WinGPIB - Windows GPIB App (multi-purpose)

[IanJ]

Hi all,

V1.38 available in post #1.

- Interlocks added so that Temp/Hum cannot be enabled until the Temp?Hum sensor is started.
- Interlock added so that you cannot RESET teh CSV file without first checking the box next to it (saves inadvertent resets!)

Any issues with the CSV loading into the Plabck Chart please let me know......there's lots of interlocks in teh code now so I'd be interested to hear any issues.

Ian.

[Messtechniker]

V 1.38:
Logging and curves are OK with Windows EN enabled.
However, playing back the recorded log.csv results in:

Conversion from sting "" to type 'Double' is not valid
 

Both sample .csv files play back correctly

Here an excerpt from the log file in question:

1,34465A,2018-12-19_07:02:15,10.0052315
2,34465A,2018-12-19_07:02:20,10.0051795
3,34465A,2018-12-19_07:02:25,10.0051923
4,34465A,2018-12-19_07:02:30,10.0052188
5,34465A,2018-12-19_07:02:35,10.0050439
6,34465A,2018-12-19_07:02:40,10.004845
7,34465A,2018-12-19_07:02:45,10.0049304
8,34465A,2018-12-19_07:02:50,10.0049489
9,34465A,2018-12-19_07:02:55,10.0049471

All lines are like that. No blank line.

[IanJ]

V 1.38:
Logging and curves are OK with Windows EN enabled.
However, playing back the recorded log.csv results in:

Conversion from sting "" to type 'Double' is not valid
 

Both sample .csv files play back correctly

Here an excerpt from the log file in question:

1,34465A,2018-12-19_07:02:15,10.0052315
2,34465A,2018-12-19_07:02:20,10.0051795
3,34465A,2018-12-19_07:02:25,10.0051923
4,34465A,2018-12-19_07:02:30,10.0052188
5,34465A,2018-12-19_07:02:35,10.0050439
6,34465A,2018-12-19_07:02:40,10.004845
7,34465A,2018-12-19_07:02:45,10.0049304
8,34465A,2018-12-19_07:02:50,10.0049489
9,34465A,2018-12-19_07:02:55,10.0049471

All lines are like that. No blank line.

Hi,

The CSV looks fine (no temp/hum sensor connected or active).
When you get the error can you copy n paste the detail from the error box and send to me.
Near the top of the error report it will mention (more than once possibly) which lines in the .vb code the error occurred…...that's the detail I need.

Ian.

[IanJ]

Hi all,

V1.39 available in post #1.

In order to fix the Playback CSV load error that some of you are getting I have made a modification to the data that is written to the CSV.
So now instead of writing nothing to the CSV if there is no Temp/Hum sensor connected I am writing 0.0 for both......I am hoping this is the source of the error.
It looks like that some non-English Windows configurations have a problem loading the CSV which results in a String conversion to Double error.

Please note that this fix will only work for new CSV files created with this version 1.39. If you try to load older CSV files they will likely still error.

Feedback appreciated.

Ian.

[Messtechniker]

Did a quick check on V 1.39. Seems to work well!
Playback loads demo files and files created with V 1.39.
Long logging in progress now.
More testing tomorrow to see if it works with
language set to German. Is EN at the moment .

[blackdog]

Hi,

I had onother crash of the program, afther 19 hour and almost 1.3MB of collected data.

I wil place nouw a lot of pictures of screenshot i made, maybe it helps...

Crashd but stil in memory.



Screen space wil nog be updated before i close the program



Taskmanager pictures in three parts, part-1



Taskmanager pictures in three parts, part-2



Taskmanager pictures in three parts, part-4


The computer has no high processor load if the program is crashd en the computer is still running fine.

Latest version problems
Crash...
I started version 1.39 of the program en created both devices, the 3458A and the 34461A.
And put this in for configuration afther power cycling of the 34461A:
*RST
FUNCTION 'VOLT:DC'
VOLT:DC:RANGE 10
VOLT:DC:NPLC 100



An a second try and no 34461A reboot, and it works...

