Program that can log from many multimeters.

[skander36]

Hi,
Thank you for your help. I have tested modified program.
Now it show me a dialog for saving the file, but I cant choose another folder, can't browse, but when try to create a director I can browse. At this moment the log record: "Failed to set director: C:\Users\Robert\Documents\TestController".
I am running the program from "c:\TestController"
In a second attempt to save the same log, again does nothing and so on. Only on first attempt it show a dialog.
I attached a full output with a short log included.

[HKJ]

Hi,
Thank you for your help. I have tested modified program.
Now it show me a dialog for saving the file, but I cant choose another folder, can't browse, but when try to create a director I can browse. At this moment the log record: "Failed to set director: C:\Users\Robert\Documents\TestController".
I am running the program from "c:\TestController"
In a second attempt to save the same log, again does nothing and so on. Only on first attempt it show a dialog.
I attached a full output with a short log included.

Try create a "TestController" directory in you "Documents" directory, it is assumed to be created automatic and I wonder why it is not.

[skander36]

It is there. Is where it save the log at first attempt with new program.

[HKJ]

It is there. Is where it save the log at first attempt with new program.

But in the screen shot it says "Documents", I suspect there is something wrong with your directories. I can be names or security.

In the testController.bat file you can add a dataDir=xxxxx option and switch the datadir to something else (You have to use the testController.bat file to start TC for this to work). For debug you will have to do the same modification in the testControllerDebug.bat file.

[skander36]

I don't know what can be.
This is the only program that make this problem.
It is a win 11 and maybe they set some security setting, I don't know.
It show Documents and can't browse as I said. It is a strange behaviour on that save dialog. At least I can save with that version ... Then if I will need to log again I need to close the program a nd start again.
Try to set datadir path but no change. What is the syntax? Just a separate line or as option to main line?

[HKJ]

I don't know what can be.
This is the only program that make this problem.
It is a win 11 and maybe they set some security setting, I don't know.

I use Win10 and have no idea why it fails on your computer.
Do the path have any special characters in it (Maybe from your name).

It show Documents and can't browse as I said. It is a strange behaviour on that save dialog. At least I can save with that version ... Then if I will need to log again I need to close the program a nd start again.
Try to set datadir path but no change. What is the syntax? Just a separate line or as option to main line?

start javaw -jar TestController.jar dataDir=C:\TCData %1 %2 %3 %4 %5 %6 %7 %8 %9

[skander36]

Hi,
No problem with my name see attached.
Tested with set path. While is still show Documents without posibility to choose another dir, it save to the location from setpath variable (C:\TestController).
Anyway if you said that it should work and nobody report this problem, that mean the problem is on my environment. I will try to clean Java instalation and check to see if this can help.
Thank you for your help!

[jmurray]

Hi HKJ,

This might be a stupid question, but what would be the recommended way to implement a delay within a #scpiCmd #pgm# routine?
I've tried runScript("#delay 1") but it does not appear to work.
TestController also doesn't like it being called more than once, for example as part of a "for" loop, as it reports: "A background job is already running, this request is ignored".

Preferably a delay in milliseconds: I know the standard #delay can do fractions of a second, but whatever method I need to employ in this case would need to work in milliseconds.
Should I be using an expression to compare the results from the seconds() function?
Or is there a simpler function I could call?

[HKJ]

Hi HKJ,

This might be a stupid question, but what would be the recommended way to implement a delay within a #scpiCmd #pgm# routine?
I've tried runScript("#delay 1") but it does not appear to work.
TestController also doesn't like it being called more than once, for example as part of a "for" loop, as it reports: "A background job is already running, this request is ignored".

Preferably a delay in milliseconds: I know the standard #delay can do fractions of a second, but whatever method I need to employ in this case would need to work in milliseconds.
Should I be using an expression to compare the results from the seconds() function?
Or is there a simpler function I could call?

