# Automotive radiator cooling fan contoller

Discussion in 'Automotive Electronics' started by tyrebyter, Oct 28, 2013.

1. ### MrAlWell-Known MemberMost Helpful Member

Joined:
Sep 7, 2008
Messages:
11,070
Likes:
963
Location:
NJ
Hello again,

I thought i had attached it but maybe not so here it is again...

File size:
129.3 KB
Views:
231
2. ### tyrebyterMember

Joined:
Oct 28, 2013
Messages:
55
Likes:
0
Location:
New England
alec-t,

I found the potentiometer on the user group and put it into my circuit schematic. I'm still runnng through some trial and error exercises trying to get the frequency of a square wave generator up to 40kHz - double the upper limit of human hearing.

My approach is to try to reduce the voltage of the waves to 5V instead of almost 10V. The data sheet for LM324/NS states a 1mHz bandwidth at 1.5V. I'm assuming that is a 1mHz input signal can be throughput effectively at 1.5V. I assume the bandwidth decreases if the voltage is increased.

I still have not figured out how to reduce the square wave voltage without reducing the supply voltage. Any suggestions ?

MrAl,

I have your drawing now and will be looking at that also, in the next day or so. Thanks for reposting.

3. ### MrAlWell-Known MemberMost Helpful Member

Joined:
Sep 7, 2008
Messages:
11,070
Likes:
963
Location:
NJ
Hi,

What is the square wave for, and where is the circuit?

The LM324 has a slew rate of about 0.5v/us, so that means it is limited as to the max peak of a sine wave that can pass through it cleanly.
The formula is:
SR=Vpk*2*pi*f/1e6

where
SR is the slew rate in volts per microsecond,
Vpk is the peak of the sine,
f is the frequency,
pi=3.14159

So for a 1MHz sine with a 1v peak we would need 2*pi which is about 6 volts per microsecond slew rate. For a 1.5v peak we'd need about 9v/us slew rate. The LM324 can only do about 0.5v/us, which is 18 times less so the max frequency would be 1MHz divided by 18 (at the same 1.5v peak).

Joined:
Jan 12, 1997
Messages:
-
Likes:
0

5. ### tyrebyterMember

Joined:
Oct 28, 2013
Messages:
55
Likes:
0
Location:
New England

MrAl,

The 5ms square wave shown is the best I can get from this circuit. I have not ahd a chance to try adding components to reduce the voltage. I've taken a step backwards to learn more about rc circuits.

Here is the square wave generator circuit I have come up with using an LM324/NS. It will probably not be what I end up using but it does make a square wave.

Joined:
Oct 30, 2010
Messages:
10,026
Likes:
1,107
7. ### MrAlWell-Known MemberMost Helpful Member

Joined:
Sep 7, 2008
Messages:
11,070
Likes:
963
Location:
NJ
Hi,

Oh ok, well that's very good. It's good to experiment with this stuff and we are lucky that today we have simulators that mimic the behavior pretty well. Not perfect, but we dont need perfect to learn from it anyway. When i first started out many years ago there were no simulators so any experiments had to be done in the lab even if they were fairly simple like a one transistor circuit. Not that we should rely too heavily on simulation alone, but it sure is a good learning tool.

Just one small point. The 'pullup' resistor is not needed with a op amp. A pullup is usually required with a comparator because they usually have open collector outputs. But then again the pullup is usually smaller than 10k because a large pullup creates a problem with the output waveshape. If you feel like looking at that you'll see the difference in waveshape on the output.

As a rule of thumb without any pressing special requirements, the min half cycle period we would want would be about 10 times the approximate rise and fall times. So as a guide if the slew rate is 0.5v/us and it has to reach from 0 to 10 volts that means it takes 20us to do that. That in turn means we want a min half cycle period of 200us. A min half cycle period is around 200us so that's 5ms, and that makes the whole output period about 10ms, which is 100Hz. So you see how the slew rate can limit the frequency severely.
Lower the voltage to 5v and we can get twice that in theory. Of course this is just a rule of thumb and there will be times when we dont need 10 times the total slew as the half period, such as when the output is followed by a Schmitt Trigger gate of some kind which squares up the output anyway.

