RECOMMENDATIONS ON 4CH LOGGING SCOPE SOLUTION W/ SIGNAL GENERATOR

[asdfs]

What I need:

A scope card that will hook up to labview or comparable with 4 channels minimum and ability to sample around 50M samples/sec or so. I will need to log and compare data from separate channels recorded simultaneously. I also will need to output another frequency that will sweep from 0-500 pulses/sec with an amplitude of .1V or so.


That .1V output along with the other channel inputs will all need to be logged and compared. For instance, I will need to compare my output signal to CH1 input at 200 pulses/sec and compare the peak offset, if any.

I will be capturing square and sine wave data.

I am hoping what I am describing is a relatively basic 4 ch scope and probably not even in the high speed category.... I am just wondering if the output signal will be a problem. I know some cards do this.

I would prefer to buy a DAQ type device for this for use with a laptop rather than buying a complete scope. I will admit that my partner is much more knowledgeable in this area but he is too busy to shop around.

When I visit NI website, I get lost FAST wiwth so many choices. I need a little hand holding I guess to learn this stuff. I am also open to other options as well. Don't have to go this route but ultimately need to send and receive frequencies and record all of it.

[John_P1]

Hi asdfs,

Welcome to the NI forums!  50 MHz falls outside the range of our standard Multifunction DAQ devices, but you might consider looking into 2x USB-5132 if you require a portable solution at this rate.  One thing to keep in mind is that the 5132 cannot perform continuous acquisition: your acquisition cannot last longer than what is available in the on-board memory (although you could take several finite acquisitions triggered at the appropriate times).

As far as the pulse output, you would need a seperate way to generate this signal.  Possible candidates might be the analog outputs of boards such as the USB-9263 or the USB-6211.  Counters or digital outputs would require external attenuation but may also be an option (the digital output should be hardware-timed).

Compromising some of your requirements, we would also have a couple of other options available:

• If the 50 MHz sample rate is not required, one other thing to possibly consider would be the USB-6251.  It can sample 4 channels at 250 kHz as well as provide the outputs that you require.

• If you don't mind going with a PCI solution instead of a portable one, the PCI-6115 would offer 10 MHz/channel on 4 channels and also offer the outputs that you require.

I hope this helps you get started in your search.  If you have any questions about it don't hesitate to post back.  Thanks for posting!

-John

[asdfs]

Thanksa very much for your reply.  Are you with NI?  IE, you know all the products that are available? 

No, I do not think 50mhz sampling is a requirement but the faster the better.  I really think 10mhz may do fine for our current needs.  I really would like to get the signal generator built into the card for simplicity.  I will take a look at your PCI-6115.  Is that really all that is available?  I see a bunch of cards.  Hard to believe I knocked them all down with my requirements.....  

Regarding portability.  We do not ultimately require a "tight package" but something that can be put on a cart is fine.  Accuracy of readings is much more important than portability.  For 99% of it's use, it will not move from the test bench if that helps...

Will I need specific line conditioning for this card or just plug and play?  I know one of our outputs will be around 1-5V at 0-500hz.  

I will look at the spec before I ask an more dumb quesitons....Thanks

[John_P1]

Hi asdfs,

I am currently an Applications Engineer with NI working mostly with Multifunction DAQ and High Speed products.  There are certainly a large number of options, I was just trying to give a representation of multiple families that might fit your needs.  The exact board is going to depend on what sample rate you require and at what accuracy.  Since you had mentioned using a laptop, I tried to stay away from PCI and PCIe devices in my original post--it sounds like you might actually want to go this route afterall.

At 10 MHz, the 6115 would be the fastest board available that also has analog outputs, anything above this would fall into our Scopes/Digitizers line. One thing to keep in mind is that 10 MHz on 4 channels would be equal to 80 MB/sec of data which is very close to the practical limit of the PCI bus (and above the practical limit on most standard hard drives).

I should also probably mention our X Series line.  These boards are available for PCIe and PXIe--we have announced a simultaneous version for PXIe that can sample at 2 MHz per channel (PXIe-6366) , but this would require an investment in a PXIe platform.

An overview of some of the products mentioned:

Scopes/Digitizers: High Rates (50 MHz -  2 GHz) but no Analog Outputs.  Products in this family will be typically numbered 51xx and use the NI-SCOPE driver (all products below use the DAQmx driver).  The PCI-5105 might also be a good board to look into, as it would have enough input channels for your task (the outputs would still have to come from another source).

S Series:  Simultaneously sampled Multifunction DAQ Devices.  Rates up to 10 MHz per channel.  Products numbered 61xx.

