Post by bmatlove on Re-create CoE online window in Visu ^{CODESYS Forge talk (Post)}
Hello, I am trying to re-create the CoE online window within a Visu for purposes of checking and interacting with device parameters occasionally for maintenance purposes. To minimize reliance on using the IDE for misc interactions. Ideally, I would like to exactly mimic the CoE online window within the IDE. I assume that it uses a few tricks to minimize overhead, such as only polling registers which are within the visual scope, for example. I'm afraid it might be too much to ask for a pre-built drop-in Visu module to exist for this task. Do I have any options aside from manually re-creating all the logic to do this, relying heavily on SdoInfo function blocks? Is there a way for me to reference the device ESI file to shortcut the process of determining how many registers there are and getting the address for each?
Last updated: 2025-12-27

Post by manuknecht on XY Chart causes extremely long cycle in Visu Task ^{CODESYS Forge talk (Post)}
I am using the visualization element "Cartesian XY Chart" in a visualization template within a library to display recorded position values. I managed to configure the chart as desired and it displays the values like it should. A problem occurs when clicking on the chart to move the cursor, zoom or pan. Upon clicking the visualization element for the first time, the task in which the visualization is run spikes to a cycle time of around 5'000 to 15'000 ms (the configured cycle time is 100 ms). Sometimes this even completely crashes the web-visualization. After this very long cycle, the visu task normalizes again and the chart can be interacted with as usual. Another spike in cycle time occurs once the visualization is switched to a frame that does not contain the chart. I did reduce the data array to only a few values and tried changing the update behaviour of the chart, but it did not change the problems I am facing. Did anyone experience a similar issue or even has a solution for this?
Last updated: 2026-06-16

Post by konstantin on Lenze i950 Servo and WAGO PFC200 ^{CODESYS Forge talk (Post)}
Hello, In our workshop we have a Lenze i950 servo drive that has an EtherCAT bus and CiA402 compatibility. The WAGO is PFC200 750-8212 and as per documentation it has built-in license for SoftMotion Light functionality. It is not capable of supporting the full SoftMotion functionality like gearing and synchronizing. I added the EtherCAT esi files from Lenze and the EtherCAT master does see the drive, then I added a SoftMotion Light Axis object. The problem is that I can only energize the axis, I cannot execute any other commands. I set the positive and negative torque limits from the PDOs to 200% and when I try to move the axis by force it does not move back to its original position. I use the SML function blocks from the PLCOpen library and the error from the function blocks is that the op mode cannot be changed. Then I tried using a standard SoftMotion axis object and everything works fine, except the MC_Home function block. The axis moves back to its position when forced out of it. I follow this guide from Lenze - Lenze i950 CODESYS V3 What could cause this issue with the SoftMotion Light axis object? Best Regards Konstantin Kolev
Last updated: 2026-02-11

Post by yannickasselin on Lenze i950 Servo and WAGO PFC200 ^{CODESYS Forge talk (Post)}
In order to use SoftMotion Light, a servo drive needs to support CiA 402 operating modes: Profile Position (PP), Profile Velocity (PV) and Homing Mode (HM). https://content.helpme-codesys.com/en/CODESYS%20SoftMotion%20Light/_sm_light_requirements.html According to the Lenze i950 documentation you linked in your post, none of these modes are supported. Not even Homing Mode which explains why MC_Home does not work even with the full SoftMotion. The full SoftMotion uses Cyclic Synchronous Position Mode (CSP) but still uses Homing Mode (HM) when using MC_Home. You can monitor PDO 0x6060 (Mode of Operation). When using SoftMotion Light you will see it set to 1 (PP) for a MC_MoveAbsolute_SML and MC_MoveRelative_SML and to 3 (PV) for MC_MoveVelocity_SML and to 6 (HM) for MC_Home_SML. When using the SoftMotion instructions, it will pretty much always be set to 8 (CSP) except when using MC_Home, then it will be set to 6 (HM).
Last updated: 2026-02-13