Kind regards,
Bram

[IanJ]

Hi,
I had onother crash of the program, afther 19 hour and almost 1.3MB of collected data.
I wil place nouw a lot of pictures of screenshot i made, maybe it helps...
Crashd but stil in memory.
Screen space wil nog be updated before i close the program
Taskmanager pictures in three parts, part-1
Taskmanager pictures in three parts, part-2
Taskmanager pictures in three parts, part-4
The computer has no high processor load if the program is crashd en the computer is still running fine.
Latest version problems
Crash...
I started version 1.39 of the program en created both devices, the 3458A and the 34461A.
And put this in for configuration afther power cycling of the 34461A:
*RST
FUNCTION 'VOLT:DC'
VOLT:DC:RANGE 10
VOLT:DC:NPLC 100
An a second try and no 34461A reboot, and it works...
Kind regards,
Bram

The pop-up displays an error code = BFFF00A6, and looking this up the description is:-
"The connection for the given session has been lost - VI_ERROR_CONN_LOST"
http://literature.cdn.keysight.com/litweb/pdf/E2090-90040.pdf

The next time in crashes can you open the Keysight Connection Expert and see if the 34461A is still connected there.......in fact leave the connection expert running on-screen.
You could also try sending commands to it via the Connection Expert to see if it reacts when it crashes.

You can also use th Keysight IO Monitor and capture all messages.

I have set up V1.39 with my own 3458A & 34461A exactly like yours so will let it run. I should manage 20hrs straight.

If it's a Keysight issue then maybe we need to play with the PRE RUN ans AT RUN commands to help it.

Other ideas:
1. 34461A firmware version?....I am running A.02.14-02.40-02.14-00.49-01-01
2. Is your 34461A on the same sub-net as your PC......could your Lan be dropping packets, or something interfering with the TCP/IP connection?
3. Windows OS?
4. I am using Keysight Connection Expert 2018
5. Is your 34461A displaying an "error" at top right corner of the LCD when it happens, or any other message?

I have never had a 34461A crash problem......so something is different between your and my setup. The only thing I notice is that sometimes READ? can take a few seconds to get a reply, but it usually settles down.

Ian.

[pwlps]

Hi Ian and Bram,

ah yes, "werkt niet meer", it is obvious: it doesn't work 

VI_ERROR_CONN_LOST  means a lost TCP connection but it shouldn't blow off the whole thing.
Now that the error message window is enabled it looks like the code handled the problem at least once without crashing, does it mean the crash went from inside a Visa function when the command was sent next time?  Even with all those taskmanager windows we can't know it without a debugger trace.

Another question (for Ian): how was the dev1/dev2 Visa interlock option configured in this version? If it was disabled it might be interesting to repeat  the test with interlock on.   

Pawel 

[blackdog]

Pawel,

This is not funny! => ah yes, "werkt niet meer", it is obvious: it doesn't work 

Got you 

Maybe comming weekend i have some time left to play again with the software.

Ian
A have the latest firmware in my two 34461A DMM's from the beginning of this month.
The OS of my measuring computer is W7 Prof, up to date.
The IO package is the latest of KeySight.
The IP address of the meters are in the same range as the measuring computer and my main workstation conected to my workbench switch and short good quality patch cables.
I use fixt IP adresses for my measuring gear.
Last crash no error messages on the display of the 34461A that was in use.

Kind regards,
Bram

[IanJ]

Hi Ian and Bram,

ah yes, "werkt niet meer", it is obvious: it doesn't work 

VI_ERROR_CONN_LOST  means a lost TCP connection but it shouldn't blow off the whole thing.
Now that the error message window is enabled it looks like the code handled the problem at least once without crashing, does it mean the crash went from inside a Visa function when the command was sent next time?  Even with all those taskmanager windows we can't know it without a debugger trace.

Another question (for Ian): how was the dev1/dev2 Visa interlock option configured in this version? If it was disabled it might be interesting to repeat  the test with interlock on.   

Pawel

Pawel,

Interlock in this version is per your last email passing me the .vb file. I inserted it 'as-is' and took the checkbox off the form.

Ian.

[pwlps]

Ian,

Interlock in this version is per your last email passing me the .vb file. I inserted it 'as-is' and took the checkbox off the form.

Then it is not activated by default.  To activate it (in a simple way - hardwired in the code) replace, in the main form (in CreateDevice) :
dev = New VisaDevice(name, address)
by
dev = New VisaDevice(name, address, true)

(see explanations in my email on Dec 12th)

As I said somewhere in my mails  the performance loss due to interlock will not be perceptible as long as you only transfer short strings (this would be different with e.g. long scope traces).  Therefore in this app you can probably leave it like this for the moment. (Later you can add a checkbox (in the main form) which would select the desired behavior if you want to be perfectionist).

Pawel

[IanJ]

Ian,

Interlock in this version is per your last email passing me the .vb file. I inserted it 'as-is' and took the checkbox off the form.

