Water Level Gauge for the "Sump" Below my Observatory

[Gina]

Following the failure of the bilge pump (or "sump pump") that is supposed to pump water out of the hole under my observatory, I have decided I need a water level gauge and alarm to warn me when it fails.  This only applies to times of exceptional rainfall but that seems to be happening ever more frequently in recent years   I won't go into all the details here but suffice it to say that drainage is not an option.  Hence an automatic bilge pump as used in boats, pumps water up and out of the hole, over the bank and across to a ditch.

Automatic Submersible Boat Bilge Water Pump 12V 750GPH w/ Float Switch.

And this is the sort of thing I'm thinking of to measure the water level.

I don't need to know the water level to the nearest millimetre - something like 4 or 5 states should suffice.  The diagram below shows the relevant levels plus a graph of how water level might change with time, with the pump working.  Turning on at the Pump Start level and turning off when the water has gone down to the Pump Stop level.  Further up is the level of the bottom of the thee main support beams on which the joists sit with the floorboards on top of them.  The ground level at which water overflows from the hole is a few inches above the bottom of the beams.

Thus the states I would like to indicate are :-

1. Bottom to Pump Stop (or slightly above)
2. Between Pump Stop (or slightly above) and Pump Start (or slightly below)
3. Well above Pump Start - indicates pump failure
4. Above Bottom of Beams - indicates pump failure and observatory timbers at risk

And in addition I will have a light that shows when the pump is operating.

This leads to the following diagram.

[Garethr]

Gina.

Looks a good solution

As ever a practical way of dealing with a serious issue

1st new project of the year.

Gareth

[A40farinagolf]

Gina,

Looks like a good idea, (hopefully I'm not teaching you how to suck eggs) buy you'll need to factor in the "rate of fill" compared to the "pumping capacity" otherwise the high level and high, high level limits could be reached.

Also there is a PFD (probability of failure on demand) figure that should be considered, this is the chance of the system failing to operate for whatever reason and what the consequences and  tolerability of the failure are.

In the chemical process industries we sometimes have dual pumping systems that are totally independent including separate electrical supplies but I think that's a bit overkill.  

Would be good to hear how you get on.

Regards,

Mark

[Gina]

Thank you Gareth and Mark

I don't think the rate of fill has got anywhere near the rate of pumping out but without a water level gauge I can't be sure.  Also, the pump has failed a couple of times so TBH I really don't know.  The consequences of pump failure are not too dramatic - the water overflows into the surrounding area well below actual floor level.  It's just the bottom couple of inches of eight inch high beams that sit below the water level in these circumstances.  These beams had several coats of real creosote when I was building.  All the same, I want to keep them dry - no point in DPC if water rises above it    Hence this latest project - to provide visual indication of the water level beneath my observatory and maybe audible as well if the water reaches timber hazard level.

[Gina]

Before I can do any actual construction I need to look at and measure up the actual space around the pump where the ground is lowest.  I think I may have to provide a reasonable sized hatch to make the job practical.  ATM I have two overlapping holes of around 5" cut in the floorboards which only provides a hand hole and view of the pump, adjacent pier base and surrounding ground area.  I think I need to take my circular saw to the floorboards and cut a hole something like a foot or 18 inches square.

[Zakalwe]

Depending on the dimensions of the sump, but would a vertical water level gauge not be better?

http://www.ebay.co.uk/itm/Vertical-Stainless-Steel-Pool-Tank-Water-Level-Liquid-Sensor-Float-Switch-100mm-/291497127187

[Skipper Billy]

The item that Zakalwe links to 'looks' like a simple on/off switch - a visit to your local car breakers might get you a petrol tank sender unit which will give you full scale output rather than on / off and its recycling!