Post by konstantin on CAM Motion Recover after STO ^{CODESYS Forge talk (Post)}
Hello Mr. Seidel, both the master and the slave axis are modulo. Basically the situation is as follows: One master cycle is 100 units. Slave's modulo length is 2000 units. In the CAM table I defined 10 master cycles on X axis - 1000 units and on the Y axis is the slave - 2000 units. Then depending on the master position 0 to 1000 units the slave's position is plotted on the Y axis - 0 to 2000 units. For example, while the master is executing the 3rd cycle i.e. above 200 units, I trigger the STO function of the slave and it stops. The master executed his cycle but the slave was stopped during the execution. How can I set the slave to continue the CAM sequence from the point where it stopped i.e. third cycle of the master? When I reset the slave error and sync again, everyrhing starts from master's 0 position. Best Regards Konstantin Kolev
Last updated: 2026-02-13

Post by fajean on Issue with library updates ^{CODESYS Forge talk (Post)}
We make heavy use of internally developed libraries, which are often nested on multiple levels. I routinely work simultaneously on several libraries along with a project that use them. I make changes to a library and immediately perform a download (online or otherwise) in the project. For the longest time, we had zero issue with this. But starting a good while back, with 3.5 SP19 I think, CODESYS started not noticing library updates in other instances most of the time. When the update is missed, the project sees no need to update the code, and any download ignores the new code. A "Clean all" does not solve the issue. But if I close and then re-open the project, then the new version is noticed and the download works. The other solution is saving the library to the registry multiple times (say, 20-30 times, less if lucky), and then at some point the project notices the changed library. I am now on SP21, and it is still not fixed. This is driving me mad, any tasks probably takes me double the time or more due to this problem. My colleague has the same problem on a separate computer. Is this a known issue? Is there a setting somewhere that changes how updates are checked? Thanks for any help!
Last updated: 2026-02-13

Post by suyash on Device User Logon ^{CODESYS Forge talk (Post)}
Hello. Basic Info : I am using Codesys Control For Raspberry Pi SL. Version 4.4.0.0 Engineering version : CODESYS V3.5 SP19 Patch 1 (Its a old application) I am facing a difficulty with the "Device User Logon". The window (screenshot attached) pops up during login. When i enter the credentials (which i feel are correct), the windows pops up again no matter how many times I enter. To ascertain, I tried entering the incorrect credentials during which its clearly specifying that i have entered them incorrectly (which is why I am quite certain about the credentials being correct). I even tried going into Communication Settings > Device(drop down) > Change Runtime secuirity policy . However it again asks me for credentials which keep popping up again and again. I am thinking of upgrading the runtime to the latest version of 4.20 Will it cause any issues with the paid license? Please suggest me a solution. BR, Suyash
Last updated: 2026-05-13

Post by hhermsen on Codesys v3.5.22.10 - Internal error:System.AggregateException...Value cannot be null...Parameter name: expLValue ^{CODESYS Forge talk (Post)}
I pinpointed down 4 possible causes for this compiler error: Bare/empty statements (;) standing alone inside an IF/CASE branch — just removed two instances. The optimizer that tries to inline/replace expression statements may choke when a statement has no actual assignment to extract. A VAR_IN_OUT parameter used directly as a FOR loop's own control variable (FOR X := X TO Y DO) General deep-nesting complexity — every one of my decode routines has the same shape (IF busy THEN ... FOR loop containing nested CASE/IF chains ... END_IF), and the crash trace consistently shows for loop → sequence → ifst → sequence → ifst → sequence regardless of the code change. This raises the possibility that it's not one specific statement but a structural depth/complexity limit in SP22.x's compiler optimizer — which would mean no small code change fixes the issue. Only restructuring to reduce nesting depth, or downgrading the compiler will. Something entirely outside my library files — which cannot not ruled out since the exact stack trace was triggered by a Visualization object, in a totally different codebase. This means the same crash signature can come from unrelated causes. IMHO cause 3 seems the most probable one..
Last updated: 2026-07-11

