06-02-2009 04:39 PM
I'm using Signal Express to record Load vs Displacement data and export it to a format our engineers can work with (in this case ASCII is okay). It would seem that by selecting the Next Available File Name from the drop down arrow it would do just that. For instance a typical save path for me would look like C:\....Desktop\Project Number and Description\Run_1.txt within that Run1.txt file would be all the data points for that run. When I hit record again Signal Express would (SHOULD) create a Run_2 since it's the Next Available File Name.
But instead what it does is creates a single txt file for every single sample point being read. Needless to say, If I'm recording 6 second of data at 1khz I end up with thousands of txt files!
The first thing that comes to mind is, why would anyone want this?
Second is how can I record multiple individal runs for the same project and have the file name increment?
06-03-2009 10:52 AM
OKors,
This can be done but in my opinon it is confusing. I understand your sentiments on the Save to ASCII file but I think the option I will discuss later covers the hole that it leaves. I have however filed a product suggestion for you and hopefully the developers will find merit in it.
The way to do what you are looking for is to go to Tools>>Options>>Logging. You want to set the Default storage directory to the place you would like the ascii files to go and enable the Automatically Export Log to ASCII option. Then rather than using the Run button you should use the Record option.
06-03-2009 11:33 AM - edited 06-03-2009 11:41 AM
Thanks for the reply Jason,
I've done what you've suggested before but the major problem with that is when I hit record it prompts for a new log name (which is good, this is what I want) but when it exports the log and other associated files (which I don't need) it creates a seperate folder. So if I have Run_1, Run_2, Run_3, etc. SE creates that many individual folders. The actual data file resides INSIDE that folder and is not named Run_1, it's whatever the channel description is. This means our engineers (read; "me") has to rename each of the data files to its appropriate 'run' name then copy it into a common folder. During the duration of a cycle fatigue test it is not uncommon to take up to 100 data plots. This is 100 files I have to rename, copy, and paste into a common folder.
Unless I'm missing a step this seems like a very convoluted way to address a simple task.
I do appreciate you filing a product suggestion for me. What are the chances the developers will implement this change?
I believe the way Signal Express handles it's output files will be a major delay in us transitioning into this software.
06-04-2009 02:38 PM
OKors,
You mentioned taking 100 data points being in 100 files. This is absolutely only the case if you store each data point in a file by itself. What sampling rate are you using? What is the samples to read in your project? Samples to read determines how many points are being read from the buffere and thus stored in each file. If you are truely just taking 100 points then simply set this higher and you will get it all in 1 file with the original method. If you meant 100 tests then as long as these tests are reasonably small you could also get these in 1 file.
06-04-2009 03:23 PM
I'm recording a load vs displacement spring plot. This is a cyclic test that is running at .5Hz. I only need about 5 seconds worth of data every 1000 cycles or so (slow speed test but I need to capture the extreme min/max load values). Currently I've set it up to record Contiuously with Samples to Read being 100 at a rate of 1k (these are the default values and seem to work for what I'm doing).
The actual amount of data I record varies; all I'm really doing is hitting Record, watching the chart to make sure I've plotted two or three complete cycles, then Stop Record. It could be anywhere from 3-5 seconds of data.
I'm confused with how signal express handles sample rates so with the settings mentioned above I do not know how many actual data points 3-5 seconds gives me. But basically when I open the directory to my saved data file I have TONS of .txt files. When I open one of the files it show's the values for one sample point.
06-05-2009 02:46 PM
OKors,
I have just set up a test using a simple 1 channel continuous acquisition. Sample rate (like you mentioned) of 1k and samples to read of 100 for 5 seconds. I recieve 50 files like the one listed below at 100 samples each (1of50).
I have another test which runs for 5 seconds at 1k samples to read and I get the text file below (1of5) 1k samples each.
Finally I have a text file that that is at 5k samples to read and holds 5k samples runing for 5 seconds at 1k rate. It is 1of1 holding 5k samples.
Are you not seeing this behavior? I am back to using the run button and save to ascii (record will alway create the folder). If this is not working maybe you can post your project and I can try it on my machine?
06-05-2009 05:31 PM
Interesting...
Your .txt files are more in line with what I would expect to see. This is not the behavior I'm experiencing w/ the Save to ASCII function.
I've attached some images of my test settings
06-05-2009 05:52 PM - edited 06-05-2009 05:57 PM
...and more attachments of the ASCII save path, before and after acquiring 4 seconds of data along with one of the files from that folder.
Again, this is 4 seconds of two-channel spring plot data at 100 Samples to Read @ 1k Rate (Start Run.... wait 4 seconds or two full test sample cycles... Stop Run).
~EDIT~
The .txt file would not attach (I think it's too small). Here's what it looks like if you were to open it:
Load vs Displ - Displ (inches) Load vs Displ - Load (lbs)
3.736323 273.751906
Also, for some reason it won't let me attach my project file. It's a .seproj extension but the forum thinks it's 1k in size and "empty"
06-08-2009 07:36 PM
Hi Okors,
I have looked at your code and you are receiving a separate file each containing a different point because you have Next Available File Name selected. I would like to elaborate on what is occurring thus far. If you select Append to File, then you will end up with one file of averaged values.
Currently, your code acquires x number of voltage samples. After acquisition is complete, these x samples are put into the Amplitude and Levels function which takes an average of your acquired samples and returns a single value. This value is saved as ASCII and saved in a text file and the process iterates. This is why you are only receiving one value in each file.
Reading Jason's test does not seem to include the Amplitude and Levels step and he is writing the voltage samples to the ASCII file. What data do you want to save?
06-09-2009 10:42 AM
h_baker,
The data I want to save are the Load/Displacement channels (scaled and with channel descriptors in the header info.).
I've added the Amplitudes and Levels step because that's the only data view that allows me to view the graph as a Strip Chart. 99% of the data we take is viewed as time-history and Xvs.Y in the case of spring plots. I have not found a way to view my scaled voltages as a Strip, Scope, or Sweep chart other than using the Amplitued and Levels.
This brings up another question based on what you're telling me; you mentioned the Amplitude and Levels step "averages" the signals? I don't want averages; I need the absolute peak signals in my data file. I've noticed this also as I watch the graph in real time. Unless I select a really high sample rate I cannot seem to capture the absolute end values.
But... before this thread becomes splintered I'd like to bring us back to my original question;