uTracerJS - software for uTracer (works on mac, win, linux and more)

[Ihor]

Hi All,

Recently I thought that it would probably be a good idea to make my webinterface that works in combination with ESP32 module (see the thread on this forum about "uTracer goes wireless" and also that project page http://boffin.nl/wp/utracer-and-esp32/ ) also available to other users, who do not want to deal with any hardware extensions (even though they are super easy to do). I had a bit of time and I rewrote the whole C++ backend part (which was run on ESP32) in javascript (using NodeJs) and now the same frontend that some of you already used in your web browsers can be run just locally on your desktop/laptop.

As a result, you can now try this new software, uTracerJS, without any modifications to your setup, in exactly the same way as you already do with Ronald's software or uTmax/uTracer, which means that uTracer has to be connected with a computer using serial connection, either via COM-to-COM or USB-t0-COM cable. But that's exactly waht you have already, if you do not own ESP32 yet:) 

The project page is here:

http://boffin.nl/wp/utracerjs/

You can see more details about the installation and usage there. For now, I compiled it for the most used platforms (win, mac and linux). In my tests it runs successfully on Intel and M1 Macs, as well as, windows 10 machine. Because everything is javascript/nodejs based, I can easily create executables for any platform (alpine, linux, linuxstatic, win, macos, (freebsd)) and any architecture (x64, arm64, (armv6, armv7)), which means that uTracerJS can in principle run or all possible devices including Raspberry Pi and so on.

The software is fully functional and will be slightly updated in the near future. The front-end part is already extensively tested by many users who have ESP32 module (the front-end part is exactly the same). I would appreciate any feedback and reporting of bug and inconsistencies. A few of them (for example implementation of a more advanced QuickTest which is present in ESP32 version, monitoring Debug window during acquisition, etc.) will be fixed in the near future, I just have to find more free time to do that ).

One note: with a modern approach of "signing" all the executables and binaries, I am not sure how it will work for other users. Please let me know if you experienced any problems. It might be not a problem at all and this note can be ignored. I have not experienced anything like that, but at the same time all my win and mac machines are probably already in some "developer modes" meaning that I might activated some checkboxes and options on the way, during the development, and do not remember anymore what I did.

Cheers,

Ihor

[BHdeC]

Very nice! 

Got it to work at first try on windoz, amazing isn't it?

Quick setup:

Unzip the file, 

Change the port to COM3 in the uTracerJS.txt file, 

ran it successfully

Point my browser to  http://localhost:80

And that's all. Went to debug and saw that the UI was communicating with the tracer nicely with a ping.

Then I tried to load one of my uts file and I couldn't tell the UI: look in that directory that's where I have all  my setup files. 

The UI is confusing for me. I tried to set my directory  path to the Config: entry and the UI locked up.

I am almost there :)

Thanks for sharing.

[Ihor]

Thanks for the feedback! Indeed, probably this part is not really clear, about the file location. So because the whole interface is exactly as the ESP32 version, it assumes that all the files are stored in the 'data' folder, which is in that uTracerJS folder. One cannot reassign it, and I am not sure if I will be changing much because then there would be two versions, one for ESP32 and one for desktops and this is not something that I want to support. SO, with ESP32 all the uploading and Downloading of existing or acquired .utd and uts files is done via tab "File" in that webinterface (there is also a "Upload" button). 

Now, for the desktop version, uTracerJS, you can skip that and then just place your old .utd files directly to the 'data' folders with a file browser of your choice. Also that .utd files that you acquire are all saved to that "data" folder and then it is up to a user to take them from there if you need. So, in other words, uTracerJS only sees "data" folder inside the folder that you unpacked. This is also done because using uTracerJS on Raspberry PI or other headless computes will not give an opportunity to choose another folder anyways. 

By the way, I already improved all those small leftovers (handling of breaks in communication between the desktop and uTracer, in case they happen), and so on, so the new version will be released this weekend. 

I also made a real test with uTraces for the situation that I described before, when for example Raspberry Pi (or in my case C.H.I.P. computer with arm based cpu) is connected to the utracer with a USB-to-Serial cable and then the user opens the webpage on an ipad and ask the uTracerJS on Raspberry PI to do the measurements, you can see the video here:

https://youtu.be/X1NVb7xgvbk

Cheers, 

Ihor

 

  

[Ihor]

I remembered now another another option for uploading the existing .utd files. It was actually there already, I just never use it. In the "Import" menu there a button "Choose file". So if you do not want to put your files into 'data' folder (which is a folder inside uTracerJS main folder) then it is possible to just upload the existing .utd file into the plot-window using that button. All the files that present in 'data' folder are automatically picked up and shown in that dropddown menu "select .utd files on SPIFFS" (see the screenshot). 

Cheers, 

Ihor

 

[BHdeC]

Thanks for the explaination.

I've created a data dir under the C:\Program Files (x86)\UTTracerJS where I installed the exe.

Now the browser complains with a message: Error: ENOENT: no such file or directory, stat 'C:\Program Files (x86)\UTTracerJS\data\fulltest.html'

I probably messed it up.

The program's console shows:

-------------------------------------------------------------------------------------
----------------   Starting uTracer.js 0.93 (build 20220218)   ----------------------
----------------   (c) 2022, Ihor Smal, m...@boffin.nl           ----------------------
-------------------------------------------------------------------------------------
[INFO] Found uTracerJS.txt, getting the COM port name and localhost port from there ...
[INFO] Serial port                    : COM3 open
[INFO] baud rate                      : 9600
[INFO] portnumber on http://localhost : 80
[INFO] path __dirname                 : C:\snapshot\nodejs
[INFO] path to 'data' folder          : C:\Program Files (x86)\UTTracerJS
[INFO] process.cwd()                  : C:\Program Files (x86)\UTTracerJS
Error: ENOENT: no such file or directory, stat 'C:\Program Files (x86)\UTTracerJS\data\fulltest.html'

[Ihor]

Hi!

You should not have created the 'data' folder, because it was already there, and the folder which was there already contained the code for uTracerJS and also some example files with .utd files. By creating a new 'data' folder, the main executable is now missing all the other program files, so the easiest way is to get the uTracerJS.zip once again from the webpage and unzip it, from example into that UTTracerJS folder but not to modify the structure inside but just use that 'data' folder (inside UTTracerJS) for storing, copying and deleting your new and old .utd files. 

When you unzip everything anew, you have to modify once again that uTracerJS.txt file and put there your COM4 port once again, but you have to do int only once. In the future version I will remove all the program files from the 'data' folder, so it will not be possible to delete those important files which are now in 'data' folder.  

Cheers, 

Ihor

[Nebojsa Tomic]

I just tried your cute program on my WIN 10, instalation was successful, but when I tried to do a ecc88 quick test, I get this message: "Lower the anode voltage point, becasue with the current stepping it exceeds 50V", anode voltage is set to 90V.

[BHdeC]

Thanks it is now working perfectly. I have to get familiar to it before I decide which one I'll now be using.

One question, the calibration info are saved where? I thought it was in a file with  the old UI, but it seems that it is saved in a EEprom?

[Ihor]

The calibration parameters are saved in a file called _eeprom.csv in the 'data; folder in order not to mess up the original tracker files, even though the calibration values are exactly the same as in the original windows software, and you have to save them only once.  

[Ihor]

It would help if you would post of the screenshot similar to one below. Did you chose Ecc88 from the dropdown menu? Do you have 400V version of uTracer or a modified low-voltage 50V version? 

The error more or less says that 90+9 and 90-9V test points are not going to work (if you see my screenshot) so probably instead of "9" you have some large number in those fields? 

The quickest page is the whole software has been tested for many users already for 2 years and I have never heard such errors, so I suspect some string problem. Is calibration values are also properly set? The screenshot there would also help. With this extra information we can easily figure it out and make it work for you! :) 

Cheers, 

Ihor

[Ihor]

I just double checked the code and that error could popup only if you wrongly put in the calibration field the version of the uTracer "uTracer LV (50V)". If you did that intentionally, then of course you cannot test tubed at anode voltage larger than 50V, and you are trying at 90V.

Ihor

[Nebojsa Tomic]

Hi Ihor.

Please check attached pitures. I have utracer 6.

[Ihor Smal]

Hi!

