Touch screen for a 60-year old injection molding machine
HTML-код
- Опубликовано: 7 янв 2023
- Support me on Patreon: / johnsl
I decided to pivot with my project on making a touch screen for a 60-year old injection molding machine. Previously I used a Nextion display and a micro controller. While this worked, it wasn't as easy or flexible as I'd like. Therefore, I decided to pivot to using an inexpensive Android tablet and an ESP32.
Previous videos using the Nextion display and Rust:
* Nextion Display for my Injection Molding Controller: • Nextion Display for my...
* Writing to the Nextion Display in Rust: • Writing to the Nextion...
* Responding to Nextion Button Pushes: • Responding to Nextion ...
Amazon Affiliate links:
* The ESP32 that I'm using: amzn.to/3X14KGX
* The 10" tablet I'm using: amzn.to/3QoRWaO
* The 7" table I've ordered: amzn.to/3jR7VCi
I suggest to put all control logic into the esp and do just the visualisation on the tablet. Also emergency stop should be a hardware switch directly to the controller - otherwise you will get in trouble if the connection between the display and controller gets broken
well, i am pretty sure you are aware that he is not taking the "usual" industrial design approach. Besides your suggestion there is plenty more that should be done
I grew up in an injection molding shop. My grandfather owned a small plastics plant and he was still using this style of equipment up into a few years ago
It's absolutely awesome that you can modernize an old machine like that. Really, when you think about it, the only difference between that machine and a modern one is the controls. That, and it's probably built to last a thousand years. I haven't gotten into injection molding yet, though it is planned in my future. I did just get into 3D printing a couple months ago. I converted an old industrial cabinet into an enclosure for my Ender 3 S1 Pro, and I currently have a control panel with a bunch of toggle and rotary switches to control the power, lights, fan, heaters, and fume extractor. I was happy with it, aside from wanting to replace the PID temp controller for one with a timer so I could have the heaters shut down when a job was done, but now after seeing this I will have to rethink the entire project. Controlling everything, including the printer, with a touchscreen would be so much better.
Difference from a 30 year old injection molding machine and today is lightyears, to make a old machine up to date you need to change all hydraulics and electronics and that would cost an enormous amount, on a big machine a few hundred thousands of dollars could be worth it maybe.. I work with this since 2003.
Great work. I need this for my device.
Great! I'm looking forward to see how you integrate those electronic commands into the hydraulics of the machine, and in a broader sense how you transform a relatively cheap single shot machine into semi or full auto.
Thanks. Initially I'm going to use this for my Arburg C4b, which is pneumatic and designed for running fully automatic. I've been thinking about getting a Buster Beagle 3D Mk3 to retrofit as well.
@@JohnSL One of the Travins would be a great starting point. One of them is already semi-auto
I guess my Travin TP-1 might be called semi automatic. However, the toggle clamp right now is human powered, so hard to make automatic. It might benefit from replacing the temperature controller (which is just a dial and a light) with a PID and add a pneumatic valve for consistent injection duration. But first up is my Arburg C4b.
LOVE THIS PROJECT! Thanks for sharing so many details. I would love to learn a little more about the code. I've been looking for a cheap way to get into building and learning automated equipment.
I'm working on a more detailed video.
I really like the direction this is headed. If you haven’t researched it yet, you can also do something very similar and cheaply using an arduino with Bluetooth, and make your screen/interface a web app so customers can use their phone or tablet as the interface
I did something like that about 6 years ago with an ESP8266 using JQuery and REST APIs on the ESP. Having a cross platform application that runs on Windows and Android is far more productive for me than writing a web app. And I just received a $45 7" tablet, which is hard to beat.
I programmed up a similar controller for the machine I bought. It runs on a really small 4 line LED screen with a single knob. If you wind up making this commercially available we may upgrade what we have. One extra function might be the number of ejection cycles. I've seen other machines that bump the part multiple times to ensure the part is ejected. Also, if you have a robot arm grabbing parts, you may want to have an input for a sensor that registers when the robot cycle sends a "complete" signal.
I do hope to make this available commercially. I'm not promising when it will be available, as this is a hobby for me. I have another perosn who wants to have the ejector pins run off a separate, smaller pneumatic cylinder and it totally makes sense to be able to set the number of eject activations per cycle. I have some ideas on how I might be able to work with a "complete" signal, as it's similar to what I have in mind for a part gate to detect if/when a part drops.
Interesting!
