MG-Rover.org Forums banner

Alternative Expansion Tank Level Sensor

5K views 21 replies 10 participants last post by  MGJohn 
#1 ·
Following on from Incony's work on detecting water leaks on the coolant system, and my previous thermal imaging of the coolant tank, I have now completed testing an alternative sensor that I hope to use to detect low coolant level.
The sensor is made by Balluff, a well known automation sensor manufacturer, and was kindly given to me by the company's local rep as a free sample. They retail at about £40.
It is an ultrasonic device and will detect liquids through non metallic containers, using an effect similar to radar.
I have made a mock up system using a spare expansion tank and a spare instrument console (kindly given to me by Darren at Bitz).
The device works great on 12v, so it can be used directly on the car without any voltage conversion, however the signal is normally open (off) when liquids are not detected, so I used a relay to inverse the switched output of the sensor so it will light up a lamp when the sensor turns off.
I used the gearbox fault indicator lamp on the instrument console, as it does not have the low level warning indicator, being from a 2003 car. The instrument console on my TF is a similar unit, so I will probably wire this the same way when on the car.
When the coolant level drops to about 5mm below the low level line the sensor switches off and the alarm light comes on.
It is connected to the IGN terminal on the console for +12v and GND for 0v so the whole circuit switches off when the ignition is turned off, saving the battery drain that the sensor would cause.
As the gearbox sensor lamp is connected permanently to the IGN supply and switches 0v to light the lamp, I used the relay to switch a 0v to the lamp when the level drops.
Here is the sensor on the tank, sitting in a useful rig knocked up out of Meccano

This is when the coolant level is ok....note the green led on the sensor for power on and the yellow led for liquid sensed.....

Now drain a bit of coolant out with my turkey baster coolant tester...

And then when the level drops below the sensor trigger level, the warning lamp comes on.

The next step is to now fit on to my TF and hopefully the sensor will be able to stand the ambient temperatures around the coolant reservoir. The sensor is rated at 85 degC but the rep said it will take more, so it will be interesting to see how it performs.
The main advantages of this sensor is that if the sensor fails, the light will come on automatically, saying there is a problem, also the sensor clips onto the tank, so can be fitted without removing the tank and draining out coolant.
On the minus side, I have to get a cable to my instrument unit, which is not that bad a job really and I have to fetch out the unit anyway, as I have to replace the binnacle soon as mine has a few cracks in it.


Sundance
 
See less See more
4
#3 · (Edited)
Vibration and corrosion are not an issue, the device is fully solid state, and corrosion resistant too, resin packed with stainless steel connector. The connector cable is highly resistant PUR cable and will not be affected by heat or oil.
If its gonna fail, it will be by heat, but my previous thermal pictures tell me that it just might be ok.
The Balluff rep says if anything the heat may shorten the lifespan of the sensor and it may fail after three or four years service instead of twenty. At forty quid I can live with that. Now I need to fit it to my TF. I am also looking into ways to monitor the coolant pressure, like Incony, and have two possible ideas on how to do this, using a digital switch with a pressure readout incorporated in to it, and a pressure transducer which will send data to a bargraph disply. This could possibly double up as a check for coolant pump effective operation, if Incony's data is correct regarding pressure to RPM relationship.
(See Incony's thread on monitoring coolant loss). http://forums.mg-rover.org/mgf-mgtf-sponsored-mgfntfbitz-12/mgf-header-tank-leak-test-method-485338/


Sundance
 
#4 · (Edited)
"When the coolant level drops to about 5mm below the low level line the sensor switches off "

i sense the level will change due to incline (the angle of the car when on a hill or cornering and the effect of motion ) and pressure. since its a sealed system. i would want more than 5mm level control...?

you could add a delay... ie.. it doesnt trigger unless it stays at 5mm or more for defined time? thats more likely if the level stays low, when the header tank sits level and stationary.

bear in mind the level changes due to pressure... and temperature... and incline about 360 degrees... if one is looking just at level, its influenced by many things.. its not specific.. the level isnt just how much fluid there is..

think of how a bubble level works... and how the bubble must be in the centre to be level.. now add pressure, now add temperature.. now add motion..


