Resistor-matching in HSDIO 6544 and SCB 2162?

[Mystogan]

hi everyone, I have a question regarding impedance matching (http://zone.ni.com/reference/en-XX/help/370524M-01/siggenhelp/impedance_matching/) by HSDIO NI 6544 and SCB 2162. What should I consider in designing my PCB to prevent impedance-matching? From the datasheet of NI6544, it says the impedance on DIO (red color in the pict) is 50 Ohm (ON) and 50 kOhm (OFF). Im going to connect the DIO's to my PCB (blue color in the pict). Anyone know about this?

thanks, Yan

[Duffy2007]

Hi Yan

on this side you can find the manual for the CB-2162 in this is some opportunities are shown how 

to wire the resistors and yes the impedance on the 6544 is 50 KOhm.

My question is what signals would you like to connect? what voltage?

[Duffy2007]

Hi Yan

as an additional information you looked into the help for the 54xx devices.

Please have to look into the NI Digital Waveform Generator/Analyzer Help this is also 

installed with the HSDIO driver for the device. In this manual you can search for Source Impedance then you find this 

The NI 6544/6545/6547/6548 data, clock, and event generation channels have a 50 Ω source series impedance. For applications where the full voltage swing is required at the DUT, a parallel termination resistance of 1 kΩ to 10 kΩ is recommended. With a system terminated by 10 kΩ, much of the signal reflections are eliminated by the source 50 Ω termination and the parallel termination. Thus, the voltage seen at the termination resistor is 99.5% of the configured voltage.

Because the NI 6544/6545/6547/6548 interface cable (NI SHC68-C68-D4) is a 50 Ω transmission line, when you use 3.3 V Logic, you can build matched impedance systems with a 50 Ω parallel termination as the load. While a matched system is beneficial because reflections are eliminated at the load, the voltage at the termination is one-half of the generation voltage level. The voltage change is caused by the voltage divider that is created by the source and termination impedance, as shown in the following figure.

For example, if RSource = 50 Ω and if the termination resistance is also set to 50 Ω, the voltage level seen at the termination is one-half the source voltage.

Always calculate the maximum current that the NI 6544/6545/6547/6548 in your test system can source and sink. You can calculate the maximum current using the following formula:

Max current = VOH (max)/(50 + RTerm)

For example, if the device drives 1.8 V into a 50 Ω load, the maximum current is calculated as follows:

Imax = 1.8 V/(50 Ω + 50 Ω) = 18 mA

I hope this helps

[Mystogan]

hi alex, thanks for your detailed response.

what I'm trying to achieve now, how do I prevent the impedance matching to occur in my circuit without losing any voltage output from 6544?

So, what I understand to prevent impedance matching, I need to have 50Ω R-termination like the one in your picture, which means, I would have 50Ω overall (R-source, R-cable, R-termination).

But, with using 50Ω for R-termination, I would get only one half of full voltage swing, right? How do prevent this? 

I'm using 1,8V output from 6544 (with frequency 10MHz). Which means, I would get only 0,9V with using impedance matching circuit, right?

I want to have full swing voltage without impedance matching issue. How to do that?

thanks,

Yan

[Jesse_O]

Yan,

In your configuration the output impedance of the PXIe-6544 is 50 Ohm.  This is matched with the SHC68-C68-D4 cable which has a characteristic impedance of 50 Ohm.  This connects to the 2162 adapter module which is designed to have traces that also have a characteristic impedance of 50 Ohm.   This makes the system matched from the PXIe-6544 right through to the 2162.

From the 2162 your cable should ideally have a 50 Ohm characteristic impedance and connect to your PCB where the traces are routed with a 50 Ohm characteristic impedance.  This trace would go to your device pin.  You can then place a 50k Ohm resistor at your pin if you like (although it is not required).  This would create a matched system where you get a clean signal at your receiver with a full voltage swing.

I hope this helps,

[Mystogan]

hi Jesse, thanks for your response.

so all I need is just a cable with 50Ω characteristic impedance to connect the digital output from 2162 to the device pin on my PCB?

Do I need to use the termination connectivity which available in 2162 board? Or should I just ignore it and take the digital output directly to my device pin in my PCB, though I'm placing my device pin with 50Ω resistor (to get matched impedance). I would get full voltage swing by doing that? Or should I not placing the 50Ω resistor in my device pin at all, but instead I should place 50kΩ? 

Thanks,

Yan.

[Jesse_O]

Yan,

If you can get a 50 Ohm characteristic impedance cable to connect from the 2162 to your PCB you'll get the best performance for that setup.  You do not need to use the termination on the 2162, and you will either want to use a 50K resistor or no resistor on your PCB at your device pin to achieve a full signal swing.  

If you use a 50 Ohm resistor at the device pin you will only receive half the generated voltage.

Regards,

[Mystogan]

Hi Jesse,

so I'm assuming that I don't need to worry to have impedance matching issue when I get a 50 ohm cable to connect my PCB to 2162, when I place 50k ohm or nothing on my pin.

By that setup, when I use high frequency (up to 100MHz, is the max. freq. that 6544 can achieve), 1 meter long RG-58 coaxial cable from PCB to 2162, wouldn't be a matching problem? Cable length doesn't matter?

thanks,

Yan.

[Jesse_O]

Yan,

Cable length does matter when you are trying to generate high data rates.  However the 6544 should not have a problem running at 100MHz through 2m of cable.  When the digital signal travels through the 2162 to your 50 Ohm RG-58 cable then to your PCB it will eventually hit the high impedance input pin of your device.  This will cause the required reflection to hit a full signal swing to occur.  This reflections will travel back to the PXIe-6544 where the source series termination prevents this reflection from bouncing back to your PCB.  This is normal for a source series termination.

Any impedance discontinuity on the cabling (normally found at connectors) will cause reflections.  However it sounds like you are on the right track.  I hope that helps.

[Mystogan]

Thanks Jesse,

really appreciated.

Regards,

Yan.