[Gina]

There's a joist 8" to the right (south) of the current hole and a noggin 15" to the west (bottom of photo).  I can cut down to the joist immediately to the left of the current hole and to the right.  A noggin crosses the hole but the next one is 15" the other side of the hole and I think that may be too far so I may add in another to support the west side of the hatch.  The brown piece of wood across the hole has been pushed aside from its normal position to allow access - normally at right angles to the noggin and held in place by wood screws through the pieces I cut out.

Here's a couple of photos.  The hole and the view through it.  The ground below the hole is almost dry ATM.  The noggin to the left side of the hole (to the north) is just visible through the hole.  The ground level is 18" below the floorboards.

[Gina]

Depending on the dimensions of the sump, but would a vertical water level gauge not be better?

http://www.ebay.co.uk/itm/Vertical-Stainless-Steel-Pool-Tank-Water-Level-Liquid-Sensor-Float-Switch-100mm-/291497127187

This is one option I have considered and may still be my choice once I get to see the "lay of the land".  Not just a single switch but a series as I said above so that I know what's going on.  A problem I envisage with a vertical gauge is the possibility of mud gumming it up though it might be possible to keep the guides above worst water level.

[Gina]

The item that Zakalwe links to 'looks' like a simple on/off switch - a visit to your local car breakers might get you a petrol tank sender unit which will give you full scale output rather than on / off and its recycling!

Yes, it's a simple switch.  I've thought about a petrol tank sender but I think I'm dealing with a greater depth of water than petrol in a tank though I guess the float bar could be lengthened.  Again, still a possibility but could be more difficult to adapt to remote display 20m away than building my own.  Nearest car breakers is 5 miles away - another consideration.

[Gina]

I have several things to decide :-

1. Lever or vertical mechanism
2. Magnetic or optical sensor (or possibly something else)
3. Four or five set levels or higher resolution

By magnetic sensor I mean a magnet and series of reed switches or hall effect devices.  This is relatively inaccurate as the magnet strength is difficult to determine.  So I would tend to favour optical.  The encoder would have to be above water level to avoid contamination with mud.  A Gray coded encoder could be 3D printed in ABS and damp-proof.  The thing about resolution is that with 4 or 5 set levels these will need to be measured and that won't be easy.  OTOH a Gray encoder could easily have 4 bits giving 16 equally spaced levels and I could see where the pump starts and stops by watching the dial in very wet weather. 

I can work out where the bottom of the beams is viz. bottom of hole is 18" below floorboards, beams are 8" high and joists are 5" high so bottom of beams are 13" below floorboards and 5" above bottom of hole.

[Gina]

Taken some more photos.

Towards the nearest beam.

Towards the south side showing the outflow pipe and just showing a joist.

Access hole showing joist.

[Gina]

Having taken a more prolonged look under the observatory floor in the region of the bilge pump I think a vertical water level sensor would be more suitable than a lever type.  There is a small area beside the pump that is the same level as the ground the pump sits on.  I have some measurements so may be able to design a float water level sensor.  Another possibility might be an ultrasonic distance sensor but I need to do more research on that.

1. Depth of hole below floorboards = 18"
2. Height of bottom of beams = 5"
3. Estimated water level above bottom of beams for overflow onto outside ground = 2"

So I reckon the water reaches about 7" deep and distance from this level to floorboards = 18 - 7 = 11" and that a measurement distance of 8" would well cover it.  That leaves 10" above the bottom of the float to the floorboards - should be plenty to house the works even taking off the height of the float.  I have also Googled the bilge pump details and found the start/stop water levels.

I think it would be a good idea to mount the pump on a piece of wood or something and provide a solid mounting.  Also, a relatively simple way of removing the pump for maintenance such as cleaning the filter and general servicing is needed.  The original installation was a rushed job to get water out from below the building.  Now is a good time to revisit the installation and make a proper job of it.

[Stub Mandrel]

Ultrasonic sensor is cool and plenty of Arduino-ready ones out there.

[Gina]

Now I have the measurements I've drawn up a modified water level diagram with dimensions to scale.  Ive shown two options - discrete set levels shown coloured and a 3 bit binary representation decoded to decimal in the right hand box.