I'd suggest not relying on Bluetooth connections. We work with a couple of Bluetooth devices in our brick and mortar stores and they're such a pain to debug and maintain. I'd suggest looking into something like odroid or orangepi. They are powerful enough to run Android and will have the IO to control devices. That way you can eliminate the Bluetooth requirement.
I think this depends on the variant of Bluetooth, the device design, etc. I'm using Classic Bluetooth. There are other products I'm aware of that also use classic Bluetooth serial and I know they're reliable.
Nice! I've been getting "flash backs' whilst watching your videos to a time in the late 1980's I was just out of my apprenticeship working in the UK for a factory which mainly made car electrical switches. I was kind of lucky being in the prototype/ small toolroom area we would sometimes make Ally moulds and run small batches on a manual injection moulding machine. I think it was called a Manumould machine
I've heard of that machine. I remember a person in one of the Facebook groups that I watch acquiring or selling one of those machines. It looked pretty interesting...
@@JohnSL quite a crude/old machine ours was. Simple straight hydraulic plunger (no screw) horizontal machine. Had a crude manual ejection device. Man I can still remember the smell of overheating the plastic (nasty!!)
What a cool idea for a video. Maybe add a way to manually run the extractor pins if your press can do that.
That's a good idea, and should be easy to implement.
If you are not familiar, Luke Malpas on the Angelsix channel has been doing a multi-part dive into Avalonia.
I absolutely am. I binge watched his videos a few weeks ago and that helped me decide to try Avalonia UI. They also helped me get back up to speed on WPF, as my last time using WPF was about 10 years ago.
Awesome video, this has me considering making my own injection molding machine but i dont think i know enough to get started yet
Is there any chance you would be able to make these or something similar on a commission basis?
Nice project. This is not really a feature as I've only just stumbled across this channel, but I'd be interested in seeing techniques for handling fault conditions, like lost connectivity, software / os crash, sensor failure, etc. It might be beyond the scope of this project, just thought I'd put it here for consideration anyway. Thanks for sharing.
I plan to change my design to put a little more of the logic into the ESP32. That way if the tablet looses connectivity, the current cycle will finish, but it will wait for the tablet before it resumes when in fully-automatic mode.
I never used a nextion display even though it looks easy to work with, but this sound even better. So the ESP acts as just a bluetooth enabled input/output module and the logic runs on the tablet behind the UI? The ESP code is also part of Avalonia, or that is your custom code?
I wrote custom in C++ for the ESP32. It's very simple code that implements a simple text-based protocol I created between the ESP32 and the program running on the tablet. I plan to cover this in a future video. But the idea is that the tablet sends codes to the ESP32 telling it how to set up the various pins, then sends "pin" commands telling it change the output of various pins. I have the read (for the button) working, but not yet integrated with the UX. Avalonia UI is just for the UX.
@@JohnSL Thanks. And where did you implement the actual logic? Like counting the parts, timing, etc. Is that a separate class, service running on the tablet?
I haven't implemented all the logic yet (like part counting). But the idea is to implement most of the logic in the tablet. The timing is all done in the tablet, for example. This keeps the code in the ESP32 very simple. It also allows me to write unit tests easily and run them on my desktop.
For USB on the tablet, there are quite cheap OTG adapters with a port to charge and a port to connect periphery, i use it for Octoprint on the phone and a connection to the 3d printer and power, as well as a drawing pad (for the pen that my tablet lacks) while charging..
If it weren't for the issue with Android tablets and CDC drivers (apparently, that's left an an exercise to application developers), I would likely have gone that route.
@@JohnSL Question is, if in Android 12 it's not yet been fixed, will it ever? xD maybe someone on XDA Dev has a solution in some custom ROM, but probably not worth the hassle
I know it can be done because the Serial USB Terminal Android app works with both the Raspberry Pi Pico and an STM32. I just decided I didn't want to go down the rabbit hole of figuring out how to make it work given that Bluetooth was super easy to get running.
yay
Did you ever get the Arburg c4 working? Looking forward to updates because I have such a machine that I would like to update.
Hey. Great project. im building something similar, but im struggle at using powerpoint for graphics. is it possible to get that powerpoint file to learn how it works?
Great video! Is that an Arburg injection molding machine?
Yes, it's an Arburg C4b. I have a little more about that machine here: ruclips.net/video/07xNXQCgk1Y/видео.html
Looking forward to this project. One Thing i want to mentioned is, you show the breadboard with one switch for the out falling Part. One the clamp of the c4 is another one who tells the Maschine that the clamp is closed. Want to ask if you plan to sell the code?
