Activity for Percy Zahl
-
Percy Zahl posted a comment on discussion GXSM Installation Knowledge
Glad to hear. Anyways I do not understand the actual cause, I suspect some dconf schemata missing for the plugin in the version you are using. May I ask: You use a pre-build .deb package, the live image system install or the latest git-hub and build it yourself?
-
Dear Takayuki-san, I am thinking about what may be causing this startup crash after "PCS init: looking for org.gnome.gxsm3.pcs.plugin-libinet-json-external-scandata.rp-pacpll-phase-hold-am-noise-limit array_flag=0 pcs_count=0" but do not have a good idea. Nor I have ever see this problem. This "External PAC-PLL (RedPitaya control) " plug-in is always loaded, used or not. If the error is as of this one (may be a missing schema, but that usually is been indicated clearly). You can try to simply delete...
-
Dear Takayuki-san, I am thinking about what may be causing this startup crash after "PCS init: looking for org.gnome.gxsm3.pcs.plugin-libinet-json-external-scandata.rp-pacpll-phase-hold-am-noise-limit array_flag=0 pcs_count=0" but do not have a good idea. Nor I have ever see this problem. This "External PAC-PLL (RedPitaya control) " plug-in is always loaded, used or not. If the error is as of this one (may be a missing schema, but that usually is been indicated clearly). You can try to simply delete...
-
Dear Takayuki-san, I am thinking about what may be causing this startup crash after "PCS init: looking for org.gnome.gxsm3.pcs.plugin-libinet-json-external-scandata.rp-pacpll-phase-hold-am-noise-limit array_flag=0 pcs_count=0" but do not have a good idea. Nor I have ever see this problem. This "External PAC-PLL (RedPitaya control) " plug-in is always loaded, used or not. If the error is as of this one (may be a missing schema, but that usually is been indicated clearly). You can try to simply delete...
-
Hi, what do you need to plot? At the moment the only plots I am able to make are amplitude-distance but am still unable to generate the expected amplitude-distance curves, ??? If you need any custom signal generation you may also use the "GVP" tab to program any action schemes. -P
-
Oh OK that Retract CI value. Leave that positive, some where 100..300. Q: Is your Z-Scan going the direction as required when you simple test the feedback? Do this test: 1) Just connect Bias (OUT6) -> Tunnel Input Signal (IN0) to simple emulate a current via Bias. 2) Put Current Setpoint to 1nA or so (assuming simple 1nA / Volt for testing) 3) Put Bias = 0V => This should move Z-Scan all the way to fully extended (+10V for positive polarity) 4) Put Bias > "1V", say 2V -- what is > 1nA, this should...
-
A few points: A) make sure you Z-Polarity is set correctly, this is what determines if + or -10V is fully extended. (Gxsm3 -> preferences) Must restart Gxsm after adjusting. Then it will ask you to confirm to reconfigure the DSP (MK2/3) accordingly. (DO NOT RESTART THE DSP!) B) All CP, CI must be positive. Else you cause internal "confusions"... only for vert special purposes or testing they may be negative in linear transfer mode. C) "Retract CI" is a "speedup" using negative CI (CP=0) temporary...
-
This is never happening on any of my systems in years. Not while regular scanning. Please list your setup details -- at minimum this: Computer type/CPU/memory/Linux System Version/Swap Size/Free disk space Gxsm version/build/install type? (via Package/Ubuntu GXSM ISO/self build and installed/etc.?) Hardware: MK2 or MK3? Scan setup, size, scan speed... PS: The very only thing may happen or go wrong is if you push the scan speed to the combined limits of the USB bandwidth -- what is a complex figure...
-
This is never happening on any of my systems in years. Not while regular scanning. Please list your setup details -- at minimum this: Computer type/CPU/memory/Linux System Version/Swap Size/Free disk space Gxsm version/build/install type? (via Package/Ubuntu GXSM ISO/self build and installed/etc.?) Hardware: MK2 or MK3? Scan setup, size, scan speed... PS: The very only thing may happen or go wrong is if you push the scan speed to the combined limits of the USB bandwidth -- what is a complex figure...
-
This is never happening on any of my systems in years. Not while regular scanning. Please list your setup details -- at minimum this: Computer type/CPU/memory/Linux System Version/Swap Size/Free disk space Gxsm version/build/install type? (via Package/Ubuntu GXSM ISO/self build and installed/etc.?) Hardware: MK2 or MK3? Scan setup, size, scan speed...