I do not have uT 6 so all the work was done with a kind of emulator. Your screenshots looks good so probably it is something with uT 6 part and then I will fix it later today. Doing the complete curve acquisition should work though. But it is good go have your feedback with real uT6 :) Thanks! 

Ihor

--
You received this message because you are subscribed to a topic in the Google Groups "uTracer" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/utracer/seaFp5pCEIA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to utracer+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/cd0ff4d8-d097-4e57-ab01-fdf3fe952463n%40googlegroups.com.

[Nebojsa Tomic]

Hi Ihor.

Thank you so much for that.

Heater command is working but I think there is some problem with HV, HV never started, it looks like my utracet board is not recognizing that command.

Please let me know if I can help to resolve this bug.

Thanks!

[Ihor]

I fixed the problem with uT6 and quicktest. Please re-download the updated version of the software from the website.

Taking into account that a user could also accidentally delete the 'data' folder and because of that also some part of the code, I also repacked the whole executable, so it contains the javascript side and the 'data' folder is clean now rom al the .js and .html files, so there are only .utd. ups and some config (.csv) files inside.

The new version also deals better with possible hangs of uTracer and interruption of communication but I will continue to improve that part as well, which is difficult because it basically never happens, so I have to create all those fake problematic disruption with Arduino emulator :) 

Cheers, 

Ihor

 

[Nebojsa Tomic]

Hi Ihor.

Just tried a new version, previous error is cleared, but I have a new one :(

HV doesn't start.

Check attached images

[D.A.R Achterberg]

Hi Ihor,

Is just saw that you are running your software for the uTracer in ARM?

And which software is that?

In the past a spend a quite some time, trying to run uTmax on a raspberry…

Because the RPI is very small and rock solid device…

But i gave up after a while, since I figured out that the soft was written in x64

And, i’m a noob as it comes to programming…

Op 25 feb. 2022 om 19:02 heeft Nebojsa Tomic <nesk...@gmail.com> het volgende geschreven:

I just tried your cute program on my WIN 10, instalation was successful, but when I tried to do a ecc88 quick test, I get this message: "Lower the anode voltage point, becasue with the current stepping it exceeds 50V", anode voltage is set to 90V.

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/f280c4ad-5fda-4be6-935f-b2d985ed9d33n%40googlegroups.com.

[Ihor]

Strange, it looks like the communication breaks after the second command. After sending two bytes the uTracer stops responding. So there is a connection over serial....What kind of cable are you using ? is it com-port to com-port or ubs-to-com? Could you also try to go to Debug Tab and press ping and see if you get the response on that page and also if in the terminal everything is ok. Strange.... I just tested myself the last build on windows machine and it worked. 

Cheers, 

Ihor

[Nebojsa Tomic]

Hi Ihor.

Thank you for doing this.

I'm using FTDI USB to serial adapter.

Please check attached files.

[Ihor]

I did the test on CHIP computer which has so called armv7, so it is the same as raspberry pi as far as I remember. I do not have any RPI anymore. But in any case, on that platform I do not compile executables, but just run uTracer.js with node. What it means is that you just have to be able to install Nodejs from the official website (https://nodejs.org/en/download/ and that one exists for all possible platforms) and then just execute uTracerJS as "node uTracer.js" from the shell after ssh'ing to your RPI. that uTracer.js is not available on the website, but I can prepare the version like that as well. I cap also try to compile for RPI but I actually need RPI to do it, and I did not want to explain all those details how to use it directly with Nodejs instigator.  

[Ihor]

Thanks! It looks like "ping" works, but actually in the previous screenshot when it broke, it happen exactly at sending ping during HV measurements.... could it just be unlucky case, and now it works?? What about some simple Quickest?

Cheers, 

Ihor

[Nebojsa Tomic]

Unfortunately it doesn't work

Check attached txt for Quick Test

[jgx plato]

Excellent app! I am making tests with utracer 3+, FTDI USB wire, and all seems work fine using quick test and measuring curves, but in the debug section all the time is connecting, there is no possibility of make a ping in my case. I have tried several times with no positive results.

Other thing that I don’t know if it is mistake or not is that, even if you doesn’t connect the heater in the app, is possible make a measure using the quick test and curve measure.

Jgxbos

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/39205850-0c05-4418-a40f-a20f3cdc9f8an%40googlegroups.com.

[Ihor]

Nice screenshots! 

- the possibility to make measurements without Heater ON before that, is a decided choice. It is mainly for those people (like me and many others here) who have external power supplies for heating the tubes, and so we do not have to activate the internal heater in any case (as in the original windows software). So, it is a responsibility of the user (who uses an internal heater) to turn it on before the measurements (and there is also a message about that on the top each time). 

- the ping is done using a bit specific protocol, so I wonder if you are using windows ? if you have any firewalls or antivirus software or so, which might block the access. Could you check if you see something similar to my screenshot in the terminal window when you start uTracerJS and then go to the Debug tab in the browser. It should say that a new client is trying to connect. Having that not working (the Debug tab) is not a problem at all, you basically never have to monitor it if things are working well :) 

Cheers, 

Ihor

[Ihor]

I just tried quite exhaustively test it with my uTracer3+ and with com-cable but also usb-to-com cable. It worked in all the cases that you tried, but I actually found one inconsistency which is a bit on a uTracer hardware side. It looks almost like you have, but it happens in a different moment. So when I do quick test multiple times with a delay of 5-10 sec., everything is fine. But, when I run one quickest and when it is finished I start the new one, uTracer hangs. It sends two bytes back in response but then the communication is frozen and one has to power it off and on. I think it happens because after the test, uTracer still discharges caps from the high voltage, so cannot execute the new quickest which wants to set some different voltage immediately, or something like that. In any case, I can reproduce it all the time. I will see why there is such issue in general but it is still strange that users of uTracer3+ (you also saw here the other user) have no problems with uTracer and uT6 has that. It would be interesting to hear something from another user with uTracer6. If it repeats, probably I just have to put some delays in communication somewhere....

Cheers, 

Ihor

[Ihor]

Would it be possible to test this build, only Quickest and for Triode mode. The archive contains only the executable that has to be replaced) 

http://boffin.nl/files/uTracerJS-win-x64.zip

I added a small delay between the first and second commands which are sent during this test. in your case, uTracerJS was hanging during that moment. It could not echo the second command. 

Cheers, 

Ihor

On Saturday, 26 February 2022 at 18:37:31 UTC+1 nesk...@gmail.com wrote:

[jgx plato]

Indeed I am working for the utracer with a Windows machine, and as you say, when I try to use the debug tab, there is no message new web socket connection.

For this reasson I have try run the app in a MacBook Pro M1 and it had been imposible. I have repeated the procedure as I did in the windows machine, but without success!

I have uses the utracerJS-macos-arm64, and I have modified the utracerJS.txt (see attached image)

Any idea to run the app in the Mac?

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/d7268cb2-7c7a-4007-b654-31bddee7db0en%40googlegroups.com.

[jgx plato]

The hang in the app, not only happen in your app, in the original app of the utracer occurs too.  As you said, it is mandatory that the HV led is off before ( capacitors discharge) you start other quick test.  I don’t know, what is the reason, but it occurs since I use utracer3+.

Regarding uT6,  I have one mounted, but I have to put into a chasis, so I can’t make tests right now :(.

As you said, is a reproducible error easy to avoid don’t  touching test button until the HV led is off! :)

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/e9c730c8-a71b-4eb2-82fb-ebbef0215644n%40googlegroups.com.

[D.A.R Achterberg]

Ooooh Ihor when you talk about SSH i’m getting back nightmares when i think about all that wasted time trying to cross-compile that software to ARM64.

I don’t mind spending loads of time into a project, but it’s so frustrating when you have to give up and ending up with nothing…

But if i get it right, you made the uTracer.js GUi…

And with that node you can you execute the GUi on a RPi , but you have to connect to connect with RS232 or Internet to the uTracer.

I’ll keep that in mind, because i have to many projects running at the moment.

But anyway keep up the good work!

