06-06-2008 04:23 PM
06-07-2008 10:11 PM
Hello Benjamin,
You have the right idea about building your matrix, but here are a couple of things to keep in mind:
-- The number of rows in a matrix define the number of independent simultaneous connections you can make. In a 40x100 matrix, you can make 40 simultaneous connections without necessarily shorting any of your instruments or test points together. If you want to be able to connect any of the 40 instrumentation signals to any of the 100 test points, but you will only need to connect a couple at a time, you could configure a 4x140 matrix. Place all of your instruments and test points on the columns and then use the rows to connect columns together. For example, C29->R1 and R1->C68 connects the instrument on C29 with the testpoint on C68, using Row 1. If you only needed 4 or fewer simultaneous independent connections, this method is much more cost-efficient. A 40x100 matrix has 4000 crosspoints. A 4x140 matrix has 560 crosspoints.
-- Consider the type of signals you'll be routing (voltage, current, frequency, differential or single-ended, etc.) when choosing the switch configuration. You mentioned having a 30 MHz signal. Routing that through a 1-wire matrix is like using a 10-inch piece of wire instead of coaxial cable from your Arb to the load. The impedance mismatch will be significant and your signal integrity will really suffer. You may do better with a 2-wire switch to carry the reference along with the signal line, or a coaxial switch to maintain a constant impedance from the cables through the switch.
-- If your project allows you to partition some of the instruments to certain test points, you may be able to define a few smaller matrices instead of one large matrix. For example, if 12 of your test points will only need to connect to your DC supply, maybe you could have a small 12x1 multiplexer for those lines and decrease the size of your large matrix.
I hope this is helpful. Good luck with your project!
Charles
06-09-2008 12:20 PM