-
OK. BTW: You actually can reconfigure the limits (lower included) via Gxsm I found: Open the configure Dialog and type in the value you want, BUT do not hit Enter or do anything else, just click OK and it takes it! PS: I have not yet figured out why it complains for all limits at -10 when hitting enter, or do anything else... and set it back to -10 then. PPS:: Fixed the config dialog issue. In git now. The Entry Spinner-Limits/updates were interfering, removed spin feature from config as not needed...
-
OK. BTW: You actually can reconfigure the limits (lower included) via Gxsm I found: Open the configure Dialog and type in the value you want, BUT do not hit Enter or do anything else, just click OK and it takes it! PS: I have not yet figured out why it complains for all limits at -10 when hitting enter, or do anything else... and set it back to -10 then. PPS:: Fixed the config dialog issue. In git.
-
OK. BTW: You actually can reconfigure the limits (lower included) via Gxsm I found: Open the configure Dialog and type in the value you want, BUT do not hit Enter or do anything else, just click OK and it takes it! PS: I have not yet figured out why it complains for all limits at -10 when hitting enter, or do anything else... and set it back to -10 then.
-
Send me what all "unusal limits" need and I'll have a look.
-
There is some odd issue, a more generic gxsm wide issue effecting all neg limits, been reset to -10 when edited... For a work around, do NOT use the dialog to adjust the limits, BUT use the system tool dconf Editor. See screenshot attached. Find the setting as shown and adjust the limits as desired (I put +/-20 in for a test). Do this before starting gxsm3 -- or restart it after adjusting as it does not get a change notification if you edit it there!
-
Try starting gxsm3 with debug option -P 10 (may use -D 10 -P10 for even more details) as shown below and see if you see messages from "pyremote" listed -- see a potential missing lib error at load time: ~~~ percy@ltncafm:~$ gxsm3 -h no -s -P 10 GXSM3 comandline option parsing results: => xsmres.HardwareTypeCmd = no => xsmres.DSPDevCmd .... = (null) => xsmres.UnitCmd ....... = (null) => xsmres.force_config .. = 0 => force gxsm defaults .. = 0 => debug_level .......... = 0 => pi_debug_level ..........
-
This sounds like the related plugin is missing -- possible was not installed as of a compile issue, may be missing lib, etc? Have a look if you can find the libpyremote.so plugin build and the lib installed, depending on your install mode in usr/local or usr: /usr/[local]/lib/gxsm3-plugins/common/libpyremote.so
-
I have to check if there is (must be, just not sure it it is at -10V) limit for the "neg limit" setting if I understand you right.... hang on...
-
I have a feeling your device permissions are not set to allow user access. check if $ ls -l /dev/sranger* looks like with rw for all: crw-rw-rw- 1 root root 180, 2 May 14 15:41 /dev/sranger_mk2_2 if not, you can adjust as root manually "chmod a+rw /dev/sranger*" or better permanently fix and automatize via having the rules file installed/copied here: /etc/udev/rules.d/49-sranger.rules This is provided via the sranger package/SVN and can be found here: SRanger/SR_MK2_MK3Pro/kernel-module/src/49-sranger_mk23.rules...
-
never heard about this or seen such an issue. make sure your computer and storage is not lagging... maybe?
-
See also my post above "This is very hard to debug..." getting lil messy here sorry... But one more easy check, see screenshot attached here. In DSP Manager App scroll down until you see the task ID018 "Request McBSP transfer". This should be "on Timer" IF NOT SCANNING (automatically) for idle monitoring purpose. (While scanning it is read in realtime matching the scan speed.) The number/counts shall change.
-
This is very hard to debug for me via this forum... all connections must be right to get started. a) 1st question is: did this ever worked? Proof? b) as I have not made those nor QC tested -- I very sorry, this requires RF measurement skills and dedicated RF / very high impedance measurement /test equipment to debug the actual signals without disturbing them... and this is beyond me to support. I have the equipment at home. Not cheap unless home made -- correctly. c) I can offer you a quality tested...
-
if in the GXSM Pan-View the dFreq is matching the number in the RP control – and sure when you get data while scanning into those McBSP-xxx channels! In the default startup configuration after connection all is set, you only need to enable the controllers. Particular: OpMode to "Streaming Operation" and make sure channels are set to : "Phase, Freq, [Am,Ex]" (all default) but if you switch to tuning or monitor IN1/2 in Scope mode the transfer is stopped. Gxsm may needs to configured right also, but...
-
OK glad to hear. If you see data transmission error/spikes -- this is related to incorrect impedance matching for RJ45 -- but may be the PostDoc who build it realized it and did it right (minor adjustments)... I do not have the number in my head, but the inputs need a matching load resistor... can be placed over the IC SMD pins directly. ideally also add in-line resistors also, but works without OK up to 4m cable or so! UPDATE and FYI: I do not know your interface / driver build, if using the PCB...