You can directly include delay in SCPI commands by using [123] (123ms) or [*OPC] that will use a special SCPI function to wait until the device is done processing commands (This do not work on all SCPI devices).
You can find it here in the documentation: https://lygte-info.dk/project/TestControllerConfigDevice%20UK.html#Timing_issues
With some drivers you can also add a fixed delay between commands.
In programming you can use sleep(123) (123ms) and #delay 0.123 in scripting.
In all cases caching can throw you delay off a little bit.

Also be aware that TC is multithreaded, it will generally not affect delays, but a delay may open up for another thread to use the device during the delay (this will not happen with the [...] delays). With standard use of TC there will usually only be two threads accessing a device: The one fetching measurements and the one doing configuration (There are exceptions to this), but with scripting this can easily be increased.

"A background job is already running, this request is ignored"
Only one regular script can run at a time, use runScriptAsync() to start a independent background job. With this you can also crash TC, because there is no limit on background jobs and even the fastest computer will have trouble running a few million background jobs.

[jmurray]

Hi HKJ,

This might be a stupid question, but what would be the recommended way to implement a delay within a #scpiCmd #pgm# routine?
I've tried runScript("#delay 1") but it does not appear to work.
TestController also doesn't like it being called more than once, for example as part of a "for" loop, as it reports: "A background job is already running, this request is ignored".

Preferably a delay in milliseconds: I know the standard #delay can do fractions of a second, but whatever method I need to employ in this case would need to work in milliseconds.
Should I be using an expression to compare the results from the seconds() function?
Or is there a simpler function I could call?

You can directly include delay in SCPI commands by using [123] (123ms) or [*OPC] that will use a special SCPI function to wait until the device is done processing commands (This do not work on all SCPI devices).
You can find it here in the documentation: https://lygte-info.dk/project/TestControllerConfigDevice%20UK.html#Timing_issues
With some drivers you can also add a fixed delay between commands.
In programming you can use sleep(123) (123ms) and #delay 0.123 in scripting.
In all cases caching can throw you delay off a little bit.

Also be aware that TC is multithreaded, it will generally not affect delays, but a delay may open up for another thread to use the device during the delay (this will not happen with the [...] delays). With standard use of TC there will usually only be two threads accessing a device: The one fetching measurements and the one doing configuration (There are exceptions to this), but with scripting this can easily be increased.

"A background job is already running, this request is ignored"
Only one regular script can run at a time, use runScriptAsync() to start a independent background job. With this you can also crash TC, because there is no limit on background jobs and even the fastest computer will have trouble running a few million background jobs.

sleep(123) is what I was needing! Thanks!

[HKJ]

sleep(123) is what I was needing! Thanks!

I expected that, but I like to explain a bit more (when I have time). I hope a few people can learn a bit more about TC that way.

[jmurray]

sleep(123) is what I was needing! Thanks!

I expected that, but I like to explain a bit more (when I have time). I hope a few people can learn a bit more about TC that way.

I realised that was the case. More information is always better!

I recently added a Mecmesin BFG1000 Basic Force Gauge to my inventory - so here's a new definition file you may include with TC.
I've done my best to accommodate every function and compensate any lacking features within the definition file.
I've also pre-filled everything to accommodate all other models in the BFG range - it will require some minor corrections for the lower capacity models but otherwise everything is there.

EDIT: There's always another bug once you hit submit! Corrected some readformat issues with the streaming mode implementation.
I've updated the attachment.

[HKJ]

I recently added a Mecmesin BFG1000 Basic Force Gauge to my inventory - so here's a new definition file you may include with TC.
I've done my best to accommodate every function and compensate any lacking features within the definition file.
I've also pre-filled everything to accommodate all other models in the BFG range - it will require some minor corrections for the lower capacity models but otherwise everything is there.

EDIT: There's always another bug once you hit submit! Corrected some readformat issues with the streaming mode implementation.
I've updated the attachment.

Thanks, it will be included in next release. If you post more updates, please make another post.

