wt1600 gpib continous measurement

[ping1]

Hello!

I'm trying to measure some items with LabVIEW (8.20) at a Yokogawa WT1600 by GPIB.

There is a possibilty to read a preset. So I defined the values of te preset and with :Num:Norm:val? I'm able to read the data.

My problem is there is no time-base or a specific interval of the read data.

When I'm using a loop two ask every 200ms the values, there exist an unspecific response time.  (between 5 and 40ms)

How is it possible to read a couple of values in a defined time interval?

Is there a possibility using a queue or something like that?

Thanks for your help!

Lukas

[amcolosk]

Hello,

Would something like I attached work for you?  It keeps track of the time, and when the difference between the last read and the current time is greater then or equal to 200 ms, it reads. 

[ping1]

Thanks for your proposal.

That's what I had already with my loop.

But I don't think it will prevent the problems.

The responsetime of the reading is still different and it would be possible, that I skip/loose one value.

My WT1600s is able to update the numerical values every 200ms (very often 😉 and its impossible to read every 200ms-period at the same postition (e.g. in the middle).

That's because of Windwos (XP) ist not a realtime-OS.

So I can't find a possibility to measure live or online to cover precise values.

Perhaps it's better to store the values in the internal memory of wt1600s and read them afterwards value by value via GPIB.

If there is another possibility I can't see, please answer.

Lukas

[amcolosk]

The first question that I have for you is why do you need the values every 200ms?  I understand if it is a message that must be sent precisely that often, but for measureing values, I don't think a 5 - 40 ms variance will make that big of a difference.  When you loose a value, how can you tell, do you get a 0, or an outrageously high number?  if the WT1600s has enough memory to store the info that you need, then it is probably best that you store it on it.  If you don't have enough memory, maybe you can do a quick read of multiple values periodically?

[ping1]

I need the values as often as possible. 200ms is the lower limit of the WT1600s.

The problem of the 5-40ms variance is, that I'm measureing from another device and this should be nearly at the same time.

Another question is the quality of the "Tick Count" in LabVIEW. Does it makes sense to use Tick Count for the loop-timing? I read there is a problem using this function for an equation because of its limit (2^32).

How pecise are these 200ms from a Tick Count when I'm using WindowsXP? Will it differ a lot?

Thank You.

Lukas

[amcolosk]

Good Question,

You made me do some research.   Yes the limit would possibly make a difference if you ran it that long, roughly 13.6 years....  So hopefully not.  Quality is a concern, since you want to be so precise.  Here's what I suggest you try, the wait for next multiple.  I would try to explain it, but I found a post by tnewton who does a lot better job than I could.  http://forums.ni.com/ni/board/message?board.id=170&thread.id=8641&view=by_date_ascending&page=2  In the forum, they talk about many of the same problems you have, and they came down to using this setup. (read the first page for the discussion about the tick count inconsistencies) Maybe this will work better for you?