Op 26 feb. 2022 om 18:21 heeft Ihor <iste...@gmail.com> het volgende geschreven:

I did the test on CHIP computer which has so called armv7, so it is the same as raspberry pi as far as I remember. I do not have any RPI anymore. But in any case, on that platform I do not compile executables, but just run uTracer.js with node. What it means is that you just have to be able to install Nodejs from the official website (https://nodejs.org/en/download/ and that one exists for all possible platforms) and then just execute uTracerJS as "node uTracer.js" from the shell after ssh'ing to your RPI. that uTracer.js is not available on the website, but I can prepare the version like that as well. I cap also try to compile for RPI but I actually need RPI to do it, and I did not want to explain all those details how to use it directly with Nodejs instigator.  

--

You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/2c74ce2d-7928-41a1-87c3-eff2b074db13n%40googlegroups.com.

[Ihor Smal]

Two things I would check on your Mac M1 ( I am also running it on Mac m1 by the way):

The port in your uTracerJS.txt (so the first line there) should be: 

/dev/tty.usbserial-FT90VQMB  

And please try to run it in the shell not as /Users/jgxbos/uTracerJS/uTracerJS-macos-arm64 but first go to that folder, where uTracerJS-macos-arm64 is located and then run it there. I see I have to add that to the description on the website :) 

So basically doing this:

Run the terminal and then:

cd /Users/jgxbos/uTracerJS/

./uTracerJS-macos-arm64

Cheers, 

Ihor

On 26 Feb 2022, at 20:09, jgx plato <jgx...@gmail.com> wrote:

Indeed I am working for the utracer with a Windows machine, and as you say, when I try to use the debug tab, there is no message new web socket connection.

For this reasson I have try run the app in a MacBook Pro M1 and it had been imposible. I have repeated the procedure as I did in the windows machine, but without success!

I have uses the utracerJS-macos-arm64, and I have modified the utracerJS.txt (see attached image)

Any idea to run the app in the Mac?

<image_50384641.JPG>

El El sáb, 26 feb 2022 a las 19:24, Ihor <iste...@gmail.com> escribió:

Nice screenshots! 

- the possibility to make measurements without Heater ON before that, is a decided choice. It is mainly for those people (like me and many others here) who have external power supplies for heating the tubes, and so we do not have to activate the internal heater in any case (as in the original windows software). So, it is a responsibility of the user (who uses an internal heater) to turn it on before the measurements (and there is also a message about that on the top each time). 

- the ping is done using a bit specific protocol, so I wonder if you are using windows ? if you have any firewalls or antivirus software or so, which might block the access. Could you check if you see something similar to my screenshot in the terminal window when you start uTracerJS and then go to the Debug tab in the browser. It should say that a new client is trying to connect. Having that not working (the Debug tab) is not a problem at all, you basically never have to monitor it if things are working well :) 

Cheers, 

Ihor

<websocket.png>

On Saturday, 26 February 2022 at 19:03:37 UTC+1 jgx...@gmail.com wrote:

Excellent app! I am making tests with utracer 3+, FTDI USB wire, and all seems work fine using quick test and measuring curves, but in the debug section all the time is connecting, there is no possibility of make a ping in my case. I have tried several times with no positive results.

Other thing that I don’t know if it is mistake or not is that, even if you doesn’t connect the heater in the app, is possible make a measure using the quick test and curve measure.

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/d7268cb2-7c7a-4007-b654-31bddee7db0en%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/CALFq-68iEVmt%2BNUKhY9heXT%3DhoedWynSX9vHPNh7YEyjdAV%2BCQ%40mail.gmail.com.

[D.A.R Achterberg]

Maybe it’s time for a mod with 2 extra caps and a flipflop? Or a relay? Or lower the resistor for decharge the caps?

Just a quick thought…

Op 26 feb. 2022 om 20:25 heeft D.A.R Achterberg <d.a.r.ac...@gmail.com> het volgende geschreven:

Ooooh Ihor when you talk about SSH i’m getting back nightmares when i think about all that wasted time trying to cross-compile that software to ARM64.

[Ihor]

Thanks! 

Indeed, initially it was a project for ESP32 module where ESP32 would communicate with the uTracer (via serial) but the user would connect to ESP32 via wifi and run this GUI in a browser. So ESP32 had a C++ server-like code and the frontend was written in javascript and ran in a browser. Quite some people built that ESP32 add-on but also many people did not want to go that way and were fine with using a cable connection :) So I decided to rewrite that "server" part from ESP32 in javascript as well and keep the frontend as it was. Now, one just needs that Nodejs installation which is a kind of a small local web server in javascript and that's it. 

I found out that with nodejs one can nicely package the javascript code as executable which has only small part of nodejs installation and it works good for many platforms. Of course there are restrictions there as well. I can for example compile arm64 and x64 versions on my M1 Mac, but I cannot compile arm64 version on my Intel Mac. So, with Raspberry pi (armv7) it is the same problem. What I read, it should be possible to compile armv7 on arm machine but I get lots of problems and probably one day will try to spend more time on that. 

Now, with Nodejs it is also possible just to run the javascript version of uTracerJS which does not require any compilation, but for that one has to install NodeJs on his/her machine beforehand (couple of hundreds megabytes), and that works nicely on all existing platforms including RPI and so on. It is almost as if when you want to run something in python, you have to install python. The "running" of course is not so user friendly. one has to open the terminal application, go to the specific folder and execute a file from a command line.... and for some it is not as easy as it sounds. 

Cheers, 

Ihor

On Saturday, 26 February 2022 at 20:25:19 UTC+1 d.a.r.ac...@gmail.com wrote:

\

[jgx plato]

The port in the txt file is ok.  I can’t try run in the shell directly, it make sense!, right  now I am not at home, tomorrow morning I will try and I will  give you feedback!

Again, thanks for the excellent job with the app!

Jgxbos

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/4620732B-FB16-495B-8109-3D1174BF68DA%40gmail.com.

[Nebojsa Tomic]

Hi.

This version is working in Quick test! 

Still problem with regular test.

[Ihor]

Great!! 

It was also a trial with a quicktest so now I introduced the same small delays in the code for full test and quicktest (pentode). So the lesson is that one should not flood uTracer with commands over serial without any delays because it cannot keep up with it, and apparently uTracer3+ is more robust and less sensitive to that, compared to uTracer6 :)

I updated all versions on the website, but you can also get the archive here: 

http://boffin.nl/files/uTracerJS.zip

and then just replace the executable in the folder that you already have. 

[Nebojsa Tomic]

Hi Ihor.

It's working now!

Thank you so much for your effort and this beautiful app!

My only concern is that I'm getting different results running your app vs original utracer app. I tried several times with different tubes.

Please check attached pictures.

[Ihor]

I am going to test it on my uTracer3+ and see I there are differences. I did it long time ago and it was fine. For uTracer6 I implemented everything in the same way as it was in the windows software for uTracer6 but mostly by reverse engineering, so matching the requests and those were the same on my simulator but it was a year ago so things might change. I will to some tests with the latest software from Ronald during the next days and let you know. Just to be sure, are the Calibration values correctly inserted to the Calibration Tab, from the windows software? 

Ihor

[Nebojsa Tomic]

I copied  manualy calibration values from utracer 6 to your app in the calibration tab

[Ihor]

Strange, I just tested the latest version of uTracer6 software and uTracerJS using my uTracer emulator and then results are exactly the same. What I mean is that I have an Arduino which pretends to be a uTracer and sends proper responses to uTracerJS or original uTrcaer software during the heating, quicktest or full tests. When I use both softwares, I get the same commands send and the same responses and the same measurements. I also tried with uTracer3+ software and it also works as before. I would have to think what is happening, because you should not get different results. Tomorrow I can try to run two softwares on ny real uTracer3+ and see how it behaves there, but it would be good to have someone else with real uT6 to test. 

Another but the way, did you also change the sensing resistors to 4.7 ohms in the Calibration tab, for uTracer 6?

Cheers, 

Ihor

[Ihor]

Another small nuance that might slightly change things is to remove that checkbox "Adjust Va/Vs", then it should really make the same quicktest as the windows software (in theory). Now with that "adjust" it tries to correct the voltages a bit, so it might have some influence, but I would doubt that.

[Nebojsa Tomic]

Q: did you also change the sensing resistors to 4.7 ohms in the Calibration tab, for uTracer 6?
A: yes, 4.7R is set

remove that checkbox "Adjust Va/Vs", then it should really make the same quicktest as the windows software

done, check box removed but I get the same thing
Check attached images

[jgx plato]

I have had success running the app in the MacBook M1 running the app in the shell directly!. However, the debug section remain “connecting” in the same way that with the windows machine. As Ihor said before, if everything works fine, who wants debug section :) !