8. ### tyrebyterMember

Joined:
Oct 28, 2013
Messages:
55
Likes:
0
Location:
New England
KISS,

The Velleman kit does not go to the 40kHz frequency, I'm aiming for (as per your suggestion which I 100% agree is required). It's limited to 5kHz. According to the data sheet, the SG3525 ic goes to 100kHz, which makes it worth consideration and I'll put some time into studying the Velleman schematic. Maybe I can figure out how to incorporate some of the kit and do a minor redesign to get it to run at 40kHz. The price for the kit is certainly attractive.

MrAl,

Please try to remember my 99.9% neophyte status as an electronics designer. Which resistor is the "pull-up" resistor ? I've read the term elsewhere and the term "pull-down" resistor also but I really don't know what they refer to. I think as used for pulling up - or down - the voltage in a voltage divider arrangment but that is just a guess.

I was unable to figure out how to make the 324 put out a lower voltage square wave except to reduce the supply voltage. Is that the only way ? I can see that a lower voltage could be "switched" faster and thus reach a higher frequency.

One of my next steps as I blunder along this path of learning what will and what won't work, is to look into small signal transistors to see if I can incorporate an "astable multivibrator" circuit to produce a 40kHz square wave.

I hope you guys don't get tired of helping me. What you've given so far has been invaluable.

9. ### MrAlWell-Known MemberMost Helpful Member

Joined:
Sep 7, 2008
Messages:
11,070
Likes:
963
Location:
NJ
Hi again,

A pullup is usually a resistor connected to a terminal of a device that will apply Vcc to that terminal through that resistor (and thus current limited). The reason for a pullup is because that device's terminal can not reach up near to Vcc by itself, yet needs to be about to get to near Vcc in certain modes of operation.
For many comparators, they have a open collector output and so the output can not source any current, it can only sink, and so it needs a pullup resistor so it's collector can get up near Vcc when the transistor turns off.

A pulldown is just the same except it would be used when the device needs to have a ground level DC signal applied during certain modes of operation. It cant get down near ground by itself so it needs a little help from a resistor.

You might be able to use feedback to limit the output voltage level for an op amp, but why go through all that when you can use a 555 timer IC and have it be done, if all you want to do is generate a square wave.

10. ### alec_tWell-Known MemberMost Helpful Member

Joined:
Jul 10, 2011
Messages:
9,420
Likes:
1,247
Location:
Cardiff, Wales
If you want to experiment with square-wave generators, here's a couple, simple in the extreme, which will easily do >100kHz.
Approximations of triangle waves (good enough for use with PWM generators) can also be taken from across the capacitors.
The second circuit effectively extends the hysteresis of the Schmitt inverter to increase the amplitude of the triangle compared to the first circuit, and provides complementary outputs from its two inverters.

File size:
4.1 KB
Views:
206
File size:
1.3 KB
Views:
124
11. ### KeepItSimpleStupidWell-Known MemberMost Helpful Member

Joined:
Oct 30, 2010
Messages:
10,026
Likes:
1,107
The datasheet says the chip is good to 400 Khz. Changing a few components should get you there. Things are tough, right now, on the home front".

12. ### tyrebyterMember

Joined:
Oct 28, 2013
Messages:
55
Likes:
0
Location:
New England
I have an astable multivibrator circuit that makes square waves at 40kHz ... at least in the simulator. I had to make it operate in a narrow voltage range to get the 2N2222 transistors to switch that fast. Maybe I can find faster transistors but for now, I'm going to try to narrow the pulse width "on" time to see if I can increase the voltage swing.

I've already made some attempts at having this 40kHz circuit drive a MOSFET and met with some success ! Problem is that I cannot figure out how to get the MOSFET to turn on and off fully with the square wave (pulse) voltage swing being only 1.3V

alec_t, I'm going to be using this pulse generator to drive a MOSFET. Can the digital ICs typically drive a MOSFET to the 40kHz, I am aiming for ?

KISS, I'll have another look at the Jameco kit data sheet and the SG3525A data sheet. Wouldn't be the first time I've misread something.

13. ### alec_tWell-Known MemberMost Helpful Member

Joined:
Jul 10, 2011
Messages:
9,420
Likes:
1,247
Location:
Cardiff, Wales
Not directly. At that frequency you would need a push-pull buffer stage between the IC and the MOSFET to supply enough current to charge/discharge the gate capacitance quickly.

14. ### ednjaNew Member

Joined:
Feb 3, 2013
Messages:
1
Likes:
0
Location:
Vancouver, BC
What every became of this project? Did you get a working circuit built?

Joined:
May 14, 2008
Messages:
456
Likes:
3
Location: