Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Low Coolant Pressure Alarm Project - LM393 - your feedback is appreciated

Status
Not open for further replies.

Picaxxxu

New Member
Hi guys!
This is my very first post here.
I've been following you for ages, now decided to register and share with you my design, to hear your thoughts and ideas about it, and maybe improve it.

History:
So, I've allready had a couple of cars that went to scrapyard because a radiator hose blew and engine overheated. No light or alarm sounded, and temp gauge was normal, since it measures water temperature, and there was no water...

Idea:
I need an effective way to be warned if a coolant leak happens. I though this could be done in a couple of ways:
- My first Idea was to dip two probes in the water reservoir, and measure it's resistance. When no water would be present, resistance should be 0, and an alarm would sound. Decided not to go this way, because there would be corrosion due to electrolysis, maintenance would be required.
- Second idea was based on my first, but using a stainless steel float switch in the coolant reservoir, but I'm afraid that it malfunctions overtime due to rust and grease present in the water.
- So I decided to measure the coolant circuit pressure. When pressure drops below some level, means there's a leak somewhere, and an alarm will be turned on.

Design:
So this is what I've got so far:
rtjr13.jpg

WPS1 will be a 5v 15PSI water pressure sensor, VOut ranges from 0.5v (0 psi) to 5v (15 psi).
LD1 - Red alarm LED
LD2 - Green OK LED
I'm using a LM393 dual comparator, with same reference and threshold but inverted between comparators, meaning that when comparator A is 'on', comparator B will be 'off', and vice versa. So when LM393's pin 3 voltage (coming from pressure sensor) is below reference the red LD1 will light up, and LD2 will go off. If voltage on pin 3 goes above reference, then LD1 will go off and green LD2 will be on.
Reference voltage is set using VR1, wich will be set to normal pressure when car is running idle.

I've breadboarded this, and it seems to be working fine.
I still don't have the pressure sensor, so I've just been testing with random input voltages, and it seems to be working as intended.

Still, since I'm a newbie enthusiast, I think there's always room for improvement.
Will it work as intended once assembled on car? Should I add any kind of protection or noise filter?

Any comments, thoughts or suggestions on this will be appreciated.

Thank you all! :)
 
Welcome to this place...

Due to alternator ripple in the supply line(s) and/or common-mode alternator noise (ground-loop) coming from the vehicle ground system, the High/Low indications will likely "chatter". Each comparator needs to have a bit of hysteresis (positive feedback) to make for clean switching.

As a minimum, the entire circuit needs to connect to the car's chassis in only one place. If the water-pressure-sensor is the kind that screws into a pipe fitting (i.e., connects to the engine block), then the body of the sensor must de-facto become the single-point ground for the circuit (e.g. the bottom of the pot, the Vee pin on the comparator, bypass caps, the 5V regulator, etc, all must be returned to the body of the sensor).
 
Last edited:
Hi MikeMl
Thank you for your feedback.
I forgot to mention that the pressure sensor I will be ordering, is a 3 pin kind, one of them being ground, wich will be connected to circuit ground.
I plan to get 5v using an 78L05 voltage regulator, so either way, circuit will share vehicle's ground.
 
..., so either way, circuit will share vehicle's ground.
The circuit has to share the vehicle's ground. At issue is how many places will the circuit/sensor connect to the vehicle's chassis?

I am advising against connecting the sensor's ground to the engine block (for example, and bringing only two wires from it to the circuit), and then building a circuit containing the VR and Comparitors and connecting the circuit to a separate grounding point (like the dashboard).

I am proposing that if the sensor can be completely isolated from anything in the engine compartment, then bring three wires back to the circuit through the firewall, and you can use a single-point ground at the dashboard. Alternatively, if the sensor is intrinsically grounded to something in the engine compartment by its mounting method, you have no choice but to keep the circuit totally isolated from a local ground at the dashboard, and to connect the circuit common back to the body of the sensor so that there is only a single-point ground in the engine room.
 
Last edited:
I admit I first thought of connecting it using vehicle's ground, but then I realised that I'll have to bring 2 cables to dashboard, no mather what, so it's just easyer to bring the 3 cables alltogether and avoid a bunch of problems.

