Automated Dust Collection

Automated Dust Collection

As part of the wood shop team’s upcoming plan (yet to be voted on) we plan to start automating the blast gates and dust control in the wood shop. Before we get too far though, we wanted to give an overview of the project so that people could give some input. Specifically, we would like feedback on what people want to get out of the automated dust collection and input into how to integrate it within the wider HSBNE infrastructure. Of course, other feedback and technical design suggestions/review is more than welcome.

After feedback and working out the exact technical solution I’ll put together a bill of materials and budget to be voted on at a general meeting.

The Problem

A critical part of keeping the wood shop clean and the air breathable is the dust extraction system. Right now, blast gates (valves) are fitted to each machine that members must manually open and close to control the flow of air. Members must also manually turn on the dust extractor. There are several problems with this manual system.

  • People forget to open the blast gate when they are using the machine
  • People forget to close blast gates when they are done with a machine, reducing the effectiveness of dust collection throughout the system.
  • People forget to turn on the dust extractor.
  • Several blast gates are difficult to reach. This makes the system inaccessible for many and disincentives its use.

Success Criteria

In order to solve these problems, a system must be developed that meets the following the criteria:

  1. Blast gates must be easily operable from where an operator would stand to use a machine.
  2. Blast gates on all other machines should be able to be closed while standing at any given machine.
  3. The dust extractor should be turned on automatically whenever a blast gate is opened.
  4. The dust collector should be able to be switched on and off manually.
  5. When the dust extractor is turned off manually or has been off for more than 1 hour, all the blast gates should be closed.

There are also a number of requirements stemming from technical considerations, mostly to avoid damaging or wearing out the dust extractor.

  1. The dust extractor should never run with all blast gates closed. To avoid this, at least one blast gate should be open at all times.
  2. The dust extractor should not be cycled on and off rapidly. When a blast gate is opened the collector should operate continuously for at least 10 minutes (in anticipation that it will be needed again soon), even if all blast gates are closed. During this 10 minute period, the dust extractor can be switched off manually.

In expectation of future interlock designs featuring current sensing (to tell if the machine is currently in use):

  1. The system must contain a mechanism by which a device measuring current use on a machine can be used to open and close a blast gate.

The Solution

Currently compressed air lines and 24VDC power is already run to each blast gate. Each blast gate is fitted with a pneumatic cylinder which can open or close the gate when paired with a solenoid. To control each solenoid a Sonoff SV will be used (similar to an Arduino). The Sonoff SV is an ESP8266 board with integrated power circuitry (5-24VDC), relay (for triggering the solenoids), and WiFi. The user interface will consist of two momentary push buttons; one to open and close the blast gate and the other to close all other blast gates. The remaining automation will be handled in code.

Note that the pneumatic cylinders fitted to the blast gates will make it impossible to manually open/close the blast gates.

The user interface would look like this, made up of two momentary push buttons.

The buttons do what they say. The green one has the additional function of turning on the dust collector if it is not on already.

Technical Details

The full technical details have not been fully fleshed out yet as I want to give the opportunity for people to review the plan before committing to it.

The Code

As far as is practical the Sonoff SVs will be kept ‘dumb’ and will have limited internal logic. To this end, the MQTT protocol will be used to communicate with a central controller service which will execute the core logic of the system and direct the peripherals (Sonoffs and dust extractor) as appropriate. Such a system could be updated later if interlocks gain current sensing abilities.

As a rough layout for the MQTT design, each Sonoff (with its own UID) will engage with the following topics:

  • Subscribe to “DustCollection/[UID]/Commands” and open/close it’s associated gate when the controller publishes to the topic.

  • Publish to “DustCollection/[UID]/State” to inform the controller when the gate is opened/closed (regardless of if the it was commanded to by the controller or by pushing the button). This topic will also be used when the ‘close all other blast gates button’ is pressed.

The Electrical Design

Excuse the MS Paint.


As a further stretch goal, blast gates may be added that cut off entire branches of the system when they are not in use. This would improve the performance of the system.

I don’t want to be a dick, but I feel this would be an over engineering of facilities. If an argument against it is appropriate read on. Otherwise enjoy your evening

It seems when ever we have a flow of electronics, they fail. Weather it is doors regularly not opening, interlock being temperamental or down, and recently a failure in security cameras, windows sensors breaking or disappearing, kiosk not working, kiosk being out of service.

There has been a number of people who try to swipe on to equipment, with out signing into site. I don’t feel this human failure is something worth enforcing with signs and light popping up via a sensor that registers new people on site and telling them to bloody sign in. Some time human failure is simply what it is, and some times an opportunity to learn.

Money to produce the system.
Manpower to assemble the system.
Manpower and money to fix the system when it inevitably break or is damaged, therefore incurring financial costs over time.

I don’t feel its a good use of money or human resources. Save for the odd vent that is unreachable being made reachable

You are describing a policy of the “IDIOCRACY” movie isn’t?


Not being a dick at all, you make some good points. I agree with you that these kinds of systems can fail and when they do it’s a total pain in the ass. With that in mind, the core functionality of ‘hit button, open blast gate’ is very simple and should be very reliable. It’s been working great on the CNC for a while now.

You have however spurred me on to think about some fault detection in the system. It would be worthwhile to add in a system that cuts the compressed air if the system loses power or screws up. Doing that allows everything to operated manually and will hopefully prevent users from ever being screwed over by a broken system.

Ultimately whether or not it’s worth the money comes out to how much it will cost and the vote at the general meeting. Personally, I think it would be a huge upgrade, especially if we get new fancy interlocks.


Sorry, not sure what you mean. I haven’t seen the movie.

Is their a mechanical system that can open a valve if the pressure gets too low?

That way if everything else fails air is let into the system and the system is not trying to create a vaccum.

If it can still be operated manualy as a failure safe, I don’t mind so much.

The whole concept that we need to automatism the whole thing bothers me. Eg, a reminder in my phone to call Mum. If I actually give a shit I’ll call mum regardless. And thusly people will remember to or will learn to open vents.

But we may disagree on those terms. But I am happy to accept disagreement.


Probably, although I don’t think one will be required. Unless I’m mistaken, the main reason the collector shouldn’t be operated with no gates open is that it will cause the impeller to spin too fast and put strain on the motor. Currently there’s no protection against all gates being closed and, from time to time, the dust collection has been run with all the gates closed.

The automated system would prevent all gates from being closed at once. Even if this failed, the dust collector will be switched off automatically after 10 minutes.

This looks great!

Just for recollection. I think that the part where you’re saying “The dust extractor should never run with all blast gates closed” is not correct.
When all the blast gates are closed, you will have “vacuum”, “zero airflow” and “no load”. So in terms of power, the dust collector will be better that even (eg. having to not putting any work).
Because is an induction motor it gets cooled separately from the system itself.
What I’m trying to say, is that nothing will happen if the extractor runs in vacuum/no load.

Edit: i think the motor can’t run any faster than it’s RPMs by design since it’s an induction motor.

This would be completely different with a standard vacuum, where the motor is an universal motor and it will spin as fast as the universe allows for it!

1 Like

The main reason you’d want to keep some airflow/gates open in the system is to keep the extractor from crushing or damaging pipes/gates etc from the higher than usual vacuum. The more closed the system is, the less load on the motor so it won’t be strained. Once you close all the blast gates, the extractor is moving no air and hence doing no (well, much less than normal) work. Impeller won’t spin faster unlike a vacuum cleaner, it’ll be spinning almost the same speed as under full load.

On the same note, the more gates open and hence higher airflow, the harder it’s working.

Depending on how strong your linear actuators are, if the control system closes all gates while it’s running, and tries to open one again, the friction caused by the higher vacuum on the gate might stop it from opening, or break something with the higher than usually required force.

I’d suggest just setting the brains to always have one gate open at a minimum so you don’t run into issues like the above, or just have no way to close all the gates. E.g. “close all other gates” button automatically opens the gate where the button was pressed.

You’re right. I think that at the moment, the max pressure of the actual piping doesn’t do any damage to itself. I think as well, that the pcv piping is better in this aspect that steel pipes. I’ve heard stories about pipes collapsing in TAFE, but right the system wouldn’t do that.

The blast gate for the cnc works very well, so if all the gates are as reliable as that one, I feel like it is a massive upgrade and I think worth the effort.

My suggestion is to rethink the run time however. It’s frustrating when the dust collect turns off after 20 minutes while you’re in the middle of a job with the current timer. Sometimes it takes a while to realise it’s turned itself off

The run time of the extractor is just code, so if its deemed to be changed its on github :wink:

However, my long term plan with all that was to get current sensing data from interlocks to detect the tool is on and running, to open the blast gate and ensure the extractor is on. But thats a while away just yet.

I started this project like… i dunno, 2-3 years ago. I designed the blast gates, the pneumatic system, suggested the use of sonoff svs, the integrations with interlocks etc etc. Its just something the woodshop has been working towards for a long time. So really Ryan is just talking about what we’ve already been doing for years. The stuff that hasnt been done is the sonoff and the switches, everything else exists currently. although i’d like to do new metal gates now we have the yag, but thats neither here nor there.

the gates all have valves on them so you can disengage them from the greater compressed air. this is how you flip them to manual mode afair. but you could put a solenoid on the whole system and disengage it if you wanted.

This is the main concern of closing all gates. I figured in the initial design we might put a gate up high somewhere as the ‘default gate’ and use it to cycle the air in the shop, get any particulate thats in the air.

Any reason we’ve moved away from the already chosen and bought buttons and the enclosure i did? not having a stress just curious for the change.

@Zac_Crow makes an important if embarrassing point - the systems produced by The Space to be used regularly by members are sometimes not reliable enough for their purpose and therefore defeat the purpose for which they were built and installed. He has listed in his post several important examples.