Next days I will make more tests!

[Ihor]

Thanks for the screenshots! Today I also did real test with my uT3+ and the results are not so bad. So I testes 6n3p at Va1=Va2=150V and Vg=-2 and I first tried with the windows software after warming up for 3-4 mins and I got this currents:

Ia1 = 6.88, 6.88, 6.9, 6.88

Ia2 = 5.44, 5.42, 5.4, 5.42

Then I tried with uTracerJS, so 1-2 minute later

Ia1 = 6.93, 6.95, 6.96, 6.92

Ia2 = 5.51, 5.46, 5.54, 5.47

Then I tired with the windows software once again, so 5 mins or so later:

Ia1 = 6.91, 6.92, 6.92, 6.9

Ia2 = 5.48, 5.46, 5.47, 5.46

So the results are very comparable. The first results are slightly lower because apparently the tube was not preheated properly. I have an external heater that was heating the tube nonstop, so I did not have to turn it on and off all that time. I believe you are using builtin heater, which means that the tube was on and off with heating for your two tests, and probably if that was on different time scale it might affect the results. 

I will have to think about other tests how to check that, but it is getting tricky. My main test was to use a simulator that always responds with the same values to uTracerJS and windows software, and the conversion to the final current measurements is exactly the same, so it should work with a perfect matching. Probably one test could be to use not a tube but a resistor so we can measure the current and we do not have to heat it (so to eliminate that component). The points of suspicion would be that the mathematical conversion has some inaccuracy on a very small scale because java and Visual Basic (of Ronald) have some differences in accumulation of errors. I observed that already with C++ and Javascript (for ESP32). Those are minor, but who knows. The other suspicion is also about the resolution in current measurements. I would have to check what are the ADC resolution for uT6 and uT3, but if it is the same, then presenting 200mA (max current for ut3+) and 1000mA (for uT6) will have different discretisation in the results (the values after floating point), but at the same time, the quick test is done with automatic gains for Ia and Is, so the range has to be adjusted nicely by the PIC during the acquisition to get the best performance. 

Cheers, 

Ihor

[Ihor]

So to be complete, I just did the experiment with a 6.8K resistor connected between the Anode and Cathode inputs and Va of 100V, which sagged to 95V (measured) during the quick test. The measured currents in 4 experiments:

WIndows software: 13.93, 13.96, 13.95, 13.94mA

uTracerJS: 13.93, 13.92, 13.93, 13.94

So, at least now I know that for uTracers3+ both softwares works exactly the same :)

Cheers, 

Ihoe

 

[Nebojsa Tomic]

Hi Ihor.

Just tried to test 10K resistor and I'm getting very similar results

Check attached pictures

I'm not sure where is the problem when I'm testing real tubes. BTW I'm using external heater supply

[Martin Manning]

Here is some data from QuickTests of a resistor using my uT3+ that shows stabilization of the measurement over time.

MPM

[Ihor]

@Nebojsa

Somehow I remembered your debug log and that you were setting the heater ON there, but it was probably just for the test. If you have an external heater, then with uTracerJS you do not have to activate the build one as in the windows software, but you are probably not doing that anyways. 

The comparison with the resistor is indeed interesting and nicely matching (especially the quicktest). To compare the curves, you can use a feature of uTracerJS for that (I recorder a small screencast for that). You can acquire .utd file in windows software and then upload it to uTracerJS (either by putting nt he data folder, or with the button Choose file in the Import tab. When you do that you can "copy it co Clipboard (as in the screencast) and then either load another datafile or just acquire one with uTracerJS, and after that it is possible to select one curve on one axis of the plot and the other one (out of 4 memory slots/clipboards) on the right axis and then you get a perfect one to one comparison. I also use that functionality to match the tubes. 

Screencast:

http://files.smal.ws/d0giNH

@MPM

The results with the resistor look interesting but also quite worrying. 1% of error is more than we observed just now (today in our experiments) but of course it is possible. I am just not sure what would cause that. So the resistor definitely also just slightly heats after all those consecutive (even though short pulses of) measurements, but also uTracer should heat (electronics) and the measurements should be floating/drifting. Interesting to know that it can happen.

[Nebojsa Tomic]

Thank you Ihor for great support and this outstanding software!

[Martin Manning]

"The results with the resistor look interesting but also quite worrying. 1% of error is more than we observed just now (today in our experiments) but of course it is possible. I am just not sure what would cause that. So the resistor definitely also just slightly heats after all those consecutive (even though short pulses of) measurements, but also uTracer should heat (electronics) and the measurements should be floating/drifting. Interesting to know that it can happen."

I ran the same trial with a tube (6L6GC beam pentode, bench supply for the heater), and saw nearly 2% total drift. It's a classic exponential curve, and I see that consistently running a saved calibration setup and built-in 10k calibration resistors. Re the test resistor heating, I don't think it's significant. For a 100ppm part that would be 0.1% for 10C. With the extremely short pulse duration, and one minute interval between measurements, there is very little temperature rise. 

[jgx plato]

Maybe we are everybody wrong  trying to get the exact measure, I don’t know if someone has try  to provide the mean and standard deviation ( a kind if error in the measure) of the measure. Maybe, when we push the test button, 5-10 measures could be done and the mean/median and standard deviation/IQR of the distribution of data shown in the app. Maybe we get more reliable and repetible measures... the prize to pay a measure done during a little more seconds....

Probably this idea has been implemented by Ronald or Ihor and rejected  ( I don’t know how the softwares works...) in the apps, if not, I think that is a way to explore!

Jgxbos

--

You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/7c0bd32f-f76e-444a-98e3-edc1cfc6b2b5n%40googlegroups.com.

[Ihor]

Indeed, the error is not so bad and quite acceptable in such measurements and indeed the resistor should don't heat much from short pulses and having delay of 1min in between, but there is still "improvement of the accuracy" over time and that's strange, because it means that either the measuring device changes it is characteristics over time or the device under test (resistor) or both. And from that plot it is clear that it indeed happens. The only suspicious (temperature sensitive) parts are usually semiconductors (transistors). So one experiment to complement that plot is to measure the resistor only in the beginning and then after each 10-20 mins (so really independent measurements) and see if that a property of uTracer that it has to be "preheated" like old school voltmeters and so on :)

It is very easy to implement in the quick test the option to do 5 or 10 measurements. Actually the reported currents are just measurements of one set voltage, so it does not cost much time to acquire more measurements. Additionally during the quicktest, uTracer measures 4 extra set points to derive all the derivatives, mu, Ra, Gm and so on and those do not have to be repeated. The problem here (again looking at the plot) that if one waits long before the measurements, and then acquires 5 or 10 consecutive measurements,  they will not change much, all of them will be very accurate. Inaccuracy is only in the first half an hour of operation or so. One would expect that measuring more times reduces the variance, and the average is unbiased and nicely estimated by such averaging procedure, but apparently there is a drift in parameters, so such averaging is not effective. 

Cheers, 

Ihor

  

[jgx plato]

Of course you are right about taking 5-10 measurements does not affect to the drift shown in the graphic. I was not talking about the drift with time in the measurements in the resistance show in the above mentioned plot ( moreover I think that a drift of 1% in this kind of    equipment  is completely acceptable). 

I am talking about the results  in the quick test, the variance is small, but I never get the same reading (I haven’t  calculated the mean error, sure that is low). I was proposing a more robust way to estimate the results shown in the quick test ( the way that I used to annotate the values of my tubes, maybe it is due to my engineer formation :) ).

Regarding the drift in the plot above mentioned, I have no idea, maybe the electronic components of utracer doesn’t reach the “stationary” temp o whatever be the cause of a drift until 30 min after switch on. It would be solved preheating the utracer 30 min, although  it make no sense for me ;)

--

You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/29c6b48c-2827-4acc-b0f1-9144d62b50b1n%40googlegroups.com.

[Martin Manning]

In recent measurements of resistors looking for the explanation for a difference in screen current I have been running traces up to 300V through 470 ohm 1/2W resistors, 30 measurements at ~1/sec, topping out at over 600mA... 170W! After such a trace, the resistors are just warm to the touch. In the above plot showing the 1% drift after 30 minutes, 4k7 resistors were measured at 250V, resulting in 13W instantaneous dissipation. There is no significant change in resistance due to heating, I'd bet, and the drift is all down to temperature stabilization of the uTracer hardware.