The push button was just to get the code working that could read a digital input, and I was planning on having both clamp closed and open micro switches. As to what I'm going to do with the code, yes, I'm currently thinking that I'll sell a custom circuit board (or boards) and include the software as part of that (the TouchDRO model).
I'd be interested in seeing links to the tools and an outline for the general use of ESP32 and an Android tablet for a general control project, not necessarily Injection moulding. I have other devices I'd like to control in a similar manner, but have no idea how to proceed.
I plan to publish such a video soon.
Raspberry says they will restock again end of Q1 2023. They just had a large issue with their supply chain. Have you looked into M5 stack? It would be an interesting platform for you I think because they also have a UI builder and the form factor is ideal
The Raspberry Pi (or a competitor) is more expensive than this solution when you add the power supply, SD card, and touch screen. There is no way you can get close to a 7" Android table at $45 plus the $8 ESP32 board. Plus, I can build the application far faster using C# because that's what I've used at work for years.
Interesting way to do this. Graphics look good. Here is a different way. Perhaps you make the microcontroller serve up a web page. Lots of ways to do this. You run the server code on your microcontroller, it could be in C, with the html right in the c code, or in python,,,You have one program, no app on the tablet. It is then independent of your display tablet, it could be an iPad, you phone etc…android, the mac, pc or a raspberry pi… open any browser and then WiFi to your local controller. And no cable needed. In the long term, you are dealing with one program to maintain….
I used that approach 5-6 years ago with an EPS8266. While it worked, it was still a lot of code running on the ESP8266. And I find it takes me quite a bit longer to write a web application than a desktop application, both because of the rendering (HTML vs XAML) and the language (Javascript/Typescript vs C#). By the way, the only cable needed for the tablet is a charging cable.
I am in about the same state with a 1962 Arburg 25T Allrounder. I designed and built a PCB for a generic molding machine controller with 24V coil drivers for Arduino Mega2650 with the intent to interface to a laptop for user interface. After 2 years no progress...... I just designed a ESP32 S3 board for LCD touch display for another protect. You should look at the SquareLine Studio and KVGL graphics library, May be a good marriage. Maybe time for me to get back on this project too......
I've done LCD programming in the past from an ESP8266, so I knew what would have been involved in going that route. Using an Android tablet is so much faster for several reasons. One is that WPF (which is basically what Avalonia UI is) and C# allow me to go a whole lot faster, as they take care of things like layout, rendering, interaction, etc. I can also easily have profiles for different molds that are saved and therefore that you can reload. Doing this all with a desktop application is a lot less work than from C++ (or Rust). I also have an Allrounder 100 that I need to bring back to life, and figured I could use this as well. That will likely require some additional electronics, but I'll be able to drive them from the digital outputs of the board I plan to design.
@@JohnSL You are welcome to use my design for the controller. It has drivers for 8 solenoid coils, 8 switch inputs, up to 8 thermocouple inputs, up to 8 solid state relays for heater drivers and plugs into the Arduino Mega 2650. Using a $10 touch graphic LCD and the SquareLine Studio and KVGL graphics library you would have a controller for less than $50. A micro SD card would allow a much more advanced controller with expanded setup storage and flexibility.
Can you send me an email? You can find my email address by clicking on the button here: www.youtube.com/@JohnSL/about
@@chuckinwyoming8526 i m planning to rewire my 80 ton machine using similar concept. Atmega 2560 and cheap tft touch screen and relay board to drive contactors. Do u think atmega is reliable to use in industrial machines like molding machine which run 24x7.
@@clappgames5474 Yes the Atmega 2560 has several fault traps you can use like the "watchdog" that will interrupt or reset the processor if it gets lost and a timer times out. Jams or mechanical problems are much more likely. You want to have switches and sensors to catch these mechanical problems and write code to deal with it. You can write fault shutdown and recovery code for processor and other faults and nearly eliminate the odds of a bad crash. In the unlikely event the processor does get lost the worst you may get is a jam, short shot or partly cooled bad part. The PCB that I designed will allow all likely heater PIDs, input and control functions most molding machines will ever need.
If you are truly interested in a cost reduced version the esp32 can easily drive a touch screen and you wouldn't need the android tablet at all.
I did look into that. A 7" touch screen from a source I trust in China is $41. The 7" Android tablet I just purchased is $45. On the other hand, I can build the Android app far faster than a C++ (or Rust) app on the ESP32. There is also a cost to writing the code. When I look at all of that, I think I've chosen the right balance.
You could use AI2 or kodular to do an Android GUI.
