Coolant booster pump control

[91 XR7]

So I'm putting the previous project on hold for a bit and going to try this cooling system thing in a different approach ?

Components on hand:
LM311
VDO temperature sensor (P/N: **broken link removed**) (A NTC sensor, 17000 to 10 ohm))

The Thermostat is open at 78'C
The Cooling Fans turn on at 96'C
The Pump in my Mind should turn on around 80-85'C

With that said, and with the sender above, that would be a 70-60 ohm resistance

A built up a circuit on my bread board, and technically it is working which is nice, but i'm also using 2.5K Pots so the adjustment in the lower Range isn't that fine. also using a 10K resister on both the Reference and "input" voltage dividers. All testing had been using a 12V power supply

Still trying to figure out the Hysterysis for it too, and how to implement it to the LM311 since it has it's own built in Transistor (open collector)

Questions i have

1) Now the sender itself is also hooked to the coolant temperature gauge (Coil style? +, -, S terminals) so i'm wondering how that will effect the whole process? and the gauge itself?
2) Since the range i want the pump to be turned on is when the Temp sensor is in the 200-0 ohm range, what would be the best way to get better (finer) adjustment in that basic range?
3) If i run say a 9Volt regulator (Have one on hand) to power the LM311 and for the reference voltage should i also do that for the input too?

Thank You

 

[Diver300]

The temperature gauge and sender are likely to be run from a regulated supply, and I would think that the voltage would be mainly dependent on the thermistor's resistance. What I suggest that you do is have a voltage comparator, using the thermistor voltage as your input.

If you want to find the voltage first, replace the thermistor with a resistor of value equal to the resistance of the thermistor at the temperature you want, check that the temperature gauge reads about right, and measure the voltage. I suggest that you repeat with various temperatures and various voltages.

 

[KeepItSimpleStupid]

The sender is likely connected to a lower voltage supply, so i would measure the open circuit voltage. i.e. Disconnect the thermister.
Determine if the thermister is attached to ground.
Monitor the voltage across the thermister.
Alternatively, you could monitor the differential voltage across the meter.

Your reference could be ratiometric. i.e. a buffered version of the likely already present reference. You could buffer, voltage divide (with a variable resistor somewhere) and buffer again. Generally the comparator will need a low Z input for the reference.

 

[91 XR7]

Maybe abit more info?

The sender itself is self grounding. So only one terminal on it.
The gauge itself isn't regulated, the postive (or high) that's going into it can range 14.4volts and down, whatever the alternator is putting out.

But think of it this way. From the fuse (12.2-14.4volts) to an coil in the temperature gauge to the thermistor which is grounded to the block. So basically the thermistor is altering the magnetic field of a coil inside the instrument to counter act the other coils magnetic field to move the needle.

In a sense it in itself is a voltage divider circuit, but not sure if one and truly use it as such? The coil will be a set resistence that's tied to high, with the thermistor tied to low inbetween the two should be the voltage divider?

Or can I just make my own voltage divider tieing into the wire going to the thermistor and using say a 10k resistor to high to create a independent voltage divider just for this circuit? But would two voltage divider alter the resistance signal of the single thermistor?

 

[KeepItSimpleStupid]

The gauge is usually a milli-ammeter that has a high resistance, The resistance of the meter can be measured "indirectly" using a voltage source and a potentiometer.

Without damaging anything, you adjust the resistor until the guage reads half-scale. Then measure the variable resistor. That's the resistance of the ammeter.

You MAY want to then use a DC-DC converter to make that voltage across the gauge constant. i.e 12 V

 

[91 XR7]

Usually a Ammeter is only a two terminal meter is it not? The gauge in question has Three terminals, with the Sensor wire taped in between two coils with in the gauge itself?

With knowledge of what Resistance the sender is registering at the temperature i like the fan to turn on i was able to setup the gauge with that resistance in hand, and was able to determine what voltage would be at the sender wire. Sadly i didn't transfer the Word file to this computer so that information will need to wait.

 

[Diver300]

How old is the car? The make and model might also be useful bits of information.

