Switch Hardware and Software

cancel
Showing results for 
Search instead for 
Did you mean: 

PXI-2575 faults randomly

Hi,

 

  I am using PXI-2575 mux with PXI-1044 chasis. I am using .NET C# to control 2575 during a year. In my setup, I am using the mux to read DC voltage or 2-wire resistance generally.  It seemed work successfully up to last 2 mounts. Suddenly, some of the channels of mux did not work. I read wrong DC voltage(like 1V instead of 4V or sometimes nothing). Mux is passing the self test in Measurement&Automation, path to COM seems correct but read voltage is wrong. I also hear the swicthing sound. Then, i changed it with different 2575. After a while, some channels of new 2575 were broken. I replaced the broken 2575 with the old broken 2575 and all channels worked successfully.(!) Why this situation happens?

 

   Finally, last day, all the channels of the mux were broken. I could not read anything from COM. I also heared swiching sounds. How can i fix it? When i replace the new one, the new one works succesfully. What is the problem?

 

Any help would be appreciated.

0 Kudos
Message 1 of 7
(6,892 Views)

Hey unver,

 

The self test feature only verifies that the digital backend of the module is functional.  The 2575 does not have path resistance circuitry. 

 

Most likely, your relays have reached end of life.  You can check the number of cycles in the relay tab of the Switch Soft Front Panel.  Note that the number of cycles before failure is largely a function of the reactance of the circuitry plugged into the module; relay cycle life before failure is listed for a purely resistive load, so it's likely that the number of cycles is much less than those figures listed in the specification sheet.

 

For example, if you're muxing through channels on the 2575 taking voltage measurements, the ~200pF capacitance of the DMM pluss the 50pF per foot of cable is going to store whatever currently connected voltage potential is and then rapidly dissipate the stored charge while the next channel is connecting.  When this occurs, there is a relatively large inrush current.  Likewise, if your DUT is inductive, the current running through a particular connection will continue to flow as the relay opens due to the same current pushing across a higher resistance (an open relay, instead of closed)...

 

An easy method to reduce the connect capacitance is to add a couple hundred ohm series resistance to your voltage channels.  Likewise, adding a flyback diode to each channel will prevent overvoltage on inductive loads.  The flyback diode will also work for the 2-wire resistance, but note that the series resistance won't (it will add to your measurement).  To solve this problem, I recommend 4-wire resistance measurements with 200 Ohm series resistors (or whatever) on each leg.  The additional series resistance does not affect the measurement when performing 4-wire resistance measurements.

 

Of course, I've jumped the gun a bit, so first, a little troubleshooting.  You obviously have a DMM, so let's set it to resistance and plug it into either side of the suspect channel.  Then cycle the relay a few hundred times and see if the resistance is consistent on each closure.  If it varies greatly, the relays have probably reached end of life.  For the voltage readings to be off 3 volts, the relays are going to need to exhibit some pretty enormous closed resistance values... something like 10MOhm... if this is the case, then the relay current/voltage limits have likely been exceeded.

 

I also recommend scoping the inrush current with an oscilloscope and a series 0.01Ω resistor.  Likewise, measure the flyback voltage when opening the relay with a different channel on the same scope... obviously, you'll want to do this with your 2575 plugged into the DUT and operating in its normal environment...

 

From your description, I'm almost certain the troubleshooting steps will indicate one or more relays have failed.  If I had to guess, I'd say it's K15.

 

 

-John Sullivan
Problem Solver
0 Kudos
Message 2 of 7
(6,886 Views)

Knights Who Say NI wrote:

Hey unver,

 

The self test feature only verifies that the digital backend of the module is functional.  The 2575 does not have path resistance circuitry. 

 

Most likely, your relays have reached end of life.  You can check the number of cycles in the relay tab of the Switch Soft Front Panel.  Note that the number of cycles before failure is largely a function of the reactance of the circuitry plugged into the module; relay cycle life before failure is listed for a purely resistive load, so it's likely that the number of cycles is much less than those figures listed in the specification sheet.

 

For example, if you're muxing through channels on the 2575 taking voltage measurements, the ~200pF capacitance of the DMM pluss the 50pF per foot of cable is going to store whatever currently connected voltage potential is and then rapidly dissipate the stored charge while the next channel is connecting.  When this occurs, there is a relatively large inrush current.  Likewise, if your DUT is inductive, the current running through a particular connection will continue to flow as the relay opens due to the same current pushing across a higher resistance (an open relay, instead of closed)...

 

An easy method to reduce the connect capacitance is to add a couple hundred ohm series resistance to your voltage channels.  Likewise, adding a flyback diode to each channel will prevent overvoltage on inductive loads.  The flyback diode will also work for the 2-wire resistance, but note that the series resistance won't (it will add to your measurement).  To solve this problem, I recommend 4-wire resistance measurements with 200 Ohm series resistors (or whatever) on each leg.  The additional series resistance does not affect the measurement when performing 4-wire resistance measurements.

