Non-standard 555 Astable Multivibrator Problem

[SteveUK]

As part of a project, I need to make a square-wave with variable frequency, but 50% duty cycle. The 555 timer seemed like a good idea, but with the normal astable multivibrator circuit you see everywhere, changing the frequency always ends up changing the duty cycle a bit. However, In my 20 year old "IC 555 Projects" book I found a circuit which it specifically states has a variable frequency but 50% duty cycle.



However, after building the circuit (see attached) (and driving an led with it), my oscilloscope shows a duty cycle well above 50%, more like 75%.

Not shown on the diagram is my decoupling of the power rails, and my decoupling of pin 5 (the control pin) to ground with a 0.1uf capacitor (this is often recommended if you are not using it).



So is this circuit flawed or could it be my 20 year old NE555 timer I'm using?


The ultimate plan is a stroboscope that can go well into audio frequencies with one pot controlling frequency and another controlling duty cycle. I will send the square wave from the 555 through an integrator, producing a triangle wave, then stick that through a comparator so I can vary the duty cycle completely independently of the frequency (which doesn't seem possible with the 555 by itself).

[IanB]

That circuit should produce a 50% duty cycle.

You can further investigate what is happening by probing the voltage on the capacitor, at pins 2 & 6. You should see a sawtooth moving between 1/3 and 2/3 of the supply voltage. (Make sure the capacitor value is large enough that the probe capacitance does not distort the circuit.)

You may also have a problem with the LED loading the output down too much. Remember the output is providing current to charge up the timing capacitor, and if you make it also drive the LED you might overburden it. Put a high value resistor in series with the LED to limit the current, or put a high impedance buffer like an op amp between the 555 and the LED.

In the same line of thinking, don't make Ct too large and Rt too small. You want to keep the output well within its current source and sink capabilities.

[SteveUK]

The resistor is 1K. Capacitor is 66.2 nF.

The "Sawtooth" is from pin 6 (riding a bit high since I hadn't fully zeroed the trace).

The other shots are pin 3 (output) with and without the resistor load.

One curious thing is that the book gives the period as 1.4CR, giving 10789.81Hz, or 10kHz.

However, from the scope trace, it comes out as about 6 KHz. The timebase is 50us, and the period of the waveform is about 3.2 divisions 1/(3.2* 50x10^-6) =  6250 Hz.

[westfw]

The circuit I usually see for modified duty cycles uses a diode: http://www.kpsec.freeuk.com/555timer.htm#dutycycle

[Rerouter]

alternativly, you could do it using an op amp http://hyperphysics.phy-astr.gsu.edu/hbase/electronic/square.html#c2

[IanB]

The resistor is 1K. Capacitor is 66.2 nF.

The "Sawtooth" is from pin 6 (riding a bit high since I hadn't fully zeroed the trace).

The other shots are pin 3 (output) with and without the resistor load.

Clearly you have a problem with the operation of the circuit as the capacitor is taking much longer to charge up than to discharge. We need to ask why that might be? I don't immediately know, but something is clearly wrong there.

What is the supply voltage to the circuit? Also what is the voltage scale on the various traces? Is the output pin voltage changing between 0 and Vcc as you would expect?

[SeanB]

Try using the CMOS version of the 555. It has the advantage of a rail to rail output at low currents, and at least has equal source and sink capability. Otherwise use a 10k or higher resistor and connect it to pin 7 instead of 3, and use a 470R to 1k resistor from 7 to the supply. That should help with loading.

The big problem with the bipolar 555 is the high output voltage is not well defined, being rather load dependent. Sink is a beefy transistor for both pin 3 and 7 ( they are the same size transistor on the die and are driven the same) so you definitely need some active pull up to the circuit. Better is to have the load be switched to ground rather than source current.

[qno]

Double the frequency and divide the output of your 555 by 2 with a Flip Flop.
guaranteed 50% duty cycle.

You do not really need a 555 then. Use your favourite multi vibrator circuit.

[SteveUK]

Clearly you have a problem with the operation of the circuit as the capacitor is taking much longer to charge up than to discharge. We need to ask why that might be? I don't immediately know, but something is clearly wrong there.

What is the supply voltage to the circuit? Also what is the voltage scale on the various traces? Is the output pin voltage changing between 0 and Vcc as you would expect?

The vertical scale on all those pictures is 0.2V/Div, but accounting for the x10 probes I'm using, it's 2V/Div. The trace is slightly too high in those pics. I'm supplying 5V to the circuit and the output pin is going between 0 and 3.6V.  The Capacitor charge seems to be going from 1.8V to 3.5V.  Since the capacitor is supposed to be charging and discharging through the same resistor, it's odd that the times are different.

I've just tried replacing the resistor in case that was behaving oddly, but there is no change.
I've also just tried replacing the 555 timer (with another 20 year old one). No change.

I've just changed the capacitor to an "identical" ceramic disc capacitor, and realized that the one I was using was 65.1nf which is 38% off from its stated value of 473nf. I've tried several different capacitors, even different types, and the charge time is always much longer than the discharge time. The only change is that the output square wave now seems to go between 0v and 4v rather than 0 and 3.6.

