Low voltage cut in/cut out

[Palmer Pease]

Greetings all, first post. I did a search to no avail.

I would like to build a circuit that will close (or open) contacts when input voltage falls below a specified value. It will need to be quite precise, within a tenth of a volt DC or less. In the 2-2.5 volt range.

Here's the project. I would like to convert my '04 GMC to electric cooling fans for the radiator/AC condenser/transmission cooler. The engine and transmission temp sensors are provided a 5v reference. When the sensors are cold the resistance is low so the return voltage is relatively high. As the fluid increases in temperature the resistance increases and the voltage drops. When the voltage drops to a specific value, I need to start triggering fans on. Most E Fan setups are two, two speed fans. I plan on four or five temperature/voltage trigger points to manage noise and electrical usage, plus provide a bit of redundancy.
Wiring the relays to control the fans is the easy part for me. Reading wiring diagrams and schematics is easy. Soldering circuit boards, resistors, diodes, relays, I'm good with all of that. Capacitors I understand but don't know the math to size them. I should have paid more attention to transistors. But I'm good with logic these days as opposed to my teens when I did the Heathkit DC electronics course

Any suggestions?

 

[Tha fios agaibh]

You could make a comparator circuit fed with a Voltage divider input and a transistor to drive a relay for your fans but why not do it the simple way;
Thermal switch to relay to drive fans?

If you need to manage noise, why not consider suppression

 

[AnalogKid]

acoustic noise.

 

[Tha fios agaibh]

Instead of 4 or 5 trigger points, why not go right to a Pwm circuit that'd send a varying output right to the fans?

 

[Palmer Pease]

Managing with thermal sensors would require adding thermal sensors. A couple of issues. There's no simple place to add an engine coolant sensor, adding a trans temp sensor would require welding a bung in the pan which is succeptible to damage.
In addition, I would like to use the current voltage outputs of the factory system, keeping my add on very stealth.
As for noise, I'm simply talking audible noise of pulling two high volume fans on high at once when that is rarely required. The only time I would need that is when the AC is on and vehicle sped is zero.

 

[Alec_t]

How are you proposing to control the speed of the fans if you don't use PWM?

 

[AnalogKid]

The fans have 2-speed controls built in.

ak

 

[Alec_t]

Then I don't see how the TS proposes to use all the trigger points :
["I plan on four or five temperature/voltage trigger points to manage noise and electrical usage"]

 

[Palmer Pease]

As the temp increases, the voltage will continue to drop and I will need more cooling.

So maybe the first fan comes on at 2.5v
Second at 2.4v
Third at 2.3v
Fourth at 2.2v

 

[duke37]

Look at the LM3914 bar graph driver circuit, this has ten outputs and will add outputs as the voltage rises.
You will need output driver circuits for each fan and an input circuit to give the required range.

 

[AnalogKid]

Then I don't see how the TS proposes to use all the trigger points :
["I plan on four or five temperature/voltage trigger points to manage noise and electrical usage"]

Two fans with 3 states each (off, half, full) equals 9 combinatorial states. Use an LM339 or LM3914 to create 5 ascending states, and diodes to matrix them to any of the 9 possible fan states. For example:

Sensor Fan 1 Fan 2
State State State
0 off off
1 1/2 off
2 1/2 1/2
3 1/2 full
4 full full

Those are supposed to be three columns, but the site deletes the spaces.

ak

 

[duke37]

Depending on the way of dropping the fan speed by half, it may need four relays with drivers and some logic. If the fans just use resistors to drop the voltage and so the speed, then I would favour putting both fans on full speed and controlling the speed with a pulse width modulator to drop the voltage as has been suggested. (#4)
It would not have steps in the control but be continuously variable from off to maximum.
I would think that both fans running gently would be quieter than one going fast and the other doing little.

 

[AnalogKid]

I would think that both fans running gently would be quieter than one going fast and the other doing little.

I agree. That's why my example increments total fan power in evenly spaced 1/2 fan steps. Notice that for the "1 full fan" condition I have two fans running at 1/2 power each.

ak

 

[Palmer Pease]

The fans have two sets of windings. No resistors or PWM required.

 

[duke37]

The fans have two sets of windings. No resistors or PWM required.

Just leave the fans on maximum and use PWM.
Have you worked out how many expensive relays and driver circuits you will need and compared with a cheap PWM module?

 

[AnalogKid]

It is not clear what the signal source impedance is. It sounds like a constant voltage with a varying resistance in series. More clarity is needed, but based on that assumption:

1 - LM324
4 - 1N914 diode (routing)
4 - power MOSFET
4 - 1N4004 diode (protection)
1 - voltage regulator (optional)

$10-$15, no programming, no 555, no PIC, no generated EMI.

ak

 

[Palmer Pease]

This signal source is varying VDC 0-5 VDC.

 

[Alec_t]

You said (post #1) that the signal is 2.5V dropping to 2V.