Air compressor left on

[rustin.atkeisson]

Hey guys,

When you are done for the night, PLEASE DO NOT FORGET TO TURN OFF THE AIR COMPRESSOR.

Checking that the air compressor is off should also be something members do when shutting down the building each night.

When @squarenuts came in this morning, the air compressor was running at full tilt since one of the air lines sprung a leak. He immediately shut off the air compressor that was getting untouchably hot.

This situation could have easily meant we needed to replace the $1500+ air compressor again if it had been left on much longer.

As I have said several times during the metalshop safety class: I am not the best plumber in the world. What that should translate to is: do not trust that the lines will hold together when you are not there.

We were able to find the leak and get the system back up and running, but I will be keeping a closer eye on the compressor to make sure that there wasn’t any damage done.

Here’s how the compressor should look when you’re done for the day.

Here’s where the damage was and what happened to the little blue pipe.

And to reiterate: PLEASE DO NOT FORGET TO TURN OFF THE AIR COMPRESSOR.

…I will be posting signs and instructions soon on how to shut off the air compressor. In addition, we are also looking into installing a timer to make sure the system gets turned off after a couple hours.

[whateg01]

I’ve been working on a design for a control for my air compressor that you turn on, it times out in an hour, and if you still need air, you turn it back on. The thing that is holding me up right now is I want to include a 12 hour override for times when I might be doing something needing air all day.

Somebody is going to think i manufacture contactors, but using a contactor to control that, you could use a timer like that on the exhaust fan.

[rustin.atkeisson]

It sounds like home Depot does have a 220v timer that might work. The timer would have to be wired in between the tank and the wall, but at least it sounds like a quick and easy protection for the air compressor.

At the same time, I’d rather not add an extra step to the process of running the machine, so I might check with our amazing friends at Ingersoll Rand to see if they have any recommendations to replace the on/off switch with a timer switch.

[whateg01]

Not sure I follow you. Running what machine? What is the longest most people use any machine? Maybe 3 hours? So set it to run for 3 hours and then require a reset. In general, that should make it easier to run machinery as it will remove the step of turning off the compressor.

[rustin.atkeisson]

I think the tormach is the one machine that would run the longest. Everything else seems to be used for less than an hour.

My concern is having two inline ‘on’ switches for the machine. It’s an extra failure point in the system. I’m a bit concerned about the old switch being left on, and then the timer failing in the on position like I’ve seen happen with older mechanical timer switches.

It’d be nice if we could find a digital timer switch that would fail in the off position. …and something that doesn’t involve reed switches like when we fixed the switch on the Logan lathe.

[whateg01]

While it does happen, contactors welding themselves closed is not a problem I have experienced in any of my implementations. Part of that is that I do use higher current rating than is required. using a contactor, much like the dust collection in the woodshop, multiple items can control the power to the machine. If the Tormach is running, it can keep the compressor turned on even if another timer has turned off. Short of relying on people to manually turn off the compressor, the circuit has to get more complex to accomplish the goal of making sure it gets turned off when not in use. My preference is for control circuitry to all be low current, so the timer would only have to deal with the coil on the contactor. The contactor would be the real workhorse. I would much rather rely on a circuit built that way than a COTS 220V timer rated for 20A and used near it’s max current rating.

[james.a.seymour]

I see it typically when switching a capacitive load. The instantaneous current is super high and welds the contacts. The motor is an inductive load in general but may use a starting cap depending on motor topology. You can probably use a ptc thermistor to mitigate that.

[whateg01]

Inrush current on a motor can be many times the rated current of the motor. In the case of a compressor with an unloader, in theory, it should never start under a load, but even without a load, there is inertia of the crankshaft and all of the other parts of the rotating assembly.

Thinking back on the Logan lathe and the welded contact on it, I suspect that somebody may have stalled the motor by trying to take too big of a cut, or maybe tried to start it with the gear train/back gear configured such that the spindle was effectively locked. That excessive current, if going through a device not rated for it, can cause a welded contact. Of course, with anything electronic or electrical, there are always going to be cases of “just a bad part” but more often it’s something like this that causes failures.

[squarenuts]

I was given this

Courtesy Rich st. Aubin

[whateg01]

I don’t know what “spring relay” means. Every relay that has a NC and/or NO contacts has a spring in it to force it back to that position when not energized. Same goes for the contactor. Unless there is voltage applied to the coil, it has a spring to push it back open. I’m not saying redundancy is a bad thing, just that any kind of relay is going to do that. In reality, a contactor is just a relay designed to handle high amounts of current. They aren’t magical.

[squarenuts]

I think that’s just the name grainger was using

[jeff]

Just a lazy idea, but why not skip a timer or complex design and just have a signal wire pulled off the light switch. When the light go off they disable the equipment you don’t want left on when no one is around. It would fail (to protect the air compressor) if members leave the lights on, but a motion sensing light switch could be the fool proof solution to that secondary failure point (of safety system)….

[whateg01]

The compressor provides air to multiple spaces. Woodshop and both metal shops are fed air. Tying it to the lights would require somebody to turn on the lights in the hot shop. As we already have problems with people not checking windows and doors, I expect this would lead to somebody leaving the light on in the space they aren’t in, then somebody else (last person) seeing the light on and assuming somebody is still there, thus leaving the alarm not set.

