NI Home > Community > NI Discussion Forums

Switch Hardware and Software

Reply
Member
vikasg
Posts: 7
0 Kudos

2576 signal routing on test board

Attached is the schematic for the test board I am building.  I am routing signals on the test board.

 

J12COM2, J11COM1 and J3COM1 are parts of DSUB - 50 pin connectors on the board connected to PXI-2576 using LFH160 cable.

5404p is the output from PXI-5404.

 

So essentially signal path is 

 

PXI5404 -> Test Board connector -> J12 connector -> PXI 2576 (BANK 14) -> J12 connector -> J11 connector -> PXI 2576 (BANK 9) -> J11 connector -> Test Chip

 

Signal from 5404 is going to be around 50MHz 1V ptp, and the impedance looking into the test chip is around 100ohm.

 

Do I need to be concerned about anything in routing the signals this way?  i.e. Signals going back and forth from the test board through cables to 2576.

Or should I avoid routing 5404 signal through 2 set of relays? 

 

board-question-ni.jpg

Member
Frank-L
Posts: 159
0 Kudos

Re: 2576 signal routing on test board

[ Edited ]

Vikasg,

 

The biggest problem that I can see regarding the routing of your signals through two sets of relays is going to bandwidth issues. Assuming you are using the 100 Ohm system with a 4x1 or 8x1 configuration, the specifications show that the bandwidth is going to be >60MHz which for your signal would be fine. However even though the bandwidth is 60MHz, you are going to have some attenuation at any frequency. This attenuation is going to be doubled due to the fact that we are going through the relay twice. If per say the signal has 10% attenuation at 50MHz (not actual values of attenuation at 50MHz), then when we reach the output of the switch we have .9V ptp. By going through the relay a second time, that attenuation is doubled resulting in an output signal of .81 V.

 

You also will have to consider the crosstalk and isolation on the 2576 using due to the high frequency signals that you are transmitting. I am not saying that this is a dealbreaker, but something that you might want to consider.

 

Regards,

Frank

 

Frank,
National Instruments
Staff Software Project Engineer
Member
vikasg
Posts: 7
0 Kudos

Re: 2576 signal routing on test board

 

Thanks Frank for the response. 

 

I am not worried about the attenuation as much as the crosstalk. Do you know where I can find crosstalk specifications? I am routing my digital signals on a relay bank that is far away from this analog signal (with assumption that the physical distance of bank 0 to bank 11 is the higher than bank 0 and bank 1 on the 2576 pxi board and cables).  

 

I am trying to measure frequency response of my test chip at different frequencies and 50MHz is the highest frequency signal I was going to use. I can compensate for the attenuation by doing the measurement twice (with and without the IC).  I am also testing for responses at 5MHz and 15MHz.

 

 

 

 

 

Active Participant
Now_With_Underscores
Posts: 425
0 Kudos

Re: 2576 signal routing on test board

Hey Vikasg,

 

I assume the 50MHz signal you're planning on routing from the 5404 is a sine wave, correct?  If you're planning on routing a digital clock at 50MHz, you're going to need a lot more analog bandwidth or else your signal is going to come out looking like a sine wave :smileywink:.  In general, I recommend that your analog bandwidth is 10 times the clock rate... you can skimp by with maybe 3-4x, but the lower you go, the less your signal is going to appear as a square wave on the output. 

 

Take a look at the PXI-2593, which is better suited for high frequency digital edges.  If your goal is to have 100ΩDC resistance, you'll need to add some series resistors.  Let us know if the 2593 will work for your application we'll go from there.

-John Sullivan
Analog Engineer
Member
vikasg
Posts: 7
0 Kudos

Re: 2576 signal routing on test board

[ Edited ]

Yes the 5404 is for sine wave generation.  

I am using the 6552 for digital output, my digital signal freq is less than 1MHz.

 

I found the crosstalk specification - bank to bank crosstalk @ 10MHz <-35db .. do you know if its between adjacent banks or this is the best case crosstalk between the farthest banks?

 

edit: I should have seen the link Frank posted in the first response

Active Participant
Now_With_Underscores
Posts: 425
0 Kudos

Re: 2576 signal routing on test board

The 4x1 10MHz bank-bank crosstalk specification is a typical value indicative of the worst bank to bank crosstalk.  Some modules will not achieve this performance, but a typical module will from any bank to any other bank.

-John Sullivan
Analog Engineer
Member
Frank-L
Posts: 159
0 Kudos

Re: 2576 signal routing on test board

Vikasg,

 

I agree with John, based on the information that we have available to us, if you are starting a new system it might be best to use several of the PXI 2593's for your application. This is due to increased signal integrity, cabling options to improve signal quality as well as much better performance with prevention of crosstalk on your system.

 

Frank,
National Instruments
Staff Software Project Engineer