Some cars, and I thought it was only older cars, have gauges with two coils, where one coil is in series with the thermistor, and the other is between the supply and ground. The needle movement depends on the ratio of current in these coils, so if the battery voltage increases, both coil current increase together and you don't get much change in needle position.

It could also be that the gauge has an internal regulator.

It seems that it would be easiest to measure the voltage on the thermistor, with everything connected as normal, at various temperatures and battery voltages. You can probably just measure with and without the engine running to get around 14.5 V and 12,5 V supplies.

 

[KeepItSimpleStupid]

See here: https://www.americanmuscle.com/autometer-watertemp-install.html There is a "good chance" that there is a low voltage regulator inside the gauge.

Again, I''m suggesting measuring the voltage at the sender terminal to ground with the sender disconnected. This is only one data point. I'm suspecting you might find a lower voltage than 12 V there.

 

[KeepItSimpleStupid]

Driver: Try a 1991 Mercury Cougar XR7? Don't ask me how I know.

 

[KeepItSimpleStupid]

Could be this: http://www.panteraplace.com/Electrical/SOBill Veglia Temp Fig 3.jpg which makes it novel.

Ford always did things different.

 

[Diver300]

Driver: Try a 1991 Mercury Cougar XR7? Don't ask me how I know.

XR7 doesn't indicate a model to most people of car on this side of the Atlantic. Off topic, "Reliant" does indicate a type of car, but a somewhat different one to a Plymouth Reliant.

I also found that diagram for the Veglia gauge, and it is the design that I was thinking about. If that is used, the thermistor voltage will change as the battery voltage changes.

 

[KeepItSimpleStupid]

No biggie. Just means he has to do it differentially and deal with the ratiometric stuff.

Hey, the pond didn't make any difference: I just Googled "1991 XR7"

 

[91 XR7]

Some cars, and I thought it was only older cars, have gauges with two coils, where one coil is in series with the thermistor, and the other is between the supply and ground. The needle movement depends on the ratio of current in these coils, so if the battery voltage increases, both coil current increase together and you don't get much change in needle position.

It seems that it would be easiest to measure the voltage on the thermistor, with everything connected as normal, at various temperatures and battery voltages. You can probably just measure with and without the engine running to get around 14.5 V and 12,5 V supplies.

The gauge in question is like the one, almost exactly other then Replace the "Term" with a "S", that KISS posted above. **broken link removed** is a description on how it basically works; YES this is describing a fuel guage, but they both work on the same principle here.

Magnetic Gauge (fig. 4-4). The sending unit in this fuel gauge contains a sliding contact. As the fuel level in the tank changes, the position of the contact changes on a rheostat winding, varying circuit resistance and resulting current flow. The unit on the instrument panel contains two magnetic coils (limiting coil and operating coil) and a permanent magnet that is attached to the gauge needle. When the fuel tank is empty, the limiting coil is stronger than the operating coil, thus the magnet is drawn toward it and the needle reads EMPTY on the gauge. As the tank is filled, the operating coil becomes stronger, attracting the magnet and moving the needle toward the FULL position.

As stated in my third post in this thread, I did what you said just above. I hooked the gauge to a power supply (only 12Volts, all i have), Used a Resistor to corresponding to the temperature i like the "pump" to turn on at tied between the "S" terminal and Ground to get the voltage that is present at the "S" terminal, which works out to be 1.53 Volts.

 

[91 XR7]

Could be this: http://www.panteraplace.com/Electrical/SOBill Veglia Temp Fig 3.jpg which makes it novel.
Ford always did things different.

technically it's not just ford that uses this same basic setup. a bunch of European cars, Almost all aftermarket gauges do (unless mechanical)

Also we aren't dealing with my personal car here. the gauge in question is a VDO gauge (my '91 XR7 does have VDO gauges from factory thou) for a late 1950's Porsche 356 repica that i build

 

[91 XR7]

I also found that diagram for the Veglia gauge, and it is the design that I was thinking about. If that is used, the thermistor voltage will change as the battery voltage changes.

Well if that's the case, then there is a good chance i'll need to get a 14.4 volt power supply Also i'm taking it when you guys are saying the voltage at the thermistor is when the gauge is hooked up? Or between it and the positive battery terminal?

