buffer overflow

[sergueiK]

TWGomez

but how you  maximize buffer size. Just by using maximum rate?

I saw the finction that allows to adjust buffer size manually. Not automatically.

I am posting image of error I am getting when the program crashes.

It says Buffer underflow not Overflow

[TWGomez]

Is that new error?

To adjust buffer size you modify the samples per channel, not rate.

That error makes it seem like the bus can't handle outputing that fast and you need to decrease the rate.

[sergueiK]

No I think that is an old error, I just did not see this screen as the device is sent to our customer.

When I played with rate some time  it behaved really strange: for some rate settings the frequency on the vibrator was double-presetted 13Hz and  vibrting with this frequency but some other that sounds like 100Hz or similar so motion of vibrator was like on a stepper motor ; for some settings the frequency was far from what i preset 3-4 Hz instead of 13Hz; or sometimes it was vibrating with some time intervals.

So I have chosen some Rate settings at which it did behave normally.May be there are some settings that are interconnected with rate and they should be changed together?

Message Edited by sergueiK on 04-10-2008 10:01 AM

[sergueiK]

We are a bit suspicious that this error could be connected with Antivirus software.

I have asked our customer disconnect the PC from internet abd disable antivirus program.

If that the case how this could be treated?

[TWGomez]

Hmm, Anti-virus shouldn't really be run when other things are running (personal opinion, not technological or scientific fact by any means). It tends to slow things down tremendously. However if the software where running so slow that things were getting over written in the buffer you should get something that says information is being over-written. In anycase as long as the buffer is set to the maximum size (as I should earlier) then you should have a fair amount of time to remove items from the buffer. I just got back from lunch, so I'll be checking more into it in a bit and see what I can find.

[sergueiK]

I'll get my hands on hardware tomorrow night(customer brings it back for 3 days).

So it would be possible to play with it Monday-Wednsday.

So you suggestion to change buffer size to the max. I just did not get how exactly it should be done.

Regarding Antivirus, I agree that it is better in this world to disable it completely, but unfortunately it is not safe and not always possible.

If it will cure my problem, OK in that case for this application we could just demand from our customer to disable Antivirus.

[TWGomez]

At the very beggining top left of your VI you call DAQmx Timing and set everything...in the "Samples Per Channel" input, right next to "Sample Mode" (that you have continuos wired into wire a constant with a 1M value. That will set the buffer to max size.

[sergueiK]

Ok, i'll try it.

[sergueiK]

I have got my hands on hardware.

I have tested it for over weekend and it worked with no problem at all.

But what we have faced today is: program stops to work (signal generation) when the window with the program is minimized and then maximized again. The program itself continues to run and increases output voltage(without output) and when I stop it, it records some big value for the generated voltage.

Another problem that program at the start uses about 60 MB of memory but then the memory usage by this program increases gradually. 

I saw about 90 MB used by it.

Message Edited by sergueiK on 04-14-2008 09:57 AM

[B.Wagner]

Hi all!

Sorry for jumping into your conversation but I experience a similar problem as SergueiK and after reading a lot of forum messages concerning this topic I thought maybe I can profit from your experience since I'm quite new to Labview.

to the problem:

I wrote a simple 1 to 3-channel (optional) acquisition .vi (see attached). For low acq-rates (till 10KHz) the continuous acq. works fine. But when I mount up to 80k - which is necessary for my experiment - I soon get the FIFO-buffer-overflow error message. I tried to raise the buffer size by setting a high number of samples per channel (p.ex: 240k) but this slows Labview that much, that the vi is not able anymore to stop after the set recording time  (up to several seconds later, if not even creating another buffer overflow).

I'm afraid that some poor programming from my side might be the main reason for this situation so I would be most thankful if you could give me some hints how to ameliorate.

Thanks in advance

Bernhard

---

my system:  DualcoreIntel 2.13GHz, 2Gigs RAM, WindowsXP prof., LabView 8.20 basic license, PCI-6259 with BNC2090A