Use PXI-2630 terminal block in matrix configuration?

[Jeffrey_Zola]

Apologies in advance for the long-windedness of this post! 😄

I am using PXI system with several PXI-2530 switch modules connected to a series of UUT using PXI-2632 (8X16 1W matrix) terminal blocks and one PXI-4071 DMM for each switch module. My application, uses the PXI system for external current and voltage measurement to verify operation and/or qualify reliability stress equipment. A requirement of the application, therefore, is that there must be a continuous current path through each UUT with minimal impedance changes as the application switches between its "bypass" mode and its "measurement" mode.

I have used this setup with matrix terminal blocks in conjunction with one of our test systems and I am satisfied with the results. I started working with Test System A, which does not have easy connections to the DUT sockets, I needed to construct a way to interface the PXI system and a resistive load to the DUT sockets, so it was not difficult to build in wires that attach to the screw terminals of the 2632. It did turn into a rat's nest of color-coded wires that I did my best to keep neat and tidy in various bundles, however. Thank goodness for cable-ties!

My next task is to use this application in conjunction with Test System B, which has an interface of male header pins from which every signal that goes to or from the UUTs can be obtained. No soldering or running wires through gaps where the designers didn't intend wires to be stuffed. I intend to construct a break-out board that will allow for simple connections between the PXI modules and the number of Test System B's that we have or will have in our lab. In order to simplify setup/installation, I would like to reduce the number of screw-terminal connections. Preferably, I'd like to eliminate screw terminals altogether, and use lever-based connections where I cannot have mating headers. The PXI-2632 terminal blocks unfortunately use screw terminals.

In the 8X16 matrix mode, the PXI-2530 closes switches kcom1/3/5/7 regardless of what matrix points are connected. A connection between row R and column C is made by closing the switch corresponding to k(16R+C). I verified this using the Soft Front Panel.

I also have a number of PXI-2630 terminal blocks. These are intended to be used with the switch module in any of its MUX modes, and include 8 banks of 2X9 header pin connections. From the documentation of the 2530 and 2630, I have identified that switch kX is associated with the output pin for chX, with ch0-15 associated with pins 1-16 of bank 0, ch16-31 associated with pins 1-16 of bank 1, etc. X=16B+P-1. Pin 18 of each bank is used for comX of the independent MUX topology. The sixteen multiplexed pins appear to correspond to the sixteen columns of the matrix, with the eight common lines corresponding to the eight rows.

Here is what I would like to do, but thought I'd ping the forum to see if anyone has tried anything similar and has wisdon to share:

• Fabricate custom cables that interconnect pins 1-16 of all eight banks of the 2630's header connections with a single 16-wire ribbon carrying signals from the interconnected banks (column connections!).
• The custom cable harness will also include one wire connected to pin18 of each of the eight banks (row connections!)
• The 24 total wires in the harness will be terminated in header connections that will likely be paired by the lines that I currently connect to each UUT.
• Fabricate additional harnesses that interface with the header pins of Test System B.
• Fabricate a breakout board using strip board or similar material to provide header pins to connect both custom cables above and allow for connection of other elements like resistors using level-based terminals.

I verified this concept using the 176-pin cable assembly and four terminal blocks, as well as a bunch of little pieces of wire. Are there other issues I need to configure, such as elements of either terminal block that establish some of the physical components of the switch topologies? Do the innards of the PXI-2632 perform more functionality than the interconnecting of the eight sets of sixteen pins? Do the innards of the PXI-2630 connect elements that wouldn't allow my proposed scheme?

I appreciate any and all input and suggestions!

Thanks,

Jeff Zola

[BenRobinson]

Hi Jeffrey,

As far as I can understand your plans for a custom connector, it should work. Based on testing that it sounds like you have already performed, the wires appear to be grouped in the correct manner. Creating custom connectors for switches is fairly common and all of  the grouping that you refer to is done within the switch itself. Good luck with the harness and your project!

[Jeffrey_Zola]

Thanks for the reply, Ben (or not?).