Thank you for your considerations. ;)
 
I admit I first thought of connecting it using vehicle's ground,...
Everybody does, and then wonders why they get hundreds of mV of alternator/other noise as a common-mode signal at the input of the comparators.:(

Some of the 1000W car-stereo types finally figured this out (common-mode noise, aka ground-loop, aka not having a single-point-grounding scheme, between the head-end and trunk-mounted power amps)
 
Welcome.
My fiat, the wifes citroen have low water detectors on the header tank, its common nowadays.
They look like capacitive sensors, the presence of water affects the frequency of an oscilator and gives an o/p.
Your circuit looks as though it will work, though theres no hysteresis, so the leds will flicker at the switch point.
Dont forget the water pressure cold is different to warm.
Land rovers have a silly problem where pressure and temp is normal at parts of the engine and overheating elsewhere.
 
Dont forget the water pressure cold is different to warm.
Actually, that is my main concern. I haven't received the pressure switch yet, so I am only assuming that as soon the engine is running, there will be pressure.
I fear however that a cold idling engine may not have enough pressure to be mesured or distinguished from a leak.
 
Hi Pica,

Afraid to say that there is a flaw in your cooling system pressure monitoring approach.:wideyed:
(1) The cooling system can be fully pressurized and still have a low water level: small water leak, head gasket blowing slightly into a water galley.
(2) An engine can overheat with a full water level: thermostat stuck. radiator blocked by debris or ice, radiator fan failure, water pump failure

Our automobile had an overheating problem a couple of months ago when stuck in a traffic jam for ten minutes after a high speed run, on a very hot day. Luckily, the dash display warned me to stop, even though the temperature gauge was only reading high, but not in the danger zone. When I stopped and opened the bonnet (hood) the heat was overwhelming and there was steam blowing from the water cap.

The automobile took half an hour to cool down sufficiently to remove the water filler cap. Incidentally, the system was fully pressurized all the time, which you could tell by the standard check of squeezing the top hose. Practically ever drop of water had bee lost, judging by the amount of water it took to refill the system.:arghh: The overheating was caused by one of the radiator fans failing.

It seems ridiculous that the engine temperature warning systems on automobiles are so inadequate. On my automobile, for example, not only is there two temperature sensors but the whole cooling system is controlled by a computer.

The classic scenario on many automobiles is that you get a small water leak and the water is slowly lost until it finally gets less and less in contact with the temperature sensor, so the temperature sensor does not indicate a problem.

Apparently cooling fan failure accounts for many breakdowns, in the UK anyway, according to the recovery mechanic who attended our automobile. He told us about an automobile which had blown its head gasket because a bird had got stuck in the radiator fan and the automobile had been waiting in a queue of traffic on a sweltering day.

Like you, I thought about a better system for warning about an impending engine overheat and came to the conclusion that a head temperature sensor, as used on piston aircraft engines, would be the best solution. In addition, a fluid level monitor would be wise and a cooling fan check would also be a good idea.

spec
 
Last edited:
Hi spec,
Thank you for your input. You'r right, this is not a flawless method to prevent overheat.
I'll take your head temperature sensor suggestion into consideration, maybe it'll be a better aproach ;)
 
Hi spec,
Thank you for your input. You'r right, this is not a flawless method to prevent overheat.
I'll take your head temperature sensor suggestion into consideration, maybe it'll be a better aproach ;)

No problems Pica- if you intend to build a head temperature monitor, can you make that two.:)

There are a few things to consider:
(1) The head temperature will vary according to driving conditions, much more than the water temperature.
(2) The temperature sensor must be capable of monitoring relatively high temperatures.
(3) Finding the optimum position on the head for the temperature sensor. By comparison, a piston aircraft engine has the temperature sensor position and fitting designed in.

spec
 
Depends where you measure it, pressure at the pump will ne highest when the 'stat is shut, as the pump will be dead headed (except for the heater & block bypass wiggler doofer), whereas in the head pressure will be lower with the stat shut.
Stands to reason that you'd measure pressure at the lowest pressure point.
 
Depends where you measure it, pressure at the pump will ne highest when the 'stat is shut, as the pump will be dead headed (except for the heater & block bypass wiggler doofer), whereas in the head pressure will be lower with the stat shut.
Stands to reason that you'd measure pressure at the lowest pressure point.
In practical terms the pressure is equal at all parts of the cooling system. The water pump does not generate significant pressure.

spec
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top