[Ihor]

I am talking about the results  in the quick test, the variance is small, but I never get the same reading (I haven’t  calculated the mean error, sure that is low). I was proposing a more robust way to estimate the results shown in the quick test ( the way that I used to annotate the values of my tubes, maybe it is due to my engineer formation :) ).

Yes, that idea is nice and at some point I was thinking about that, but this is for example what I get if I measure a tube in quicktest 4 times in a row, even even with 10-20sec delays:

a1 = 6.88, 6.88, 6.9, 6.88

Ia2 = 5.44, 5.42, 5.4, 5.42

Then I tried with uTracerJS, so 1-2 minute later

Ia1 = 6.93, 6.95, 6.96, 6.92

Ia2 = 5.51, 5.46, 5.54, 5.47

Then I tired with the windows software once again, so 5 mins or so later:

Ia1 = 6.91, 6.92, 6.92, 6.9

Ia2 = 5.48, 5.46, 5.47, 5.46 

So the error would be very small and in this cases I would be happy with saying that that tube has 6.9 and 5.5mA emission taking into account that the other similar tube will have a difference of 5-10% (just from the factory specs) :) 

[Ihor]

A small update about the latest version that I put on the website ( http://boffin.nl/wp/utracerjs/ ). 

- Now, all the parts mentioned in the first post (a third way of doing quick tests (by searching the the grid voltage which corresponds to the given/set anode current) and handling the breaks in communication between the uTracer and a computer) are implemented. 

- Also, the way to configure uTracerJS during the first run is also simplified, no need to create uTracerJS.txt config file yourself, it is possible to configure the port directly via webinterface

- uTracerJS is now also automatically checks if there is a new version available, and reports with a message below the menu bar.

Cheers, 

Ihor

 

[Nebojsa Tomic]

Hi Ihor.

Thank you so much for this update.

Only problem I can see is when I dowload zip file from your site it contains version 0.94 intead of 0.95, at least for Windows.

[Ihor]

Thanks for reminding me about that. I was debugging that "new version update" functionality and did not change the version number inside. I uploaded the new version already, but the one that you downloaded is as good, without any changes, except that version number. 

[Nebojsa Tomic]

everything works now :)

just one question, I noticed during the plot everything goes smoothly to the point 9 of the curve and then stop,  HV LED is off and then after 1/2 sec it turns on and plot continue to point 10, it looks like a lag, is this normal?

[Ihor]

Strange behaviour. There is nothing really specific about point 9, 10 or any other. Some delays might happen due to compliance errors (but those you should be able to see on the webpage) or, for example, if you acquire let's say two curves (for two values of grid voltage) and then you measure let's say 10 points per curve (for each of Vg), then after acquiring the last (high voltage point) of the first curve, there is always a delay (of 1-4 sec) when uTracer discharger high anode voltage in order to continue with the next curve for new value of Vg and low anode voltage. I am not sure if that's what you are observing. If you could share a screenshot of that webpage, so I can see the parameters and name/type of the tube, probably I can do similar test with my uTracer.   

[Nebojsa Tomic]

I ploted only 3 curves right now, 0 -1 and -2.

Same behaviour, I have small lag between point 9 and 10 on ploting the curve.

Check attached file.

[Nebojsa Tomic]

I just tried to plot only one curve and there is no lug.

It looks like there is some software colosion between discharging the cap and ploting the end point of the curve.

Anyway not big problem for me.

[Ihor]

Thanks for pointing out that behaviour. Now, in the latest version there are no more delays anymore, as we already figured out. Such behaviour was due to the fact that uTracerJS always measures a point but then sends that measured value to the browser only when the next point is measured (so a kind of one point delay). The problem was that after acquiring the last point of one curve (with high anode voltage) and going to the low anode voltage for the next curve, uTracer waits a bit before returning the new measurement (discharging caps), and that's why the last point of the previous curve was also not updated/transferred. Now the measurement update routine is slightly changed, so there are no more delays. 

[Nebojsa Tomic]

Great, now evrything looks OK.

Thank you.

[Nebojsa Tomic]

Hi Ihor. 

After Windows 10 update to 21H2 application is not working :(

Updated the Java to the latest version but no change.

Strangely it works on Windows 11 on another laptop

[Ihor Smal]

Hi!

Strange, I have not checked it myself and apparently my windows is older, so I am now updating to 21H2. But, if you can, could you share what is written in that black windows when you run it. If it closes too quickly then you can open the terminal (command line) window manually and run uTracerJS there like this:

Just open a command prompt at the location of your .exe file, and manually key in the name of your .exe file to run it within that window.

1.Navigate to the folder where your executable resides2.Shift-Right click and select "Command Window from here"3.type in the name of the executable and hit enter4.The process should run, but the window should stay open

Updated java version should not help because the code runs using NodeJS which already has “java” plus the nodejs is embedded into the uTracerJS.exe so it is not possible to update like that.

 
Cheers,

 

Ihor

-- 
You received this message because you are subscribed to a topic in the Google Groups "uTracer" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/utracer/seaFp5pCEIA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to utracer+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/23e2dcc1-67f5-414a-a0a5-bd6d7fd40407n%40googlegroups.com.

[Nebojsa Tomic]

C:\Users\Nebojsa\Downloads\utracer\uTracerJS\uTracerJS>uTracerJS-win-x64.exe
pkg/prelude/bootstrap.js:1740
      throw error;
      ^

Error: The specified module could not be found.
C:\Users\Nebojsa\AppData\Local\Temp\pkg\0d3279c2c717847d4ab775afe917aa864764820336bd4291efd50791a8db0118\@serialport\bindings-cpp\prebuilds\win32-x64\node.napi.node
    at process.dlopen (pkg/prelude/bootstrap.js:2114:28)
    at Object.Module._extensions..node (node:internal/modules/cjs/loader:1183:18)
    at Module.load (node:internal/modules/cjs/loader:981:32)
    at Function.Module._load (node:internal/modules/cjs/loader:822:12)
    at Module.require (node:internal/modules/cjs/loader:1005:19)
    at Module.require (pkg/prelude/bootstrap.js:1719:31)
    at require (node:internal/modules/cjs/helpers:102:18)
    at load (C:\snapshot\nodejs\node_modules\node-gyp-build\index.js:21:10)
    at Object.<anonymous> (C:\snapshot\nodejs\node_modules\@serialport\bindings-cpp\dist\load-bindings.js:10:46)
    at Module._compile (pkg/prelude/bootstrap.js:1794:22) {
  code: 'ERR_DLOPEN_FAILED'
}

You received this message because you are subscribed to the Google Groups "uTracer" group.

To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/B082AB4B-96D5-46F8-895E-F5D92AAC025B%40gmail.com.

[Ihor Smal]

Ok I saw such problem also on my Mac one tie, and it is not windows specific but apparently some problem of a specific computer, where some temporary files are stored by nodejs (I am not sure why). In any case, the solution is simple, you just have to delete that “pkg” folder with all the content there. So go to C:\Users\Nebojsa\AppData\Local\Temp\ and delete everything that is in “pkg” folder. 

Cheers, 

Ihor

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/CAOMW_vuBiRQEVnapWvR--6fUMaAfZ8_gOHeYzazy9%3Da-SYh%3DjA%40mail.gmail.com.

[Nebojsa Tomic]

It's working now :)

Thank you so much for your help!

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/0EB55B35-6CAC-44F0-A98A-521996F68632%40gmail.com.

