Switch Hardware and Software

cancel
Showing results for 
Search instead for 
Did you mean: 

Variation in Thermal EMF of PXI-2530 modules

Solved!
Go to solution

This post/question is a companion to my most recent post in another thread.

 

In addition to having observed certain higher-than-expected resistances that affected current measurements using PXI-2530 modules in the 4W multiplexer topology, I also saw systematic channel-to-channel variation in voltage measurements. The voltages would steadily increase across the sixteen channels in a setup measuring the voltage across a 1.5 kOhm resistor with 0.5 uA passing through (75 uV). I identified that the thermal EMF of the reed switches in the PXI2530 module is on the same magnitude of these measurements and set out to quantify channel-to-channel Thermal EMF differences between my three modules.

 

Test Method: I have a TB-50 that is configured to MUX voltage signals to a DMM. I connected each of the four DB-50's of a single 176-pin cable to this block and collected voltage readings with a PXI-4071 configured for 7.5 digits accuracy in the 100 mV range and >10 GOhm impedance. For many of the channels, it took several minutes for the voltage to stabilize- or at least appear that it was stabilizing.

 

I have attached three graphs. Note that the vertical scale is the same on each.

 

The data that triggered this investigation was collected with MUX1, using the P2 connector for voltage. The magnitude was not quite the same- probably related to the stabilization time phenomenon, but this obviously the worst group of channels of the three MUX modules.

 

The three modules were all purchased at the same time (approximately 2 years ago), but had only limited usage during the first year or so. The three now have varying "mileage" based on my usage. But MUX1 clearly is behaving differently from the other two modules. The

 

I have modified my test conditions to pass the 0.5 uA through a larger resistance to make the thermal EMF less significant. The PXI-2530 spec sheets indicate that the Thermal EMF should be less than 50 uV. In most of my measurements, it is. But not for MUX1!

 

Any thoughts?

 

THanks,

Jeff

Jeffrey Zola
0 Kudos
Message 1 of 5
(6,615 Views)
Solution
Accepted by topic author Jeffrey_Zola

Hey Jeff,

Can you verify all three modules are PXI-2530, not PXI-2530B (just making sure, as the PXI-2530B exhibits slightly higher thermal emf)?

 

The PXI-2530's 50uV thermal emf specification is a typical value, so it is not a guaranteed spec.  Seeing a few channels greater than the spec is not a cause for alarm, but it does show that we need to account for this in our measurement error.  Note that the industry standard for thermal emf spec measurement is to close the relay, wait several minutes, then take a voltage measurement.  Thus, if you're scanning through a switch faster than one relay per minute or so, the thermal emf is going to be less predictable/stable.  That one module performs worse than another at these low voltages isn't indicative that that module is failing, etc... the module is fine.  Unfortunately, reed relays have more thermal emf than armature relays, mainly due to the numerous metal layers in a reed compared to an armature (each metal junction is a source of emf if those metals aren't the same).

 

Since thermal emf is proportional to temperature, it might be worth noting the chassis position of the worst performing module.  Placing it away from warmer modules (digitizers, arbs, RF, etc) will reduce the thermal emf. 

 

 

-John Sullivan
Problem Solver
Message 2 of 5
(6,595 Views)

Thanks for the info, John.

 

Yes, I am using the original PXI-2530 modules, not the updated 2530B. What is the reason that the new style exhibits higher thermal EMF, as my understanding is that the only difference was the change from the 176-pin connector to the LFH-160? Are there additional dissimilar metal interfaces in the LFH-200 that are absent in the 176 pin version?

 

Is there something about the layout of the relays on the PCBs that could cause the increasing thermal EMF that I am seeing on P2 of MUX1?

 

I have attached a screen shot of the MAX screen that shows my system configuration.  HP_MUX1 is indeed next to a digitizer that sits idle. I rearranged the chassis (PXIe-1065) about a month ago to space out the module I use exclusively (2530's, 2569, 4141, 4071's- with a third one that was pulled from slot 6 to be calibrated). I will see if there are compatible slots to move the digitizer away from the MUX. If so, I will repeat these tests.

 

Jeff


@Now_With_Underscores wrote:

Hey Jeff,

Can you verify all three modules are PXI-2530, not PXI-2530B (just making sure, as the PXI-2530B exhibits slightly higher thermal emf)?

 

The PXI-2530's 50uV thermal emf specification is a typical value, so it is not a guaranteed spec.  Seeing a few channels greater than the spec is not a cause for alarm, but it does show that we need to account for this in our measurement error.  Note that the industry standard for thermal emf spec measurement is to close the relay, wait several minutes, then take a voltage measurement.  Thus, if you're scanning through a switch faster than one relay per minute or so, the thermal emf is going to be less predictable/stable.  That one module performs worse than another at these low voltages isn't indicative that that module is failing, etc... the module is fine.  Unfortunately, reed relays have more thermal emf than armature relays, mainly due to the numerous metal layers in a reed compared to an armature (each metal junction is a source of emf if those metals aren't the same).

 

Since thermal emf is proportional to temperature, it might be worth noting the chassis position of the worst performing module.  Placing it away from warmer modules (digitizers, arbs, RF, etc) will reduce the thermal emf. 

 

 


 

Jeffrey Zola
0 Kudos
Message 3 of 5
(6,591 Views)

Hey Jeff,

 

From a customer usability standpoint, the 2530 and 2530B share the same pinout, support the same topologies, and use the same software API, code, etc.  However, the 2530B is substantially different once you get inside the black box. In fact, the entire analog path - connector, PCB, relay* - is different.  Notice that the specifications for thermal emf - 50 uV, typical - have not changed... I just happen to know the thermal emf is slightly higher on the 2530B.

 

*The 2530B uses the Coto 9012 relay, whereas the PXI-2530 used the Coto 9097.

-John Sullivan
Problem Solver
Message 4 of 5
(6,587 Views)

Very interesting.....

 

I moved modules around to get the idle digitizer away from the MUX modules, and remeasured the Thermal EMF of MUX1 and MUX2. I have attached graphs of MUX1 before and after the rearrangement and the second measurements. Definitely appears to have been some heating from the neighboring slots.

 

Jeff

 


@Now_With_Underscores wrote:

Hey Jeff,

 

From a customer usability standpoint, the 2530 and 2530B share the same pinout, support the same topologies, and use the same software API, code, etc.  However, the 2530B is substantially different once you get inside the black box. In fact, the entire analog path - connector, PCB, relay* - is different.  Notice that the specifications for thermal emf - 50 uV, typical - have not changed... I just happen to know the thermal emf is slightly higher on the 2530B.

 

*The 2530B uses the Coto 9012 relay, whereas the PXI-2530 used the Coto 9097.


 

Jeffrey Zola
0 Kudos
Message 5 of 5
(6,544 Views)