-
My house has a central vacuum system. It consists of a big motorized suction unit, connected to multiple inlets throughout the house via a network of pipes. Along with the pipes is a pair of signal wires; when the homeowner plugs into one of the inlets, the wire pair is shorted together, and this activates the suction unit (via a relay within the unit). I've experienced a malfunction where the signal wires get inadvertently shorted together, causing the suction unit to run indefinitely, putting a lot of wear on the motor and perhaps even presenting a fire hazard. I thought it'd be amusing to create a little interface between the signal wire pair and the signal inputs to the suction unit, which will shut the unit down if it runs continuously for an unreasonable period of time.
See the circuit below." alt="" class="bbc_img" />
Since the relay within the suction unit operates on 24vac, I thought it best to isolate my interface circuit from the suction unit using an add'l relay, operating off DC; when activated, this relay shorts the signal inputs to the suction unit. The 227-ohm Rcoil represents this relay. The LM311 consists of a comparator whose output drives the base of an NPN transistor; when the comparator diff'al input cmpH/cmpL is positive, the transistor is turned on and activates the relay. So the crux is generating the desired voltages on cmpH/cmpL. (Note the comparator is inverting, so the cmpH/cmpL inputs are swapped on the inputs to the LM311).
When the unit is not in use, the big cap C1 on cmpH is charged to near Vdd by R1, R2, and D1 (R5 is unnecessary ?); cmpL is guaranteed to be driven to to a slightly higher voltage (there will be a small current flowing through R2, thanks to R4) and thus cmpH-cmpL is negative and the system is off. When the user plugs in a hose and shorts the signal wires together (this is represented by switch SW), cmpL is pulled to approximately Vdd/4 and cmpH-cmpL goes immediately positive and turns on the unit. C1 begins discharging through the 1 megaohm resistor, and when cmpH drops below Vdd/4, the relay is deactivated and the system turns off. The time constant of R3-C1 is approximately 1000sec, or about 15 minutes. Once the user disconnects the hose, cmpH is recharged in about 1sec and the system will again turn on when the hose is plugged back in. The whole thing is powered by a small 5vdc wall wart.
I don't see a simpler way to do this. Obviously the time-out period is imprecise since typical capacitor tolerances are 10-20%, but the goal is simply to insure the suction unit doesn't run for hour(s), so this shouldn't be a problem. It does raise the interesting question of whether there's an easy way to measure time accurately (within a few %) using analog components.
-
I think this is a perfect case for a micro, like maybe the ATTINY. Trying to get analog timing of 10's of minutes will be rather difficult, and likely will drift over time as electrolytic caps degrade. The code required should be just a couple lines of c code.
Jon -
I would go with a time-delay relay.
-
Yeah 1M and 1000µF is not a good combination.
Big caps have leakage or will develop leakage.
In the olden days they used to do long delays with a higher frequency RC oscillator.
Then they'd stuff that into 8 to 14 flip flops to make a long delay.
It was much more reliable and accurate.
Me? I'd go with the ATTiny, but that's because I'm all set up to do that.
Still, it's a good idea.
How long would you set the timer for? 15 minutes?
(Since you'd probably move the hose to the next room.)
Still, what if it sucked in a plastic bag and the doorbell rang.
You could have the thing straining for 15 minutes.
How about rethinking the problem?
You don't care so much about how long it's running, you really don't want the thing straining itself.
You could monitor the current on the motor and make sure that it isn't at peak for over one minute.
A current transformer would work fine for that.
You could monitor the vacuum and make sure that it isn't at peak for over one minute.
My favorite pressure sensor? Go to the auto parts store and ask the oldest employee there:
"I want the least expensive manifold pressure sensor that you sell."
I bought one for a project, 3 terminal, +5 V, Gnd, linear output.
-
Digital probably is the way to go, I was just spitballing about whether it could be done analog. The ATTiny would still require at RTC, no ? An oscillator into a counter (chain of flip-flops), or just counting cycles in the ATTiny (or some minimal Arduino), would certainly be no more precise than a capacitor - at least notwithstanding the issue of leakage increasing with age.
As far as it straining itself, when something like a plastic bag sucked in ... After the incident that provoked this project, in which signal wires shorted while I was not at home and motor ran for hours, I called the manufacturer. They said that, contrary to intuition, when the thing was sucking against closed inlets, so no air flow, it was actually working LESS hard than under normal conditions. Because the power exerted is flow * pressure differential, right ?, so when there's nearly zero flow there is very little power. That's why you hear the motor running faster under these conditions, because it's doing less work; so it's not actually straining, in fact quite the opposite.
-
Digital probably is the way to go, I was just spitballing about whether it could be done analog. The ATTiny would still require at RTC, no ?
The internal CPU clock is certainly good enough for this type of application.
Jon -
They said that, contrary to intuition, when the thing was sucking against closed inlets, so no air flow, it was actually working LESS hard than under normal conditions.
Oh, that makes sense.
Then the vacuum sensor makes even more sense.
My project needed positive pressure so I had to buy a MAP sensor for a turbo engine.
You only need a MAP sensor for a normally aspirated engine.
A RTC can tell the difference between day and night.
We don't need that, only to measure relative time.
Very often these little µP's count ticks at every millisecond.
So, just count the ticks until you get to 0x36EE80 (3600000) and you have an hour. -
Put a temperature sensor on the motor and trip a protection relay if it goes over. Try measuring motor temperature during several normal vacuum use cycles to get an idea of what would be a good trip point. (Make it adjustable and start at, for example, 10C above the maximum observed. You can always increase it if you get false trips.)