[Gertjan]

Hi HKJ,

I am working on a definition for a Philips / Fluke PM2811 programmable power supply. It is SCPI compliant, and I am using the R&S HMC8043 definition for guidance.
This power supply was sold in several versions: 30V/10A and 60V/5A, front panel and back panel connectors etc. All these versions are having a bit different type number, and thus give a different answer to *IDN?.

Philips PM2811 30V / 10A:  PHILIPS,PM2811/01 Philips PM2811 60V / 5A :   PHILIPS,PM2811/11

1. Obviously the 30V/10A and 60V/5A versions need a bit different definition.
Would this work:

Code: [Select]

; Meta definition for PM2811 30V/10A version
#metadef
#idString PHILIPS,PM2811/01
....

; Meta definition for PM2811 60V/5A version
#metadef
#idString PHILIPS,PM2811/11
.....

Would TC automatically pick the right #metadef, depending on the *IDN? Answer?
I suppose the easy way would be to create just two model names: “Philips PM2811 30V/10” and “Philips PM2811 60V/5A”. But that seems to be a less elegant solution.

2. The second digit after the slash can change too…
Probably it means "front- or back panel connectors". But it could also be "Country Version" (In the full Type Number there are 3 digits after the slash…)
Anyway, the value of this last digit is not important for the definition. (No changes in the definition are needed when this digit changes)
But with a different *IDN? Answer TestController will not recognize the device…
So, is it possible to “mask” this last digit when TC does the device recognition with *IDN? Or is there another solution?

3. These power supplies were sold by Philips, but after the Fluke/Philips T&M alliance, they were sold under the Fluke brand. (In practice, I see more Fluke PM2811 than Philips PM2811).
However, also with the Fluke version, the *IDN? Answer remained PHILIPS,PM2811/xx
So, would you like to have both the Fluke and Philips version in your supported list?
And would coding like this work, as both #metadef’s are having the same #idstring:

Code: [Select]

; Meta definition for Philips PM2811 Power Supply
#metadef
#idString PHILIPS,PM2811/01
#name Philips PM2811
#handle PM2811
....

; Meta definition for Fluke PM2811 Power Supply
; As far as known the later Fluke versions kept "PHILIPS" in de *IDN? answer string
#metadef
#idString PHILIPS,PM2811/01
#name Fluke PM2811
#handle PM2811
....

Thanks in advance & regards, Gertjan.

edit: added the info that PM2811 is a SCPI instrument.

[HKJ]

1. Obviously the 30V/10A and 60V/5A versions need a bit different definition.
Would this work:

Code: [Select]

; Meta definition for PM2811 30V/10A version
#metadef
#idString PHILIPS,PM2811/01
....

; Meta definition for PM2811 60V/5A version
#metadef
#idString PHILIPS,PM2811/11
.....

Would TC automatically pick the right #metadef, depending on the *IDN? Answer?
I suppose the easy way would be to create just two model names: “Philips PM2811 30V/10” and “Philips PM2811 60V/5A”. But that seems to be a less elegant solution.

When you select you equipment on the "Load devices" page you select the correct *idn? answer, TC can also automatic pick a matching *idn? answer in some cases.

2. The second digit after the slash can change too…
Probably it means "front- or back panel connectors". But it could also be "Country Version" (In the full Type Number there are 3 digits after the slash…)
Anyway, the value of this last digit is not important for the definition. (No changes in the definition are needed when this digit changes)
But with a different *IDN? Answer TestController will not recognize the device…
So, is it possible to “mask” this last digit when TC does the device recognition with *IDN? Or is there another solution?

TC will compare the specified text, if you leave out the second digit it will not be checked.

3. These power supplies were sold by Philips, but after the Fluke/Philips T&M alliance, they were sold under the Fluke brand. (In practice, I see more Fluke PM2811 than Philips PM2811)
However, also with the Fluke version, the *IDN? Answer remained PHILIPS,PM2811/xx
So, would you like to have both the Fluke and Philips version in your supported list?
And would coding like this work, as both #metadef’s are having the same #idstring:

Code: [Select]

; Meta definition for Philips PM2811 Power Supply
#metadef
#idString PHILIPS,PM2811/01
#name Philips PM2811
#handle PM2811
....