I would make all running code on the esp side, in the event that the tablet disconnects or similar potentially catastrophic situations, the system retains all of its safety touchpoints.
Additionally, you could have manual buttons, leds and regular lcd displays on the machine its self. Then, using an http api, interface the app on the tablet to update and display things . This means that if you had multiple machines, you could use a pi or similar device to build a full web interface that connects the individual machines for scaling up.
I understand it is a small project, but if you're thinking retail, I would at least account for the direction it could go to minimize rework down the road and afford the ability to ramp scale quickly.
On the esp side, use preferences.h to store your settings, like temps, times etc. You could even store by part number and subsequently that parts settings.
cool esp32 stuff, what did you do to run it locally ?
When I run it locally, I don't currently have Bluetooth support, so it's just for quick layout/interaction testing. But now that I think of it, it would not take much effort to support Bluetooth for local testing as well. That would speed things up a little!
I just implemented support for talking to the ESP32 from my desktop app. Thank you for the idea! Now I can do most of my testing from my desktop, and then occasionally from the tablet for a sanity check.
John's custom bread. Any flavor you need.
As long as it's a flavor I like 😉
chip shortage sucks
I want a tutorial with avalonia and esp32!!
I plan to do a more in depth video covering both the Avalonia UI (I'm using Visual Studio 2022 Community) and ESP32 side (I'm using PlatformIO in Visual Studio Code).
Hi, I also dream to restore old KUASSY (this is the one seems to a lot of ppl here has) machine one day. Though I doubt it can actually mold what I want - thin wall small parts.. Can you tell what is the amount of pressure machine should output to be able to inject smth like small thin lego brick, or keycap for keyboard for example? Walls should be only smth like .6-.8 mm wide, I guess this requires high pressure machines, and does it even possible to use simple vertical injection machine, or it is horizontal with screw machines only job? Thx.
It all depends on the size. The smaller the part, the easier you can have small thicknesses. The best way I know of to determine the pressure required is to do a mold flow analysis. I just discovered that Fusion 360 has a pretty good one. I did a simulation today, which cost me 6 cloud credits (I think that was about $20) and about 2 hours. It did tell me how much pressure I needed.
Your Amazon Affiliate link to the 10" tablet is the same as your ESP32 link...but great video!
Fixed. Thank you for bringing this to my attention.
do you sell this code and screen set
A cheap plc and hmi would be way easier than this and a thousand times more reliable, definitely worth investigation.
I think easy is relative. I have no experience with PLCs, so there would be a learning curve there, both to learn the editors/language and to research and decide on what PLC and HMI to purchase. More importantly, however, is that I aim to create a configurable solution that can be used on different desktop machines, and the app will have a simple configuration screen. With a PLC, as I understand it, you'd have to create different programs for different configurations. But the reality is that I built what I showed here probably in less time than it would have take me to get up to speed with ladder logic, etc.
is it an old arburg maschine looks like it
Yes, Arburg C4b. I have a couple of previous videos about the machine and will have more soon.
you can do same job using MIT app inventor to make android app more easy than you write code in C# or C++
Probably not anywhere close as easy for me, as I'm very experienced with C# and C++. The C++ code I have running only took me an hour or two to get running. And C# is what I use at my day job, so I can build applications very quickly.
If you want to save costs, just make it an iOS and Android app and skip the dedicated tablet. Users can add their own device if they want, plus, depending on the minimum platform version you target, they can use an older iPhone, iPad or Android SmartPhone or tablet. As for the microcontroller, you should create a HAL to make it portable across other microcontrollers that support either WIFI, Bluetooth or both. This will allow you to switch to newer or better Microcontrollers. Stay away from USB because it's more work than it's worth and Apple doesn't like small third party USB products. I'm also experienced in writing Xaml for WPF and UWP and you made an excellent choice selecting Avalonia. .NET MAUI looks like vaporware (marijuana vapor). It still has a lot of bugs and with no apps using it. Microsoft mislead developers into thinking MAUI was a entirely new from the ground up, but it's just new code on top of Xamarin. Speaking of MAUI are you Hawaiian?
I like the project BUT it doesn’t really address the safety issues as it really needs to have CAT 3 saftey features
What does it need CAT 3? Most of these machines, without this controller, have no safety interlocks of any kind. This is designed for small, desktop machines, not the large industrial machines. As an example, check out the Buster Beagle 3D. In any event, I do plan to have a separate E-Stop button that is directly tied into the system (and not connected to the Android tablet).
