Fully automating dust collection

[TX_Lenador]

While there are several postings about methods to start a DC remotely, this posting covers my implementation of the system I created to fully automate the operation when a connected tool is turned on.

A little background
When I built my separator (http://www.jpthien.com/smf/index.php?topic=1054.msg5839#msg5839) I implemented a current sensing relay to start the DC when current was detected on 1 of 2 circuits (120V used for router and other tools and 220V for table saw). While this worked well it still required to be sure the blast gates (BG) are properly configured and fell short when using the CSMS as it caused the DC to cycle frequently (not the best for 2 HP motors) (@retired2 solved this by using a current sensing relay with an off delay timer). I decided to look around and see what options there were to fully automate the operation (open the BG and start the DC) by simply turning on the tool. After a bit of searching it was apparent the commercially available options were not reasonably priced and geared toward commercial operations. I then started working on an idea to use the current sensing relay, interposing relays and pneumatics to operate the BG and DC when a tool turned on. I posted a diagram on this (http://www.jpthien.com/smf/index.php?topic=1179.0). This started to get quite involved wiring wise and felt a bit antiquated. I refined the idea by using a microcontroller instead of relays. I also decided that the pneumatics were not as cost effective as I had desired and lacked granular control of the gate position. This led me to use a servo (like in RC planes) which I had seen in an online video. This idea of one microcontroller still required a bit of wiring and a considerable amount of i/o for one microcontroller. I then looked at using wifi with the microcontroller and decided that a microcontroller at the DC and the BG would work using wifi to communicate with each other. This would also make wiring easy as the microcontroller would be close to the DC or BG and also required minimum I/O for each. I proceeded to build a prototype for testing and soon realized that the servo operation was not consistent, so I switched to a stepper motor. A slight redesign of the prototype with this change and the subsequent testing went well so I proceeded with this idea.

System design
The system consists of Dust Collector Controller (DCC), Gate Controller (GC) and Gate Controller Extender (GCE). The DCC and GC use a ATmega2560 microcontroller (Arduino) with a wifi interface to allow communications between the GCs and DCC. The DCC is mounted on the DC and connects to a power relay that is used to turn the DC on/off. The GC and stepper motor are mounted on a bracket attached to the BG. On the other side of this bracket is an electrical box with an outlet and a current sensor. This creates a fully contained BG with a GC (or GCE) to control the BG and the outlet for the tool to be associated with this BG. Details for each of the components is given below.

Dust Collector Controller (DCC)
The DCC receives commands from the GC when to turn on and off the DC. The on command is immediate while the off command is delayed based on the delay interval assigned to the BG that requested the DC to turn off. This delay allows the DC to continue to run while the tool cycles on/off in short intervals (i.e. CSMS operation). The DCC also reads amps, differential pressure and records run time. The differential pressure (DP) reading will be used (waiting on new DP sensor as the one I have Is not working) to open a relief gate (preassigned to a GC, will be the table saw BG) when high DP is sensed. Once this happens the DCC will send an open BG command to the appropriate GC. The BG will be opened in small increments until the DP drops below the threshold and will remain open till the DC is turned off. This is being done to prevent collapsing the dust collection bin when using a tool with a small port. The DCC is programmed to handle 8 BGs but could be expanded to handle more as there is still quite a bit of memory and processor time available for additional BGs.



Gate Controller (GC)
The GC uses a current sensor to monitor the amps on the circuit the tool is plugged in to. Once the amps go above a preset threshold the GC sends an on command to the DCC and actuates the stepper motor to open the BG associated with that circuit/tool. When the amps drop below the threshold the GC sends an off command with the off-delay interval to the DCC. The GC then waits for the DCC to send back a message to close the BG which happens after the DCC turns off once the delay has expired. If the tool starts again before the delay expires then the GC sends an on command to the DCC which keeps the DC on and resets the delay timer. A GC can manage 1 to 3 BGs; one attached to the GC and each of the other two are attached to a GCE. The GC is pretty much maxxed out at 3 BGs due to the processing required and I/O available.



Gate Control Extender (GCE)
The GCE is simply a physical extension of the current sensor and stepper motor connections on the GC through a ten-conductor interconnect cable. This allows the wiring for a BG to be kept short and only needing one cable back to the GC. All processing is done by the connected GC.



Here are a few pics showing the installation in the shop.










I have been using this now for several weeks and all is working well. I think it will hold up over time as one of tests I ran was to cycle a BG (open, wait 15 seconds, close, wait 5 seconds, repeat) for 8 hours with no failures and no missed steps. The cycle time from closed to fully open is under 0.7 seconds (for 5? BG). This happens while the DC is coming up to full speed so no delay in operation. The only issue I found was that the BG would not fully close if the close command was issued at the same time the DC was turned off. I determined that there was too much drag on the BG slide as it closed and the servo was not strong enough to overcome the drag. I solved this by adding a delay between the DC turning off and the time the gate close command is sent. This allows the DC time to wind down and thus not create as much drag on the BG slide. Once I get the DP sensor working I may find that the relief BG stepper motor may need to be stronger to overcome the drag imposed on the BG slide in a high DP situation.

[retired2]

That is a great piece of work.  If I were starting over, I would duplicate your design.  Unfortunately, I am not seeing the photos you included , not sure why.

[TX_Lenador]

@retired2 - I think I have the photo problem fixed.

[retired2]

@retired2 - I think I have the photo problem fixed.

Photos are indeed working.  Your DC system is an award winning piece of engineering.  I don?t think anything I?ve seen on the forum compares.  Spiral wound pipe and matching fittings.

I do have one question.  Does your wife know how much this system cost? 

[TX_Lenador]

@retired2 LOL, probably not as it has been assembled over the last few years. Thanks for the compliment. It has been an interesting project and am happy with how it has turned out.

[bbain]

Wow, this is awesome.  I like the idea of motors to open the blast gates, I have seen one recently that used pneumatics, but then you always have to have air.

[TX_Lenador]

Here are a couple videos that show the GC and DCC in operation. When the gate number is backlit the BG is open. On the GC you will see the amps the tool is drawing and then when the tool shuts off the off delay countdown timer will start. Once it goes to zero there will be a delay before the gate closes (this is the gate close delay at the DC). On the DCC you will see the status backlit and once the off command is received the delay off countdown timer will start. The small number counting down under the gate number is the delay before the gate close command is sent to the GC (this allows the DC time to wind down and let the gate close).