Given this history, the design rule ought to be that the extractor motor and the dust gates should all be able to be manually operated, should the overall gates system or any of its components go offline or fail.

As to the problem of excessive vacuum in the system, there are standard vacuum relief valves which will open when the vacuum in the pipework is too low and close again after that. Google “vacuum relief valve” for the gorey details. Finding an off the shelf valve that is suitable might be a bit of a problem as these valves seem to be used mainly in the petroleum sector which is notoriously expensive. @Thermoelectric1 the incorporation of this valve removes any flow on problems with the main extraction motor, do you think?

(I saw, working on another dust system, the photos of that TAFE system which sucked itself to death; the bloke who designed it was proud of the photos and the system - not sure about the view of his insurance company).

This system has been in gestation for 3-4 years. It is to the credit of @Ryan2 that he is soliciting comment from the membership about the system before the design is finalised. A designer should always remember that he/she spends very little time with the design relative to the time the users do over the life of the system. The designer has to “design the shit out of the system” for the benefit of the users. Good luck with that, @Ryan2. :upside_down_face:

@Steve the woodshop team would really like assistance with the dust extraction and blast gates. Check with them for their meeting day and help design the shit out of it.

I’ve been working on the automated dust collection and I thought I’d give an update.

The Blast Gate

I’ve completed my first prototype for the blast gate itself.

I’m quite happy with the design but there are a few improvements that are going to made to the second revision. Thanks to Ale and Josh for feedback on the design.

  • A shorter piston that doesn’t poke out
  • Strengthening of the mechanical linkages
  • Adding bearings and bushings where they’re needed
  • Adding a handle to the sliding portion to make it easy to open manually.
  • Tweaks to the part that clamps the pipe for ease of assembly
  • Decreasing the radius of the cutout on the sliding part to prevent dust build up.

The User Interface

I’ve changed my plans on the original design somewhat which necessitated a rethinking of the user interface. Originally I was not planning on detecting if a machine was running so I couldn’t automatically open and close blast gates but have since decided that the complexity added by automatic detection is totally worth it.

This is what I propose the user interface be like:

Most of the time the selector switch would be left in AUTOMATIC mode and the blast gate will open and close automatically as the machine is turned on and off. If, for whatever reason, you want to manually open or close the blast gate then you can set the selector switch to MANUAL mode and then use the second switch to select open or closed.

I’ll go into this in more detail later but selecting MANUAL completely bypasses the micro-controller/code so you can always be sure that the system will work.

Designing for Reliability

As some of you rightly pointed out, automatic systems that don’t work are a pain in the ass and there has to be a a way to override the when they fail. Excluding instances where something actually breaks (e.g. the blast gate explodes) I’ve broken down what the system relies on and what will happen if that thing fails:

  • Compressed air. The system uses compressed air to drive the piston that moves the blast gate. If the compressed air cuts out then blast gates will have to be moved by hand (like what happens now).

  • 24V power supply. The system uses electrical power for everything, including pneumatic valves. If the power supply cuts out, a pneumatic relief valve will automatically bleed all the air in the system which will allow blast gates to be moved by hand.

  • WiFi. The system will use WiFi (in automatic mode only) to turn on or off the dust collector when a machine is being used. If the the WiFi cuts out then the dust collector will have to be turned on manually (like what happens now).

  • The code/ micro-controller. Obviously the code on the micro-controller needs to be functioning correctly for the automatic mode to work. If this fails, the blast gate can be set to manual mode and operated using the switch.

Here’s my sketch of the electrical and pneumatic systems. I’ve left out the current sensing circuit since it’s a little more involved. I might look into doing a JLC PCB order for the current sensing circuit but I don’t really know anything about making PCBs so we’ll see.

1 Like

I have very little to add except my thanks for your hard work on this mate. Having automated gates will be amazing and I really appreciate the hard work you are putting into it. My only input is to make sure that you set a delay of 5-10 seconds after the machine turns off for the gate to close as you need to make sure the dust collection runs for a little bit after we turn off the machine.

Thanks again mate - great work :slight_smile:

Thanks Bobby, good point.

At the moment I’m planning to keep the gates open for 1-2 mins after the machine turns off to prevent them from opening and closing needlessly.

I’m also planning to add a delay to shutting down the dust collector to avoid wear and tear, maybe 20 minutes.

This is current dust extractor code:

I think as part of this, swapping the extractor to an esp board and shifting out the logic from the code to mqtt and hass/nodered makes sense.

Thanks for the flowchart Josh, that will be useful.

My current plan with logic being on/off the devices is to keep just enough logic on the device itself that it will work mostly normally even if the MQTT broker is unreachable. What this means in practice is logic for the gates to open/ close and the dust collector state must be handled mostly internally. However, none of the devices will ever talk to each other directly. Coordination between the devices can be handled by HASS, nodered, etc. I have to admit that it’s kind of a shitty middle ground but I think the compromise to simplicity and maintainability is worth it for the boost in reliability.