[Gina]

More Googling has brought up an ultrasonic sensor for distance measurement.  HC-SR04 Ultrasonic Range Finder Distance Measuring Module Sensor Arduino

Working range is 2cm - 450cm ie. 20mm to 4.5m with 3mm accuracy.  Sounds ideal   Think I'll get one and connect it to an Arduino and play with it   Don't know if it's suitable for damp conditions though.

[Gina]

Seems it's fine for water level measurement - many examples on the web by Googling including the Arduino forum

[A40farinagolf]

Gina,

Seems to be taking shape.

Other thing that needs to be considered are what action (by someone or something) is taken when the pump failure warning is detected and how much time is there to carry out the response before the bottom of the beams from become submerged.

Also what happens if the alarm occurs during the night or when you are away from the property?

Sounds like an automated 12v back up pumping system is required so that it will work during a power cut- process control can become very expensive - bit like astronomy

Keep up the good work.

Regards,

Mark

[Gina]

Thank you Mark

You make a good point.  I've been considering installing a second pump and running this off the main observatory battery backed power supply is a good idea.  The second pump could be positioned above the main pump so that it isn't run unless the primary pump fails.  Having a second pump would give me time to fix or replace the primary pump and I wouldn't much like to venture out in torrential rain to try to fix a failed pump

A bit later...  Decision made and I've ordered another pump (same type) - due to arrive in a couple of days.

[Gina]

Took the circular saw to the floorboards and cut out a square hole.  Also, took the pump out, disconnected it and took it indoors for a good clean.  Covered in a thin layer of mud but nothing untoward.  Have put it back in position with a lump of rock on top to hold it down prior to making up a proper mounting and now going out again to reconnect it.

Photos to follow...

[A40farinagolf]

The 2nd battery powered back up would also kick in if the fill rate is higher than the pumping capacity of the mains powered pump thus adding to the extract rate.

Keep suction and pressure drop to a minimum by:

- keep any pipework to / from the pumps as short as possible and as straight as possible,

- the bore of the pipework should be smooth rather than rough to reduce turbulence

- pipework should be larger diameter rather than smaller (not too big as the pump might not prime effectively though)

- reduce elevation differences to a minimum 

Have you thought about trying to achieve a syphon effect so that the discharge flow is used to suck the water out?

Regards,

Mark

[Gina]

Just the one photo showing pump back in hole and connected up.  Dry test fine.  I've tucked the terminal strip connecting 3 core cable to pump wires out of the way above the pier base for now - later it will be screwed to the mounting framework when I've made it.

[Gina]

Checked the observatory before I went out this morning and the pump was gurgling nicely every few minutes.  And no wet got in   The ultrasonic sensors came this morning so I can connect to an Arduino and see how well they work - when I get some time - I have a roll off roof to check out to leakage.

[Scooot]

Hi Gina,

I have a small sump hidden underground next to my garden pond to catch the overflow. The pump is operated by a normal swing float switch which kept getting stuck because of lack of space. The float is now strapped in the on position and I use a vertical fish pond magnetic switch to turn it on when it reaches a certain height and off at the bottom. One if these

http://www.ebay.co.uk/itm/Fish-Tank-Liquid-Water-Level-Sensor-Vertical-Float-Switch-27-5cm-Long-/191688500289?hash=item2ca1866441

I installed it some time in 2014 and has worked well ever since. My neighbour whose an electrician wired it for me which I think uses a relay but for someone that knows what they're doing I think it's quite straightforward. I saw another vertical switch suggested early on which I'm sure would work just as well but I thought I'd just show you this one.

[Gina]

Thank you Richard   The pump is already equipped with its own float switch.  I want to know when it fails and how high the water is so that I can service it and avoid the support timbers "getting their feet wet".  If I can keep water below the top of the foundations, the DPC will keep the timbers dry.  I'm going to provide a backup pump but I would still like to know what the water level in in wet weather.