how do i create AO voltage waveforms of about 1MHz with high speed M series DAQ?

[scarface]

hi kevin,

thanks again for all the help.basically, what needs doing is:

1.  send 2 (to start with, then 3) seperate voltage pulse trains out of the DAQ  card - these pulse trains must both be of the same frequency (~100KHz-1MHz), the individual pulses themselves must have a specific width in time which can be varied (0.1 microseconds - 1 ms), and they must have the same amplitude (i now know about the 0-5V imit and that is fine) - s this last requirement is not a problem.

     also, we need to somehow implement an initial time delay between the two pulse trains - on the order of 100 ns - 1 ms.

2.  this is where the important bit of the experiment happens.after sending our two pulses down some wire, they are then supposed to interfere with eachother creating a superposition signal along the wire.

3.  we then need to measure the voltage signal with time at different points along this wire (this is where the the digital inputs to the DAQ card come in).

is this at all possible?

steve

[Kevin_Price]

---

1.  send 2 (to start with, then 3) seperate voltage pulse trains out of the DAQ  card...

2.  this is where the important bit of the experiment happens.after sending our two pulses down some wire, they are then supposed to interfere with eachother creating a superposition signal along the wire.

3.  we then need to measure the voltage signal with time at different points along this wire (this is where the the digital inputs to the DAQ card come in).

---

1. 2 counter pulsetrains should be fine, but a 3rd is impossible using counters because your board has only 2 counters.

2. Are you planning to connect these 2 independent outputs to a common wire?  If so, DON'T.  You will NOT get a summed "superposition" signal and you risk damaging your board!   What are you expecting as a "superimposed" output -- a logical OR, a logic AND, a voltage SUM, or something else?

3. Are you trying to measure signal propagation delays?  I'd think your sampling positions would need to be quite far apart along the wire to have any hope of quantifying such an effect.  Like 10's or 100's of meters.   I'm not an electronics expert, but as you make your wire longer to separate the sampling positions, you also increase the capacitance you're trying to drive.  I'm not so sure the trade-off will work out to your advantage.

-Kevin P.

[scarface]

hi kevin,

thanks for the help again.theres no problem with sending multiple outputs down our wire, as its not an orthodox wire - its a nanotech fabricated wire that allows such superposition possible.

the main things i need to know are:

1.can i generate a 2 digital pulse trains (defined frequency ~ 1MHz-100KHz; defined temporal width - 0.1 microseconds - 1 ms; and a defined amplitude - 0-5V: where these values can all be changed within these limits for different runs) from two counters?

and 2. is it possible to generate an initial delay (100 ns- 1ms) between these two counters, and still have both pulse trains written to the digital output with no problems?

steve

[Kevin_Price]

I'll take your word for it on the superposition business.  To address your specific questions:

1. Yes you can generate 2 pulse trains with your 2 counters.  You can vary the frequencies and widths in the ranges you want.  The amplitude can't be controlled, but will be well-defined and fixed, cycling between ~0 and a~5 volts.  Note that because the freq is generated based on a 80 MHz timebase, then you will only be able to produce certain discrete frequencies.  1.00000 MHz is possible based on (80 MHz / 80).  The next lower possible freq is (80 Mhz / 81) = 987.654321 kHz.  Nothing between them is possible.

2. Yes the two pulse trains can be offset from one another in the range you want. They'll need to be triggered off a common signal while using a different value for "initial delay".  The minimum possible delay would be 12.5 nanosec

-Kevin P.

[scarface]

hi kevin,

thanks for all the help with that.i have managed to create a program that outputs 2 counter pulse trains with variable width and frequency, which also can be seperated by a variable initial time delay.the program then aquires whatever data i need through a couple of analog inputs and shows the data in a waveform graph.

i was just wondering, seeing as it now seems that we will need to use variable voltages for the input pulse trains (ie.not just 5V), would it be possible to set up such a program which does what this 2 counter program does but allows the output pulses generated to have variable voltages (between 0 and 5V)?i am a bit worried that if i use the counters as triggers for generating analog waveforms or digital pulses i will not be able to attain very high frequencies.

could you give me any help with this?

regards

steve

[Kevin_Price]

I don't have any experience trying to generate MHz-realm analog outputs.  Offhand, I'd say you'd really have your work cut out for you, especially if you try to do it with 2 analog output channels on a single board.  All of the timing intervals (pulse widths, phase offsets, periods) would be restricted to be integer multiples of the fastest analog output sampling interval.

To generate a single 1 MHz analog square wave, you need 2 MHz output capability.  To generate 2 of them with 0.1 microsec offset, you need 10 MHz output capability.  And you'll have to define your analog output buffer very carefully.

I wonder if there are any special purpose external signal conditioner devices that would accept 5V TTL inputs and then output a variable amplitude non-TTL pulsetrain <= 5 V? 

-Kevin P.

[scarface]

hi,

i'm trying to create a program that writes a finite number of samples from a waveform (created from a general digital waveform generator which i made- that works fine) to an analog output channel and then reads in the same number of samples through an analog input channel and saves all this data to file.however, when i run the program, it comes up with an error and says:

Read failed, because there are no input channels in this task from which data can be read.

can anyone suggest any possible solutions to this problem?

steve

[TheWay]

Hi Steve,

Please refer to this post for a solution to your question:

http://forums.ni.com/ni/board/message?board.id=250&message.id=27790

Way S.

NI-UK Applications Engineer