Coolant warning light tested and failed

[Hacksawbob]

So I did the coolant warning light test and its failed. Initially the light flashes on start up (this is just the test on the PCB in the dash) but when I pull the plug on the coolant bottle the light doesn't come on. I have been doing a bit of electrical work on the van recently so I wonder if I have knocked something. Any pointers where to start? I am guessing that if the light doesn't come on then either the module 43 is faulty (not testable) the dash connector block need attention, one of the earths is missing off the crown of earth under the dash, break in the wiring?

I had a quick look yesterday to but I having some work done today so I don't have the van til this evening. Planning on heading off tomorrow so suggestions welcome!

[ghost123uk]

Bob, how long did you run the engine with the sensor unplugged ?

I ask because on our new van it takes a longer to trip the warning than our other van, like 30 seconds or so.

[Hacksawbob]

OK this was running for about 10 minutes or so. I had the coolant warning light come on about a month or so ago, but when I pulled over it all seemed OK. light has been off since.

[ghost123uk]

Ummm, smells like a faulty or in need of a wiggle unit 43.
Perhaps borrow a known good one ?

I know you know about earth crowns etc, so I won't mention them.

[Hacksawbob]

Yeah I gave the 43 a wiggle... I think Angeloevs said you can detect the ac pulse with a meter I might try that, I guess it needs to be on an ac setting?

[ghost123uk]

I think Angeloevs said you can detect the ac pulse with a meter I might try that, I guess it needs to be on an ac setting?

E D I T = the text in red below has proved (By Angelo) to be duff information, read the rest of the posts below.

No I doubt it (my electronics background speaking), the pulse will not show on a meter, either AC or DC setting, it is too short a pulse with too long a gap before the next one. (known as a long mark space ratio)

It would show on an oscilloscope, but even I don't have one of those any more !

As a break in the wiring would cause it to come on all the time, it can't be that. Oh, unless the break is between 43 and the tiny LED circuit in the dash of course

If it were I, first test would be temporary replacement of 43 with a known good one.

[AngeloEvs]

The ac signal across the probes is a steady frequency, I can't remember the exact frequency or the voltage amplitude. The peak to peak voltage is a few volts as far as I can remember. I have only ever used an oscillsocope but you should be able to detect it with a mumtimeter set to ac. The Module 43 does send pulses to the temeperature guage via the wire that connects the temp gauge to the temp sender in the thermostat housing. These would register on a meter set to 12V DC by a sudden drop in voltage reading every second or and can be seen by placing a multimeter set to 12V DC on the left nut of the temp gauge itself. These negatiive low frequency pulses are what actually triggers the temp gauges internal circuit. With an analogue meter you would see the needle kick once every second or so. In a nutshell - Module 43 sends a steady ac waveform to the probes continually and a series of short duration pulses to the temp gauge if and only when coolant is lost.

I would remove the Module 43 and check that you have +12V at the vacant socket connection (ignition on) and that the socket does connect to earth. The diagram shows the vacant socket connections for reference. The earth is at the crown cluster next to the fuse box.

Chek that the socket connection marked 'temp sender' does connect to the left nut of the temperature gauge or you can short circuit the socket cnx 'TEMP sender' to Earth - this should trigger the coolant led into continuous operation and confirm that the socket is indeed connecting to the temp gauge.

The other socket marked Level sender connects to the probe but its a hell of a long run to measure with a multimeter unless you attach a long piece of wire to the sender plug. The other side of the level sender probe connects to earth but, if 43 was working and you had an earth fault, you would have the continual flashing coolant LED problem and not the one you currently have.

[ghost123uk]

Is Angelo's expertise (on this) in wiki yet ? ( )

[AngeloEvs]

When the weather improves I will measure the signal across the level sender with an oscilloscope and a typical DIY voltmeter and post the readings, also, time permitting - 'how to bench test module 43'..........

[Hacksawbob]

When the weather improves

I'll not hold my breath then!
Great work Angelo,
I tested the wires for continuity between the no3 slot behind the relay and the positive side of the probe (with a long wire) and that checked out OK That;s the blue green wire that goes to the probe. As this is diesel the earth instead of being at the coil is in the connection box on the LHS of the engine bay. It shares earth with a few other things from what I can see. But one is marked as Earth point engine bay left not sure exactly where this is maybe syncroMcAndy can help? But the earth from the probe clip checked out OK to the battery so I can assume that the wiring is alright.

Where is earth 12 on this diagram? (diesel)


I have been uncovering a few electrical nightmares on my van by the PO, the more you look the more you find!

[Hacksawbob]

OK looks like I may have fixed it. the 14 pin connector under the dash wasn't too clever looking so I have given it a shine and coolant warning seems to be coming on as it should be. lost my glow plug warning light in the process but there you go, a job for another day!

[AngeloEvs]

Glad you sorted it! I measured the output across the probe socket using a digital multimeter set to 2V ac (with it disconnected from the level sender) and the reading is approx 1.8V The frequency of the pulse is 100Hz so nothing special there. Same for both types of Level sender modules 42 and 43 - either can be used.

The only bench test that can be carried out using a multimeter is to connect a 12v supply and earth to the relevant terminals and measure the ac voltage across the level sender terminal to earth on the rear of the module itself - should be around 1.8v. Not a lot of point in bench testing since you can do this in situ unless you suspect a break in the wiring to the level sender itself from fuse box to engine compartment. There is no simple way of bench testing the trigger output without setting up a dummy load and an oscilloscope.

However, these modules generally either work or don't - if you have the steady ac frequency at the probes then the chip is working and should also provide the trigger pulses to the temp gauge if coolant is lost. The amplitude of the pulses drops significantly when the probes are immersed in coolant - as low as 0.2V rms which is how the system detects the prescence or absence of coolant (by monitoring the amplitude) so take the reading with the coolant sender disconnected at the soket terminals. The only other fault is the microchip in module 42/43 going short circuit (another cause of permanent led flashing syndrome).

[Hacksawbob]

Thanks again for your help Angelo, I will get in the wiki at some point, unless some one beats me to it!

[ghost123uk]

I think Angeloevs said you can detect the ac pulse with a meter I might try that, I guess it needs to be on an ac setting?

E D I T = the text in red below has proved (By Angelo) to be duff information, read the rest of the posts below.

No I doubt it (my electronics background speaking), the pulse will not show on a meter, either AC or DC setting, it is too short a pulse with too long a gap before the next one. (known as a long mark space ratio)

Glad you sorted it! I measured the output across the probe socket using a digital multimeter set to 2V ac (with it disconnected from the level sender) and the reading is approx 1.8V. The amplitude of the pulses drops significantly when the probes are immersed in coolant - as low as 0.2V rms. The frequency of the pulse is 100Hz so nothing special there.

Ah, hands up, I was wrong
I have edited my post (as per in the quote above) to avoid confusion for future readers.

I had tried to measure it (a good while back) with the probes in the coolant and as Angelo says, the voltage then was so low I decided it was inconclusive and "assumed" it was a mark / space ratio thing. Note to self = don't assume things !!

~~~~~~~~

I too am glad you got it working, I guess all this was sparked off by my "Test your coolant sensor" post
So, whilst it brought you a job to do Bob, at least you will now know when (sorry, "if") you ever start to lose coolant Plus we have all learnt some stuff thanks to Angelo, and the wiki will be a little richer too