On Tue, 2020-06-30 at 01:23 +0100, David Slipper wrote:
>
> I'm running the latest version (downloaded today) on a 32bit version
> of win 7 and I am seeing a problem when recording a large (2GB)
> number of samples at 4MHz.
>
> It tends to stop at a varying point and getting nowhere near the 2GB
> limit that I set, sometimes it is almost immediate which is
> irritating when trying to track down a problem.
>
> The logic analyser is a 16bit fx2 based unit and I'm recording all
> 16 bits with a GPIB decoder.
>
> Is this a known problem or am I (more likely) doing something wrong ?
>
> The LA should be up to doing 12MHz for 16 bits, would have thought.
Given the popularity of the devices, and this very issue, I was
surprised to not find a FAQ item for it. Since the question must
have been answered a hundred times before (in different places).
Here is an attempt to answer the question. Feel free to add more
information, background, or activities that an affected user
could carry out. I really feel there should be a FAQ item, since
linking to the answer instead of "re-inventing it again" must be
easier for those who answer, and shall result in faster help for
those who experience the issue. Ideally users could find it even
before having to wait for somebody else to dig it up.
(FX2 based) logic analyzer terminates acquisition
before the specified limit.
Got trouble getting captures with an FX2 based logic analyzer
at higher samplerates. Acquisition terminates before the
specified amount of samples or time. Logs contain a message
that the device "... only sent ..." (a smaller amount of data).
That's a known constraint of the ubiquitous FX2 chips which are
found in many cheap and thus rather popular logic analyzers
(and also in some oscilloscopes). The high rates of 24MSa/s for
up to 8 logic channels, or 12MSa/s for up to 16 channels, are
near the theoretical bandwidth limit of the USB 2.0 connection
when communication overhead gets considered. In addition the
FX2 chip only has little memory for to-get-transmitted data
(covering less than a millisecond). That's why successful
communication heavily depends on the PC's capability to drain
the data which the FX2 chip provides. The slightest hiccup
makes the PC lose sync. Lost data cannot get recovered, and
it's uncertain which period of acquisition time was affected.
So the only remaining option is to terminate the acquisition.
Things to check: Sample data at lower rates. Pick proper cables
(those shipped with the cheap devices often are not up to their
task). Make sure USB bandwidth is not shared with other devices
(ideally put the logic analyzer on a separate port so that
nothing else occupies that bus). Disable features which could
stall the acquisition (separate the data acquisition from the
processing of that data). Reduce the PC's workload during the
period of acquisition. Some operating systems are said to
suffer more often from that issue than others.
virtually yours
Gerhard Sittig
--
If you don't understand or are scared by any of the above
ask your parents or an adult to help you.
_______________________________________________
sigrok-devel mailing list
sigrok-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sigrok-devel