Then it is not activated by default.  To activate it (in a simple way - hardwired in the code) replace, in the main form (in CreateDevice) :
dev = New VisaDevice(name, address)
by
dev = New VisaDevice(name, address, true)

(see explanations in my email on Dec 12th)

As I said somewhere in my mails  the performance loss due to interlock will not be perceptible as long as you only transfer short strings (this would be different with e.g. long scope traces).  Therefore in this app you can probably leave it like this for the moment. (Later you can add a checkbox (in the main form) which would select the desired behavior if you want to be perfectionist).

Pawel

Ok, I agree......the voltnuts can easily live with it so I will amend tonight.

Ian

[IanJ]

Hi all,

V1.40 available in post #1.

This version activates an interlock in order to try and stop an issue where some user(s) have experienced the app crashing when a connected device appears to go offline, i.e a VI_ERROR_CONN_LOST error reported. This also only seems to happen when dual devices are connected and running.

Feedback appreciated.

Ian.

[blackdog]

Good morning Ian, 

Since last night your version 1.40 runs here with the 3458A and the 34461A, which is currently about 10 hours log time.
I used again the same setting.

Yesterday I also played with the new web interface of the 34461A.
I think there are some big bug's in it, I will test it later today with my second 34461A DMM.

All you have to do is go to the IP address of the 34461A and then choose the Interactive IO,
Then the measuring instrument gives even if you only type 1 character, directly an error!
So giving an enter is not necessary, just enter a character and ERROR...

If you choose a command from the "Execute" drop down menu, the instrument performs this command without error.

I used three different browsers from my main workstation to see if the browser influenced this.
These were the latest versions of Firefox, Opera and Internet explorer (No Edge and also no Google, which are not appreciated here)
I also tested a virtual, latest version of MINT Linux with Firefox, same result.

Later today I will test this with two computers on the 34461A I currently have available for this purpose.
Why am I telling this here, because I thought it would be useful to test the commados I would like to use via GPIB.
Before i put them in Ians software.
I will keep my comment short about this so as not to pollute this topic.

Kind Regards,
Bram

[IanJ]

Hi all,

V1.41 available in post #1.

Relatively minor updates:-
- Data Log & Event Log window data formatted so that it is more aligned/readable (small text, but it's readable).
- Further checks that the temp/hum sensor data is valid before writing to Data Log and CSV. This is necessary since if the sensor/data is lost it was creating CSV entries that the Playback chart couldn't handle.

Ian.

[IanJ]

Hi all,

V1.42 available in post #1.

- Added a TEMPERATURE MAX DIFF RECORDED display and RESET button below the chart on the main display.
  This complements the Device 1&2 Max Diff Recorded displays already there.
  I also increased the responsivness of the RESET button there.

Ian.

[IanJ]

Hi all,

V1.43 available in post #1.

- Bug fix - Device 2 AT STOP commands were not being sent.
- I have changed the way the AT STOP commands are sent, i.e. changed to blind commands that don't expect a reply.
- Bug Fix - If you hit STOP on any device and then hit RUN again then the software would not return to a working RUN mode again.
  The idea is that you can hit STOP, then adjust some PRE-RUN or AT RUN parameters then continue on again as it will re-send the PRE-RUN commands.

Ian.

[IanJ]

Hi all,

V1.44 available in post #1.

- Log scale functionality added to the PlayBack Chart

Please note that the Log scale functionality won't work with negative data......I have taken some precautions, but if there is negative data anywhere in the CSV then it may cause the app to crash.
A checkbox on the PlayBack Chart enables Log scale.

I have left 1.43 available for download as there's quite a bit of mods done to make 1.44......a bit of a tidy up, and adding the Log scale functionality.

Ian.

[Grandchuck]

Returned from a holiday trip and found nice upgrades to the logging app!  Thanks much Ian and a happy new year to you and yours.

PPM scale is not showing on the saved screen so I also attached a screen capture.

Happy new year to all here.

[IanJ]

Returned from a holiday trip and found nice upgrades to the logging app!  Thanks much Ian and a happy new year to you and yours.

PPM scale is not showing on the saved screen so I also attached a screen capture.

Happy new year to all here.

Yep, it's a quirk/bug in the way vb saves a bitmap......and not much I can do about it except replace 1 line of code with a whole routine......which I will get around to eventually. You'll notice it's not just the PPM scale, other random stuff is missing.

Ian.

[deepskyridge]

Ian, I have your windows app running with my HP 3457A and seems to be working well. Will run it for a couple of days next to check reliability.

Is there a way to make the instrument id and gipb address retain between restarts.

Never Mind, I figured it out.

Thanks for the nifty software.

Gary

[IanJ]

Ian, I have your windows app running with my HP 3457A and seems to be working well. Will run it for a couple of days next to check reliability.

