digital pulse train output frequency limit for USB 6229?

[lcdos]

The equipment is a USB 6229

I am using a counter to generate clock. Then use the generated clock for digital output to make a finite pulse train.

My code is imilar to the example of "Multi-Function-Ctr Pulse Train Generation for AI Sample Clock.vi". The pulse train I am generating is around 350 kHz, thus I want the clock to be somehow 10-20 times higher.

The problem is, it seems the clock can go to pretty high frequencies, such as 10MHz. Yet the the "digital Wfm 1Chan NSamp" give out error on any clock frequency higher than 1M. My question arises here, am I reaching the up frequency limit on the digital output for USB6229?

If the eqipment is not adquate, would you suggest another NI card?

[Jennifer O]

Hi lcdos and welcome to the Discussion Forums.

If you refer to the Specifications for the NI-6229. The clock rate for digital waveform measurement/generation (Port 0 only) is 1MHz maximum. 

Are you generating an irregular waveform from the digital channel (i.e. it is not just a square wave)?  If you are just working with a square wave, the counter can be used directly.  Assuming you are not generating a basic square wave, I am curious as to why you want to generate at 10 or 20 times the rate of the pulse?  Generating at 2x should enable the high and low  transitions.  The faster generation rate is typical with analog waves that require the smooth curve update, but digital signals which only register at two levels (high or low) do nor require it.

If you do have reason to run your digital task at .3.5MHz-7MHz you could could consider using an NI-625x board.  These are available in USB, PXI and PCI forms and are capable of up to 10MHz when using the onboard FIFO.  If you need to stream from memory the USB option is limited to 1MHz.  This timing information is also system dependent which means that the card can generate at the full speed so long as data is available.

Hopefully this helps,

Jennifer O.
Applications Engineer
National Instruments

Message Edited by Jennifer O on 06-11-2007 01:00 PM

[lcdos]

Thank you Jennifer,

I am generating a square pulse train which I need to alter duty cycle and period between the consecutive pulses. That is the reason of having higher clock.

For example, when specified before hand, I need to generate a square pulse train with 3 pulses of 10% duty cycle and 5us period, followed by 2 pulses of 15% duty cycle and 6us period, then followed by 10 pulses of 54% duty cycle and 8 us period. I would need 6 different pulse trains simultaneously, yet correlated in time.

Probably I should call what I need digital waveform, rather than square pulse train. I am Mechanical Engineer by training, so please bear with me.

I figured out I can not get what I want by just using timer. What I need to use are DO lines, thus the 1M bandwidth of USB6229 is insufficient.

I don't have to use USB, I am weighing between NI 625x and PCI 6534 now.

Thank you very much for your reply.

lcdos

[Jennifer O]

Hi lcdos,

Either board should do the job.  Here are a few things to consider:

1) The NI-625x has a variety of functionality including analog input, analog output,  digital I/O and  counters which may benefit if the application expands.
2) The NI-6534 has more hardware timed digital lines in case the application requires more digital in the future. This board also has programmable power-up states.
3) If you have a PCI Express slot available, the new NI-6536/6537 would be recommended over the NI-6534 for its selectable voltage and higher throughput.

A counter board would be able to output various frequencies at changing duty cycles, however you would not have control over the number of pulses at each specification.

Let me know if there is anything I can help you with in regards to this decision.

Regards,

Jennifer O.