Bornloser

Here is what I need.....An astable or monostable circuit, runs on 12Vdc. It should "oscillate" at approx 10KHz (not critical). During each cycle, the circuit needs to go HIGH for 1 to 50 % (using an adjustable pot.) of the duty cycle, then return to ground potential until the next cycle (10KHz), where it repeats.
I prefer it to use a 556 timer.
Any gurus out there willing to help an old geezer?
Thanks!

Torben

Hi BL,

I'm certainly no guru but I'll try to lend a hand. Is this for a project or assignment of some kind?

Anyway, I believe that a 555 changes its frequency as you change the duty cycle. To keep a stable frequency you need an astable *and* a monostable, which is easy because a 556 is just two 555s in one package. I'd use half the 556 as an astable to generate the 10kHz signal, and use its output to trigger the other half, configured as a monostable, to generate the output pulses. Use a pot across the timing resistor on the monostable to affect the duty cycle.

Does that help at all?


Torben

Bornloser

This is for a personal project. I figured it would take both, which is why I had suggested a 556, not a 555. I just didnt have the background to design (calculate) the values needed.

Torben

Gotcha. OK, I don't have access to my shop right now so I can't build this to test it in Real Life (TM) but attached is what I would do.

There is a good tutorial on 555 calculations at http://www.doctronics.co.uk/555.htm and another at http://www.uoguelph.ca/~antoon/gadgets/555/555.html in case you hadn't found those yet.

Anyway, the circuit I've attached uses the first half uses Ra = 130k and Rb = 500, and C = 1nF, for a frequency of:

f = 1.44 / ((Ra + 2Rb) * C)
f = 1.44 / ((130k + 1000) * 0.000001)
f = ~10.9kHz

So my math above gives ~11kHz. In the simulator it show closer to 10kHz. You did say it doesn't need to be exact. The values might not look like a normal 555 astable, but that's because I wanted this one to have a very high duty cycle, in order to not affect the minimum off time of the following stage too much. (Note: this relates to a problem with the circuit--see below).

The second 555 takes the incoming pulses and produces output pulses from ~1us to ~50us. This is a problem (mentioned above), since the tutorials indicate that you don't want to try to get output pulses of less than ~10us out of them, or else risk having troubles with the 555 retriggering itself. At 10kHz your pulses occur every 0.1ms; 1% of 0.1ms is 1us. Too short. Note that I have found some forum posts indicating that 1us is OK--although right on the edge of the 555's capabilities.

So in summary: this might work. Common wisdom says the pulses needed are too short for a 555 to handle; some people appear to maintain that it's OK but be aware that it's potentially pushing things. Anyway, I'd just build the thing and scope it and see what I got.

If this can't be made to work, perhaps you can describe the problem you're trying to solve (instead of how you're trying to solve it) and we can help you come up with another solution.


Hope this helps,

Torben

Attached Images

556_adjustable_duty_cycle.png (8.7 KB, 10 views)

Torben

Hi BL,

Just another thought. In order to get around the potential for trouble with the short monostable pulses in my above circuit, I've just modified a simple 555 PWM circuit from http://www.dprg.org/tutorials/2005-11a/index.html

This version uses only one 555. The output frequency is not completely stable but it does stick fairly close to 10kHz. In the simulator it varies between 9.4kHz and 9.6kHz. The 100k pot adjusts the duty cycle and R4 and R5 limit the duty cycle from ~1% to ~50%. Again, I can't guarantee the stability in the lower end of the range; perhaps one of the real gurus around here can answer that question for you. If it's not a suitable solution you may have to look at another IC to handle the task.


Again, hope this helps!

Torben

Attached Images

555_pwm.png (10.0 KB, 11 views)

kell

You can make a 555 into a fixed-frequency adjustable duty cycle astable oscillator with a pot and a 1N4148 diode.
String the ends of the pot across pins 8 and 6 (keeping pin 6 connected to pin 2 and the timing cap, as per usual). Connect pin 7 to the pot wiper, and run a diode from pin 7 to pin 6 with the cathode toward pin 6.
And you need an extra resistor between pin 8 and the high end of the pot so you don't short the power supply through pin 7 if you turn the pot all the way to one end. Make that protection resistor 1k (or whatever minimum the data sheet indicates), and make the pot a higher order of magnitude (like one hundred or several hundred k) if you want to be able to get to a really low duty cycle.

Bornloser

Thanks a LOT for the circuits. I GREATLY appreciate the time you put into this for me. I'll probably breadboard one this weekend and scope it to make sure it will do what I need.
I will post back with the preliminary results as soon as I can.
Again, thanks!

Bornloser

Well, seems like there has been a slight change to my plans. After looking at the circuit this is going into, it seems I need for the unit to be HIGH and pulse LOW, not the way I originally posted it. The output of the multivibrator is driving a SMALL solenoid, so I am not sure if the timer will be able to handle the current without a transistor to switch it. Maybe keep the timer circuit as is, and put a pnp instead of the npn???

Gayan Soyza

I'm using a transistor via a resister from 555 output to drive relays, solenoids, etc.....

Most of the time I'm using NPN transistors.

__________________
Gayan
Forum Supporter

Torben

Can you post the circuit you're going to use this in? And the specs of the solenoid? Also, which timer circuit are you using?


Torben

Bornloser

I am using your 555 (NOT the 556) circuit above. This is being put in an auto, it is driving a single throttle body injector, so not much current, but unsure of exactly how many mils.

Torben

OK, I have absolutely no experience with driving those. Sorry. If I had to guess I'd say beef up the output transistor (to maybe a TIP31 or bigger), and either invert the signal with another transistor before the output transistor or use a PNP (TIP32). At this point I should warn you again that I am officially guessing. Or else swap the timing resistors in the 555 circuit so that you're using the other half of the duty cycle (i.e. 51% to 99% instead of 1% to 50%)--but I would still look into using the bigger output transistor.

Do you have the make and model of the injector? I've found some forum posts about hotrodders using 555 PWM circuits to drive their injectors but haven't really dug much deeper than that.


Torben

Bornloser

No, I dont have any model on the injector, but it cant draw much current, I would be surprised if its more than 10 mils. Do I need to do anything else besides replaceing the npn with a pnp?
Thanks

Torben

Well, if the thing really is only going to draw very little amperage (by 'mil', do you mean 'mA'?) then what I would do to invert it is shown below. As long as the PNP can handle the solenoid current it should be OK. I just used a 2N3906 in the simulation since it was available; you might need something bigger (although for 10mA the 2N3906 should be more than adequate).

I'm no expert but I must confess I'm a little skeptical that the solenoid draws under 10mA.

I've added a resistor to the PNP to ensure that it's completely turned off when it's supposed to be. I've also added a protection diode in reverse-parallel across the solenoid, in order to keep any back EMF from kicking back and cooking the transistor when the solenoid is turned off. Your solenoid might include a protection diode; I don't know how common that is. I've shown a 1N4001 which I would think is a bare minimum; I don't know whether you need higher or not (i.e. 1N4002 or such).


Hope this helps,

Torben

Attached Images

555_pwm_pnp.png (2.2 KB, 6 views)

Bornloser

YOU ARE THE MAN!!!!
Many thanks!