?

so i sense you will be fitting it to the tank with some overall level limit based on those things.. ie they dont matter its the absolute limit that is important..

and i agree.. i has promise in that respect..

so 5mm below absolute zero... should work.. and you need to define the absolute zero.. :) not a problem .. but it will need more determining, else you will be plagued by false alarms.. each of which need you to see why the alarm is triggered.

of course you can link its output to other sensors and get an average... or a median .. ie.. level sensor says low... temperature says high, pressure is high... car is stationary and engine not running and level.. so.. the signal adds up on average that something is wrong..

i understand the process.. if a + b + c =< d.. tell me.. :)

if only a =< d then what?
 
#5 ·
I have thought about the time delay, I think the MG-Rover fitted units have 30 seconds delay time. As its only possible to fit the sensor below the centre rim of the tank, it naturally makes the sensor read 5 or 6 mm below the actual low level mark on the tank. This on its own may help emit erroneous reading from inclines or hard braking, I am thinking first of all fit it and see how it performs, then if needed, I can put a delay in. I have a few spare 555's knocking about, or may even have some dedicated timer relays in my spares boxes somewhere (handy being an automation engineer).
Now I am contemplating whether or not to drill a hole in the top of my expansion tank to take a 6mm push in bulkhead connector to connect to the programmable pressure switch or pressure transducer if I buy one.
Incony, what pressures did you max out on when you did your tests with the pressure gauge?
Incidentally the pressure switch I looked at only works above 15v, so I have bought a DC/DC step up converter to boost the supply for it to 24V.


Sundance
 
#7 · (Edited)
Qoute Sundance:
"Incony, what pressures did you max out on when you did your tests with the pressure gauge?"

i got up too late today ( first day off after four - 12 hour day shifts.. plus the the clocks going forward ) so i havent got the fittings i need yet to make it more permanent.. but ive seen 13 plus psi.. just trying it out.. and that was with a warm engine and just using the accelerator to increase the revs, but its very hard to do that and see the gauge in the boot.. and i didnt have the the throttle exposed, so i couldnt just rev it from the boot and see the rev counter :)


this isnt the post to discuss more about it really.. this is about level sensors.. so any further posts i make about the pressure sensor i will post here: http://forums.mg-rover.org/mgf-mgtf-sponsored-mgfntfbitz-12/mgf-header-tank-leak-test-method-485338/
 
#8 ·
From the original SCU Design Document.....

The SCU shall provide a warning to the driver if the coolant level falls below a safe level for a set period of time.
When the ignition is switched on, the SCU shall illuminate the low coolant warning light for 4 seconds as part of
an initial systems check. It will then monitor the input from the low coolant warning sensor and if this input is
short to ground (ie coolant level low) for a period of 40 seconds, it will illuminate the warning bulb. It will
maintain the illumination for as long as this input is low. The SCU will need to see the input open circuit (ie
coolant level acceptable) for a further 40 seconds before the lamp is extinguished - thus providing an
anti-slosh capability for the system.



.
 
#11 ·
Philhex2-
If a hose explodes at 70 on the motorway all your coolant will vanish in a huge cloud of steam. The honking of other amused drivers will help alert you.
Even if you can stop quickly in busy traffic the engine is going to be so hot the head is likely to fry anyway- just think how long they stay hot after a good run.
Better to have early notice of a slow leak (as they happen often) than hope for instant warning of catastrophe.

C'est La Vie for F/TF drivers.
 
#12 ·
Philhex2-

Better to have early notice of a slow leak (as they happen often) than hope for instant warning of catastrophe.
Surely it's better to have both.

I had a hose fail while on the M25, the first I knew of it was when I pulled off and was enveloped in steam, that's why I developed my own low coolant alarm.

This gives an immediate LED indication followed by a buzzer that beeps once at 5 seconds, twice at 7.5 seconds and continuous at 10 seconds.

Of the 2 hoses and 1 radiator I've had fail all failed with a leak as opposed to exploding, the alarm would have picked these up before catastrophic engine failure.

Event if a hose explodes I think I'd still like to know about it (quickly) before grinding to a halt with seized engine.


 
#14 ·
Hi mgbob.
I wasn't suggesting that I wouldn't love a system that offered a warning both in the event of slow leak and sudden catastrophic loss. However, this comment-
"could the sensor not just be lowered further down the bottle to allow for the slosh factor so it would trigger with say only an inch of coolant left?"
prompted me. If the system has to be biased one way or the other, I personally think I'd rather know early about small losses, as they are more likely and seem to happen more often. I'm very impressed by the work your doing, and if it saves someone from disaster well done.