Post by imdatatas on CFC editor: enable inline value entry on unconnected input pins (as in Ladder Logic editor) ^{CODESYS Forge talk (Post)}
In the CODESYS Ladder Logic editor, there's a really practical feature available while logged in (online) to the PLC: for an existing function block (e.g. TON or MC_MoveAbsolute), double-clicking on an unconnected input pin directly opens a value entry window right in the editor — with no need to assign any variable — and you can type a value directly. Likewise, if the input pin is of an ENUM or STRUCT type, it lets you pick a value of that type straight from a dropdown list. It also doesn't require assigning a variable or adding it to a watch list, etc. Unfortunately, in the CFC editor you can't write values directly to those unconnected input pins the way you can in Ladder — CFC only supports monitoring/display there. I wish this existing Ladder editor feature were also available in CFC — it would make the CFC editor far more usable. I think this would be a relatively small addition to the CFC editor, but it would make a real difference in daily commissioning and debugging work. It would be great to hear the development team's thoughts on whether this could be considered for a future release. Thanks in advance for reading!
Last updated: 2026-09-10

Post by darwin03 on Mismatch between coordinate monitor and actual motion using SMC_GroupJog2, MC_MoveLinearAbsolute (6DOF) ^{CODESYS Forge talk (Post)}
Hi Georg, Thank you very much for your help and for taking the time to look into my issue! After reversing the directions of A2 and A3, the robot now moves correctly according to the coordinate system. That solved the main issue with the joint rotation directions. However, I still have two issues: When I command the robot to move along the X, Y, and Z axes, the movement seems slightly off, and the trajectory appears somewhat diagonal instead of following a straight line along the commanded axis. The A6 axis also rotates during the movement, whereas I would expect it to maintain its orientation and remain unchanged. Do you have any suggestions on how I could correct the Cartesian linear movement and keep A6 fixed during the motion? Thank you again for your support! Best regards, Darwin Video link: https://drive.google.com/file/d/1tnokDYoNUEDy--eQtX_JZavMVSn-a9Le/view?usp=sharing
Last updated: 2026-10-15

Post by kevinrn on Github Actions CI/CD tasks - development topic ^{CODESYS Forge talk (Post)}
Hello community, I just want to inform you about our plans and the current status of automating the build process for CODESYS libraries. So it might be helpful for some people who are in the same situation. Background story: I am a software engineer at powerIO GmbH, normally I use high level languages for the products we develop and offer. Sometimes I work with my team on CODESYS libraries, which can also be found in the CODESYS store. Most of the time I spend with them on architecture tasks and tasks that are not directly related to our products. Every time I see the manual processes for releasing a library etc., I am very surprised about this time intensive process. Current situation: I know there has been a scripting interface for years. CODESYS Git was released a few years ago, but the scripting interface was only released a few months (weeks?) ago. We developed a pseudo-automated solution a few months ago, but it was all very hacky. So now there is a better Git implementation and also more modularization and installation options. For example, the CODESYS installer has a full CLI, which makes it very convenient to install CODESYS installations headless. Plans: I think it's the right time to develop a better automation solution. Most of our software projects are hosted on Github and we use Github actions very intensively for other software projects. Therefore, we have decided to use Github for our CODESYS library projects as well. Following tasks should be implemented in automation for CI/CD tasks: - Automated setup of CODESYS installation (Already released: https://github.com/marketplace/actions/setup-codesys-installation) - Checkout CODESYS libraries - Execute tests scripts - Sign CODESYS libraries - Extract Library documentation - Create CODESYS package - Sign CODESYS package Side Note: - This is mainly a side project, and we do not provide support for the CI/CD part. - This topic is here to help and also to get help from other software engineers. - This project and idea is not prioritized as I am currently the only one working on it and I hope it will help us and also others and increase productivity. - We are putting this out there as an organization, but priorities can change quickly and the CI/CD tasks I am developing are mostly done in my spare time. So please don't expect this to happen in a short time or even be fully completed. I would be very happy if some ppl will join the development process and might be open to discuss some technical details for this. I hope this topic is not to off-topic, but I think the CI/CD part is very important today and it can increase a lot of quality and push productivity. Thanks :)
Last updated: 2024-03-28