-
OK glad to hear. If you see data transmission error/spikes -- this is related to incorrect impedance matching for RJ45 -- but may be the PostDoc who build it realized it and did it right (minor adjustments)... I do not have the number in my head, but the inputs need a matching load resistor... can be placed over the IC SMD pins directly. ideally also add in-line resistors also, but works without OK up to 4m cable or so!
-
You got the right one, this is it debug.log -- my bad, not the server log... that is only important fort networking issues. So did that happen after startup or after adjusting some settings on your side (what)? I see a Freq adjustment. And controller parameters been changed. CP,CI are in dB -- so make sure not to enter odd values.... up to about -190dB (max slow) and possible not over -10dB at any time needed. I think it should and is not possible to enter invalid parameters causing a "signal 7"...
-
You got the right one, this is it debug.log -- my bad, not the server log... that is only important fort networking issues. So did that happen after startup or after adjusting some settings on your side (what)? I see a Freq adjustment. And controller parameters been changed. CP,CI are in dB -- so make sure not to enter odd values.... up to about -190dB (max slow) and possible not over -10dB at any time needed. I think it should and is not possible to enter invalid parameters causing a "signal 7"...
-
a) are there any numbers fluctuating or changing, status cpu %... working? To make sure it is a life first. b) send me a screenshot with OK connection Unrelated to you current issue: I am curious: did you (or who did) build the McBSP interface or where is that from -- as I did not provided it nor anyone else from the Gxsm team -- I first thought I missed some thing here. If it is a 1:1 copy of the in SVN / Gxsm provided -- there is a minor patch required to make it work stable with longer (than very...
-
While connection monitor on the rp the nginx log file, this may give you and me hints: root@rp-f05aa9:/opt/redpitaya/www/apps/pacpll# tail -f /var/log/redpitaya_nginx/ws_server.log you may adjust the verbose/logging level -- can be done via gxsm GUI, but obviously not when loosing connection right away. So you may need to adjust temporary in the source code the default settings for that. percy@ltncafm:~/SVN/Gxsm3/plug-ins/control/RedPitaya-Support/pacpll/src$ less main.cpp find and adjust (both lines)...
-
Sorry I missed that detail before in the output: "Web socket connection closed externally" (line of the log) This is not OK. This can mean various things, as the socket connection was dropped on the RP side after the server exiting for some error condition, any of those: a) the FPGA bit file is not the correct version, non existent, or the wrong FPGA/RP type/model. What exact RedPiatya are you using? Must be the 14bit version. Can be either regular Z10 or bigger Z20. The default FPGA bit file in...
-
Sorry I missed that detail before in the output: "Web socket connection closed externally" (line of the log) This is not OK. This can mean various things, as the socket connection was dropped on the RP side after the server exiting for some error condition, any of those: a) the FPGA bit file is not the correct version, non existent, or the wrong FPGA/RP type/model. What exact RedPiatya are you using? Must be the 14bit version. Can be either regular Z10 or bigger Z20. b) In general means the RP support...
-
Do you think the problem could be in the McBSP connection between the RP and the MK3-A810 or with the installation of the RP FPGA and support software? The output signal and all settings will work even without the McBSP connection -- only the DSP/Gxsm won't get data while scanning. Everything else will work. However, when I choose a Volume and a Frequency, the In1 Offset is 0mV, the DDS Freq is 0.0000 Hz and when I connect Output2 of the RP to an Oscilloscope I see nothing (just noise). No idea right...
-
Hi, OK, so you got that interface via Thorsten... did not know. Please have a look and follow the steps in the pdf I attached -- there are detailed instruction to get started! Do you really have set a amplitude right (put in 300 mV) and set the Frequency to 32000 Hz or so. Keep both regulators Ampl Phase OFF (not checked) to get started -- manual control. I see in your screenshot operation mode = "INIT BRAM TRANSPORT". That should not normally set to "STREAMING OPERATION" (to get data to GXSM via...
-
Hi, OK, so you got that interface via Thorsten... did not know. Please have a look and follow the steps in the pdf I attached -- there are detailed instruction to get started! Do you really have set a amplitude right (put in 300 mV) and set the Frequency to 32000 Hz or so. Keep both regulators Ampl Phase OFF (not checked) to get started -- manual control. Why this: ??? " Ref. frequency in the MK3-PACPLL control is set to zero."
-
Hi, OK, so you got that interface via Thorsten... did not know. Please have a look and follow the steps in the pdf I attached -- there are detailed instruction to get started! Do you really have set a amplitude right (put in 300 mV) and set the Frequency to 32000 Hz or so. Keep both regulators Ampl Phase OFF (not checked) to get started -- manual control.
