PCI 6601 Counter and speed problem

[Andrea]

Hi everibody,

I developed a system for frequency (period) measurement based on a PCI6601
for acquiring 4 different signals in a range 5Khz - 2Mhz.
For reason of speed I used only the DMA channel counter multiplexing the 4
input.
The PCI 6601 has a 20Mhz timebase used as source in a buffered measurement.
I have the following problem and questions:
- I noticed that for low freq signal the max number counted reachs 5 or 6
sample for run, instead on 2Mhz i can overcome the limit of 500 sample for
single run! Why? is it possible force the minimum sample or it depends only
on hardware limit?
- I need to reach a relative error of Hz so I modify the TIO buffered period
mesurement VI in the library adding another clause for exiting the cy
cle
setting a minum in order to produce an average. The problem here is after
the first loop for some error of downsampling I cannot reach this number.
Labview tells me that is due to HW limitation, but is it sure that
increasing the CPU something will change? (actually I have a 450Mhz Celeron
with win2000 OS and 256 Mb ram).

Thank you in advice for any help

A kiss
Andrea


__________________________________________________________________
Andrea Ing. ICQ#: 145715261 Current ICQ status: + More ways to contact me i
See more about me:
__________________________________________________________________

[Randy Schmidt]

Hi,

If you are willing to email me the code then I can help. i have done a lot
of work with the 6601/2/8 for a 32kHz signal with a resolution of 3 counts.

daqtek@charter.net

Randy

"Andrea" wrote in message
news:6wiA8.91646$vF6.2743330@news2.tin.it...
> Hi everibody,
>
> I developed a system for frequency (period) measurement based on a
PCI6601
> for acquiring 4 different signals in a range 5Khz - 2Mhz.
> For reason of speed I used only the DMA channel counter multiplexing the 4
> input.
> The PCI 6601 has a 20Mhz timebase used as source in a buffered
measurement.
> I have the following problem and questions:
> - I noticed that for low freq signal the max number counted reachs 5 or 6
> sample for run, instead on 2Mhz i can overcome
the limit of 500 sample for
> single run! Why? is it possible force the minimum sample or it depends
only
> on hardware limit?
> - I need to reach a relative error of Hz so I modify the TIO buffered
period
> mesurement VI in the library adding another clause for exiting the cycle
> setting a minum in order to produce an average. The problem here is after
> the first loop for some error of downsampling I cannot reach this number.
> Labview tells me that is due to HW limitation, but is it sure that
> increasing the CPU something will change? (actually I have a 450Mhz
Celeron
> with win2000 OS and 256 Mb ram).
>
> Thank you in advice for any help
>
> A kiss
> Andrea
>
>
> __________________________________________________________________
> Andrea Ing. ICQ#: 145715261 Current ICQ status: + More ways to contact me
i
> See more about me:
> __________________________________________________________________
>
>

[Andrea Pede]

Hi!
I have exactly the same problem.
Also I discovered that for low frequencies (below 200Khz) the maximum number of measure read (count read) doesn't go over 100 and looks like inversely proportional to frequency; i.e. for 9 Khz only 6 count read! Someone knows why ?

[bman]

Are you using the 'Count Buffered Edges VI'? I have tried it out and have not had any problems with it with a 660x board.

There are two parameters to set, 'counts to read' and 'time limit'.

The 'counts to read' parameter sets up how many values to read from the counter at a time. If, for example, you want to know the counter value for every pulse on the Gate, you would set this to one. However, for faster Gate frequencies, it is helpfull to set this high.

By default it is set at 100, due to this, it will read 100 counts at a time.

The 'time limit' parameter sets how long to wait for 'counts to read' number of counts. If you set the time limit to 10 seconds, though you don't get 'counts to read' number of counts, it will just return however many
counts it does have.

Let me know if this does not make since or you are seeing different behavior.

Brian