Post by scoob on ModbusFB - Slow Response Time ^{CODESYS Forge talk (Post)}
Hello, I have been trying to use the ModbusFB functions so I can put some code into libraries, but it seems to be very slow for me. I have a Modbus device with 100ms registers. I previously setup 10 channels in the 'traditional' Modbus Slave with channels and mappings - and set a cyclic trigger at 100ms - this worked fine. I then tried the ModbusFB example, and setup reading the same 10 blocks of modbus addresses, copying the example and putting all of the requests into an array and triggering the requests sequentially. I timed how long the requests are taking to get round to each one, and it is around 1s 450ms. How do I speed this up to match the cyclic time? IF NOT(init) THEN init := TRUE; // Set the required IP address: ipAddress[0] := 192; ipAddress[1] := 168; ipAddress[2] := 1; ipAddress[3] := 10; // Pass the required IP address to the clinet FB: client_NetworkSwitch.aIPaddr := ipAddress; client_NetworkSwitch.udiLogOptions := (ModbusFB.LoggingOptions.ClientConnectDisconnect OR ModbusFB.LoggingOptions.ClientReceivedValidReplies); // Try to connect the client client_NetworkSwitch(xConnect:=TRUE); // Configure all the channels to read connecting them to the client: portStatus_Request(rClient := client_NetworkSwitch, uiStartItem := 4096, uiQuantity := 32, pData := ADR(portStatus), udiReplyTimeout := udiReplyTimeout); portSpeed_Request(rClient := client_NetworkSwitch, uiStartItem := 4352, uiQuantity := 32, pData := ADR(portSpeed)); flowControl_Request(rClient := client_NetworkSwitch, uiStartItem := 4608, uiQuantity := 32, pData := ADR(flowControl)); linkUpCounter_Request(rClient := client_NetworkSwitch, uiStartItem := 5888, uiQuantity := 32, pData := ADR(linkUpCounter)); txPacketCounter1_Request(rClient := client_NetworkSwitch, uiStartItem := 8192, uiQuantity := 100, pData := ADR(txPacketCounter1)); txPacketCounter2_Request(rClient := client_NetworkSwitch, uiStartItem := 8292, uiQuantity := 28, pData := ADR(txPacketCounter2)); rxPacketCounter1_Request(rClient := client_NetworkSwitch, uiStartItem := 8448, uiQuantity := 100, pData := ADR(rxPacketCounter1)); rxPacketCounter2_Request(rClient := client_NetworkSwitch, uiStartItem := 8548, uiQuantity := 28, pData := ADR(rxPacketCounter2)); txErrors_Request(rClient := client_NetworkSwitch, uiStartItem := 8704, uiQuantity := 64, pData := ADR(txErrors)); rxErrors_Request(rClient := client_NetworkSwitch, uiStartItem := 8960, uiQuantity := 64, pData := ADR(rxErrors)); // Trigger all client requests initially FOR clientRequestsCnt := 0 TO (SIZEOF(clientRequests)/SIZEOF(clientRequests[0]))-1 DO pClientRequest := clientRequests[clientRequestsCnt]; pClientRequest^.xExecute := TRUE; END_FOR // Prepare sequential trigger / control of client requests. clientRequestsCnt := 0; pClientRequest := clientRequests[clientRequestsCnt]; END_IF // Call the client to do request processing: client_NetworkSwitch(); // Now we trigger client request sequentially ... IF NOT pClientRequest^.xExecute AND NOT pClientRequest^.xDone AND run AND client_NetworkSwitch.xConnected THEN pClientRequest^.xExecute := TRUE; END_IF // .. and check result/error IF pClientRequest^.xExecute AND run AND client_NetworkSwitch.xConnected THEN IF pClientRequest^.xDone THEN // Prepare next trigger of client request (a rising edge of xExecute) pClientRequest^.xExecute := FALSE; IF clientRequestsCnt < SIZEOF(clientRequests)/SIZEOF(clientRequests[0])-1 THEN // next client request clientRequestsCnt := clientRequestsCnt + 1; ELSE clientRequestsIterationCounter := clientRequestsIterationCounter + 1; clientRequestsCnt := 0; END_IF pClientRequest := clientRequests[clientRequestsCnt]; END_IF END_IF I did try a semi-coded way using the IoDrvModbusTCP library, and setting the slave com settings, then 10 commands and 10 requests, then using a TP on xDone as a pause, before triggering another request - this is time the delay is around 120ms - so the device is fine with the speed, just something I am doing wrong in the ModbusFB method I am sure.
Last updated: 2024-04-26

