Digiscope impedance match to 50 ohm

[dantm]

Hi,

I'm using a PCI-5102 digitizer to sample two signals that are 10 MHz eaech coming from a 50-ohm SMA connector.  These are I/Q signals coming from a downconverted RF system.

I noticed that the PCI-5102 has a 1 Mohm input impedance.  Given that my cables connecting the 50-ohm SMA output to the PCI-5102 are relatively long (~3 ft), would there be any issue in sampling them?  I assume the signal would show twice as large as when matched into a 50-ohm load, but what else would cause an issue?

On a separate note -- are there any NI sampling/scopes which are 50 ohm matched?

Thank you!

[DFGray]

Any problems with using the 5102 will depend on what is at the other end of the line.  You will get a reflection due to the impedance mismatch, but it may or may not hurt you.

You can get around this issue by putting a T connector with a 50Ω terminator on the input to 5102.  This will give you an impedance match and scale your signals to the right level.

The 5102 was the first "scope" card produced by National Instruments, and is essentially an overclocked E series DAQ board (and can be programmed with the DAQ API).  As such, it lacks many features expected in a scope.  50Ω termination is one of those.  Most of the other National Instruments high-speed digitizers have a 50Ω input setting (the lower cost ones, such as the USB-5132, lack it).

Disclaimer:  The terminator and T connector links above are for reference and do not constitute an endorsement on my part.

[dantm]

---

@DFGray wrote:

Any problems with using the 5102 will depend on what is at the other end of the line.  You will get a reflection due to the impedance mismatch, but it may or may not hurt you.

You can get around this issue by putting a T connector with a 50Ω terminator on the input to 5102.  This will give you an impedance match and scale your signals to the right level.

The 5102 was the first "scope" card produced by National Instruments, and is essentially an overclocked E series DAQ board (and can be programmed with the DAQ API).  As such, it lacks many features expected in a scope.  50Ω termination is one of those.  Most of the other National Instruments high-speed digitizers have a 50Ω input setting (the lower cost ones, such as the USB-5132, lack it).

Disclaimer:  The terminator and T connector links above are for reference and do not constitute an endorsement on my part.

---

Thanks for your feedback on this.  Is there another relatively low cost card that you'd recommend instead of the 5102?  I'm looking for something fairly inexpensive to be used as a 2 channel (simultaneous) digitizer, with a 10 MHz input frequency.  Sampling rate could be 1 MSps or in that range.

On a second note -- can you mention some of the other limitations of the 5102?

Thank you!

[DFGray]

The least expensive oscilloscope card that National Instruments sells which meets your criterion is the 5114.  It starts at about $3,000.  There may be another type of DAQ or DSA board which would work for you, but these typically do not have 50Ω input impedance options.  Much as I would love you to buy a shiny new National Instruments High-Speed Digitizer, you would probably be better off getting T-connectors and 50Ω terminators.

Note that I have not worked with the 5102 for six years, but here is what I remember are some features it lacks:

1. 50Ω termination
2. Dual ported RAM (cannot stream data while acquiring)
3. Enough ranges (it has three, if I remember correctly, and they are very widely spaced, resulting in about 6 effective bits of resolution before switching to the new range)

None the less, I found it to be a clean digitizer that worked well for what it did.

[dansch]

Hi dantm,

DFGray explained everything very well. but just to add more information to what he mentioned, the closest upgrade we recommended was the USB-5132 which has very similar features to the 5102 despite being in a different form factor. The closest PCI/PXI card we offer as an upgrade is the 5114 as he mentioned, which has many more features than the 5112. Since this is an SMC-based device, it will have all the advanced features of data streaming, increased memory options, and T-Clk synchronization (plus the 50 Ohm impedance you were interested in). However, the PCI version of the card is quite large and much longer than the PCI-5102, which can sometimes be a problem in some PC chassis.

Regarding the USB-5132, here is a comparison to the 5102:

1. The 5132 does not have 50 Ohm input impedance (same as 5102).

2. The 5132 can continuously acquire and stream data (which the 5102 cannot), but it cannot perform multi-record acquisition (which the 5102 cannot do either). Obviously, streaming throughput is limited due to the USB bus.

3. The 5132 has 50 MS/s sampling and up to 50 MHz bandwidth, compared to 20 MS/s and 15 MHz on the 5102.

4. The 5132 has 4 MB and 32 MB memory options.

5. The 5132 has 0.04, 0.1, 0.2, 0.4, 1, 2, 4, 10, 20, and 40 Vpp ranges compared to just the 0.1, 0.5, 2, and 10 Vpp ranges of the 5102.

6. The 5132 cannot perform RIS sampling (5102 can RIS up to 1 GS/s).

If you are only interested in something that is 1 MS/s, you could potentially look at some of the simultaneous sampling DAQ cards as he mentioned as well. Of course, these are not digitizers, and they generally include other features such as Digital I/O, Counters, and/or Analog Output channels which may not interest you. Also, the programming is different as they use DAQmx as opposed to NI-SCOPE (may or may not be a big deal). Finally, they will not have the bandwidth that you are probably interested in (I saw you mentioned 10 MHz input frequency) or the 50 Ohm input impedance, but I just wanted to mention them.

Hope this helps,

[dantm]

Thank you both for your comments.  I'll mention a little bit about the application.

We have a transmit and a receive microwave source which use a baseband signal (tone) at 10 MHz and which is upconverted on the Tx side and down-converted on the Rx side back to 10 MHz, with an I and Q output (hence dual channel requirement).  We want to sample this signal at 1 MSps (or slightly less) at the output of the Rx which is a 50 ohm SMA connector pair (one connector for I one for Q).

At this time the vision would be that we could use a 5102 (with or without solving the impedance mismatch issue) and sample this signal concurrently on two channels.

I looked at some of the multi-function cards but the understading there was that the signal bandwidth had to be << (much less) than 10 MHz -- in many cases the datasheets do not specify this at all.

Are there any other cards in the family lineup which will allow this to be done?

Thank you!

[dansch]

I think you would be best suited with the 5102, especially since you already have it. There are obviously many more feature-rich digitizers (such as the 5142, which has onboard signal processing for just this sort of application), but these are quite a bit more expensive than the low-cost 5102 or 5132, and they will come in the much larger PCI form factor.

Best Regards,