; Meta definition for Fluke PM2811 Power Supply
; As far as known the later Fluke versions kept "PHILIPS" in de *IDN? answer string
#metadef
#idString PHILIPS,PM2811/01
#name Fluke PM2811
#handle PM2811
....


I believe the best solution is to add both Fluke & Philips definitions, people may not know it is the same device. In cases where TC automatic pick device it may select the wrong one, but because it is the same definition it will work.

On serial devices TC will try to automatic pick devices if "Scan serial ports" are checked, but first after added the devices listed and checked below that.

[Gertjan]

Hi HKJ,

Thanks. Most things are clear  .
But not the first question. Probably mostly because my question was not clear....
Lets show a bit more code:

Code: [Select]

; Meta definition for PM2811 30V/10A version
#metadef
#idString PHILIPS,PM2811/0
#name Philips PM2811
#handle PM2811
....

; Meta definition for PM2811 60V/5A version
#metadef
#idString PHILIPS,PM2811/1
#name Philips PM2811
#handle PM2811
.....
So there is only only one device name, but this device can return 2 different *IDN? answers, depending on whether it is a 30V/10A or 60V/5A version. (I adapted the #idstring already to your answer on the second question )

Will TestController, with the above code, select  the correct #metadef depending on the *IDN? answer?
If not, is there an other solution?

[HKJ]

Code: [Select]

; Meta definition for PM2811 30V/10A version
...
#name Philips PM2811
....

; Meta definition for PM2811 60V/5A version
...
#name Philips PM2811
.....
So there is only only one device name, but this device can return 2 different *IDN? answers, depending on whether it is a 30V/10A or 60V/5A version. (I adapted the #idstring already to your answer on the second question )

Will TestController, with the above code, select  the correct #metadef depending on the *IDN? answer?
If not, is there an other solution?

I did not see that, there can only exist one definition for each #name (i.e. the #name string must be unique), you must add /0 or /1 to them.

[Messtechniker]

Thinking about getting an UNI-T UT622 LCR meter after a tax return.
Anybody already doing a definition file? It has a Virtual COM interface ,
no HID thank God. So I might be able to produce a definition file myself
once I have it.

[Gertjan]

Hi HKJ,

I wrote a device configuration file for the Fluke / Philips PM2811 programmable power supplies.
It is based on the R&S HMC8043 definition.


IMG 1676  Philips PM2811-2000pix.jpg

These supplies exists in 2 versions: 30V/10A and 60V/5A. I wanted to combine both versions in one definition. The PM2811 has SCPI commands to query the maximum output voltage and current. Idea was to use the answers from those in variables, and use them in the menus as "maxVolt" etc. But I was not able to get this working. So now, as a bit less elegant solution, there are two different #name for the two versions.

Although this is an older power supply (introduced in 1991), it is quite speedy over GPIB. Using AR488 I get 5 read-outs per second. (V, A and W)

Divergent from the R&S definition I choose to NOT turn on the outputs with the quick select U and I radio buttons. I feel that one should first set up a power supply, and only after that, and when needed, switch its output on. Mostly to prevent accidents  .


PM2811 CV menu screenshot-1000pix.png

The PM2811 protection works different from the R&S, and is more extensive. For Over Current Protection, one can choose between just current limiting, or switching the output off after over current.
The protection status indicator will show in detail what is wrong: Over voltage, Over current, Over temperature, Out of Sense or Calibration invalid.

There are also 2 and 3 channel versions: PM2812 and PM2813. These are big modular 19” units, with several output channel options. So these are too complicated to include in this definition with #metadefs. Especially without any means of testing...

It should be easy to create a definition for these, based on this definition, and using the 3ch R&S HMC8043 definition as a guide. To facilitate this, all the Output Channel selection commands are retained in this definition. (Here all CH1, the only Output... )

Please find the definition file here: Philips PM2811.zip

Regards, Gertjan.

[Oleksii]

Request to bump jSerialComm to recent release v2.10.3 and adopt TC code.

