Workshop blast gate automation

Cheers @byron.

I’m on to prototype 4 now which corrects/improves the design now that the plastics are included. I’ve made a drilling template for the base plate, so I’m hoping most of the bugs are resolved. There are a lot of holes, mostly tapped, and marking them out each time is a pain in the arse. 
I tried 3D printing a template that went very wrong and damaged the print head by bunging it up with melted plastic. It actually ruptured the hot end silicon boot. 🙄

The arm design in F360 isn’t that complicated, it’s created from simple 2D sketches rather than 3D sketches that are more tricky. The parts are also designed keeping in mind that I will be 3D printing them myself. If you ever need any help with F360, let me know. I’m not an expert, but might be able to help.

It’s been a while since I’ve given an update on the Blastgate Automation project. I have been busy, honestly, but thought the previous updates may have been a bit too long winded for some.
Phase 1 is coming to an end. Three gates are completed and installed and have had some use; a fourth is built and ready to install. A fifth is to be build next, to complete the first phase.
There have been quite a few developments and design changes along the way, but as previously mentioned I didn’t want to bore everyone with the nitty gritty.
So, I‘ll summarise three sections over the next three posts. Electronics, mechanical, software.

When the modules are powered up they go through a self check to make sure the motor and limit switches are working okay.

Electronics
I had originally intended to design a pcb for the modules, but the vero board versions have come out quite well and seem to work well, so I probably won’t bother.

The following have been added since the last time of posting.

Hardware addressing
The prototypes had different firmware in each due to the need for each having a unique network/mqtt names, as well using different GPIOs.
I considered using wifi configurable names, but decided on hardware addressing instead. A DIP switch is connected to three GPIO pins and is read on boot up to define one of eight possible addresses. The address is then used for the wifi network name as well as the mqtt name.

COB LEDs
The original design had a RGB LED to show the status of the module. Connecting to wifi, mqtt, gate state errors, etc. Then I had the bright idea of adding individually addressable LED's to make a fancier display. It was quite a performance and needed much software experimentation, but worked nicely in the end.
The cheap COBs I could find were all 5V, so I added a CYT1076 level shifter to go from 3.3V to 5V. It has four channels and I thought I might use other channels for additional features. I haven't, and could probably have used a single MOSFET instead?

Voltage regulator
The original RGB LED and ESP32 didn't use much energy. I picked an 800mA regulator/module which was adequate. Since then I've added a COB LED strip that uses much more power and so switched to a L7805 linear regulator. The modules are powered by their own plug in 12V PSU so the inefficiencies of the regulator aren't a problem. Plus they seemed fairly easy to use for a novice.

Stepper motor driver
The original stepper driver was an A4988 driver module. Apart from some initial difficulty in setting the current level, it was easy to connect and use. One thing I didn't like about it was how loud the stepper motor was. I hadn't appreciated that this was down to the driver module until I stumbled upon the TMC2208. I'm pretty sure the A4988 would have been quieter if I had used the micro step modes, but I'm happy with the TMC which interpolates micro steps on board. (I think I've described that correctly?).
I did have to make the software go faster to get the required gate speed, as I guess I would have had to with the A4988 micro step modes? More on the software later.

Voltage isolator
I read/saw on a youtube video that it is unwise to connect the ESP to USB and an external power supply at the same time. I didn't test this out as I have already blown a few ESP modules with other problems. The 5V isolator wire loop deliberately covers the ESP32s USB port so I have to remove the link, which isolates the ESP32 from the PSU, to plug in the USB lead.

Optical limit switches
Prototype 2 had optical limit switches and those made it into the final design. The only change I made was to have one with a three way Dupont connector and one with a four way. This made the tracking simper and stopped me plugging them in the wrong board position.

External trigger
A last minute thought was to add the external trigger connector. The next phase of the project will be to add a current sensing module to some of the machines so the appropriate gate opens and the extractor comes on when the machine starts, and shuts off again when the machine is turned off. The software has been added and tested for this feature so I just need to work out the current sensing module. The current sensing module will have a volt free relay which grounds the ESP32 GPIO pin. When the relay drops, the GPIO pin is internally pulled up which turns the extractor off.