Post by dwpessoa on CNC Jumps G20 - SMC_NCInterpreter and long time to process ^{CODESYS Forge talk (Post)}
I am studying and developing a Softmotion+CNC system for a machine that executes multiple pieces. The G code program is written by the machine operator and each cycle execute 1 piece. The programs are large, exceeding 1000 lines and using up to 8 axes (X, Y, Z, A, B, C, P and Q). The machine needs to run cyclically, executing N pieces (selected by the Operator)... so I tested it using Looping and counters (G36 G37 and G20) and it worked, but it takes a long time to process, and the more pieces I need, the longer the processing time and this is totally impracticable. I found this solution which was very good, and for a few cycles it works well, but for 99999 pieces of a program with 1000 lines, it doesn't work very well... Another solution I tested is to maintain the interpolator with an automatic restart, that is, I load the program without looping (without G20) and give it another start as soon as it finishes. This partially resolved it, but there is still a delay in processing SMC_NCInterpreter in each restart :(. Another solution I thought of is to manually create the SMC_GEOINFO structure and then reuse it, avoinding the Interpreter, but reading the documentation and checking the structure filled by standard blocks, I noticed that there doesn't seem to be a "JUMP" function in the structure! In other words, the SMC_NCInterpreter actually keeps copying and copying the program section for each jump (G20)... If I repeat a 10-line program 1000 times, I will have a structure with more than 10000 lines... possibly this is the cause of take so long to process. Has anyone ever had a problem like this? I believe the same thing happens with typical applications with manipulator robots using Codesys in continuous cycles, and I would like to know if there is any solution, or even if I am misinterpreting the G20 question in SMC_GEOINFO. Thanks!
Last updated: 2023-09-20

Post by bbm1995 on Camera RTSP Feed ^{CODESYS Forge talk (Post)}
Hi dgrard, I had the same issue for a long time, but on a WAGO webvisu. Now I'm glad that I can share my solution. I don't know if you are trying to use the webvisu or the target visu, but this works on webvisu: Get "go2rtc" and set it up according to the documentation. I'm running the Windows binary. Here's my example of the go2rtc.yaml config file (contains working example streams): api: listen: ":1984" # default ":1984", HTTP API port ("" - disabled) origin: "*" # default "", allow CORS requests (only * supported) static_dir: "www" # default "", folder for static files (custom web interface) tls_listen: ":443" # default "", enable HTTPS server tls_cert: "./SSL/fullchain.pem" tls_key: "./SSL/privatekey.pem" streams: # Streams with multiple links will fall back on the next link. ABUS TVIP48511: - rtsp://<username>:<password>@<hostname>:<port>/ch1/main - rtsp://<username>:<password>@<hostname>:<port>/ch1/sub # Diagnostic connections Kirchhoff Institute for Physics - Germany: http://pendelcam.kip.uni-heidelberg.de/mjpg/video.mjpg Blanton Bottling, Kentucky - USA: http://camera.buffalotrace.com/mjpg/video.mjpg Tokyo - Japan: http://61.211.241.239/nphMotionJpeg?Resolution=320x240&Quality=Standard Tampere Hacklab - Finland: http://tamperehacklab.tunk.org:38001/nphMotionJpeg?Resolution=640x480&Quality=Clarity Soltorget Pajala - Sweden: http://195.196.36.242/mjpg/video.mjpg Kaiskuru Skistadion - Norway: http://77.222.181.11:8080/mjpg/video.mjpg webrtc: listen: ":8555" #ice_servers: # - urls: [ "stun:localhost:3478" ] # username: "" # credential: "" Access the webinterface of go2rtc and get the link of your stream. Use the link as for your browser frame in the visualization. Depending on your device, you need to be able to access the file /etc/lighttpd/lighttpd.conf or /etc/lighttpd/webvisu.conf and change the contents similar to this one, otherwise you'll get CORS and CSP errors in the browser and you won't be able to view the stream on the webvisu: # Webvisu specific settings $SERVER["socket"] =~ port_webvisu_used_any { url.redirect += ( "^/webvisu/?$" => "/webvisu/webvisu.htm" ) $HTTP["url"] =~ "^/webvisu/?" { var.response_header_policy_webvisu = ( # CSP for WebVisu, allowing inline sources. "Content-Security-Policy" => "default-src 'self' 'unsafe-inline'; media-src *; frame-src *", # CORS for WebVisu, allowing any origin to access. "Access-Control-Allow-Origin" => "*", # Tell older browsers that this page can only be displayed if all ancestor # frames are same origin to the page itself. "X-Frame-Options" => "SAMEORIGIN" ) # Response header policy for WebVisu setenv.set-response-header = var.response_header_policy_webvisu setenv.set-response-header += var.response_header_policy_common }
Last updated: 2023-11-14

