PCI 6534

[sahara agrasen]

hi to all,

i m looking for one high speed digital I/O card. i seen the data sheet of NI PCI 6534 and adlink PCI 7300A. i seen there is not much change in specs when i compare both the specs of card. only two specs differs from each other. onboard memory which NI offers is 64Mb as well minimum pulse width detection capability of NI PCI 6534 is 10 ns. but cant be able to judge becuase of not any knowlege of importance of these two parameters. can any one please explain it to me.

[Neville D]

Having more memory onboard means you have to query the card less often when reading the data off of it.

Also try to get a card that supports DMA (and the more DMA channels the better). This helps with simultaneous AI and AO if you need it.

But really the biggest advantage with using NI is the seamless support with Labview/Daqmx. You won't get that with any non-NI cards, and for that reason alone I would stay away from them.

Obviously, if you need a specialized function that is not available on an NI product then you would be forced to deal with all the integration issues of getting the non-NI product to work well with the rest of your LV code.

Neville.

[Louis Manfredi]

QUOTE (Neville D @ Oct 17 2008, 02:05 PM)

But really the biggest advantage with using NI is the seamless support with Labview/Daqmx. You won't get that with any non-NI cards, and for that reason alone I would stay away from them.

Neville.

I wouldn't exactly call the LabVIEW/Daqmx support of the 653X series of cards "Seamless" -- I've found the documentation for 653x DaqMX a little cryptic, and for things like figuring out how to use external sample clock input, hardware triggering, and the like, you need to refer also to the traditional nidaq documentation, which is more thorough, but unfortunately seems to use different signal names, from the Daqmx. Count on needing a little patience to understand this hardware's finer points.

Just my perspective, and in any case, Neville's point is still valid-- even if the support isn't seamless, the good folks at NI support will do their best to get you across the seams. If you buy hardware from someone else, they may lack the moral obligation, motivation, and probably even the knowledge to help you out.

Best Regards, Louis

[Neville D]

QUOTE (Louis Manfredi @ Oct 17 2008, 11:48 AM)

I've found the documentation for 653x DaqMX a little cryptic, and for things like figuring out how to use external sample clock input, hardware triggering, and the like, you need to refer also to the traditional nidaq documentation, which is more thorough, but unfortunately seems to use different signal names, from the Daqmx. Count on needing a little patience to understand this hardware's finer points.

True, the old-style documentation is a lot more thorough, and carefully walks you through steps of making different types of measurements. Some of those pdf's don't even ship with Labview any more :thumbdown: . NI seems to rely more and more on clunky "help" files that are no help.

That said, the DAQmx API is indeed fast, multi-threaded and very powerful.

N.

[Gabi1]

QUOTE (Louis Manfredi @ Oct 17 2008, 08:48 PM)

Just my perspective, and in any case, Neville's point is still valid-- even if the support isn't seamless, the good folks at NI support will do their best to get you across the seams. If you buy hardware from someone else, they may lack the moral obligation, motivation, and probably even the knowledge to help you out.

first, the 6534 is a DIO card. no analog there. output rate is up to 2Mb/s, in burst or handshake mode.

the cards work similarly with DaQ and DAQmx, but the DAQmx definitely offers more control and readout options, for example for the buffer state.

the correlation of the card with NI would make sense, but i have to warn you the technical knowledge of the support engineers is basic at most, as they dont do high speed DIO... so advanced bugs or issues you encounter will not be solved. Specifically, some bug with Labview doesnt allow the card to operate at 2MHz even in Burst mode. depending on computer power i found 1MHz to relatively deliver stably, but not above.

one note: go to the 6533. much cheaper, also can be found on Ebay. no buffer on the card, but for today computers i have not found performance differences in both input and output.

note 2: NI have now developped new version of those cards, that are more reliable, and perform better in terms of configuration and programation with LV, allwoing to use only the high level functions delivered by LV. if not too expensive for you, i would go on those. they also feature higher data rate, in the range 5Mb/s and above (but then you will probably need onboard memory).

[sahara agrasen]

can anyone please let me know the importance of pulse width and onboard memory options. in what kind of application required this two points. or it is useless?

QUOTE (Gabi1 @ Oct 17 2008, 09:33 PM)

first, the 6534 is a DIO card. no analog there. output rate is up to 2Mb/s, in burst or handshake mode.

the cards work similarly with DaQ and DAQmx, but the DAQmx definitely offers more control and readout options, for example for the buffer state.

the correlation of the card with NI would make sense, but i have to warn you the technical knowledge of the support engineers is basic at most, as they dont do high speed DIO... so advanced bugs or issues you encounter will not be solved. Specifically, some bug with Labview doesnt allow the card to operate at 2MHz even in Burst mode. depending on computer power i found 1MHz to relatively deliver stably, but not above.

one note: go to the 6533. much cheaper, also can be found on Ebay. no buffer on the card, but for today computers i have not found performance differences in both input and output.

note 2: NI have now developped new version of those cards, that are more reliable, and perform better in terms of configuration and programation with LV, allwoing to use only the high level functions delivered by LV. if not too expensive for you, i would go on those. they also feature higher data rate, in the range 5Mb/s and above (but then you will probably need onboard memory).

[LAVA 1.0 Content]

QUOTE (sahara agrasen @ Oct 18 2008, 06:23 PM)

can anyone please let me know the importance of pulse width and onboard memory options. in what kind of application required this two points. or it is useless?

Depends on your needs.

If you are interested in trigger signal in the form of a short pulse than the pulse width is important.

If you want to read during long times DIOs the onboard memory is important.

However 64 MB onboard memory is a lot.

Consider 10 kHz, 32 channel, this means 320 kB per second, resulting in 2000 seconds of memory.

Do the math and see if it is worthful, any memory available over 1 second should be enough. End users expect every 100 msecs updates.

Ton

[Gabi1]

QUOTE (Ton @ Oct 18 2008, 10:31 PM)

Depends on your needs.

If you are interested in trigger signal in the form of a short pulse than the pulse width is important.

If you want to read during long times DIOs the onboard memory is important.

However 64 MB onboard memory is a lot.

Consider 10 kHz, 32 channel, this means 320 kB per second, resulting in 2000 seconds of memory.

the onboard memory is usefull only at high rate of transfer, well above 1Mb/s. below that, for any decent computer, the operating system can store the data directly on the main board memory for unlimited time (depending on operating mode, you can set new data to overwrite old data). that is why 6533 would be sufficient for you (no onboard memory).

But if we dont get more input on your application it is difficult to give effective advice. if you only want triggering and DIO on the KHz regime, there are much cheaper cards and more versatile either. the NI USB 6009 might even be sufficient for you, inlcuding several DIO and some analog ports. i recall this is on the range of merely 100$. a bargain for what it propose. the 6534 i think it is not sold anymore by NI, and they have newer version, but it cost ten fold more!