 

[KeepItSimpleStupid]

What Diver is trying to say, is that the reading presented on the face will be independent of the supply voltage.

What I was trying to get at earlier, was to determine IF there could be a "regulator" within the gauge.

The important numbers are the open circuit voltage, the power supply voltage (at the meter terminal) and the voltage across the thermistor (or simulated thermister) to ground.

I SUSPECT that the voltages across a known resistor (variable resistor set to a specific value) ( at 12 V and 8 V for the supply) will be proportional to each other.

If it is, you have a few choices:

1) Use a 9-18 DC input to 12 V DC-DC converter to power the gauge. Ripple may be an issue. Thus the gauge and your circuit sees 12V all the time, so your fixed setpoints (voltage across the thermistor) stays the same for each temperature.

2) The "other way" requires that say you adopt 5V for your circuit power and "battery voltage to the gauge". Let's say that the conditioned signal has a maximum Vout of 2.5V @ 12 V (battery) and therefore 1.25 V when the Battery is 6V. You have to then take the 12V battery voltage and divide it so it can be processed by your circuit. e.g. make 12 appear as say 500 mV. Now deal with the amplified voltage across the resistor and the (proportional voltage to battery) and subtract. You can make it higher as well, because what you care about is the difference and you'll build your comparator around that.

 

[KeepItSimpleStupid]

As stated in my third post in this thread, I did what you said just above. I hooked the gauge to a power supply (only 12Volts, all i have), Used a Resistor to corresponding to the temperature i like the "pump" to turn on at tied between the "S" terminal and Ground to get the voltage that is present at the "S" terminal, which works out to be 1.53 Volts.

And PROBABLY without the resistor, the voltage measured at the meter where the resistor was connected will be 12 V to ground?

OK, if you have a dry cell, put the battery in anti-series to get 12-(1.5 V) and measure it. Do the measurement again across the resistor and you'll probably read 1.38V. (12-1.5)/12*1.53; 87.5% less.

 

[91 XR7]

OKay i have made up a circuit that seems to work, least on the bench but before heading there, i see if i can make myself less confused?

Power source = 12.32v
Power at sensor terminal with sensor disconnected = 7.5v
Power at sensor terminal with sensor connected (85*C) = 2.064v

Power Source = 14.1v
Power at sensor terminal with sensor disconnected = 8.63v
Power at sensor terminal with sensor connected (85*C = 2.37v

The information you wanted at the beginning right, KISS? I'm thinking no voltage regulator since the voltage changed on the open sender terminal of the gauge.

Also i'm not even sure if this gauge would perform correctly if i was to wire in say a 5 to 9 volt regulator, It may be just fine and i can try it out later. But i like to keep the modifications to minimal to get this circuit to work.

 

[KeepItSimpleStupid]

Nope, there is something there, not necessarily a regulator. because it's not 12 V and as predicted: 7.5/8.63*2.37 = 2.06 which is really close to 2.06 (Identical).

The output "tracks" the input voltage.

I wan thinking of something like this: https://www.digikey.com/product-detail/en/RS3-1212S/945-1564-5-ND to power the gauge.

But, you would have 50 mV p-p of high frequency ripple at 12 V to deal with.

So, either, you setpoint has to "track" the input voltage or you have to "FIX" the input voltage. When you fix the input voltage, at some value, it won't work. In the DC=DC converter I found, that would be 9V. You would want a DC-DC converter that can take a maximum of 18 and any minimum <= 9 V.

A 5 to 9V regulator might work. Your choice may have a drop of 3V. e.g. an LM317T, so an Low dropout regulator might be more appropriate.

 

[alec_t]

Based on the figures in post #18, here's a suggested circuit. I've attached a zip file for simulating the circuit.
The comparator receives a reference voltage which, like the voltage at the temp sender terminal, tracks the battery voltage. The trip point is thus not significantly affected by battery voltage and no regulator is necessary.
The trip point is adjustable with a trimpot. R3 sets the amount of hysteresis, so may need to be tweaked.
Supply decoupling and spike suppression components are not shown but would be advisable.