the 2kOhm resistance of the PXI-2501s really screwed me. The voltage and current I create is very important on the back end of this process, and rather than build/buy another sub-system to step up the current, I elected to swap them for PXI-2503s.
Getting back to business, while testing I discovered a functional problem that may make this whole discussion moot... more than one relay is closed at one time. I assumed only one could be closed at a time, as this is the behavior illustrated in the NI Switch Soft Front Panel. This is really cool and great to know going forward, but for now I need only one closed at a time. My switching is random between trials so I guess I'll have to close->settle then open->settle each trial. Bummer!
This is unless of course you know of a way to change the default behavior of the PXI-2503!
01-27-2009 02:05 PM
01-27-2009 02:25 PM
Thank you for the response to my post!
The topology I have is 2503/1-Wire 48x1 Mux. However, my testing doesn't support the break before make theory. What I see is that once the relay closes, it stays closed until I reopen it... one to many routing. I'm switching analog voltage to an array of LEDs, and am lighting up my whole array which wouldn't be possible. Break before make would be nice, I just do not know if it is possible or how to make it happen if it is... 😉
I'll leave this post live a little longer and see if I get anymore input, but if not I'll accept the switch wear/tear and code it to close/open each trial.
01-28-2009 03:22 PM
Whether this acts as "break before make" depends on whether you are using a relay or a channel API. My suggestion would be to use the NISwitch Connect Channels along with the Disconnect All VI to make sure that all of the relays are open before closing one. When using the Connect Channels, you'll have the first channel be the relay name, and your second channel will be com0, you'll want to keep your topology as 1-Wire 48x1. Please see the attached picture for a better idea of what I am referring to. Take care!
Applications EngineerNational Instruments