Maker Pro
Maker Pro

Timer relay, to turn pump on, and then off

I see there's a somewhat similar thread on a timer relay, but it is different enough for me to ask this.

I need to replace some failed electronics that turn on a water pump. When water moves through a valve in our hot water supply line, the valve acts as a switch that closes a circuit. The valve has two low voltage wires that connect to the old failed pump controller.

I want to take those wires and connect to a timer relay that will complete a 110v circuit and turn on the pump for 2 minutes and then turn off until water flow triggers the switch again. I have plenty of experience with household electrical, but nothing of this more delicate sort.

Ideas? Thanks in advance.

(very cool forum, btw)
 

KrisBlueNZ

Sadly passed away in 2015
Hi there and welcome to Electronics Point :)

I'm afraid I don't quite understand your question.

You have a valve in your hot water supply line. As I understand it, valves control the flow. They don't detect flow, and they don't control circuits. Do you mean it's a switch that detects flow?

I think that's what you're saying. You have a switch that detects when the hot water is flowing, and you want that switch to trigger a relay that switches 110V AC through to a pump. You want the pump to run for two minutes. Is that right?

So do you want the pump to continue to run while there is flow, and then to run for another two minutes after the flow stops?

What low-voltage DC power sources do you have? Is there a 12V DC power source available?
 
Thanks for your reply! Yes, the valve is a switch that triggers the pump to turn on. The (failed) existing control box sits between the pump and wall 110v, and the valve switches the power on via a timed relay.

Ideally, the pump would turn off 2 mins after the initial trigger, not when the valve closes (e.g., water is shut off). That's because once hot water has circulated around, it no longer needs the pump. But I think it would be easier to have it start the countdown once the valve closes, otherwise it would need additional lockout logic to keep it from turning on again immediately since the valve is still open.
 

KrisBlueNZ

Sadly passed away in 2015
It's possible to do it the way you want. When the switch initially activates (flow starts), the pump runs for two minutes then turns OFF. The pump will only be retriggered if flow stops, then starts again. Would that be your preference?
 
Yes, exactly. Taking it a step of complexity further (wish list!), it would require 2 seconds of closed circuit from the valve before switching on, because often cold water use surges in the lines and pushes flow for a moment, opening the valve. It then runs the pump needlessly for the interval. But that's a minor concern.
 

KrisBlueNZ

Sadly passed away in 2015
OK, here's my suggestion. The circuit is powered from 12V DC and drives a relay whose contacts can switch the 110V AC to the fan directly.

271877.001.GIF

The circuit is based around two ICs from the CD4000 CMOS family. U1 is a CD4093B which contains four identical "gates" called NAND gates with Schmitt trigger inputs. They are shown as the four D-shapes in the diagram, with two inputs on the left and one output on the right.

One input of each gate is tied to VDD. This makes each gate operate as an inverter with a Schmitt trigger input. An inverter simply drives its output to the opposite state to its input: if the input is low, it drives its output high, and vice versa.

I won't get into details in this description as I've explained it all several times already - if you're having trouble getting to sleep, read my tutorial posts at https://www.electronicspoint.com/threads/led-flasher-help.263649/page-2#post-1580565 and https://www.electronicspoint.com/threads/led-flasher-help.263649/page-2#post-1580565. Only a few parts are directly relevant but you can learn a lot from them.

U2 is a CD4024B binary counter/divider IC and it provides the two minute timer for the pump control signal.

Your flow switch connects to CN1. It must provide a closed circuit when there is flow, and an open circuit when there is no flow.

Normally, R1 pulls the voltage at its bottom end up to VDD, the positive supply rail. This is considered a logic "high" at pin 2 of U1. When the flow switch closes, it pulls that voltage down to 0V, which is considered a logic "low". R2 protects U1A's input from possible voltage spikes from interference etc.

So, U1A's output (pin 3) goes high when there's flow in the pipe. When this happens, C3 is allowed to charge up through R3. (When pin 3 is low, D1 holds C3 discharged.) If the flow sensor remains active steadily for about two seconds, C3's voltage will reach the rising input threshold voltage of U1B (on pin 6) and U1B's output will go low.

When U1B's output goes low, the RC pulse generator made from C2 and R5 will cause a brief low pulse on U1D's input (pin 12) and its output will go briefly high. This resets U2 and starts a pump run cycle.

U1C forms a Schmitt trigger oscillator (see the tutorial posts I linked to earlier) that oscillates (alternates high and low at a regular rate). The time period of one cycle is about 1.875 seconds (it can be set fairly accurately by VR1).

This frequency is used as a clock for U2, which is a 7-bit binary counter. If you're familiar with binary numbering, this IC generates a 7-bit binary number that increments (increases by 1) on each clock pulse.

The Q1~Q7 outputs are the binary number. When U2 is reset (which happens whenever the flow switch changes from OFF to ON and remains ON steadily for two seconds), the outputs reset to 0000000 binary. As the clock pulses come along, the outputs increment, to 00000001, then 0000010, then 0000011, and so on. After 64 clock pulses, the count reaches 1000000, i.e. the Q7 output (pin 3) goes high. This occurs after (1.875 × 64) = 120 seconds.

The Q7 output is named -PUMPRUN. This means that when it is low, the pump runs. (the "-" at the start of the name shows that it's an active low signal.)

When -PUMPRUN goes high, two things happen. First, D2 conducts and forces U1 pin 8 high. This disables the oscillator, so U2 remains stuck at that count. Second, the voltage to the base of Q1 goes high, and this causes the relay to open.

Q1 is used as an emitter follower. The voltage at its emitter follows its base voltage, with a slight voltage difference of about 0.7V which is not significant. When Q1's base is low, it pulls its emitter low as well and this applies most of the 12V power supply across the coil of relay K1, making it close and activate the pump. When Q1's base is high, the relay drops out.