-
Hi, the "analog" out option is not any more enabled and obsoleted -- was only for early testing and very limited in resolution. You will need the digital hi speed / real time serial link (RP 2 MK3 via McBSP) I designed. I do have a small number of prototypes I can provide at cost. Need to check myself and get a set ready. There will be a high speed serial link with drivers attached to both sides, one on the RP and one will plug between the MK3 and A810. Then a connection with two identical RJ45 (network...
-
What do you mean with McBSP connected -- this is not that "simple" and requires a special interface on both ends to work. See other message. Also for this you do need to enable the McBSP and a few settings via the mk3_spm_configurator.... for the connection to get a life. Oh, and do not try to do anything with that "web GUI" what is totally outdated and obsoleted by now. The Gxsm "Inet-JSON" plugin is all you need.
-
What do you mean with McBSP connected -- this is not that "simple" and requires a special interface on both ends to work. See other message. Also for this you do need to enable the McBSP and a few settings via the mk3_spm_configurator.... for the connection to get a life.
-
What do you mean with McBSP connected -- this is not that "simple" and requires a special interface on both ends to work. See other message.
-
Hi, the "analog" out option is not any more enabled and obsoleted -- was only for early testing and very limited in resolution. You will need the digital hi speed / real time serial link (RP 2 MK3 via McBSP) I designed. I do have a small number of prototypes I can provide at cost. Need to check myself and get a set ready. There will be a high speed serial link with drivers attached to both sides, one on the RP and one will plug between the MK3 and A810. Then a connection with two identical RJ45 (network...
-
Make sure the Z polarity is correct. Else you may have mechanical instabilities.
-
I am not sure, can you describe that behavior more clear? Gxsm at startup per design will not touch any DSP settings (unless you confirm a polarity change) or sure change parameters, etc. Also ADC's are never on or off. They are all 0V at power up. And the feedback is off at that point. PS: Restarting or even killing or (even restarting the PC) while the DSP stayed powered on will NOT at all effect anything what the DSP does, it will even keep scanning - -only you loose the data stream and will have...
-
Just in case: you have to restart GXSM after adjusting Z-polarity in stetting . But NOT the DSP. Gxsm will only adjust the DSP settings only once at Gxsm startup (there will be a message dialog informing/warning about it).
-
Hi! Please check you Z-polarity (not all systems are the same) settings in preferences -- i.e.if the feedback on with current setpoint settings > the current noise level is actually moving your tip "Z-Scan" towards the sample all the way (when coarse disengaged). And retracts Z all the way when Setpoint = 0 (as there is always noise it will retract) -P
-
Hi, three are various options, also do not know your data/issue to fix. Nor the purpose or intend like quick view or publication in mind. primitive line by line, plane, hi pass, ....
-
Tested myself and pushed to git now.
-
Edited/fixed the code snipped... I'll put it in GIT next.
-
Here is a quick "blind" (=untested) hack add on for to read all AICs (IN0..7) via RTQuery: Gxsm3/plug-ins/hard/sranger_mk3_hwi_spm.C /* ... "M" : AIC[1,2,3] // MK2 hack "N" : AIC[4,5,6] // MK2 hack "L" : AIC[0,6,7] // MK2 hack */ //#define MK2_SCAN_STRUCT_DEBUG gint sranger_mk2_hwi_spm::RTQuery (const gchar *property, double &val1, double &val2, double &val3){ const gint64 max_age = 50000; // 50ms const gint64 max_age_as = 250000; // 250ms static gint64 time_of_last_xyz_reading = 0; // abs time in...
-
I mean the RTQuery (HwI member function, default prototype) is only providing 3 return values. But I can easy create a modified version.... But if you are OK with 3 readings at once all is good. As you see the readings (all 8) are updated at once, and with a limited update rate (buffered) to prohibit USB/DSP overload issues if some one tries to poll arbitrary fast. Thus a consecutive query via "M,N,L" will most likely give you the last reading of all value sampled at the exact same time anyways!
-
Very good! I do not see any issues if it looks reasonable and nothing obvious swapped or so. PS: If needed: with few more hacks I could make it also return a vector of all values at once -- only this RTQuery function is not designed to do so.
-
Better use "Dig2VoltIn(...)" -- but (almost) the same... I need to test it and commit to git.
-
Better use "Dig2VoltIn(...)" -- but (almost) the same...
-
Here is a quick "blind" (=untested) hack add on for to read all AICs (IN0..7) via RTQuery: Gxsm3/plug-ins/hard/sranger_mk3_hwi_spm.C /* ... "M" : AIC[1,2,3] // MK2 hack "N" : AIC[4,5,6] // MK2 hack "L" : AIC[0,6,7] // MK2 hack */ //#define MK2_SCAN_STRUCT_DEBUG gint sranger_mk2_hwi_spm::RTQuery (const gchar *property, double &val1, double &val2, double &val3){ const gint64 max_age = 50000; // 50ms const gint64 max_age_as = 250000; // 250ms static gint64 time_of_last_xyz_reading = 0; // abs time in...
-
Here is a hack add on for to read all AICs (IN0..7) via RTQuery: Gxsm3/plug-ins/hard/sranger_mk3_hwi_spm.C /* ... "M" : AIC[1,2,3] // MK2 hack "N" : AIC[4,5,6] // MK2 hack "L" : AIC[0,6,7] // MK2 hack */ //#define MK2_SCAN_STRUCT_DEBUG gint sranger_mk2_hwi_spm::RTQuery (const gchar *property, double &val1, double &val2, double &val3){ const gint64 max_age = 50000; // 50ms const gint64 max_age_as = 250000; // 250ms static gint64 time_of_last_xyz_reading = 0; // abs time in us static gint64 time_of_last_analog_reading...
-
No you do not. This requires "signal monitoring/mapping DSP capabilities" and is not an available function for the MK2. Only the MK3 can map signals and query them via rtquery("M").... what is required here. Thus you get a dummy 0 return. Sorry. I have to double check, you may be able (or I may make it possible) to access the 4 Mixer inputs. Have to check myself as not working with the MK2 actively any more. Q: I assume you do need this in context for the full SPM control? As there is a plain RTE4MK2...
-
Correct. Please use CP=CI=0 or small to slow or freeze Z. Also (advanced) consider using the "CZ-FUZZY-LOG" Transfer mode (for the mixer channel used) together with "Level" setting: Use "Level=0" for normal operation/with Z feedback (exactly =LOG mode). But once Level is set > max current possible Z will be forced to the value set in Z-Pos/Setpoint. If Level is set to say 0.1nA Z will got the Z-Pos if this max or compliance current is NOT exceeded. (Safety mode) HOWEVER, a warning here: With usual...
-
Correct. Please use CP=CI=0 or small to slow or freeze Z. Intend for this switch is: The feedback on/off is a more drastic measure and usually there is no need to "disable" it this way. The intend for this switch is mainly for user controlled "start up" after DSP power up to enable the Z axis from been set to 0V at DSP power up (and kept until enabled) manually. -- See it as a electrical safety measure! -P
-
Hi Heuy, all gxsm's "scans" are 4 dimensional data sets. However, a simple single scan usually is degenerated and only is 2d, with t = time and v = value dimensions = 1 and t=v=0. Notes: Example: when using the "movie" scan mode, frames go into "t" dimension. If using for example a automated bias (or any parameter) range scan, bias is the "v" value dimension. And if you dare to combine both you have extra t,v dimensions. Or they can be used for any other purpose. t and v dimensions can also be swapped...
-
Hi Heuy, all gxsm's "scans" are 4 dimensional data sets. However, a simple single scan usually is degenerated and only is 2d, with t = time and v = value dimensions (and index) = 0. Notes: Example: when using the "movie" scan mode, frames go into "t" dimension. If using for example a automated bias (or any parameter) range scan, bias is the "v" value dimension. And if you dare to combine both you have extra t,v dimensions. Or they can be used for any other purpose. t and v dimensions can also be...
-
In my changes, all Python scripts are running in 1 thread. Thus, I think my suggested changes break the ability to run an action script and a user script simultaneously. I suggest to keep the thread spawning for the "action scripts" and only use one dedicated thread for the console controlled script where the "kill" may be issued. BTW, you may create a copy of the github gxsm3 repo, include you changes and create a merge request. Or alternatively if you like, I could add you (you have a github user...
-
Good. To put it this way: I never had any intention for "action scripts" doing anything long or infinite or blocking... They are intended so far for automating brief tasks or a do a set of multiple adjustments. The simplest example is putting bias and current to some defaults as you see. I have more advanced actions taking a few sec may be as they do some measurements/averaging and then do some thing related... Or transition between AFM in const height mode and STM in a safe and automatic way. The...
-
Sure, let's move on with all those reworks, not saying at all to use a sub or new interpreter now. I just have that in mind for the future -- may be. As those "action scripts" currently are blocked from running when a longer working script is in progress! The menupath->... calls are exactly what I mean with "Action Scripts". Those are never called automatically yet. Only if you press one of the "hot keys" Shift-F1",.... from a scan window. Plus You do need to define/create them first, but there are...
-
Hi Nick, sounds good! Great job! So moving all python init,... code into the same thread helped? Do you re-initialize python from scratch now every time for every script run now? I.e. variables won't "stay"? Possibly a good idea to prevent confusion. If have to look for the details, but there are two ways to fire up a script: a) via the code you see in the console and the "Exec" button or b) call of the Action-Scripts via Shift-F1...12 (while focus is on any scan window) (and some more dedication...
-
Hi Nick, sounds good! Great job! So moving all python init,... code into the same thread helped? Do you re-initialize python from scratch now every time for every script run now? I.e. variables won't "stay"? Possibly a good idea to prevent confusion. If have to look for the details, but there are two ways to fire up a script: a) via the code you see in the console and the "Exec" button or b) call of the Action-Scripts via Shift-F1...12 (while focus is on any scan window) (and some more dedication...
-
Hi Nick, sounds good! Great job! So moving all python init,... code into the same thread helped? Do you re-initialize python from scratch now every time for every script run now? I.e. variables won't "stay"? Possibly a good idea to prevent confusion. If have to look for the details, but there are two ways to fire up a script: a) via the code you see in the console and the "Exec" button or b) call of the Action-Scripts via Shift-F1...12 (while focus is on any scan window) (and some more dedication...
-
Did you had a look at this above?
-
We may have to set this PyConfig.inspect to non zero! To avoid python crashing gxsm on certain execeptions! See here: int PyRun_SimpleString(const char command) This is a simplified interface to PyRun_SimpleStringFlags() below, leaving the PyCompilerFlags argument set to NULL. int PyRun_SimpleStringFlags(const char command, PyCompilerFlags flags) Executes the Python source code from command in the main module according to the flags argument. If main does not already exist, it is created. Returns...
-
From https://docs.python.org/3/c-api/init.html?highlight=pygilstate#c.PyGILState_Check sounds to me like this should solve the issue, but yes Idid played with tha tin past myself also... will check again also with my python version. I still have a feeling there is some thing fishy some where... or I still do not get the whole picture right. At least I'd say the calls to PyErr_SetString(PyExc_KeyboardInterrupt, "Abort") and PyErr_CheckSignals() shall NOT crash/segfault the program in any situation...
-
May be those calls give some hints -- try that around it: PyGILState_STATE py_state = PyGILState_Ensure(); PyErr_SetString(PyExc_KeyboardInterrupt, "Abort"); PyErr_CheckSignals(); PyGILState_Release (py_state);
-
May be those calls give some hints -- try that around it: ~~~~ PyGILState_STATE py_state = PyGILState_Ensure(); PyErr_SetString(PyExc_KeyboardInterrupt, "Abort"); PyErr_CheckSignals(); PyGILState_Release (py_state); ~~~~~
-
May be those calls give some hints? PyGILState_STATE py_state = PyGILState_Ensure(); PyGILState_Release (py_state);
-
Agree, it would be good to know what other python "idle" and GIL release scenarios could come up! About "killing" it, I certainly could kill blindly the whole thread python runs within which I do have control over.... even that would be anything but nice to it. But normally I do not kill the python session -- you may noticed, if you rerun a script, active variables stay alive! I wonder if there is any way externally to check if the GIL is present? Or make it been regained to manipulate. That would...
-
That's an excellent analysis! Thanks. So in short, when using python level time.sleep() the GIL is released and the kbd interrupt goes no where and causing an exception with the PendingCall().... :( While when gxsm.sleep() is halting the the python thread via a "usleep()" (while still releasing the CPU load!!) at the gxsm python thread level while holding on the GIL and we can in issue the kbd interrupt to it! Q: Thus is there any issue not using the gxsm.sleep() at all times we may like to interrupt...
-
That's an excellent analysis! Thanks. So in short, when using python level time.sleep() the GIL is released and the kbd interrupt goes no where and causing an exception with the PendingCall().... :( While when gxsm.sleep() is halting the the python thread via a "usleep()" (while still releasing the CPU load!!) at the gxsm python thread level while holding on the GIL and we can in issue the kbd interrupt to it! Q: Thus is there any issue not using the gxsm.sleep() at all times we may like to interrupt...
-
As I said, avoid (electronically) saturating the input(s) above +/-10V. I know if you exceed 10.00 V the DSP may see false readings or even 0... and the unobvious obvious may happen... -- this is out of specification.
-
Interesting... so this actually may be a python lib issue! I am developing on Debian 12 and Testing. (Python 3.11) However, did you tried my script above without the try...except...? As your script for odd reason also crashes when I click the abort button while mine does not! gxsm.set('script-control','1') i=0 while int(gxsm.get('script-control')) > 0: print (i) i=i+1 gxsm.sleep(10) Output: >>> Executing parsed script >>> 0 1 2 3 << here I pressed the abort button *** SCRIPT INTERRUPT REQUESTED:...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread -- possibly. What you all list as of possible options is about all I recall been testing myself long ago... See also: https://docs.python.org/3/c-api/exceptions.html I assume you located "py_gxsm_consoile::kill", line 3717 in pyremote.cpp (OK that Gxsm4, but still the same issue, see also some testing code fractions...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread -- possibly. What you all list as of possible options is about all I recall been testing myself long ago... See also: https://docs.python.org/3/c-api/exceptions.html I assume you located "py_gxsm_consoile::kill", line 3717 in pyremote.cpp (OK that Gxsm4, but still the same issue, see also some testing code fractions...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread -- possibly. What you all list as of possible options is about all I recall been testing myself long ago... See also: https://docs.python.org/3/c-api/exceptions.html I assume you located "py_gxsm_consoile::kill", line 3717 in pyremote.cpp (OK that Gxsm4, but still the same issue, see also some testing code fractions...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread -- possibly. What you all list as of possible options is about all I recall been testing myself long ago... See also: https://docs.python.org/3/c-api/exceptions.html And honestly at some point I gave up on it.... and started writing non blocking scripts what is smarter anyways and keep monitoring any of the script...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread -- possibly. What you all list as of possible options is about all I recall been testing myself long ago... And honestly at some point I gave up on it.... and started writing non blocking scripts what is smarter anyways and keep monitoring any of the script control entry values to exit correctly if requested! It...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread -- possibly. But honestly at some point I gave up on it.... and started writing non blocking scripts what is smarter anyways and keep monitoring any of the script control entry values to exit correctly if requested! It kind of worked in the past before been fully threaded -- but even then it had advert side effects/not...
-
Script forced exit: you tell me, no clue... As you see there are many attempts (if you look up the history) getting this sig int passed to python. Yes some thing is related to the thread possible, but honestly at some point I gave up on it.... and started writing non blocking scripts what is smarter anyways and keep monitoring any of the script control entry values to exit correctly if requested! I kind of worked in the past before been fully threaded -- but even then it had advert side effects/not...
-
Nick, you brought up an very old -- going back to the roots of gxsm -- quest here also: The pro and cons of storing channels in separated files! I am kind of glad you did. And wonder if any one else has some input here? Facts: * all "meta data" (minor from size for larger scans) besides the Z unit and 4D data set itself is a identical copy. Few exceptions are potentially attached user&probe data sets only been attached to the "master" scan, usually the 1st in line or Topo. * File name convention,...
-
Nick, you brought up an very old - -going back to the roots of gxsm -- quest here also: The pro and cons of storing channels in separated files! I am kind of glad you did. And wonder if any one else has some input here? Facts: * all "meta data" (minor from size for larger scans) besides the Z unit and 4D data set itself is a identical copy. Few exceptions are potentially attached user&probe data sets only been attached to the "master" scan, usually the 1st in line or Topo. * File name convention,...
-
Sure it is! I do have many scripts for various special data analysis tasks! Many even run right inside the Gxsm python console ;) And BTW, you can even swap units of scan data after the fact... but it's still a bit experimental -- using the build in python methods -- and sure if you wish, you do apply any scaling. See also: gxsm.set_scan_unit : Set Scan X,Y,Z,L Dim Unit: gxsm.set_scan_unit (ch,'X|Y|Z|L|T','UnitId string','Label string') gxsm.set_scan_lookup : Set Scan Lookup for Dim: gxsm.set_scan_lookup...
-
Hi, you must keep the "FloatField" type, as this is a hardware /HwI design requirement. you can change name and label to anything, may avoid special characters. Like the default exanple shown for ADC0-ITunnel. (No need for ADC0, but I recommend to keep it and add info like "ADC3-TestSignal" so you later know where is was originated from, no need so!) Label is used for the profile axis label (X or Y as used). Name is the long name is used also for the channel's file name post fix. You can pick a unit...
-
Hi, you must keep the "FloatField" type, as this is a hardware /HwI design requirement. you can change name and label to anything, may avoid special characters. Like the default exanple shown for ADC0-ITunnel. (No need for ADC0, but I recommend to keep it and add info like "ADC3-TestSignal" so you later know where is was originated from, no need so!) Label is used for the profile axis label (X or Y as used). Name is the long name is used also for the channel's file name post fix. You can pick a unit...
