PXI-5105 8 channel

[plibby@efw.com]

I am planning on using the PXI-5105 to do some data acquisition on some signals that will range from 1 MHz to 10 MHz and the amplitudes will either by 7 Vpeak or 25 Vpeak.

Will this board be able to measure this Bandwidth and can the analog inputs handle the 7 and 25 VDC level?

[Matt_E.]

Hi Peter,

I am assuming this is the same application you mentioned in your previous post about the 5114.  The 5105's largest input range is 30Vpp which means the maximum positive voltage it can measure is +15V.  The 5105 does not have a vertical offset feature like many of our boards so it will not be able to measure the 25V signal you have.

Many of our other boards do have the ability to measure 25V DC but you must use the vertical offset feature of the board to shift the board's measurement range to cover the area you are interested in.

For example the 5114 has a 40Vpp input range with +/- 15V of vertical offset.  If you program the board's vertical offset to 0V and the input range to 40Vp then you could measure any signal from -20V DC to +20V DC.  If you leave the range at 40Vpp but program the vertical offset to +15V then you would be able to measure from -5V DC to +35V DC.  This would allow you to measure your 25V signal as long as the signal doesn't go below -5V.

Another approach you can use to increase the input voltage range is to use a 10x probe.  The probe connects to your signal and outputs a signal equal to 1/10 of the input signal.  This effectively increases the 40Vpp range to 400Vpp (+/- 200V DC max).  If you were going to use a 10x probe in your application with the 5114 you would probably want to use the 10Vpp range which would give you a 100Vpp actual input range (+/- 50V DC).  The NI-SCOPE driver software has a value you can set to tell it what probe attenuation you are using so it will automatically scale your data so the voltage you read in the software is the voltage at the input of the probe.

If you are willing to use probes then most of our digitizers will work for your application, you will just need to pick the one that has the resolution and bandwidth you need.  The 5105 is not the best choice for applications using probes because the connector on the front of the board is an SMB and most probes are designed to connect to BNC connectors.

-Matt

[plibby@efw.com]

Matt,

Could I use a simple voltage divider before I connect to the 5105?

[Matt_E.]

Depending on your situation this could work for you.  If you only care about measuring DC or low frequency signals you shouldn't have any problems but if you have high frequency content (such as fast rising edges in digital signals) the input capacitance of the digitizer will start to lower the input impedance at high frequencies.  The easiest way to think about what is happening is if you hooked up a voltage divider to your signal the 1Mohm input resistance of the digitizer would be in parallel with the lower resistor of your divider.  There is also at least 56 pF of capacitance in parallel with this 1Mohm resistance.  At DC this capacitance looks like an open circuit so you can pretend it isn't there at all.  At higher frequencies however this capacitance looks like a lower impedance, eventually this input impedance of the digitizer will not look 1Mohm anymore but it will be something lower, the value keeps decreasing as frequency goes up.

If the equivalent impedance of the digitizer is close in value to the lower resistor of the voltage divider the digitizers impedance starts to reduce the impedance of the lower part of the divider and this changes the divider's ratio.  This means that even if you design the divider to have a ratio of 1/10 it won't necessarily have this ratio over the full frequency range.  You can help this problem by choosing lower values of resistors for your divider but you also have to be careful not to pick values so low that the resistors burn up from power dissipated in them by the source you are measuring.  For example if you wanted a 1/2 ratio and used two 100 ohm resistors you would have I = 25V / 200 ohm = 125 mA.  The power dissipated in the resistor would be P = I^2 * R = (0.125 * 0.125 * 100) = 1.56 Watts!!!.  You would need some very large resistors to handle this power and to make matters worse this will put a large load on your source that you are trying to measure.  Your source may or may not be able to supply this current but even if it can now the load on the source while you are measuring it is not the same as when the scope is not hooked up so what you are measuring may not be what the signal looks like when you aren't measuring it, somewhat of a catch 22.

10x probes take care of this problem by putting a variable capacitor in parallel with the top resistor of the divider which you then adjust to compensate for the capacitance in the bottom half of the divider.  This is known as compensating your probe.  You could follow a similar approach but the 5105 still may not be a good candidate because you will likely have a very large capacitance on the bottom half of the divider and it can be difficult to get a good variable capacitor that has enough range for this application.  Our other digitizers have lower input capacitance and also a 10x scope probe cable has much less capacitance than the coax cables you would have to use on the 5105 which is why the other boards are a better choice.

The main reason people use probes with digitizers that were designed for use with probes is that these issues have already been considered and the probe was specifically designed to not load down your source and to be able to compensate for this non-DC gain error.  It is certainly possible to create your own setup using the 5105 but you will have to consider many more issues if you do.

-Matt

[plibby@efw.com]

Thanks Matt.

I can't get away from the 5105 becuase of its high density channel count and my limitations with the test station. To make a long story short I have to monitor 32 output channels simultaneously with four available PXI slots. This is why I have chosen the 5105.

Is there IC device I could use instead of the basic voltage divider solution?

Peter

[Matt_E.]

How much of a load are you willing to put on the source you are trying to measure.  There are some fairly simple things that could be done with 1 resistor and a 50 ohm coax cable but they would look like about a 2kohm load to your source which may not be acceptable.

-Matt