discretely jumping of frequency sweep in AWG

[labmaster]

Paul, Thank you so much for the reply.

I want to correct "1" to "4". In this case, there are huge numbers so that labview is out of memory in calculating waveform before loading to PXI.

I have to say, privately, NI developing department of AWG has a responsibility to give suitable numbers to AWG users.

or guideline for the memory or procedures.

1) Can I obtain the VI code of "Analog Signal Editor" or similar?

I just asked the price of upgrading memory to my NI sales engineer.

2)Can you estimate the PXI memory required for a waveform? for example, how much the memory for the 100 points waveform?

Currently, I have to use my modified code with out of phase (around several degrees in each frequency.)

However, I just posted one more problem about DC offset which is limited to 25% of AC amplitude.

Currently, my application needs to be applied the voltage of offset separately. 

3)Is there a commercial singal adder or a way to overcome this problem for my purpose(AC:1MHz~10MHz, DC:+-10V)? or in my posting on the forum.

Thank you so much, Paul.

[Paul_C.]

Hello Labmaster,

I understand some of the frustration.  The problem is that this type of applications or more geared towards our cards that support frequency list mode. I do believe we can get something working with the 5421, but it will take some work and tweaking.  Here are answers to your questions:

1) Can I obtain the VI code of "Analog Signal Editor" or similar?

Can you please explain why you need the code for this?  In addition, I do have some modified code that you can use to reduce the size of the waveforms.  See the attached document.  Try it out and let me know if it doesn't help you resolve the issue.

2)Can you estimate the PXI memory required for a waveform? for example, how much the memory for the 100 points waveform?

The memory limits are available in the specifications document here.  The 8MB stardard option allows for 4,194,176 samples.  It is important to note that your script information does require some space on the onboard memory.  This should be small compared to the size of your waveforms, however you should spare some room for this information when determining of a memory option.

3)Is there a commercial singal adder or a way to overcome this problem for my purpose(AC:1MHz~10MHz, DC:+-10V)? or in my posting on the forum.

Can you tell me what load you are generating the signal into?  What impedance, type of load, etc.?  Is it ok for your application to do the offset digitally?  What does your signal i bamplitude and offset need to be?

Regards,

Paul C.

[labmaster]

Thanks Paul.

1,2) Forget my request. I got to know memory is more urgent to me.

Although I requested to my sales engineer, I would like to ask if I can upgrade the memory? Our PXI has a 8MB.

Can you explain me why you used especially "divided by 2" and other 1E3 (maybe for 1kHz resolution) and 1E6 numbers in the loop?

I would like to develop more your code.

3) Although I can't estimate, the impedance will be much higher than 1 kohm to the application.

My application is a bias voltage terminal in non-contactedI in very small gap.

It needs to apply a AC(1M~3MHz, 6Vpp)+DC(+-6Vpp) in a bias voltage.

I don't care of digital instrument but both signals should be added to a single terminal.

Have a good idea or alternatives?

Thanks so much again, Paul.

Labmaster.

*)Back to my original problem, I used non-ratio of 100MHz (50.23445333MHz) as a sampling rate.

Generally, is it not allowed? only multiple of 100MHz (50Mhz?)

[Paul_C.]

Hello Labmaster,

1,2) I'm not sure if you are able to upgrade the memory.  Your sales engineer will more likely know if this is possible.  The divide by 2 section is simply to determine if the result is even or odd.  For even results, you can simply divide the desired frequency by 1M and multiple that by the sample rate to get the minimun number of samples.  In addition, you are correct in the fact that the divide by 1000 prior is to adjust is for 1KHz steps.

3)  I'm still a bit confused on exactly what type of offset you need.  From the way it sounds, you may be needing something outside the capabilities of our device.  Are you looking to generate a signal that is 6Vpp with a DC offset?  I'm a bit confused on the DC(+/-6Vpp).  Can you clarify the amplitude and offset of the final signal?  A picture might be helpful.  If you are looking to amplify the signal outside the range of the card, you might need to consider an amplifier circuit (op-amp, etc.) on the output of the card.  

*)Back to my original problem, I used non-ratio of 100MHz (50.23445333MHz) as a sampling rate.

Generally, is it not allowed? only multiple of 100MHz (50Mhz?) 

Is this clock provided externally?  If not, we do have high-resolution clocking available on the 5421.  This should able to get you resolution to 1.06uHz.  

Regards,

Paul C.

[labmaster]

Paul, I appreciate for your reply.

Regarding to offset, sorry for my unsuitable notation Vpp in Dc.

please see my pictures.

I want to control both AC and DC independently.

I found a suitable wide-band op-amp but I am not sure if I can make a success.

If you get to know something about this within NI or other commercial products, please let me know. 

I will post again today for the sampling rate.

I remembered to see a limitation of internal sampling rate in FGEN and Help file.

Sungjun.

Message Edited by labmaster on 05-12-2009 08:14 AM

[labmaster]

Paul, there is another limitation in the sampling rate.

The minimum is 1E7 (see FGEN error message).

In terms of multiples in frequency, did you mean high resolution mode in PXI-5421?

I guess I can set the high resolution mode by Configure.vi.

If so, I can use any frequency(such as 10.00002MHz).

am I right?

Do I have to reintialize FGEN after Closing to change the sampling rate? or set sampling rate once more without closing?  

Does the aribrary waveform mode include "script" mode?

In specification or help files, I can't find any contents of script mode (contrary to sequence mode) in memory calculations.

 Best,

labmaster 

Message Edited by labmaster on 05-12-2009 10:09 AM

Message Edited by labmaster on 05-12-2009 10:13 AM

[labmaster]

I thought about your code you posted.

You took an approximation with the number of samples.

But, you didn't think about phase of generated waveforms because functions will be generated continously.

As your alternative, I would like to propose to check the phase (threshold of some value) inside while loop.

But, it takes a long time according to the threshold.

Anyway, thank you Paul for supplying code.

Labmaster.

[Paul_C.]

Hello Labmaster,

The code was simply meant as guidance on how to find the amount of samples.  Unfortunately, you may need to make modifications for it work exactly with your application.  The phase should be handled by the sine wave function generator based on the number of samples and frequency provided.  I would take a look at the labVIEW documentation for this Vi for more information.  In addition, you might post questions about these fuctions calls directly to the LabVIEW support forum.   

Regards,

Paul C.

[labmaster]

Hello Paul,

Recently, I met the limited situation, single cycle in AWG.

When I used some multiple cycles, it brought a serious delay in script trigger mode.

I checked this by external or software triggers.

I have to present two frequency signals (around several hundreds kHz and around 2 MHz.).

I have a degree of freedom in sample rate (high resolution mode).

Can you suggest me some routines for finding a common sample rate in one or less than two cycles of two random (above ranges) frequencies?

labmaster.

[Paul_C.]

Hello Labmaster,

I'm afraid this routine isn't trivial and may take some time to complete.  My recommendation would be to break this into the formula's required and figure out exactly how you want to make the routine.

The equations you could use are something like the following:

1st Desired Frequency = Sampling rate / 1st # of samples

2nd Desired Frequency = Sampling rate / 2nd # of samples

1st # of samples and 2nd # of samples has to be an integer or have a remainder of 0.5 and be divisible by 4.

I would play around with these formulas and find out exactly how you want to compute the sampling rate.  Once you have a formula, I'm sure the guys on the LabVIEW forum would be happy to give you suggestions on getting this into LabVIEW.

Regards,

Paul C.