My TF seems to be like the Titanic in reverse- I can't keep the water in the goddam thing.
 
#15 · (Edited)
i have an idea that i can use the temperature gauge in a dual function... pressure and temperature... i am looking at that right now... but - i think one could use it to sense level, and why i reply..

i.e a damped level sensor, that shows the level of the fluid in the header tank.. if one has the right know how...

its not an easy thing to do,, since the level sensor would need to actually see the level of the coolant like the fuel gauge uses to do that.

alternatively the sensor or level switch like sundance proposes, could just supply a known voltage to the temp sensor gauge via nothing more complicated than a switch... an automatic switch, that turns on an led, one fits in the dash near to the gauge...

the switch turns on, and the gauge drops to low... and a led comes on... :)

i just punt ideas .. :)
 
#16 ·
I put a TF header tank and sensor in my MGF a few years back, with a timer relay and piezo buzzer. With delay set to about five seconds it will buzz in a long left hander which gives me a check that it's working O.K. It did warn me of a slow leak and so paid for itself, not long after it was installed.

However the OE level sensor has never worked consistently and recently stopped triggering altogether. Took it out and after softening it a bit in some hot water then flexing it, I find I can get it to be 1. always open circuit 2. always closed circuit 3. sensing poorly 4. sensing accurately. I put it back in the car in condition 4. but don't expect that to last. They are pricey so I'm not going to try another, and have ordered some reed switches to experiment with.

It's nice to see another option, so will watch this with interest.

David
 
#17 ·
Ok so now I got round to fitting my sensor in my TF, also replacing my cracked binnacle while I had the instrument cluster out. I ended up having to remove the steering wheel, as I needed some extra access.....damn I hate removing the airbag!
Anyhow here is the Balluff contactless sensor fitted to my tank. A few dollops of Sikoflex to help it stay in place...

And the signal inverting relay also stuck down with Sikoflex...

And the connections to the rear of the instrument PCB. I used the 0v and IGN +12v to supply the sensor and the unused gearbox warning light signal connection as the switched signal input. Note I removed the connection screw from the gearbox signal from inside the unit to the PCB, otherwise a couple of internal diodes would be fried and my LED indicators on the cluster would have failed! I soldered the wire direct to the copper tracking of the PCB.

Finally put it all back together and reconnect the battery and check everything works again, great, then remove some coolant from the expansion tank and yay! the light comes on! Its the orange/yellow warning light on the right.

Now all I have to do is to roadtest it over the next week or so and see if it behaves ok, or needs a timing circuit installed to prevent erroneous signals if the coolant sloshes about too much. Hopefully the sensor is mounted low enough to avoid this.


Sundance
 
#18 ·
Having worked on the Screen Washer Fluid Reservoir on my project ZT-T recently, I noticed it has a low fluid level sensor ... which works because I checked it when the Reservoir was emptied when cleaned of the dead algae which can block washer jets.

I'm now wondering if one of these removed from a scrap yard car could be made to work for a coolant Expansion Bottle on our cars. I also wonder if it is heat resistant. That would not be difficult to check. I will explore that possibility. May work out well. I'll get one on my next scrap yard visit.

The internal low level sensor faces unseen upwards in this image. Located in the very bottom left of the Reservoir. Follow the Twin Wires to it's external connector. :~

 
  • Like
Reactions: sundanceuk
#19 ·
My reed switches arrived and, fitting one in place of the OE sensor, it works just fine. Less than $2 for a reed switch vs. $38 for the OE sensor, which was never fully reliable.

If I didn't already have the TF header tank with the magnetic float in it, ($72) then it looks like this high tech sensor would be a nice solution.

David
 
#21 ·
Just a quick update on how the sensor has behaved over the last couple of weeks since fitting. The Staffs moorlands where I live has plenty of good hills and corners to test it with.
The lamp has only come on briefly a couple of times, both times when taking a sharp left hand corner at speed. No triggering on hills or right hand corners and I do not mind the odd blink on the dash, as it tells me that my sensor is still working ok, without having to drain my coolant tank periodically to test it.
So I will not need to fit a timer circuit and the sensor seems fine with the hot temperatures around the expansion tank.


Sundance
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top