Power connector
I wanted a way of preventing myself reversing the polarity of the 12V supply. I don't have any polarised connectors and so looked at, I think they're called idiot diodes, or something like that, but couldn't quite fathom it out. In the end I went for a three way connector with gnd as the center pin and 12V as an outside pin. One outside pin connects to the 12v circuit and the other outside pin connects to nothing, So if I get it the wrong way round nothing happens.

@andydowns

Well thats very nice and its good to see a 'whole' project being documented from the design to a finished product, including all the aspects like the mechanical, 3D printing etc as well as the electronics.  I particularly like the finish of the enclosure.    

The LED strips glowing through the 3d printed case is a master stroke.  I'm going to be copying that.  We have a gravelled drive and I used to hear the slight crunch of folks approaching, but now the old age hearing, even with the hearing aids, means I don't hear that crunch any more.  I've a few sensors in place to assist notifying any driveway movement etc, but I've not really done much to bring whats happening to our attention as yet apart from a small screen display.   Now I can envision a collection of nice small enclosures, probably with a screen in the middle, a euphonious speaker in the enclosure and a glowing enclosure circumference oozing various shades indicative of the type of info/alert status.  All in a curvy 3D printed 'box' that SWMBO would not object to.   I'm off to order some of those addressable LED strips to have a play with as a starter. 😀 

Thanks for sharing your project. 👍 

Posted by: @andydowns

↑

Background
I decided a couple of weeks ago that it was about time I got back into writing software; something I haven't done for a very long time. I've always liked the idea of controlling things with micros and this seemed a good place to start.
So, I needed to pick a platform and a project to aim for. After considering the Arduino, I settled on the ESP32. I thought of making something to connect to my home automation system Hubitat Elevation hub, which mainly uses zigbee. Although Xbee could be used with Arduino, the ESP32-C6 apparently had onboard support.

So next was picking a project. I have a workshop that is mainly used for woodworking but I also do other things as well. Whilst doing a woodwork project a week or two ago, my extraction remote broke. It still works intermittently but is irritating to use. So I thought I would replace the remote that controls the extractors and add some further functionality. And so the blast gate/extraction project was conceived.

-- attachment is not available --
-- attachment is not available --

The workshop
My workshop is, what I would describe as, a well equipped home workshop. I am retired and the workshop is at the bottom of my garden in the UK, and just for fun.
Apologies for the messy state. 😊

-- attachment is not available --

There are two extraction systems:
An Axminster Craft AC153E low vacuum high flow Dust Extractor that is connected through rigid PVCu piping and used for the machines. The blast gates are cast alloy body with a steel guillotine style shutter and are all connected to this system.

-- attachment is not available --

The second is a twin motor Camvac which is a high vacuum low flow system. This is used for hand power tools and machine cleaning. There are no blast gates connected to this. I have included it because this will need to be controlled (on/off) via the system created through the project.

-- attachment is not available --

The workshop currently has six blast gates:
Mitre saw and Table saw gates

-- attachment is not available --

Floor sweep vent gate

-- attachment is not available --

Planer/thicknesser gate

-- attachment is not available --

Bandsaw gate (You'll have to use your imagination here because I removed it for the project development 😊)

-- attachment is not available --

Pillar drill gate

-- attachment is not available --

A seventh gate for the router table will be added later this year

-- attachment is not available --

System requirements
Requirements spec version 1 - a system that would:
• Open a particular blast gate, close all other blast gates and start the Axminster extractor.
• Control the Axminster extractor (on/off) independently of the blast gates.
• Control the Camvac extractor (on/off).
• Control the system via a remote control that I could wear on my trouser belt loop so it would always be with me.
• Open each gate locally which would also close all other gates and start the extractor.
• Close each gate locally.
• Each gate to show it's status via an indicator (probably led).
• Option to easily disconnect the blast gate module from the blast gate. If I'm in the middle of a woodworking project and there is a module error, I don't want to have to stop and fix it immediately.
• The modules will be retrofittable to the existing gates.