I am pretty confident that I fleshed out the details of what pins are connected to specific switches in the PXI-2530 module. My biggest concern relates to internal components of the 2630/2631/2632 terminal block that help define the topology or protect the module or terminal block in some way. For example, I have one of each block in front of me- there is a group of resistors (R1-R16) on the 2632 (8X16 matrix), but those pads on the PCB are not populated with components on the 2631 (4X32 1W or 4X16 2W matrix), which has 16 additional column connections, although these resistors appear to connect one group of column terminals to the other. These resistors are not present on the 2630 (MUX) terminal block.

I was hoping that someone would have a "war story" to share that would confirm my plan or give me recommendations what I need to change to make sure my fixture works as desired out of the block. There is still time for that....   Anyone?  Bueller?

Regards,

Jeff

[BenRobinson]

Hi Jeff,

The resistors that you are seeing are to protect the reed relay from current surges that could arc and damage or even fuse the contacts of the relay. This link shows the theory behind the reed relay protection in the terminal block. It would be a good idea to build this kind of protection resistor into the connector that you are building.

[Jeffrey_Zola]

That is good information, Ben.

It appears that the 2632 terminal block for 8X16 matrices uses a series resistor in each column. My eyes are not able to read the markings on them- do you know what size resistor is used?

The 2631 terminal block for 4X32 1W or 4X16 2W matrices does not use the resistor. Are applications using this topology at risk of switch damage because of this? I also do not see resistors on the PCB for the 2630 MUX terminal block. The white paper discusses how matrix layouts can build stray capacitance than multiplexed layouts. Are there any hidden resistors in the 176-pin cables that attach to the 50-pin terminal blocks?

I use the switch modules in applications where I am performing high-accuracy measurements of voltage and/or current. Are the resistors likely to affect the sensitivity of my measurements?

THanks

JEff

[BenRobinson · Accepted solution]

Hi Jeff,

First a correction to my previous post: The 2632 terminal block does not have reed relay protection resistors as I previously stated. The resistors you were referring on the 2632, and the ones that I mistook, are there to connect the columns of the switch. The resistors have a zero resistance value and are acting as electrical connections. The 2632 connects columns c0 to c16, c1 to c17, c2 to c18, and so on. The 2531 and 2532 switch cards do have the reed relay protection resistors onboard.

As to the resistors in the card that protect the reed relays, they are generally very small and do not significatly affect even small voltages travelling through the switch. The resistors will not affect the currents in the card at all. Any effect that the resistors do have on the voltages will be within the accuracy specifications of the switch card.

So, to adress the other question in your post, there are not resistors within the connectors as they are not needed.

[Frank-L]

Jeff,

FYI a modification of what Ben mentioned below,

Using protection resistance is recommended to increase the life of the relay (as you will prevent inrush current), but it may affect certain types of measurements. On certain modules we offer 100 Ohm methods and 0 Ohm methods for when the resistance might affect the measurements.

The 100 Ohm protection resistances will NOT affect 4 wire measurements, but may affect 2 wire measurements and general signal passing through a switch.

For example:

If you were measuring resistance of a DUT, the additional 100 Ohm resistors will increase the resistance measured in 2 wire mode, but NOT in 4 wire mode. (LINK)

If you were measuring voltage or current, the additional 100 Ohms resistors will affect the voltage drop measure (of the DUT, or of the shunt), and will give an inaccurate reading.

If you were passing current from a power supply through the module, the additional of the 100 Ohm resistor in the path will cause an increase in the supply voltage (in constant current mode) or a decrease in the supply current (in constant voltage mode) to compensate for the increased path resistance.

Hope that this helps!

[Jeffrey_Zola]

Hi, Frank.

Thanks for the additional information on the forum and on the phone earlier today.

I am thankful for the support that NI has provided me as we have explored our options for this project. The solution we found to use relays to provide the required continuous current path as the SMU is MUXed into position looks elegant and cost effective.

Jeff