Problems with Tektronix Open Choice desktop software

[Alexander Schnase]

Dear readers,

here I like to ask about experience
on a problem with Tektronix measurement software.

I am working as researcher at an accelerator lab.

For measuring, we use Tektronix oscilloscopes TDS3xxx, which have a
LAN connection.
This should be quite convenient, because data storage via LAN is
faster than storage to
internal floppy disc.

So far, I had good experience with the data (stored as comma separated
values, CSV)
stored to floppy disc. The only disadvantage is, storing of 10000
points with 4 channels
takes more than 1 minute.

Now with the Tektronix Open Choice Desktop software, it seemed, speed
can be improved.
Saving is much faster.
But looking at the data on the hard-disk of the PC where the data had
been transfered,
I was deeply disappointed.
Compared to the data on floppy, the data transfered by LAN and Open
Choice
Desktop software was at lower resolution, e.g. substantial information
was lost.

a) the time scale (x-axis) is to coarse
The accelerator needs 20ms for the acceleration.
If one wants to see what happens in the middle of the acceleration
ramp,
one uses a trigger delay of 10ms.
And the Scope will then sample 10000 points at 5Gsamples/sec.
Each point is then on a 0.2 ns timescale.
So the points are located between (10ms-1000ns) and (10ms+1000ns).
The data stored to floppy follows this correctly.
For each data point the time stamp advances by 0.2ns.

But not so with the Open Choice Desktop software.
From (10ms-1000ns) to (10ms), there are 4 points, where the time stamp
is the same, and then one point which is a 1ns step.
From (10ms) to (10ms+1000ns) it is even worse.
There are 499 points, where the time stamp does not change, and then
a big jump in one point by 100ns.
Of course, such a timescale is useless for serious analysis.
One has to replace it with a corrected one.

b) The channel data (y-axis) has reduced precision
Looking at the quantization noise from the samples, the noise
of the data stored by Open Choice Desktop is higher than compared
to the data of the same acquisition stored to floppy.
Especially, when using the MATH function, several bit of information
can be lost.


Last year, I contacted the person in charge of Tektronix in Germany
and explained him about the problems.
But he was not interested. He said I should contact the local
representative in Japan.
Talking about customer care.
The representative in Germany did not understand that nowadays neither
the scopes nor the software are localized any more.
This means it is the same scope, if in Japan or Germany or US,
just the power cord is replaced and the printed manual is in different
language. And the oscilloscope language can be set by the user.

Same for the Open choice desktop software.
There is just one program and the language setting can be changed at
runtime.
In other words, the representative did not realize that my measurement
problem was not related to being in Japan.
It can and will happen anywhere where Tektronix sells its products
worldwide.

And this is why I am posting this to this newsgroup.

Well, he forwarded my request to the Japanese office of Tektronix, and
they
contacted me and I told them my problem with examples of the data
again.

Also my Japanese colleagues talked to the representative there.
And it happened - almost nothing for more than half a year!

As I noticed yesterday, there was an updated version of the open
Choice
Desktop software (version 1.5 -> 1.6) available,
so I made a test-setup with a scope, stopped it, and saved the data
1) to floppy
2) to hard-disk with the older version 1.5 of Open Choice desktop
3) to hard-disk after updating to version 1.6 of Open Choice desktop

Still, data on floppy were good as expected.
But the loss of resolution was same between Version 1.5 and 1.6.
I was disappointed, because the announcement of Tektronix
was that many issues in the saving of CSV files had been corrected.

Well, the problems are still there, and the numerical values between
version 1.5 and 1.6 of the data are identical.

Now I am wondering - am I the only person who found out about the
problem reduced data quality?
Or are there (thousands?) of other Tektronix customers who
still believe that their data saved by LAN and the Open Choice desktop
software are correct, when actually the data is crippled?

I am waiting for comments. If interested, I can send some files which
clearly demonstrate the issue.

Sincerely,

Alexander Schnase

 

[Alexander Schnase]

Dear Robert,

thank you for your answers.

However, it seems that at best, the name "OPEN CHOICE" is rather misleading.

Well, I completely agree with that

Here are possible work-arounds:
1) Re-direct the data from floppy to RAM (drive) where the speed will be
vastly improved.

Sorry, the Floppy is inside the scope and the scope does not have a
RAM disk.
Writing the data to floppy inside the scope works, but is slow.

On the other hand, when using OpenChoice Desktop, of course the
data is written to hard disk.
It is fast, but the data is crippled in precision.

One "workaround" is to use Epics, set-up one PC as a so-called IOC,
and then access the scope using Epics routines.
This seems to work, but it is a lot of overhead.

Another workaround is to use the Web-browser interface and save
the data from there.
But it is not much fun, as one has to save each of 4 channels
independently.

2) Re-direct data from floppy to I/O port (serial, parallel, USB,
whatever).

If the scope had USB, problem would be solved.

Thanks for comments,

Alexander