[nickb]

When I try to run it under in Windows 21H1 the server fails to launch with a permissions error:

-------------------------------------------------------------------------------------
----------------   Starting uTracer.js 0.95 (build 20220303)   ----------------------
----------------   (c) 2022, Ihor Smal, me---boffin.nl           ----------------------
-------------------------------------------------------------------------------------
[....] path __dirname                 : C:\snapshot\nodejs
[....] path to 'data' folder          : C:\Users\nick\Downloads\uTracerJS\uTracerJS
[....] path to 'js'   folder          : C:\snapshot\nodejs\js
[....] process.cwd()                  : C:\Users\nick\Downloads\uTracerJS\uTracerJS
[LOAD] Found uTracerJS.txt, getting the COM port name and localhost port from there
node:events:368
      throw er; // Unhandled 'error' event
      ^

Error: listen EACCES: permission denied 0.0.0.0:80
    at Server.setupListenHandle [as _listen2] (node:net:1317:21)
    at listenInCluster (node:net:1382:12)
    at Server.listen (node:net:1469:7)
    at Function.serverListen [as listen] (C:\snapshot\nodejs\node_modules\express-ws\lib\index.js:42:40)
    at Object.<anonymous> (C:\snapshot\nodejs\uTracer.js:2269:18)

    at Module._compile (pkg/prelude/bootstrap.js:1794:22)

    at Object.Module._extensions..js (node:internal/modules/cjs/loader:1153:10)

    at Module.load (node:internal/modules/cjs/loader:981:32)
    at Function.Module._load (node:internal/modules/cjs/loader:822:12)

    at Function.runMain (pkg/prelude/bootstrap.js:1847:12)
Emitted 'error' event on WebSocketServer instance at:
    at Server.emit (node:events:390:28)
    at emitErrorNT (node:net:1361:8)
    at processTicksAndRejections (node:internal/process/task_queues:83:21)
    at process.runNextTicks [as _tickCallback] (node:internal/process/task_queues:65:3)
    at Function.runMain (pkg/prelude/bootstrap.js:1848:13)
    at node:internal/main/run_main_module:17:47 {
  code: 'EACCES',
  errno: -4092,
  syscall: 'listen',
  address: '0.0.0.0',
  port: 80

On Sunday, February 20, 2022 at 8:12:16 PM UTC Ihor wrote:

Hi All,

Recently I thought that it would probably be a good idea to make my webinterface that works in combination with ESP32 module (see the thread on this forum about "uTracer goes wireless" and also that project page http://boffin.nl/wp/utracer-and-esp32/ ) also available to other users, who do not want to deal with any hardware extensions (even though they are super easy to do). I had a bit of time and I rewrote the whole C++ backend part (which was run on ESP32) in javascript (using NodeJs) and now the same frontend that some of you already used in your web browsers can be run just locally on your desktop/laptop.

As a result, you can now try this new software, uTracerJS, without any modifications to your setup, in exactly the same way as you already do with Ronald's software or uTmax/uTracer, which means that uTracer has to be connected with a computer using serial connection, either via COM-to-COM or USB-t0-COM cable. But that's exactly waht you have already, if you do not own ESP32 yet:) 

The project page is here:

http://boffin.nl/wp/utracerjs/

You can see more details about the installation and usage there. For now, I compiled it for the most used platforms (win, mac and linux). In my tests it runs successfully on Intel and M1 Macs, as well as, windows 10 machine. Because everything is javascript/nodejs based, I can easily create executables for any platform (alpine, linux, linuxstatic, win, macos, (freebsd)) and any architecture (x64, arm64, (armv6, armv7)), which means that uTracerJS can in principle run or all possible devices including Raspberry Pi and so on.

The software is fully functional and will be slightly updated in the near future. The front-end part is already extensively tested by many users who have ESP32 module (the front-end part is exactly the same). I would appreciate any feedback and reporting of bug and inconsistencies. A few of them (for example implementation of a more advanced QuickTest which is present in ESP32 version, monitoring Debug window during acquisition, etc.) will be fixed in the near future, I just have to find more free time to do that ).

One note: with a modern approach of "signing" all the executables and binaries, I am not sure how it will work for other users. Please let me know if you experienced any problems. It might be not a problem at all and this note can be ignored. I have not experienced anything like that, but at the same time all my win and mac machines are probably already in some "developer modes" meaning that I might activated some checkboxes and options on the way, during the development, and do not remember anymore what I did.

Cheers,

Ihor

[Ihor]

Hi!

This error indicates that there is something already running on port 80 on your machine, or some firewall or antivirus is blocking uTracerJS and does not allow it to start the server on port 80. There are lots of forum discussions on that, with respect to nodejs and all possible norton antivirus and firewall, in order to make it work on port 08, but the simpler solution that you can try is to use another port, for example 8080. For that, you have to edit the text file uTracerJS.txt (with any text editor of your choice) which is in the same folder as the uTracerJS-win-x64.exe and change the port number there from 80 to 8080, then save that file, run   uTracerJS-win-x64.exe and in your webbrowser go to http://localhost:8080

Let us know here if it worked!

or

[Nebojsa Tomic]

Hi Ihor.

I upgraded my utracet to 6p2 version beacuse I installed extension board.

Now uTracerJS in not working anymore :(

Any clue?

[Ihor Smal]

Hi!

Strange, but I noticed from other users that uTracerJ sometimes stops working because of corrupted libraries. I think in your case it has nothing to do with installing the original software. It does not modify anything in terms of drivers for serial communication and uTracerJS uses totally different libraries for that. So, it can only happen that those libraries got corrupted in one of the temporary folders, or somehow uTracer’s original software keeps the serial port open after the running, and then uTracer cannot connect. The latter can be easily tested with a reboot of the PC. For the first option please check Note2 and Note 3 here: http://boffin.nl/wp/utracerjs/

In general, it would be good to take a look at the log after you run uTracerJS. If you can copy paste that piece, that would be helpful for me. In Note2 on that webpage I also describe how to keep tha tblack terminal window open while you run uTracerJS and then see all the errors and complaints that are produced. 

Cheers, 

Ihor

-- 
You received this message because you are subscribed to a topic in the Google Groups "uTracer" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/utracer/seaFp5pCEIA/unsubscribe.
To unsubscribe from this group and all its topics, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/6dabff4f-ff4c-4b89-905b-2f0171ecd337n%40googlegroups.com.

[Nebojsa Tomic]

Hi Ihor.

Please check attached txt file.

I think problem is in communication between PC and utracer, different code with a new version or something like that.

Once I installed extension board utracer stoped working with previos version 6p1.

Only when I installed new version 6p2 it started to work again.

Please let me know if I can help.

Thx

[Ihor]

I think know what you mean now. Is this an extension board with the positive grid voltage for uTracer6? Also, does that 6p2 software works for both cases, when the board is connected and disconnected physically? or you have to switch to 6p1?  

Because that board can measure grid currents uTracer sends extra measurements to the computer, which means that that communication protocol is slightly different. I remember I read on Ronald's website that uTracer6 automatically detects the presence of such board on a hardware level and changes the communication protocol correspondingly. I believe the changes to the protocol are described on the website and it is not difficult to implement, the main problem is with testing and debugging it. I myself do not have uTracers6 and that extension board, and for a moment have no plan to buy them. I can take a look at the implementation of that feature in uTracerJS but testing will take more time because I would need a volunteer with the actual device. Some time ago I was talking about that possibility with another user from this forum, but at that time he was not ready with his extension board. On a shorter time I can probably implement ignoring those extra grid current measurements, so uTracerJS still can work. 

[Nebojsa Tomic]

Hi Ihor.

That is correct.  This is  an extension board with the positive grid voltage .

I didn't try to run 6p2 version without extension board, beacuse for that I will need to disassemble whole thing to remove that ext board.

If you need anything to try or test please let me know.

[Ihor]

Thanks! I will let you know then. Looking at your log file, I see that it is possible to automatically identify if uTracer has the extension board or not (because the uTracer's response now starts with "90...." and not "10.."). I can easily make a modification so that uTracerJS "works" with that extension board but ignores the grid current measurements. Right now it is also not possible to do anything with those measurements, because all the plotting functionality is not designed for that. Right now I have two related questions that would help a bit. 

- would it be possible to have one example of .utd file with grid current measurements, just to see how everything is arranged in a text-file, because that's not described on the main website

- did you have to put any calibration values in the Calibration dialog of the windows software, if so, how many and which ones, or probably you can point me to the calibration procedure for that combination of utracer and extension board. 

Cheers, 

Ihor

 

[Nebojsa Tomic]

Hi Ihor.

You can see calibration procedures for extension board in this construction manual for ext board, from page 18

https://www.dos4ever.com/uTracer3/Manual-ExtBrd.pdf

also please check attached utd file

--

You received this message because you are subscribed to the Google Groups "uTracer" group.

To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/ad496062-571d-42aa-99e1-933022ae3e97n%40googlegroups.com.

