Hello, I''ve been working with this PWM circuit on a car's electrical system provided by Bill:

Differences are I'm powering the circuit with 5v from a 7805 plus needed caps.

The problem is that the sawtooth wave output from the 555 varies with the car's RPM. If I use a multi-meter to measure average voltage on the sawtooth wave, V goes down when car's RPMs increase. I have both the 555 and 339 connected directly to the 5v supply and ground. Other than the small caps on the input and output of the 7805 5v regulator and the one cap in the above circuit, no other caps were used.

How might I stabilize the 555 output?

Thanks,
Rey

 

I'll add that while this was happening, the 7805 ~13v input and 5.03v output were steady. As well as, all the other comparator inputs.

 

Is it mounted under the hood? Ignition noise can cause false triggering. Try adding 100nF from pin 5 to ground. Noise on +12V may also be getting onto the output of the 7805. You might also try adding 100Ω in series with the 7805 input, and 100uF from the input to ground. An electrolytic cap (10-100uF) from +5V to ground might also help.

 

Thanks I'll try those additions. (I already have the cap from 7805 input to gnd).

The circuit is in the cabin with me.

 

The voltage divider is shown connected to +12v which might not be steady.
Instead connect the voltage divider to the regulated +5V.

 

I have the circuit shown powered by 5v from a 7805. The only 12v supply is connected to the 7805 input.

 

After some testing, measuring, etc. it seems I'm getting nosie on the input signal for the comparator....

I have an additional comparator that turns the 555 on/off. The output of this comparator feeds the 555 reset pin. The inputs for this comparator is a ref V that I set with a pot (~2.1v) and a 0-5v signal from a car's sensor.

The circuit works fine if input is from another pot. The 555 output is steady and turns on/off crisply. However, when I use the car's sensor signal as the input, I get a very wide band where the 555 output varies between it's on and off points. I decided to hook my oscilloscope to the sensor signal and see noise.

How can I filter this noise and just get the average voltage, yet still allow it to vary as the sensor does it's job?

Thanks,
Rey

 

What is the sensor, and where is it mounted? You might need shielded cable, and possibly an RC filter. A filter will introduce delay. Can you tolerate delay? If so, how much?

 

Thanks, Sgt. helped me out in another thread. But I will try a shielded wire to see if it make a hugh difference. The run is long and goes across the entire engine bay.

 

Sounds like coupling to me, try putting a resistor in line with pin 4, say 10-47Ω, along with a large capacitor (something like 470µF). If this doesn't work then we can put a zener, something like a 10V, as well as upping the resistor a bit more, to regulatate and isolate the power supply to the 555.

If you can get your hands on a low dropout regulator even better.