-
Thanks for positing! Nice util. PS: a few basic assignments are possible via preferences so. See screenshot. It it helps, I could make custom per channel conversion factors to a core set of units available in future. For sure you can since ever "replace" the default names ADC1,2,... with any meaning full name there. And the file names will also have that post-fix name. -P
-
What is your l-V gain? Or what voltage dose 1nA correspond to? if 1.5V => 10V (or close to with noise) this may be your issue: As a safety note, your input voltage shall never (not even spikes) exceed +/-10.00V (better stay < 9.5V for safety). I say this as saturating the analog input (beyond +10 or below -10V) can cause undefined reading/behaviors and may even damage the inputs. This said, assuming you get a 11V peak the DSP may NOT see the saturated 10V max but anything... can obviously that will...
-
I think this is off topic GXSM. I do read "gnome shell crashed". Sorry not gxsm related, may be a bug in the library as Gxsm makes extensive use of it. Any other messages or warning/errors from Gxsm popping up at all? FYI: Gnome-Shell of the desktop/window manager.
-
Any details about the error? This does NOT look like Gxsm related error. More some thing kernel level or USB low level. If at all Gxsm related. I can say with very very high confidence that our usb_sranger_mk23 kernel module is very mature and proven to work solid. There has also been no change or update in many years now. Over a year machine uptime proven!! And this on all known machines and platforms from intel, amd to arm! What execute exactly? " initially click execute" If any USB related issues/error...
-
The approach ramp slope/speed depends on your feedback (CI) and (current) setpoint setting and yes if your current input noise level is huge or varies that will accordingly effect the ramp speed. "Usually" I would expects in a non tunnel range only marginal noise, but I do not know your system and electronics. For a "normal" room temp/air STM on metal surfaces you should be able to scan and tunnel neatly around 0.1 ... 2.0 nA typically. You disengages noise level shall by way less than << 0.1nA....
-
Sorry, I have no good idea what is causing those issues. Not normal. A few hints: If you have any uncommon "local" (language) -- try to stay with "C" or English. No not start gxsm fron any folder not within the current users's home folder with all default permission. Never start gxsm as root -- this could create a Ev-log as root for example and next time you start as non root you will get write access issues. Make sure gxsm is installed properly and ONLY in one way -- no multiple version or installs...
-
From what path or location did you started it? Looks there is no write permission. Get Outlook for Androidhttps://aka.ms/AAb9ysg From: discussion@gxsm.p.re.sourceforge.net discussion@gxsm.p.re.sourceforge.net on behalf of bugrahan guner bugrahanguner@users.sourceforge.net Sent: Tuesday, September 26, 2023 10:25:44 PM To: [gxsm:discussion] 40919@discussion.gxsm.p.re.sourceforge.net Subject: [gxsm:discussion] Error encountered after starting GXSM Hi Percy, I posted the question accidentally into another...
-
You calculate if for your system 16bits (with little extra oversampling, but ,ainly for smoothing out edges, not really precision) are doing the job.
-
Hi, can you please briefly indicate/assure your exact setup: MK2 or Mk3 (Think was Mk2) Offset mode intended to use: a) internal/digital offset adding to scan signals. Common analog signals for Scan with Offset on XYZ. (Limited to 16 bits +"2 1/2" bits oversampling/smoothing resolution). b) external (analog) offset adding via OUT0 and OU1 (2) (so you can have different gains and a larger area of access while maintaining high resolution for scanning. This requires dedicated analog pre-amplifiers in/before...
-
Uninstall all kernel header packages eventually there. Make sure to select the "default" kernel package only -- no old or pinned versions, no version number, just the "type" (I assume amd64, but even on Arm machines my modules do work ;)) and or so... update/make sure it is the latest. Then install the kernel header package. That should do it. This is the only set one you should have: (as of Debian 12) Current kernel package: "linux-image-amd64" and matching kernel-headers package: "linux-headers-amd64"...
-
Uninstall all kernel header packages eventually there. Make sure to select the "default" kernel package only -- no old or pinned versions, no version number, just the "type" and or so... update/make sure it is the latest. Then install the kernel header package. That should do it. This is the only set one you should have: (as of Debian 12) Current kernel package: "linux-image-amd64" and matching kernel-headers package: "linux-headers-amd64" And remember after updates a reboot is required and rebuild/install...
-
Hi, I wonder what is "possibly mixed up" here with kernel sources/installed packages. Speaking of a clean system, Debian or Ubuntu -- simply have the kernel-header package up-to-date with the running kernel. What shoudl be the case after updating and rebooting. NEVER select any explicit kernel version!!! { unless dedicated hardware make that necessary, unlikely (that may be tricky and you are on you own building a kernel form scratch -- you really do not want to do this today, I even get scared today,...
1 >