[Ihor]

Thanks for the files! I see that the grid current measurements were added as a new column, but unfortunately not as a last column in that table but in the middle of the of the other columns. It means that compatibility with older versions of .utd is destroyed and more "smarter" parsing is required by other developers. I will try to implement is sooner or later as well. I will also check the calibration procedure, if it is completely clear, in terms of conversion formulas (which are not usually in those pdfs, but in Ronald's description on the website).

Right now I modified uTracerJS (you can redownload it from the website) so it works with the extension board without a problem (at least I hope, I tired with my adruino based simulator, but please let me know), for all versions of utracers with and without the extension boards. The grid current measurements are read out but not converted and not used.  

[Nebojsa Tomic]

Hi Ihor.

Now is working!!!

Thank you so much for your effort.

[Ihor]

I just updated the uTracerJS with the possibility to measure and display grid current using the extension board. Feel free to download it and test it (and let me know if it works). Ronald will soon publish his conversion formulas for that board, but I tried to reverse-engineer them and at least input-output responses (on my uTracer-simulator) are the same as with the original windows software (I do not have the actual uTracer6 and also no extension board). Please report if there are any inconsistencies but in the worst case in a week we will know the real formulas :) 

The beta version can be downloaded from here: http://boffin.nl/files/uTracerJSbeta.zip

One still has to input the calibration values as in the screenshot. 

After the acquisition, one can choose Ig from the drop-down menu (as in the screenshot) to see the grid currents. 

Cheers, 

Ihor

 

[Nebojsa Tomic]

Hi Ihor.

Just tested, everything looks good!

Thanks.

[Ihor]

Good to hear!

[Martin Manning]

Hi Ihor, just got around to trying this with uT3 on Intel core MacBook Air OS 12.4. I originally opened the .zip in my downloads folder and ran from there. After setting the right USB port name (offered to me by the system, and it identifies the FTDI ser/USB adapter). I was able to run some traces, but I could not save the cal or .uts files, and the terminal session kept dropping. I found I had to move the app to the top level of my username directory, and add a folder named data at that level to get the store and recall functions to work. After that, I copied in my Win 10 cal file values and it seems to be working using normal and Vg Low grid bias. 

At one point I had to retype the Rsense values and resave to EEPROM to get the compliance limit on the main page to reflect the 6 ohm resistances I'm using. Also, I believe R Dekker's equation for the upper compliance limit is 3600/Rsense, so for 6 ohms I should be getting exactly 600 mA instead of 598.

Also, I have this error repeating, where one Vg is run twice an the adjacent one is skipped. Shutting down the terminal and the browser page and restarting fixed that, but then I had to retype the Rsense and resave the cal file to get the right compliance limit.

It's a nice interface, and much faster to get up and running vs. booting into Windows! 

[Ihor]

Hi Martin, 

Thanks for the feedback!

Indeed, one needs to have a 'data' folder next to the executables. In the main uTracerJS.zip archive, if all the files are unzipped then this folder is also there. It could happen that you tried the BETA version and there I only packed the executables. I will soon release the BETA as the main one so it should not be a problem. I will also implement a quick check if the 'data' folder is present so the user can see in the terminal window what actually goes wrong. 

It could happen that the calibration from the previous version got a bit 'corrupted'. I had to add one more Rsense for the grid, and I did it in a more consistent way for myself but that breaks a bit the previous format. So, there might be some messy values, and one has to check the calibration before using new BETA (of the new version in the future). I will also make a check before starting if the calibration file should be "upgraded".

The values for the complience are actually computed "exactly" (using the values from Ronald's blog), so that's why you see all those 'weird' numbers. In original Windows software the compliance values are nicely rounded to 20, 50, 100mA, etc. but in fact they are not so true and so nice. The actual coefficients (from the webpage) are

compliance_voltage_vals [] = {3.59, 3.125, 2.71, 2.29, 1.875, 1.41, 0.83, 0.42, 00};

and the formula is compliance_voltage_vals[i] * 100000.0 / resistorAnodeSense

I quickly searched once again and you can take a look here https://www.dos4ever.com/uTracerlog/tubetester2.html in the section 14 "Setting the compliance". It was some time ago but I remember that I was getting those numbers from such table. 

The last point is very interesting (about the missing curves)! One user already showed me that problem for only one case, which was not really reproducible all the time (but it was consistent for several runs that he had before reseting everything). I would really like to debug such problems because I have never observer it in my simulator and it is a very weird case because apparently it happens for a very specific condition which I cannot predict. I suspect it is some sort of specific settings that you set which cause some problems in the software. If you can repeat that problem and can send me the config file or parameters that you use, it would be great!  

[Martin Manning]

Indeed I am running the beta version.

I believe the top-most compliance limit is 3600/Rsense, since I have tested that against the values that appear in the GUI for different Rsense input values. I think that is done so that the max voltage into the ADC 3.6. The lower limit values could be found by dividing the max limit by 8, but I didn't check that. I started up again, and still got the default compliance limit. Retyping those inputs and saving the cal file fixed it. I recalled and re-ran the same setup and unfortunately I did not get the missing grid voltage error. When it happened before, I had been adding and deleting Vg values, and playing with the auto value sweep generator.

[Ihor]

Next time if you notice something like that please let me know, I really want to catch that behaviour.

About the compliance, here what Ronald writes in his blog "The voltage reference is however a peculiar device. Basically it is a 4 bit ladder network DA converter. It can operate in two modes, determined by the value of bit CVRR. In total 32 Vref values can be generated and Fig.14.1 shows hows, depending the value of CVRR, how Vref depends on the DA converter input. The highest value the voltage reference can generate is 3.59 V. The value of the current sense resistor is selected so that this voltage corresponds to a maximum current of 200 mA, the maximum current the uTracer is designed for. Using the excel table shown in Fig.14.1 I selected the voltage reference settings which as good as possible resulted in the compliance range 200, 175, 150,…, 25 mA. The settings are stored in a table in the GUI and send to the uTracer in the “<00> settings” string which preceeds a new measurement sequence."

So, the value is not 3.6 but 3.59, that's why there is such difference, so it is 199ma and not 200, and so on :) 

Cheers, 

Ihor

 

[Martin Manning]

I suspect I'll be using your interface regularly :^) I'll watch out for any errors that come up.

From the 400V 3+ conversion notes, located here: https://www.dos4ever.com/uTracerNotebook/Notebook.html#Current2

5. Adjusting the value of the current sense resistors
The anode and cathode currents are measured by the voltage drop over the current sense resistors R20 & R45. The standard value of the current sense resistors has been chosen such that at the maximum nominal current of 200 mA, the voltage drop over the resistor (18*0.2 = 3.6V) is well within the range of the AD converter and still allows for some “overdrive.” By adding per current sense resistor two 18 ohm resistors in parallel, the nominal current range is tripled to 600 mA. Fig. 15.3 shows how these resistors are added to the bottom side of the PCB.

In the Windows GUI, a 6 ohm input for Rsense gives the following compliance limit and steps:

600, 525, 450, 375, 300, 225, 150, 75

And for a 5 ohm Rsense:

720, 630, 540, 450, 360, 270, 180, 90

Both reflect a maximum of 3600/Rsense, which is then divided into 8 parts.

[Ihor]

Hi Martin, 

I am looking forward to your feedback! greatly appreciated! 

I am not sure how strict we should be with the compliance but here is my understanding from what I read and implemented: The thing written in the blog "(18*0.2 = 3.6V)" is correct from the theoretical point of view but "3.6" does not exists. From the previous post that I cited, one can only reach 3.59 with that voltage reference device that is used in uTracer. The voltage reference, it is a DAC device with some limited bit-resolution, which also means that we cannot get any analog voltage that we want, which also means that in practice we cannot just get divisions of 3.6 into 2, 3, 4, 5 or any other factor as they should be in theory. From Ronald's experiments, it turned out that it is possible to get only this desired range of voltages:     

3.59, 3.125, 2.71, 2.29, 1.875, 1.41, 0.83, 0.42

which (with 18 ohm resistor) results in the following currents (ma):

197, 174, 151, 127, 104, 78, 46, 23

Those are the real currents that uTracer deals with and compares against during the compliance test. I think in order not to make things look ugly, Ronald rounded them to nice numbers in the GUI:

200, 175, 150, 125, 100, 75, 50, 25

These are very close approximations in fact and those are the numbers that Ronald wanted to get in the first place with that DAC, but due to reality of DACs could not. 