I hope this post hasn't been too long winded. Development has already begun so I'll post over the next few days until the posts have caught up with the development, and then just post when there is something new to say/show.
The next post will be system design.

Wow! What a mammoth task! 

@byron 

This is the filament I used.

"https://www.amazon.co.uk/gp/product/B07TPTDPH7"

Horrible stuff to post process. 

I wonder if simply printing a fairly this white filament might not be better.

@andydowns

Hi Andy, thanks for that link.  Actually I put in an order for a PETG transparent filiament as I've always had good results with this type of filament. https://www.prusa3d.com/product/prusament-petg-clear-1kg/ and it should arrive in the next few days.  I'll let you know how well this filament works out.  

I'm also thinking of some outdoor info/alert type enclosures to flag up warnings when I'm working in the garden and PETG is meant to be OK for outside boxes.  These days I'm likely to miss a car of van arriving and perhaps someone ringing the door bell whilst I'm in the greenhouse and my wife is also outside, so its about time I upped my game on this.  

I was actually looking at ThePiHut for some addressable LED strips as your post arrived.  There seems to be some different varieties.  Which sort did you use?  I'm presuming that they can be powered directly from a power source, and the separately powered picoW (or EPS32) taps into the data line and shares the ground line.  I did not see any obvious wiring diagram examples, but than I have not dug into this much as yet.

@byron 

PTEG sounds interesting. I haven't tried that. My 3D printing isn't great, but does the job as I'm mainly interested in function over aesthetics. I will give PTEG a try.

This is the LED strip I used - https://www.aliexpress.com/item/1005005486743999.html .

I've noticed the second delivery of COB is a slightly different colour than the first. I imagine they switched suppliers. No big deal and probably what should be expected given the price.

I used a CYT1076 level shifter because I thought I might want further channels, but never did.

"https://www.amazon.co.uk/dp/B07RY15XMJ"

It shares the the 5V (HV) with the ESP32 5V supply, the 3.3V (LV) comes from the ESP32, and they share a ground.

Posted by: @andydowns

↑

Background
I decided a couple of weeks ago that it was about time I got back into writing software; something I haven't done for a very long time. I've always liked the idea of controlling things with micros and this seemed a good place to start.
So, I needed to pick a platform and a project to aim for. After considering the Arduino, I settled on the ESP32. I thought of making something to connect to my home automation system Hubitat Elevation hub, which mainly uses zigbee. Although Xbee could be used with Arduino, the ESP32-C6 apparently had onboard support.

So next was picking a project. I have a workshop that is mainly used for woodworking but I also do other things as well. Whilst doing a woodwork project a week or two ago, my extraction remote broke. It still works intermittently but is irritating to use. So I thought I would replace the remote that controls the extractors and add some further functionality. And so the blast gate/extraction project was conceived.

-- attachment is not available --
-- attachment is not available --

The workshop
My workshop is, what I would describe as, a well equipped home workshop. I am retired and the workshop is at the bottom of my garden in the UK, and just for fun.
Apologies for the messy state. 😊

-- attachment is not available --

There are two extraction systems:
An Axminster Craft AC153E low vacuum high flow Dust Extractor that is connected through rigid PVCu piping and used for the machines. The blast gates are cast alloy body with a steel guillotine style shutter and are all connected to this system.

-- attachment is not available --

The second is a twin motor Camvac which is a high vacuum low flow system. This is used for hand power tools and machine cleaning. There are no blast gates connected to this. I have included it because this will need to be controlled (on/off) via the system created through the project.