Normally a diode would be connected across the relay coil to absorb the "inductive kickback" from the relay coil and prevent damage to Q1, but with Q1 used as an emitter follower, that diode isn't needed.

So when the circuit is idle, U2 contains a count of 1000000 binary; the Q7 output (pin 3) (-PUMPRUN) is high. The relay is de-energised, so the pump doesn't run, and D2 prevents U1C from oscillating.

When flow starts, and remains active for two seconds continuously, a short high pulse resets U2 to 0000000 binary. This activates the relay (because -PUMPRUN is now low), and enables the oscillator, and 2 starts counting up towards 10000000 binary.

I hope this makes sense. Try to understand those tutorial posts I pointed you to. This circuit is different in the details but a lot of the terminology and principles are relevant.


C1 and C5 are decoupling capacitors for U1 and U2 and must be connected as directly as possible to pins 14 and 7 of their respective ICs. Use ceramic capacitors for these, and for C2 as well (it's the same value).

I have given specific Digi-Key part numbers for C3 and C4, as well as two options for K1. Other components are not critical and can be by searching for their part numbers, or through the product guide.

Stripboard is an appropriate construction method for this circuit. The relay contacts are at mains voltage and must be thoroughly insulated from the other circuitry - remove a large area of copper around these to ensure good clearance and creepage distances.

VR1 should be set roughly half way. If this produces a pump run that's too short, turn it clockwise and try again. If the pump runs too long, turn it counterclockwise. The clockwise and counterclockwise ends of the trimpot are marked on the schematic.
 
AWESOME. Thanks so much for laying that out in such detail. I have read it through and mostly get it, and will read the tutorials and study up.

It'll take me a bit to digest, search for parts, etc. I'll likely have a follow up question or two. :)

Thanks again.
 
How does water flow in the beginning if it needs the pump to creat the flow in the first place?

I have seen circuits that rely on a pulse or momentry switch closure to initiate and then the flow switch keeps the pump running in bypass mode.
Safety flow switch if you like , not difficult to attach a run time then.
First bit is confusing though.............
 
How does water flow in the beginning if it needs the pump to creat the flow in the first place?
Magic. :)

It doesn't need the pump to flow; it needs the pump to get water there in a fraction of the time.

Water flows naturally from system pressure. The problem is that the supply lines snake all over under the house and through the walls. You have to run the hot water for 10 minutes until hot water arrives (wasting >10 gallons). By running a pump, even when a faucet isn't running, it moves the water through a loop (returning through a dedicated return line to the water heater), so it's always very close to the faucet, and wastes only maybe 1/10 of a gallon. The trick, though, is to not have the pump running all the time. That wastes electricity to circulate, and propane to keep all that extra water hot as it circulates around. The idea here is to have it switch on only when it is needed (a valve senses hot water flowing). Sure, it isn't instant hot at the faucet, but all you have to do is turn on hot water, turn it off, do something for maybe 60 seconds until the circulation loop is full of hot water, and then turn on the hot water again.
 

KrisBlueNZ

Sadly passed away in 2015
all you have to do is turn on hot water, turn it off, do something for maybe 60 seconds until the circulation loop is full of hot water, and then turn on the hot water again.
Would it be better just to have a pushbutton to do that?
 
Yes, if we installed one at every fixture. Some people do exactly that. The house is rather spread out, with the kitchen a long way from the bathrooms, which are then split from each other. So we would need at least 4 or 5 of the switches. I thought about using a keyfob but then we'd have to issue one to everyone in the house.
 
I've seen water circulation systems used in hospitals and the like.
There is a massive loss involved in heat transfer (even with insulation ) so not a very efficient system as far as saving a dollar.
It does maintain hot water at each outlet but in your application, when you turn your pump off the water will start to cool again so you seem to be back to square one.
Unless I'm missing something here....o_O
Then again if circulation is the high priority, you could take a look at the pumps used for solar systems (diaphragm type) only about 10w if my memory serves me correctly ..
If just a standard size house it should have no trouble keeping up and control at far end with a heat sensor strapped to the pipe..
Trouble then would be the need for a bypass.
All seems a bit like trying to get more mpg.
Maybe just insulate your piping, fit a heat sensor to the far end, a timer for times when everyone is away or asleep and go from there. At least then your existing pump will only run when it needs to.
 
Last edited:
Circulation is important because of the >10 gallons of water wasted down the drain, multiple times per day. The electricity used with on-demand circulation is minor, so the pump efficiency isn't the waste; it's the propane heat loss in the pipes once you shut it off. But that will happen no matter what we do.
 
Yes, well you kind of lost me now with reference to electricity as energy source and propane as heat loss.
By my calculations, 10 gallons in a 3/4" pipe would run 142 metres or 466 ft so it must be a long house you have there.
As I see it you would be better off with a system at each end of the house.
 
Sorry if I wasn't clear. Electricity is to circulate the water around the loop, ~90w. If it's only on when we need it on, when we turn on hot water, call it 20 mins/day, the energy use is tiny (90w x 20 mins = ~35 cents/ month). Propane is obviously heating the water, back at the water heater at one end of the house.

Correct about the calculation of footage for 3/4". The supply line is 1" under there, which shortens that down to ~250'. But even then, we (including multiple plumbers) count about 75' less distance even when we generously measure. We haven't been unable to figure out the discrepancy, nor why we have this issue. Bottom line is we do.

Two systems would be a much more expensive solution (as in 100x the cost), given the problem of tapping into the line, cutting out linkage to the other side, and locating the heater (even a small on-demand). Plus, we'd still be wasting a lot more water than if the pump puts hot water to the faucet via a circulation loop. This is So Cal, drought land.
 
Top