 

Of course, I've jumped the gun a bit, so first, a little troubleshooting.  You obviously have a DMM, so let's set it to resistance and plug it into either side of the suspect channel.  Then cycle the relay a few hundred times and see if the resistance is consistent on each closure.  If it varies greatly, the relays have probably reached end of life.  For the voltage readings to be off 3 volts, the relays are going to need to exhibit some pretty enormous closed resistance values... something like 10MOhm... if this is the case, then the relay current/voltage limits have likely been exceeded.

 

I also recommend scoping the inrush current with an oscilloscope and a series 0.01Ω resistor.  Likewise, measure the flyback voltage when opening the relay with a different channel on the same scope... obviously, you'll want to do this with your 2575 plugged into the DUT and operating in its normal environment...

 

From your description, I'm almost certain the troubleshooting steps will indicate one or more relays have failed.  If I had to guess, I'd say it's K15.

 

 


 

Hi John ,

 

I think this is not the case. Because, highest switch usage is 117(I think this is very very low) and I am doing the voltage measurements over 10kOhm resistance. Hence, capacitive effect is also not the cause. After i had written my previous message, i prepared a basic set up that applies 5 volts one by one by one to all of the 98 input ports of the broken 2575 and reads the com port with NI6259 daq device. I did this process many times with different time periods. As I guest, the device worked successfully at all the tests. Hence, cause of the problem is different and i still need a help.

 

0 Kudos
Message 3 of 7
(6,852 Views)

How is your software configured?  Are you only closing one relay at a time? Is it possible that you have two relays closed by accident?

 

Another thing to confirms is that you are waiting for both the relays to settle and your signal to settle before you take your measurement.  Ensure "Wait for Settle" is enabled on your relay control functions to get the minimum waiting time, but I would recommend waiting an additional time for the analog voltage signal to settle before you take the voltage reading.

Eric S.
AE Specialist | Global Support
National Instruments
0 Kudos
Message 4 of 7
(6,841 Views)
My software is only closing one swicth at a time. When fault occurs, I still have the same problem although I am waiting for seconds after switching to read voltage.
Additionally, I am wondering that is there any command to reset the 2575 or open all switches?
0 Kudos
Message 5 of 7
(6,829 Views)

Unver,

 

I did some research into this and found the following information from the specifications of the 2575 (http://www.ni.com/pdf/manuals/373870m.pdf) Note The NI PXI/PXIe-2575 is not recommended for 2-wire resistance measurements.) What are the voltage / current that you are putting into the 2575, as this device needs 1 mA of current to work properly? This is due to the following, taken from Armature Relay Contact Stability:

 

 Inside of sealed relay housings there exist gases that are emitted from the plastics, adhesives, and etc. that go into making the relay. This material is organic and is adsorbed onto the contacts where it can then polymerize under these low voltage and current switching conditions. Low level contact loads do not provide enough energy to form an arc sufficient to clean off the contacts. Instead polymerization occurs which leaves behind a powdery substance on the contacts causing the resistance to rise. Switching voltage/current, and contact material also contributes to the severity of this problem. Higher voltages and currents ensure that switching arcs are of sufficient energy to clean deposits from the contact surfaces. 

 

 

I have attached links to the help documentation for the C prototype function calls for opening all switches (niSwitch_DisconnectAll) and for reseting the 2575 (niSwitch_reset). These were found within the NI Switches Help under NI Switches Help >> Programming with NI-SWITCH >> Function Reference >> Functions.

 

niSwitch_DisconnectAll

http://zone.ni.com/reference/en-XX/help/370388Y-01/switchcref/c_nisw_disconnectall/

 

niSwitch_reset

http://zone.ni.com/reference/en-XX/help/375472A-01/switchcref/c_nisw_reset/

Frank,
National Instruments
Software Group Manager
0 Kudos
Message 6 of 7
(6,825 Views)

Unver,

 

With regards to the information posted before, depending on the type of DMM that you are using, the amount of current going from the DC source to the DMM will be rather low. This is due to the high impedance on most DMM's inputs. These impedances would affect current flow greatly. For example on the 4072 there is either 10G Ohm or 10M Ohm's of input resistance which, with 5 V DC, would mean a current of 5nA or 5uA respectively flowing across the relay which would not remove the material that might build up. I have attached a drawing with will allow a higher current flow across the relay in addition to an accurate reading of the voltage level. In this image, we have attached a resistor to ground (1k Ohm) which will force a current draw of 5 mA to go across the relay reducing the severity of this problem.

 

2575 Suggestion.png

Frank,
National Instruments
Software Group Manager
0 Kudos
Message 7 of 7
(6,819 Views)