X Series:  Our newest line of Multifunction DAQ, offers simultaneous or multiplexed options, but only up to 2 MHz per channel currently.  Only available in PCIe or PXIe.  Products numbered 63xx.

M Series:  These boards are strictly multiplexed (the sample rate is divided across all channels being used).  Unlike X Series, these boards are available in PCI, PXI, and USB form factors. The products are numbered 62xx.

I would have a look at the above to help narrow down your exact requirements.   1-5V should be just fine without any external conditioning (the above devices have programmable gain that would accomodate a variety of ranges). 

Are you sure about the speed requirement?  If the signal is only 0-500 Hz you might not need to sample at such a high rate.  Are you trying to characterize the edge of a square wave?

-John

[asdfs]

Yes, John, that is precisely what we are doing.  The frequency of the square PWM signal would not exceed 500hz but we would need to accurately capture the edges of the pulse to discover the pulse width.  Possibly you could recommend some sampling speeds for something like this?? 

I realize accuracy will ultimately determine the sample speed but we are estimating about +/- .05ms error in the pulse width time elapse if that helps.   Keep in mind that is a round about figure for now but our plan was to over kill the sampling so we have more than enough accuracy for other test procedures.  I will try to get a more accurate sample accuracy for you so you can better recommend products.  

[asdfs]

John, upon further discussion in the shop, we really think 200Mhz sampling would be the absolute max with most tests being in the 10k-100K range.  It was voiced that possibly with the lower sample rate requirement, we could get more card for our money with inputs, outputs, and I/Os.  I was scolded for considering a card just for the scope requirements and that I should be considering a larger board to cover everything in one shot.  That will likely be upwards of 16ch of analog inputs and probably several digital. 

Most of our analog inputs will be IR tachs, either thermocouple or IR temp sensing, and potentiometers as well as our scope requirements at much lower sample speeds (probably 1000hz).  

We will also need programmable outputs similar to PLC type functions that will compare the sensor data and output accordingly while logging all the data.  

Hope some of that made sense and I apoligize for bouncing around a bit.  

[John_P1]

Hi asdfs,

It sounds like you want to perform a wide variety of tests--let's try to get a list of what tests you need to do.  For example, if you are interested in measuring the width of a digital pulses instead of digitizing them, you can actually use the on-board counters of our multifunction DAQ devices which can use a relatively fast timebase (100 MHz on X Series, 80 MHz on M Series)--you would just need to amplify the signal a bit to bring it up to TTL voltage levels (0 - .4 V low, 2.4 V - 5V high).  The 100 MHz timebase of the X Series would give you a pulse width measurement up to 10 ns accuracy (this could be improved even further by averaging over time).

It sounds like most of your other measurements might be close to DC.  I guess what I want to point out is that you might not need as high of an analog input sample rate as you think (but this would depend on what information you are looking for out of the PWM signal).

The control aspect of your application makes things a bit more complicated ("PLC type functions").  What do you mean specifically by this, and what kind of latency do you need between the input and the output?

-John

[asdfs]

I apologize again for my lack of definative input here.  I did not plan to aquire a multi use platform yet so a little premature I believe but hopefully you can point us in the right direction on your products. 

The PLC type functions will be controllers to control other inputs.  Basically, consider for simplicity, a car going down a variable decline hill.  The controller will need to receive tach inputs like wheel rpm and other contol factors to control the decent at a decided mph.  (Keep in mind this is just for example, this will not be mobile).   The output will need to apply either a variable sine, PWM, or variable DC to control a brake system to maintain a certain speed.  We would also want ot program other functions such as a programmed deceleration rate until the target speed is met and keep it there, decelerate to a stop, etc.  

In a nutshell, the procedures will need to be automated so as certain inputs are sensed and recorded, they are also going to be part of a logic control to output signals to keep the inputs in check.  

The outputs would be easiest to work with if analog.  Regarding latency, I don't think we have enough R&D in this yet to accurately discuss that.  

I will honestly have to sit down the the crew and get more specifics on the setup.  The counters should be just fine for tach inputs.  We would not need to look at the wave forms, just pull a frequency from the signal for output controls.  All of the input and output data will need logged.  I figured if we could do this in one card, it might save some setup hassles and maintain better accuracy.  

[John_P1]

Hi asdfs,

Sounds good, be sure to let us know if you still have any questions after narrowing down the specs a bit more.  The latency requirement can make a huge difference in the hardware required--if you only need to react within a few ms of the sensor inputs then you will likely be fine with a standard PC-based solution like the ones we have discussed so far.

-John

[asdfs]

John, would you have multifunction products that will provide the type of outputs I have described?  Controller type outputs?