-- attachment is not available --

The workshop currently has six blast gates:
Mitre saw and Table saw gates

-- attachment is not available --

Floor sweep vent gate

-- attachment is not available --

Planer/thicknesser gate

-- attachment is not available --

Bandsaw gate (You'll have to use your imagination here because I removed it for the project development 😊)

-- attachment is not available --

Pillar drill gate

-- attachment is not available --

A seventh gate for the router table will be added later this year

-- attachment is not available --

System requirements
Requirements spec version 1 - a system that would:
• Open a particular blast gate, close all other blast gates and start the Axminster extractor.
• Control the Axminster extractor (on/off) independently of the blast gates.
• Control the Camvac extractor (on/off).
• Control the system via a remote control that I could wear on my trouser belt loop so it would always be with me.
• Open each gate locally which would also close all other gates and start the extractor.
• Close each gate locally.
• Each gate to show it's status via an indicator (probably led).
• Option to easily disconnect the blast gate module from the blast gate. If I'm in the middle of a woodworking project and there is a module error, I don't want to have to stop and fix it immediately.
• The modules will be retrofittable to the existing gates.

I hope this post hasn't been too long winded. Development has already begun so I'll post over the next few days until the posts have caught up with the development, and then just post when there is something new to say/show.
The next post will be system design.

How much time did it take to finish this project?

@aliarifat 

Its taken about 3 months so far from first idea to an installed first phase. I have some software ideas to implement and install, yet, so probably another couple of weeks.

I have more phases to do.

1) House the Shelly relay in an enclosure of some sort.

2) Machine startup detection to open the appropriate machine gate and start the extractor.

3) Key fob remote control.

4) Pressure/vacuum checker to turn the extractor off if all gates are closed. An additional safety feature.

The overall mechanical design hasn't changed radically since prototype 3. As previously mentioned it was mainly based on prototype 1, but incorporated elements from prototype 2; namely the optical switches.
There have been many minor design tweaks and improvements to most parts, like addition chamfers and the way the limit arm connected to the shutter arm.

The main changes have been to the top and bottom covers.
The bottom cover only had minor changes to thickness on the front to support the fixing bolt and a beefing up of the locator mount on the rear. It went surprisingly easy to what I expected.

The top cover was more problematical and is currently on version 7. The fit around the keyboard proved to be difficult and needed a few goes at aperture size and a bezel size to make the fit good enough. A part of the problem was that I wanted the fit around the keyboard to form the front fixing. The cover fits over the front first, and is then held on by the power plug fitting into the rear. I wanted it to be easy to remove and replace and it is remarkably secure even though there are no actual fixings.
In addition, I thought of the external trigger function at the eleventh hour and had to alter the power connector bracket, which now needed to house an additional connector, and matching holes in the cover.

I will be building more of these at a later date and this aide memoire shows the parts in the design.

Favourite part
The part I'm most pleased with is the lens. I wanted a way to simply hold the LED strip in place against the lens without having to stick it in. If it needs to be changed at a later date, I didn't want the hassle of trying to unstick it.
The lens design has a channel that is 0.5mm wider than the led strip. It snuggly slides into the channel and serves to hold the LED strip in place and up against the lens outer wall.
It also has a location ridge for the top cover to locate onto to hold it in position.

What would I do differently if I designed it again?

Bolt lengths
Because the mounting plate is quite busy on both sides there are some areas where bolts and parts collide. Some parts could be moved to avoid this, but not all. The mounting plate is 4mm aluminium and most plastic parts are 5mm to give them some strength. Since 10mm is a typical length of bolt, they stick out by 1mm. Next time I will take this into account, and the fact of which bolt lengths I already have, and design to make the bolts flush.

Limit switch adjustment
The limit switches are mounted on blocks with sliding captive nuts to allow them to adjust back and forth and fine tune the position. In practice this was unnecessary as the point at which the gate is fully open or closed doesn't need to be that accurate. So a simple stand-off would have sufficed meaning less holes and less plastic parts.

Limit switch arm
The shutter arm and limit switch arm are printed in two parts because of the awkward shape to print. They are printed separately and bolted together with m3x10 bolts. This works perfectly, but with a little more messing about with support structures in the slicer, I think it would have been possible to print them as one part.

Fusion 360 model with fixings
It's not my habit to include nuts bolts and washers for simply home projects, but this nearly caught me out.
The gap between the shutter arm and mounting plate looks fine from the model, but when the 6mm mounting nuts are added, there is only about 1mm clearance between the two. It's tight, but enough and gravity works in my favour, so I'll leave it as is. Lesson learnt though.