Post by solidlogicguy on Little endian to Float from Modbus RTU ^{CODESYS Forge talk (Post)}
Hello, I got a device from which I require to read values from I am using a WAGO PLC 750-8212 and I am communicating through Modbus Master FUNCTION BLOCK from library WagoAppPLCModbus in Codesys 3.5 to this device. I already receive data from the device that is a CVM to monitor voltage from a fuel cell. The technical support of the company that makes these devices says that the data is sent in little endian form. And I want to convert it to a float value. The tech support sent me the next instructions of how to do it but I am new using codesys, so any advice or help I will really appreciate so much. Message from tech support: The process is complicated, better to do it with already implemented library in the language/program you use. Basically the process should be next: To convert the two Modbus registers containing parts of a 32-bit float in little-endian byte order to a floating-point number using mathematical operations, you first need to combine the two 16-bit integers (assuming reg1 is the lower word and reg2 is the higher word) and then interpret the result according to the IEEE 754 standard. Given: - Register 192 (reg1) = 4096 - Register 193 (reg2) = 14884 Step 1: Combine the two registers. Since we are dealing with little-endian byte order, reg2 is the high word, and reg1 is the low word: combined = reg2 * 2^16 + reg1 combined = 14884 * 65536 + 4096 combined = 975175680 + 4096 combined = 975179776 Step 2: Convert the combined value to binary: combined_binary = '1110101101011100000000000000000' Step 3: Split the binary into IEEE 754 components: Sign bit (1 bit): 0 Exponent (8 bits): 11101011 Mantissa (23 bits): 01011100000000000000000 Step 4: Convert the binary exponent to decimal and subtract the bias (127 for 32-bit floats): exponent = int('11101011', 2) - 127 exponent = 235 - 127 exponent = 108 Step 5: Calculate the mantissa as a fraction: The mantissa in IEEE 754 format is the fractional part after the leading 1 (which is implicit). Therefore, we need to convert the binary mantissa to decimal and add the implicit leading 1: mantissa_fractional = 1 + int('01011100000000000000000', 2) / 2^23 mantissa_fractional = 1 + 18688 / 8388608 mantissa_fractional = 1 + 0.002227783203125 mantissa_fractional ≈ 1.002227783203125 Step 6: Combine the sign, exponent, and mantissa to get the float value: float_value = (-1)^0 * mantissa_fractional * 2^exponent float_value = 1 * 1.002227783203125 * 2^108 Because the exponent is quite large, the resulting float value is a very large number.
Last updated: 2023-12-15