-
I would go with old-school analog controls; they are far more reliable and less resistant to EMI --- a brushed motor generates a lot of noise, and especially if the signal wires are long, they will pick it up very well.
If the switch wires draw enough current, you could perhaps use a bimetallic relay. -
A clever electromechanical way to do it would be to have a mechanical thermostat in series with the normally closed contact of a relay and the coil of that relay across the combo. (The coil current of the lockout relay must be less than the holding current of the main relay/contactor.) Then when the thermostat opens, the lockout relay opens its contacts and won't allow it to restart merely by cooling down - have to cycle power or open/close the control circuit to do that.
I would go with old-school analog controls; they are far more reliable and less resistant to EMI --- a brushed motor generates a lot of noise, and especially if the signal wires are long, they will pick it up very well.
Some higher end central vacuum units use induction motors. Because they're fixed to the building, the size and weight are far less important than for normal vacuum cleaners.
-
A time-delay relay would certainly be perhaps the simplest solution.
Here's one that might work. Minimum order is 100 though :-(
-
We had one of those central vacuum things in the house I grew up in, it's a neat idea but in practice it's a pain in the ass. The hose is bulky and heavy and there are always places it isn't quite long enough to reach. If you want a vacuum head with a beater bar you need somewhere to plug in power for that. Then if you have small kids it's common to get toys stuck into the sockets where they clog up the system. Eventually my parents abandoned it and got a regular upright vacuum, a bit noisier but overall it worked better.
We never had a case where the control wiring shorted though, I think I'd look at solving that problem rather than trying to make a workaround, although I agree that a time delay relay is probably the easiest approach. Just have it shut off after ~30 minutes and wire it inline to the motor. -
A common problem is the central vac motor failing and burning out. A brand I won't mention used undersized cheap chinese motors that roast, the replacement is a larger motor imagine that. Or people put something blocking the (top) cooling air intake.
OP for the amount of work in your project, I would think about some coverage for that, perhaps using a CT to look at motor current as well as trigger the run timer. So sensing the motor operating (current) rather than the wall plate switch, or looking/including stator temperature. Or at least see if your vac has a thermal fuse or some protection, as many people just notice the stink to alert them to a problem.
The LM311 has about 0.1-0.25uA input bias current, LM555 is 0.5-0.9uA and a CMOS TLC555 is 0.01uA which is much better for long delays, but the large electrolytic capacitor leakage still gets in the way. Cap datasheets are usually leakage current "I<0.01 CV or 3μA after 2 minutes, whichever is greater".
For 30 minute delays, maybe consider going digital with a counter like CD4060, or use a microcontroller. -
We had one of those central vacuum things in the house I grew up in, it's a neat idea but in practice it's a pain in the ass. The hose is bulky and heavy and there are always places it isn't quite long enough to reach. If you want a vacuum head with a beater bar you need somewhere to plug in power for that. Then if you have small kids it's common to get toys stuck into the sockets where they clog up the system.
Well, most of things don't apply to my situation, and anyhow it's already there. I've heard the most common issue is shoddy installation and thus poor suction (due to leaks and poorly-designed routing of the. pipes, I suppose).QuoteI agree that a time delay relay is probably the easiest approach. Just have it shut off after ~30 minutes and wire it inline to the motor.
I'm trouble identifying a relay that does the right thing. -
A common problem is the central vac motor failing and burning out. A brand I won't mention used undersized cheap chinese motors that roast, the replacement is a larger motor imagine that. Or people put something blocking the (top) cooling air intake.
Yeah, maybe detecting overheating is a better approach. I doubt the motor current rises if it runs way too long.
OP for the amount of work in your project, I would think about some coverage for that, perhaps using a CT to look at motor current as well as trigger the run timer. So sensing the motor operating (current) rather than the wall plate switch, or looking/including stator temperature. Or at least see if your vac has a thermal fuse or some protection, as many people just notice the stink to alert them to a problem.QuoteThe LM311 has about 0.1-0.25uA input bias current, LM555 is 0.5-0.9uA and a CMOS TLC555 is 0.01uA which is much better for long delays, but the large electrolytic capacitor leakage still gets in the way. Cap datasheets are usually leakage current "I<0.01 CV or 3μA after 2 minutes, whichever is greater".
I thought about digital before. Then I found this LTC6995 long-interval timer and started a new thread (spacing out and forgetting I started THIS thread back then). Kinda dumb. Maybe I should ask moderator to merge them.
For 30 minute delays, maybe consider going digital with a counter like CD4060, or use a microcontroller.
-
You could also insert a power switch you need to manually enable when you want to use the vacuum. Or a timer relay that enables power to the vacuum system for, say, an hour. Could even be Wifi enabled.
-
You could also insert a power switch you need to manually enable when you want to use the vacuum. Or a timer relay that enables power to the vacuum system for, say, an hour. Could even be Wifi enabled.
Good idea. In a similar vein, I do turn off the circuit-breaker that feeds the thing, when I go out of town - along with those for the well pump and the water heater.
The reason I'm pondering this again is that I've installed a system in a vacation house. So worst-case, the thing could be running for weeks before being discovered. So turning it off breaker is definitely part of going-home checklist.
But, all this being said, it'd be more fun (we are EE geeks, after all) to have a circuit which shuts the thing down after some reasonable time. I proposed one in the other thread I started on this.