[Brian]

In my working life as a building automation engineer, we often made attempts to idiot proof control systems on everything from aircraft paint booths to chiller plants to Bobs office…
The idiots always won. Training and retraining was the best we could hope for with contact information on who to call for help was generally the best practice.
Try not to automate in unforseen problems while automating out the problem of the day.
Best of luck.

[whateg01]

I fail to see how a simple timer plus an input from the Tormach and maybe the plasma table is over-engineering it. Aside from somebody figuring out how to automatically restart the timer, I also don’t see how this would not solve the issue of the compressor being left on. If the default state is “off”, barring failures, after a fixed amount of time and Tormach and plasma table being off, the compressor would end in the off state.

[jameslancaster]

It sounds like we need to make a relay, and minimal current sensor to monitor and turn it on/off, for that and the woodshop. At that point, Why not add a card reader as it’s easy, as a check if they should be able to, since I know people have been complaining about not having that.

If we can get the high voltage side which can take an enable signal (Frankly, it should could probably just be a relay-transistor-optoisolator on one side) and a current sensor like: Non-Invasive Current Sensor - 30A - SEN-11005 - SparkFun Electronics (One of the few in stock) as a check it can be setup to:

Authorize someone and turn it on/off.
Check voltage
Report status via something like MQTT. (Very simple messaging protocol, often used for weather reporting, but it’s bidirectional so it could potentially allow for turning off remotely)
Temperature sensor also reporting via MQTT.

I’d base it off an ESP8266, as I’ve done most of that (minus the switching) before on that platform, so it’s simple enough. At this point, I’d make it monitoring only though.

Quick parts list (est prices in packs, not necc including shipping):

• Esp8266 mini D1 ($2-3)
• Sparkfun current sensor ($11)
• BMP280 or similar sensor (Also a pressure sensor, and reads a bit high depending on how often it’s checked.) ($2)
• Card reader ($2)
• Wire
• Relay ($20?)
• Optoisolators (Not 100% technically necessary, but a good idea)
• PCB (simple enough to make or order.)
• 5V or 3.3V power supply. (Can use USB power supply + micro USB as input)

If we want to get the current sensors & relays, I think I can make a few prototypes. From what I’ve got on hand.

I’ve made a simple schematic,

Quick notes: Top ‘resistors’ are there only as switches/clamping/construction of a resistor divider, because the sensor’s documentation is iffy on which pins are which on the connector. I2C has another header in addition to the one for the temp sensor.

Left upper header is for a CP2102 debug, with reset (unpopulated by default) if needed, same with the lower header. Lower for a cable like this one: USB to TTL Serial Cable - Debug / Console Cable for Raspberry Pi : ID 954 : $9.95 : Adafruit Industries, Unique & fun DIY electronics and kits or serial communication to another device, (or just 2 IO pins & resistor switch power.)

D0 would be used as the signal activation signal. D4 is the onboard LED already, so it could be used later on as an LED. (Though due to boot modes one must be careful, might need a pull up resistor.)

I might add an optoisolator on the output header, but I haven’t found one I’ve got atm. (Which is mostly why it’s not laid out yet.)

If it works as planned, it could also be used to automate the dust collector to auto turn on, but we’ll see

Also, for the amusement: Going back I just found out the pinout via this project, but not the listed datasheet or other document that I saw… sometimes I love sparkfun documentation, sometimes I hate them, but it’s also almost what I’m looking at doing with the bmp280 even (Great minds or something): Environmental Monitoring with the Tessel 2 - learn.sparkfun.com I’ll go back later and now that I know, that should be much easier.

Eh, I just did it now.

@james.a.seymour and others: Thoughts?

[ladeana]

I love tech and had for a long time lights that would turn out when I left the room. Sometimes I wouldn’t move enough and it would turn out the lights. I fear that lights going out when someone is using a machine would be a safety issue.

[mikeb]

The R? on the second schematic across the current transformer needs to be a bleeder resistor able to withstand ~3kV. That also implies standard precautions on PCB, for spacing and slots to reduce arc over. Current transformers are conceptually like a step up transformer (usually effectively 1/2 turn ‘primary’ and 100-300 turn secondary). Perceptive readers doing the arithmetic might wonder why the requirement isn’t for up to 600 times line voltage. It’s a very poorly coupled transformer with high series resistance in the secondary and only microamp output current. It’s sometimes recommended to split the load resistance in half and directly solder to an inline (not PCB mounted jack) for redundancy and physical isolation. Regular, 600V rated vinyl wiring isn’t recommended, nor is using a flame to shrink any heat shrink tubing.

[whateg01]

I really think for the air compressor, a simpler circuit would do. The dust collector will require some smarts though.

@ladeana if the air compressor turns off, it’s not going to be instantaneous catastrophe. The pressure will just drop below whatever threshold the machines require to operate and they’ll stop. I suppose it could be setup like the car wash where it beeps at you when you have 10 minutes left so you can go restart the timer.

[jeff]

Would there be any issue pulling a signal wire from each area that the air compressor serves? With diodes or something you could isolate the signal to not power the lights in other areas, if that makes sense. Just an idea.

If I were more knowledgeable with electronics, a device that monitors equipment and could notify via the internet would be really nice… I don’t know the terminology, but we bought a cat feeder that can tell me when we are low on cat food and automatically feed the cat. That idea but for equipment…