Because those exact numbers are available on his blog and we are also into exact quantitative measurements, I thought why not to look at ugly but true fractions and values :) 

Ihor 

[Martin Manning]

Understood. Of course the ability to input new vales for Rsense came along later, and so now one can get ugly numbers depending upon what value is chosen for Rsense. It's a small point, but I do think that for display purposes Ronald is using 3600/Rsense, and then rounding to the nearest whole mA rather than using 3590. Maybe that was part of the reasoning for his recommendation to parallel 3x 18R to get 6R, which yields a nice round 600 mA upper limit. 

A comment on the plot, it's nice to limit the traces to the input Pa max, but one might want to see the curves out to the max Va sweep voltage and still show the Pa max curve. p-p amps operate above Pa max, and one might want to see the load line there along with the Pa max and grid curves. Maybe a tick box could be added to make that an option?

[Ihor]

Hi Martin,

Thanks for posting in the Amp garage forum!!!

About the Power limit, I am not sure if I understood you correctly, but I think all the options are already there. There are 3 places where one can set Pmax (max power) on the main webpage (of the webinterface). The first one is at the very top, in the acquisition parameters, specifies if the curves should be acquired by uTracer only up to specific Pmax value, or fully, without any restrictions (Pmax=0 in this case), so only the compliance violation will stop the acquisition. It will result in different pictures, one with a sort of hyperbolically "cut off" of the curves (Pmax >0) or full curves (rising up) (Pmax=0).

Now, after we acquire full curves (without Pmax limits), it is still possible to plot Pmax curves on top of those acquired plot by using two other places for the Pmax parameter. Those fields are in the specification of axes Y1 and Y2.  The values there can be changed by a user at any time and the cure will be redrawn on the fly. It is basically a visual overlay of Pmax curve with tube curves.

Also there, it is possible to change the min and max values for Va for example, and the plot range will be updated as well, so for example one can put a range from 0 to 1000 without any problems and have the tube curves rescaled on the fly. For that one has to specify "manual" in the "Scale" field. 

Is this what you mean? or you meant something else? 

Cheers,

Ihor

[mman...@fuse.net]

Hi, let me look at that again. In the mean time I have a case where the grid voltage skip occurred. What should I save so you can look into it?

---

From: "Ihor" <iste...@gmail.com>
To: "utracer" <utr...@googlegroups.com>
Sent: Friday, July 22, 2022 2:23:51 PM
Subject: Re: uTracerJS - software for uTracer (works on mac, win, linux and more)

--
You received this message because you are subscribed to the Google Groups "uTracer" group.
To unsubscribe from this group and stop receiving emails from it, send an email to utracer+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/utracer/a64a4d1b-a8a5-4e32-ab44-cf47f01f1fc8n%40googlegroups.com.

[Ihor Smal]

Hi Martin, 

Nice, let’s hope we can figure it out. One thing it would be nice to have the screenshot of the webpage, so I can see all the parameters that you use for that case. Probably also separately a copy of the Vg field (with all the values). Sometimes you can have a lot of them so not all of them will be visible int he screenshot. 

Another helpful thing is to have a copy of the terminal output, so all those messages that are appearing in the terminal window after you start uTracerJS and run that problematic test once. But probably to have the screenshot with parameters will already help. 

If all that does not work, I will create another version of binary which will log other things under the hood and help to catch that glitch. 

Cheers, 

Ihor

[Ihor]

The glitch with missing/disappearing some of the curves during the acquisition, initially reported by David some time ago and recently encountered by Martin is finally solved. Thanks for sharing the plots and trying different things that I asked. It was difficult to reproduce that problem because apparently it was happening only for the situations where a couple of things should coincide: exactly with the last acquired point within a curve we should violate either compliance or power limit or both.

The new version of software is updated on the website and ready for download!

Cheers, 

Ihor

    

[Ihor]

Small update: 

- The uTracer6 with the positive grid extension board was tested more thoroughly by Dave and the results are nicely matching (see the screenshots). 

- I also implemented one more feature, which I am not sure really useful but probably some people with uTracer6 or modifier uTracer3+ (for high currents) can dest and demonstrate the results :) As we discussed a few times somewhere on this forum, while measuring the anode and screen currents for the user-defined Va and Vs, some sagging of Va voltage is happening (which is proportional to the measured current). For example, while measuring 2A3 at 250V and currents of 50-60mA, uTracer reports that the measured Va was in fact 245V. In uTracerJS, in the Quicktest menu, there is a special option that will already compensate for that, so it will automatically (by taking one extra measurement) adjust Va (to a bit higher value, 254V or so) in order to get the exact measurement at 250V. 

Now, I thought to try such approach with the full curve acquisition. Basically for each point on the curve, for the specified range of Va or Vs we take two measurements, one to see how much we have to compensate and the second one (which will be recorded) in with this adjusted voltage. Actually, for measuring triodes no big improvements will happen. The thing is that even if we have the voltage sagging (and correspondingly a lower measured current), such measurement is still exactly on the real tube curve, it is just moved a bit lower along x and y axes (but still along the curve). I believe it is more interesting for pentodes where there is a dependence on Vs. Here it is also interesting to see if there is a benefit, because screen current is quite low, so Vs should not sag much...

In any case, feel free to test it. The new option is just a checkbox next to the "Abort" button (see the screenshot). If it is not checked - you get the standard function, if it is checked then there will be double-measurement mode (which also takes two times longer). It is possible to take two sets of curves and actually import them in uTrracerJS at the same time and compare in one plot. For those who never used that option: you can acquire the first curve, then go to Clipboard tab, click on button Clipboard 1, then go back and acquire the second curve, and after that select the Ia of the first curve stored in Clipboard 1 in Axis Y2 dropdown menu.      

The beta version of the software can be found here, you just have to replace the binary from your platform in the folder where you keep your uTracerJS and no changes to 'data' folder or any other files should be made:

http://boffin.nl/files/uTracerJSbeta.zip

Cheers, 

Ihor

[Martin Manning]

Hi Ihor, IMO doubling the acquisition time would not be worth while except when the highest accuracy is desired, so making this mode optional is probably best. 

Here is an example of the impact of voltage sag on anode curves for a KT88 using uT3+. I did not use your latest beta version, just the QT, first with Va and Vs adjustment to hit the 250V input values, and then with Va adjusted manually to get 250V actual (Vs input is 250V, and allowed to sag). The drop in anode and screen current indicates that the anode and screen current curves will be 2.1 and 2.9% low respectively at 250V Va without compensating for screen voltage sag. 

[Ihor]

I just ran comparison tests on KT88 with my uTracer3+, so the currents are reasonable, but of course they are limited to 200ma with the max. voltage of 400V, but the results are reasonable and indeed show that only small differences are present. In the plot the left Y axis (and solid lines) show the standard operation without the voltage-drop adjustments, and the right axis (dotted lines) shows the adjusted measurements (so they are always a bit to the right and up in the plots).

[Martin Manning]

Ok, so I downloaded and ran the latest. Full trace on a KT88 at 300V Vg2. Time to trace with new voltage correction checked 1:50.0, without 1:52.7. No time penalty?The two sets of curves are captured below, showing that the biggest difference is in the shape of the Ig2 curves at low Va. Overall the levels shift is as expected from my test using QT and your traces. I think this is good!

[Ihor]

Nice! Indeed, the changes are small and as predicted (but still for those who is into highest possible accuracy might be useful). 

Indeed, I also noticed that it does not take 2 times longer. Theoretically it is, but apparently the overhead in charging and discharging capacitors in the beginning of each cure is still the major time-consuming component, so measuring points twice is basically for "free". 

I will keep that option in the dialog most likely, but I would have to find a place for it, because it is going to be difficult to squeeze that check box with a text label somewhere there. 

[Martin Manning]

I vote "keep it" ;^)

[Martin Manning]

Ihor wrote: "I will keep that option in the dialog most likely, but I would have to find a place for it, because it is going to be difficult to squeeze that check box with a text label somewhere there."

A suggestion: Why not use the space currently occupied by the Voltage Correction text and checkbox? That doesn't exist on the original Windows UI, and anyway why not always report the actual voltages from the test? This would be following the same convention as you are using on the Quick Test form. The text could say something like "Adjust Va/s to match inputs."