Post by risele on dynamic modbus devices ^{CODESYS Forge talk (Post)}
Have the same issue: I'm making a test setup for modbus devices testing and calibration. One is run for every possible address, and it starts with minimal request: if the device is found, it continue to work with it, otherwise it's set to Disabled temporary. Those state machines are running under FOR cycle, and disabled are just skipped. Every instance have its own timer for not running too frequently and a start delay, so the actual requests are separate in time. I also check not for a single error, but have some timer to allow one-two timeouts. The small program disables and re-enables devices temporary (manually, since i could not iterate over MB devices). VAR Reset:ARRAY[1..GVL_MB.MD_COUNT] OF R_TRIG; i:INT; END_VAR {IF hasconstantvalue(GVL_MB.MD_COUNT, 1, >= )} i:=1; MD_1.Enable:=GVL_MB.MDPortsEnable[i]; Reset[i](CLK:=GVL_MB.MDPortsEnable[i]); MD_1.xReset:=Reset[i].Q; {END_IF} {IF hasconstantvalue(GVL_MB.MD_COUNT, 2, >= )} i:=2; MD_2.Enable:=GVL_MB.MDPortsEnable[i]; Reset[i](CLK:=GVL_MB.MDPortsEnable[i]); MD_2.xReset:=Reset[i].Q; {END_IF} There is also program to re-enable devices periodically. METHOD ScanBus VAR_INPUT END_VAR VAR j:INT; (*block auto scanning if calibration is running for any device since it's data-reach process that should not be interrupted with modbus timeouts *) CalibrationBusy:BOOL:=FALSE; END_VAR CalibrationBusy:=FALSE; FOR j:=1 TO GVL_MB.MD_COUNT DO CalibrationBusy:=CalibrationBusy OR MD_Calibrate.AutoTesters[j].xBusy; END_FOR; IF CalibrationBusy THEN RETURN; END_IF tBusScanSchedule(IN:=(NOT tBusScanSchedule.Q), PT:=tBusScanPeriod); IF (NOT tBusScanDelayer.Q) AND (tBusScanSchedule.Q OR (NOT xFirstScanPerforemed)) THEN xBusScanBusy:=TRUE; MissingCounter:=0; xFirstScanPerforemed:=TRUE; FOR j:=1 TO GVL_MB.MD_COUNT DO (* MD_Calibrate is main calibration program. Identifiers - is array of State machines mentioned above Phase_Faulted is a state of Identifiers where I'm sure device is missing on the bus *) IF MD_Calibrate.Identifiers[j].state=eStateMDIdentifier.Phase_Faulted THEN MD_Calibrate.Identifiers[j].Rescan(); MissingCounter:=MissingCounter+1; END_IF END_FOR END_IF tBusScanPauser(IN:=tBusScanDelayer.Q AND (MissingCounter>0),PT:=tBusScanDelay*MissingCounter); xBusScanBusy:=tBusScanPauser.Q; xFirstScanPerforemed:=xFirstScanPerforemed OR tBusScanPauser.ET>=tBusScanPauser.PT; Also the GUI have buttons for each address possible that re-enables the device if required.
Last updated: 2026-02-13

Post by alexgooi on Modbus TCP interface get stuck ^{CODESYS Forge talk (Post)}
Dear Forum, I'm currently experimenting with Modbus interfaces in combination with a virtual controller. I have the following setup (Modbus_Setup.png): I have an IPC with Debian installed on it (Debian 13). On this Debian machine I have installed a docker server and on this docker server a Codesys runtime container is deployed (Codesys_image.png). The Modbus scanner makes connection with and IO coupler and this works well. But when I physically disconnect the ethernet connection between the IPC and the IO coupler the Modbus interface gets stuck in a state where it gets stuck in the busy state (Modbus_Stuck.png). When I plug the cable back in, the IO coupler isn't communicating until the xBusy variable is manually set to False, after this action is performed the communication goes back to normal. I have tried to set the container network to host and bridge (Network.png), but both version gives the same result (stuck scanner). I would expect that the Modbus interface is generating a timeout after 1 sec, because this is how it is setup (Modbus_settings.png), but this is not what happens. I have thought to write a piece of code that writes xBusy to false when it is stuck but because it is an output I'm not able to write to it. I have tried this with another PC (different CPU/motherboard) but this gives the same outcome. I tried updating to the latest version of Virtual control (4.20) but this gives the same result. I also tried to do the same thing with an ethernet IP scanner, and this gives an error after the timeout is expired (so this works well). I have tried to connect to the IO coupler using the Modbus_FB lib (Modbus_FB.png) the FB's are giving the correct errors and behaving as expected. At last I have tried to use a Scanner on a Codesys controller (no IPC) and here the scanner behaves as expected (an alarm is triggered after the timeout has been expired). The version of the Codesys components are listed in the picture Versions.png Is there anyone that can give me a push in the right direction? Kind regards Alex
Last updated: 2026-04-14

Sort by relevance or date