[time-nuts] Pendulum CNT91 with TimeLab
GandalfG8 at aol.com
GandalfG8 at aol.com
Fri Oct 11 10:42:39 UTC 2013
Hi John
I can assure you that "helped" is very much an understatement:-), and again
many thanks.
Thanks too for your comments re talk-only vs native drivers, I understand
better now and will keep those archived for future reference.
There's certainly a lot less incentive now to further investigate where
I'm going wrong with the CNT91 talk-only configuration, I'm assuming it's me
rather than the counter anyway:-), but if I do ever get it sorted I will
certainly report back.
Regards
Nigel
GM8PZR
In a message dated 10/10/2013 20:05:21 GMT Daylight Time, john at miles.io
writes:
Good to hear the modified driver helped -- thanks for the confirmation!
I'll leave that option in the code for future releases.
There's no compelling reason to use the generic talk-only driver with a
counter that is also supported by a native talk/listen driver, as long as
the latter works reliably. The native drivers don't do much to configure
the counter, but they can do one or two helpful things such as ensuring
that
certain statistical features that would interfere with the measurement are
turned off. They also allow the counter to reside on a GPIB segment
that's
shared with other addressable devices, which is problematic in talk-only
mode.
The trouble with the native drivers arises from the fact that the only TIC
I
have around here is the HP 5370B, which doesn't use the same driver code as
the SCPI-era models like the 5313xA, SR620, DTS207x, etc. So when I make
a
change that affects those drivers, it's always going to be a "Customer as
Beta Tester" scenario. And we all know how well untested code usually
works.
The need to support Prologix USB, Prologix Ethernet, and NI 488.2 adapters
complicates matters further because their low-level behavior, timing
characteristics, and driver interfaces couldn't be more different from each
other. Add in the notion of counters like the Pendulum that are
almost-but-not-quite compatible with supported models, and the fact that
even the native TimeLab drivers still rely on the user to set up the
measurement parameters, and you get a pretty intimidating matrix of
conditions that are realistically never going to get tested until someone
tries to use them.
The talk-only driver simplifies the picture because I can test it with the
5370B just by flipping a DIP switch and verifying that it works with the
three supported GPIB adapter types. It doesn't rely on issuing
device-specific trigger and read commands to counters I don't own.
Talk-only measurements are still not 100% reliable due to occasional
weirdnesses in the GPIB adapter or counter firmware, but those aren't my
fault. :)
Basically, just use whichever driver works most reliably in your particular
situation, and keep the reports coming so we all know what works and what
doesn't.
-- john, KE5FX
Miles Design LLC
> Hi John
>
> Just as a quick follow up, I'm still intrigued as to why I've been
> seemingly unable to get the CNT91/FCA3100 to run in talk-only mode and
> might well
> come back to that when I have a bit more time but, from a TimeLab point
of
> view, with the HP emulation up and running would there actually be
> anything
> to be gained by doing so?
>
> Regards
>
> Nigel
> GM8PZR
_______________________________________________
time-nuts mailing list -- time-nuts at febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
More information about the Time-nuts_lists.febo.com
mailing list