I have a PXIe-5672 and want to capture the baseband on a scope. Can I do this?

[Fewoiz]

I have a PXIe-5672 VSG and want to capture the baseband with a scope to compare timing with my receiver (DUT).  I would like to connect a scope to the out put of the PXIe-5442, but not sure if this will cause any problems.  Of course I'm especially concerned about damaging the hardware.  This method would be the ideal way to verify my response timing, but if it can't be done, I'm open to other suggestions.

[mbb25 · Accepted solution]

Yes, you can do this.  The 5672 device driver configures the 5442 to use it's direct path.  Direct path max voltage is 1 Vpp into 50 Ohm scope, 2 Vpp into high impedance, so make sure your scope is set up for that.

Continue to use the 5672 as normal, but with your scope connected to the 5442 CH 0 output.  Assuming the scope doesn't drive its input (it shouldn't), there's no worry of damaging the 5610 or 5442 by doing this.

Unless you split the signal (not what I'm suggesting), you won't get a signal out of the 5610.  It isn't clear how you were planning on comparing signal timing with your DUT.  No signal may be OK if you are comparing the timing vs a digital trigger, perhaps exported from the 5442.  Be aware, however, that the 5610 delay from IF IN to RF Out is about 1.2 us.

Note also that if you set NIRFSG_ATTR_COMPENSATE_FOR_FILTER_GROUP_DELAY to true, and output a marker event on sample 0, the marker event will be delayed to correct for signal processing delays within the 5442 and the 5610 analog delay, causing the marker event to be aligned with the time "sample 0" shows up on the 5610 RF Out.  This attribute was added after the initial release, and defaults to false in order to preserve backward compatibility.

If you are looking to compare timing, setting Compensate for Filter Group Delay to true is actually what I'd do, rather than splitting out the 5442 IF Out signal.  You can verify how well things are aligned by tuning the 5672 to a low enough frequency for your scope to measure.  The 5610 delay is dominated by the IF filter, which is independent of the RF frequency.

[Fewoiz]

Thanks for the incredibly helpful response.  I'm not only generating a sample 0 marker, but also at several specific points during the transmission.  I need to measure how long the DUT takes to respond to specific data in the signal.  In one transmission I may have up to 8 markers.  In addition to using the scope I might be able to feed the marker trigger directly into a prototype DUT with special firmware that could measuring the delay I'm looking for. 

Regardless of the markers, I want to split the signal of the 5442 while transmitting because seeing the timing on the scope shows the marker timing accuracy.  Hopefully I can see the effects of the filter group delay and the 1.2 us delay between the LF IN and RF Out..

[mbb25]

You can split the signal for debugging, but there are a few things to be aware of:

   1) if you split using something similar to a BNC T junction and use high impedence on the scope, you will likely introduce some transmission line effects, which can affect your signal quality (degraded EVM, for example).  Keeping the cables short will help some.

   2) if you split using a splitter and 50 Ohms impedence on the scope, you will loose 6 dB of power from the arb to 5610.  So power at RF out will be 6 dB low.

Depending on your signal quality requirements, option 1 is probably OK for debugging/characterization.  If you want to do this long-term, that makes me wince a bit.  Option 2 is probably better, but not great.  Hopefully once you get everything characterized, you find that markers are aligned well enough with the 5672 RF Out that splitting the signal is not needed.  Checking this on your own like you are planning seems like reasonable due diligence.

[Fewoiz]

I'm only doing this once to confirm timing and only for as long as it takes get it right.  I'm hoping we have a splitter around but if not I'll have to use a T.  I'm definetly going to give this a try thanks to your help.