So, in total I've replaced every component in the circuit and checked the wiring many times, yet things are still amiss. I suppose I could rebuild it further up the breadboard, but I'm just clutching at straws.


P.S. I've only just acquired this scope. There is a slight anticlockwise rotation on the trace I think. I'll have to sort that out.

[ModemHead]

This circuit will never produce a 50% duty cycle.

The output of the 555 can swing from near 0V to two diode drops less than Vcc, by design.  The capacitor will charge to 2/3Vcc (3.33V) and discharge to 1/3Vcc (1.67V), again by design.

So the cap charges up from 1.67V by being connected to 3.60V through a 1K resistor, which is a 1.93V difference.  The cap discharges from 3.33V by being connected to 0V through the same 1K resistor, which is a 3.33V difference.  So the discharge current will always be higher than the charge current.

added:

@SteveUK, I just re-read your "ultimate plan". Just feed the sawtooth from your current 555 timebase to a comparator with an adjustable threshold from 1/3Vcc to 2/Vcc.  You should get a variable duty cycle square wave.  There will be some non-linearity between the threshold voltage and the duty cycle, but you may find it acceptable anyway.

Also, raising the supply voltage will diminish the effect of the 555's voltage drop on the output.

[IanB]

I'm supplying 5V to the circuit and the output pin is going between 0 and 3.6V.

This is the reason the circuit doesn't work. It will only produce a 50% duty cycle if the output pin swings between 0 and 5 V, for the reasons explained by ModemHead above.

[SteveUK]

Yes, Modemhead's explanation makes sense.  So the circuit design only works for the ideal 555 timer that can output 0 to Vcc and not the NE555's I've got.  How annoying.

I thought maybe the best solution is to get a more modern 555 that can actually go rail to rail on it's output, but so many of them don't seem to be able to. I am beginning to wonder whether the author tested this circuit at all!

Perhaps I should just try making a square wave generator with an opamp instead, since I will be using opamps for the rest of the project.

[w2aew]

Page 10 on this datasheet http://www.ti.com/lit/ds/symlink/lm555.pdf shows a variant of the typical astable multivibrator 555 circuit that should be able to be adjusted to 50% duty cycle due to the different charge and discharge paths.  That is, if you still want to do this with a 555...

[Zero999]

Try using the CMOS version of the 555. It has the advantage of a rail to rail output at low currents, and at least has equal source and sink capability.

Yes, that circuit requires a CMOS 555 timer. It won't work with 50% duty cycle with a standard 555.

Double the frequency and divide the output of your 555 by 2 with a Flip Flop.
guaranteed 50% duty cycle.

You do not really need a 555 then. Use your favourite multi vibrator circuit.

Yes, a divide by 2 counter will work, in which case I'd recommend a dual flip-flop IC such as the CD4013 so one half can be configured as astable and the other as a divide by 2 counter. Another option is to use the CD4060 oscillator + counter IC.

[SteveUK]

Yes, a divide by 2 counter will work, in which case I'd recommend a dual flip-flop IC such as the CD4013 so one half can be configured as astable and the other as a divide by 2 counter. Another option is to use the CD4060 oscillator + counter IC.

I can make an astable multivibrator with a D-Flip Flop?
I was looking at a circuit earlier for making an astable multivibrator with an inverter, a resistor and capacitor.

This one.


So I dug out an old SN7404N TTL inverter chip to try it out. Sadly though, it really doesn't seem to oscillate, at least not in anything resembling a square wave. I think the signal I am getting on the output is just spurious, I can't see an charge cycle on the capacitor either.

I tried another circuit using a SN7400N Quad Nand gate chip and that fails to oscillate as well.  Are oscillators made with logic chips generally reliable? I guess I'm doing something wrong somewhere.

I'll order some CD4013's though, since at least I'm getting an adjustable square-wave out of the 555.

[IanB]

Are oscillators made with logic chips generally reliable?

They should be, but as with all things it depends on the exact details of how you construct them.

In the example on the page below the circuit is constructed with CMOS logic:

http://www.electronics-tutorials.ws/sequential/seq_3.html

It may not work the same with with TTL logic.

[westfw]

Oscillators made from logic gates tend to work much better with CMOS logic.  With TTL, you have to worry quite a bit about how much current goes into the gate input, carefully biasing the gate into "linear" behavior, while CMOS gates are pretty much ideal switches.

Logic-gate oscillators also tend to vary in frequency with supply voltage and such, more than a 555 would (OTOH, they tend to be used with regulated power supplies, so...)

[lomejordeesto]

Hey what I do in these cases is take the 555 output and take it to a 74LS74 flip flop. That way you have variable frequency and 50% duty cycle. There is a PDF attached with the schematics. (You don't have to use the transistor is just that I ran out of inverters xD).

[Zero999]

Here's how to make an astable using the CD4013. I've done this before - it works perfectly.