@@JohnSL
I am a retired 67-year-old machine refurbishment tech for Unilever and it was critical that we always address any possibility of unexpected motion
While your a small shop you setting and and example for the world at large
I understand what you're saying. How does that jive with having a manual lathe or a manual mill. Both of those are a lot more dangerous than these desktop injection molding machines, and they have no safety mechanisms like you describe. I know a lathe can kill you in seconds if you're not careful. I'm asking out of ignorance and curiosity.
@@JohnSL
Again it’s a hobby project which I am enjoying but one of the bigger widespread issues I see now is not using UL approved control devices even now when places like Automation Direct I’ve come out with low cost Arduino based devices
The machine can squirt out molten plastic under high enough pressure to explode the mold so there our safety issues that should be considered when your putting in so much effort
My thoughts our these days machines need 4 physical buttons
1. Start
2.Stop
3.reset
4.Estop
The estop Should control a master control relay for total shutdown
The reset should have to be pressed each time you remove the mold
The Start should have to be pressed after each reset for the next moulding cycle
The stop should take the machine to a safe state after the cycle is completed each time it’s pushed AND this is not done correctly very often
At the least it should have a physical button to enable the high voltage power to allow the injection to function along with the Estop you already have though of
The other thing is everything that can squish your fingers or squirt dangerous stuff has to have enclosures
Even a relatively safe thing like a Low-speed stretch wrapper has to have a barrier enclosure with a light curtain and key in hand cycle start
99% of all countries use °C and millimeters, including the UK
only the US uses °F and inches
Not sure what this is in reference to, as nothing in this interface is showing temperature, and certainly not units of length. If you're referring to the mockup, the F will be something you can tap to change to C. Unfortunately, we use both temperature units here in the US.
Ever think of using a cheap china FX Compatible PLC Interface controller board and control it through Modus Ethernet,that way all interface circuits like opti coupled i/o's and transistor outputs can be all in one little PLC PCB or even more powerful controller if needed,I know CNC PC Control Software like Mach3 and Mach4 use that Modbus Protocol for controlling other devices, I found cheap china HMI and PLC all in ones TouchWin,Coolmay that have all what you would need in a Control box to run your stuff,but the software screen designer are always the worst user interface and stupid like nextion, some US HMI software designers are very good at assembling together a quick interface and controller,with bunches of icons,buttons and clip art,like from Weintek Easy Builder Pro Mitsubishi , & AB Factory Automation , Maye the china hmi all in ones would be the cheapest for under 75 to 100 US for 7" touch screen with IO's built in back to the screen housing. Just my two cents,as i've been looking for a all in one solution for a long time to use in small machine automation projects and going down the microcontroller route and designing a pcb for io isolation and all the learning stuff along with it is way to much work, Check out Weinteks Easybuilder Software its modern interface and can control any PLC you can think of. Thanks Eric
I know that's an option. There are several reasons I didn't go the PLC route. First is that it would be a learning curve for me. Second, from what I can tell, the devices would cost quite a bit more (for HMI, PLC devices, etc.) than what I'm using. And finally, I plan to make my solution configurable with a settings screen to be able to work with different types of injection molding machines. Doing the same thing with PLCs, as I understand it, would require modifying the ladder logic, so a lot harder for people who want to use my solution.
Stereo voice recording: Dont bother. Stereo sound for voices is something that isnt even done in recording studios. The worst is those Rodes are notorious for breaking or losing one channel which is very annoying I imagine. Just make sure the one channel is duplicated on left and right and everybody is a happy camper :)
I have the Rode set to "merge" so that the microphone is recorded to both tracks. My previous mic has that behavior as well. However it was far to easy to change back to sending the signal to just one track (it was a single button push, which I would do by accident).
@@JohnSL Yeah, thats a stupid design decision on their part then, in my opinion.
I agree. Which is why I decided to switch to another manufacturer (I'm not naming the one with the stupid design).
I don't think the price drop will happen until the end of Russia and Ukraine war ends... One of the heads of the PI project said they would be all fine in June but then the War came and boom still nothing.. I think some part is made in Russia so with the war and sanctions doesn't look like they will be available for a few more years... You can get Pi alternatives for like $45 but Raspberry Pi it will probably be 2 more years or until Putin ends the war...
In some ways I'm glad I looked for an alternative. I'm much happier with a 7" Android tablet at $45 and an $8 ESP32.
looking at the intro, it makes me think of Blondihacks... would it be insulating to call this channel Boldihacks?
Not at all! Thanks for the compliment (which I assume it was meant to be 😉).
@@JohnSL indeed it was!