Is there a way to make the instrument id and gipb address retain between restarts.

Never Mind, I figured it out.

Thanks for the nifty software.

Gary

I hope the software is of good use to you...…...and thanks for the donation, very much appreciated.

Ian.

[deepskyridge]

Two Things that would be nice to add.

1. Ability to convert temp from Celsius to Fahrenheit.

2. Add support for the scanning card on the 3457A, I have the 44491A card in mine.

Thanks
Gary

[IanJ]

Two Things that would be nice to add.

1. Ability to convert temp from Celsius to Fahrenheit.

2. Add support for the scanning card on the 3457A, I have the 44491A card in mine.

Thanks
Gary

The reason I chose DegC, apart from that what we use mostly here, is that since it shares the same scale as Humidity the numbers are closer together when set to DegC. Thats said, probably a good idea to have a selection option.

I don't have a 44491A card in my 3457A.....but if I pick one up I will look at integrating it.

Ian.

[Svgeesus]

On a Keysight 34465A, I'm looking for some way to automate testing and logging. I would preferably like to be able to control a PDVS2 at the same time. Details below of what exactly led me to the need to automate, for those interested, but basically I want to generate control voltages and log voltage, frequency, temperature and time for later analysis (mean, median, regression) and charting.

Compared to my old handheld meter (UT-71C) besides the greatly improved accuracy and precision I certainly appreciate being able to get a min, max, mean, standard deviation plus a visual time series/trend chart and histogram. But the limitations of that approach are becoming irksome - the output is raster screenshots saved, somewhat laboriously, to a USB stick. I want to be able to capture the individual readings, log them with timestamp, and then be able to calculate mean, sd, also median and so on. I want more flexibility in graphing and analysis (like charts with error bars).

This moved up in urgency when I was measuring some analog audio oscillators (Eurorack modular oscillators, for electronic music synthesis). These are controlled by an analog input voltage, where an increase in one volt results in a doubling of output frequency (one octave pitch change). I was using a PDVS2 to create the control voltages, covering 20V in 100mV steps (10V positive and 10V negative). This voltage went to the oscillator, and the voltage at the oscillator control voltage input went to the front panel of the 34465A (100PLC, HiZ). The triangle wave produced by the oscillator went to the rear panel of the 34465A, where the frequency was measured. So each measurement required a series of manual steps:

• Set the new voltage on the PDVS2, with the rotary encoder
• Change to front panel inputs, reset the readings, monitor the voltage for 10 seconds to get a mean and sd (the PDVS2 is very stable, so this was quick)
• Write the voltage values in a CSV file
• Note the temperature inside the synthesizer case
• Write the temperature value in a CSV file
• Switch to rear panel inputs, reset the frequency reading, monitor the frequency for 1000 readings to get the mean and sd
• Write the frequency values in a CSV file

This had the disadvantage that the voltage measurements were all taken together and then the frequency measurements all together, rather than taking interleaved measurements over the recording of each data point. The PDVS2 is very stable, the temperature was fairly stable but climbing gradually. Because the 20V range was done in two stages (with the leads being reversed to get the negative voltages)  I think this explains the fact that the two halves of the voltage vs. log frequency plot did not fully line up. It also had the disadvantage of switching manually between front and rear via the pushbutton, a lot of times. This would be better done with software control.

My next VB app will be to automate calibration of my PDVS2's without human intervention......i.e.: VB reads output voltage of PDVS2 via GPIB, VB sends instruction to PDVS2 to adjust up/down, VB instructs PDVS2 to save cal (zero & span DAC adjustment)............this will save me hours and hours of calibrating batches pf PDVS2's!!!

Because Ian wrote a Windows program to control the PDVS2, I'm first looking at his program to control my meter (and hopefully, to also log temperatures and to control the PDVS2 at the same time). Only if that doesn't work out will I bite the bullet and do the whole Raspberry Pi Linux logging station thing, which looks a lot more complicated to set up.

Questions:

• Does this require purchasing the GPIB option for my meter, or can I use the USB or ethernet connections?
• Would this be compatible with my meter or is it 3458a only (I saw mention of other meters but not my specific one)
• Could one program run both devices or would I need to run two programs with a timed series and hpe they don't drift apart?
• Also, what are my options for expanding beyond two channels (front and rear) on the meter/ I see other Keysight meters that take multi-channel scanner cards, but these don't seem to turn op on searches on this forum. And there doesn't seem to be an add-on scanner card for this meter. I can see this being my next requirement. Or do people just buy multiple similar meters?