There were major and important changed for me in v2.10.0, so now jSerial can fully see and work with PTY ports on Linux.
I mean ports created by "socat" tools or from Python.
https://github.com/Fazecast/jSerialComm/releases/tag/v2.10.0


I've tried to repack TestController.jar to include updated jSerialComm, it almost working and now can even see ports created by Python (real Bluetooth BLE connections actually).
Error happening:

Code: [Select]

;; 00:54:43.342 jSerialComm version: 2.10.3
;; 00:54:43.344 Start thread for: ttyBLE - ATorch DL24P
;; 00:54:43.344 ttyBLE: Set params: 9600
;; 00:54:43.544 ttyBLE: Tx: B1 B2 11 00 00 B6
Exception in thread "Device Pseudo-Terminal Device (ttyBLE)" java.lang.NoSuchMethodError: 'int com.fazecast.jSerialComm.SerialPort.writeBytes(byte[], long, long)'
        at dk.hkj.comm.SerialPacketInterface.writeDataInternal(SerialPacketInterface.java:114)
        at dk.hkj.comm.SerialPacketInterface.writeData(SerialPacketInterface.java:134)
        at dk.hkj.comm.SerialPacketInterface.writeReadData(SerialPacketInterface.java:180)
        at dk.hkj.devices.DeviceATorchPX100$TranslatingCommInterface.communicate(DeviceATorchPX100.java:73)
        at dk.hkj.devices.DeviceATorchPX100$TranslatingCommInterface.cmdIdn(DeviceATorchPX100.java:139)
        at dk.hkj.devices.DeviceATorchPX100$TranslatingCommInterface.write(DeviceATorchPX100.java:188)
        at dk.hkj.comm.CommInterface.writeRead(CommInterface.java:144)
        at dk.hkj.comm.CommInterface.writeRead(CommInterface.java:139)
        at dk.hkj.main.InterfaceThreads$DeviceThread.initDevice(InterfaceThreads.java:1507)
        at dk.hkj.main.InterfaceThreads$DeviceThread.run(InterfaceThreads.java:1590)

This is probably because of change in v2.10.0, quote from the release notes:

IMPORTANT NOTE: The bytesToRead and offset parameters in both readBytes() and writeBytes() have been changed from type long to int so that there is no mismatch between the parameters and the return type. Any code specifically passing a long to these functions will need to be updated.

I believe that's what causing the issue.

So, please refresh TC with current jSerialComm release.

[HKJ]

Thinking about getting an UNI-T UT622 LCR meter after a tax return.
Anybody already doing a definition file? It has a Virtual COM interface ,
no HID thank God. So I might be able to produce a definition file myself
once I have it.

I hope you make the definition.
There is one caveat with LCR meters, you need to read and set mode (i.e. L, C or R) for each measurement. You can see in the existing LCR meter definitions how it is done.

[HKJ]

I wrote a device configuration file for the Fluke / Philips PM2811 programmable power supplies.
It is based on the R&S HMC8043 definition.

Thanks, it will be included in the next release.
Can I use your picture?

[HKJ]

Request to bump jSerialComm to recent release v2.10.3 and adopt TC code.

I will do that in the next release.
No changes in TC code is necessary for that. There is mostly no issues with updating jSerialComm, but because most updates do not add anything new to TC I do not update all the time.

[Gertjan]

Hi HKJ,

You can always use my pictures, and any other graphics, for your website.

Best regards, Gertjan.

[Oleksii]

When using "block" driver I have some issues, a few questions:

1. I need to send first a specific byteSequence to force device to respond on a pool byteSequence.
How can I do that?
#initCmd works only after device Found, which is a result of the pool definition, so it does not help.


2. If I need to perform a fixed delay after the specific byteSequence, before starting the pool, how is it possible?
#cmdDelayTime and #readingDelay are not related here - do not give me a result.

3. If I need to make a few different pools (with different byteSequences) and then parse them all to have all variables to parse their values.
How this can be made?