555 50% precision.

[dr pepper]

I want to put togther a cheap nasty smps to power a project.
A 555 is used as the circuit is usb powered, the circuit works well as a 2 switch forward converter but I want a dead simple single switch so I need a 50% duty o/p from the 555 with good accuracy.
I saw somewhere a diode from discharge to trigger, is this going to work?

Edit: I shouldnt be so lazy, looking at the schem its clear to see that the timing cap charges up through the resistor to + and the forward biased diode, then the cap discharges through the resistor between trigger and discharge with the diode reversed, if if the 2 resistors are the same duty should be 50%, I must have poxy resistors or something as I get 60%.

 

[MrAl]

Hi,

If you post the circuit we could look at it.

 

[jpanhalt]

An alternative is to make the 555 oscillate at 2x frequency desired and then use a divide by 2 flip flop. That gives exactly 50% duty cycle. Check out Paisley's site for details.

Edit: You can also use the discharge pin as your output. Again, check Paisley's site.

John

 

[audioguru]

A Cmos 555 can make a perfectly 50-50 squarewave.
A CD4047 has an RC oscillator and a digital divider that also gives a perfect 50-50.

 

[tcmtech]

If you use the output pin 3 going through a resistor to drive the timing capacitor that is connected to pins (2,6) and (1) you can get a 50% duty cycle out of a 555 timer with only two extra external components.

 

[Mikebits]

An alternative is to make the 555 oscillate at 2x frequency desired and then use a divide by 2 flip flop. That gives exactly 50% duty cycle. Check out Paisley's site for details.

Edit: You can also use the discharge pin as your output. Again, check Paisley's site.

John

This sounds like the best way to go.

 

[dr pepper]

Nice idea tcm and jp.
Pin 3, 2 diodes worked.
Paisleys site is interesting.

Something in my head is saying do this properly and use a '34063 with a gdt and a pair of fets in push pull, the gdt having turns ratio to drive the fets to lowest rds on, but it looks as though I'm going to try the hard way first before doing that.

 

[MrAl]

What power you looking for?

 

[audioguru]

If you use the output pin 3 going through a resistor to drive the timing capacitor that is connected to pins (2,6) and (1) you can get a 50% duty cycle out of a 555 timer with only two extra external components.

Nope, not with an ordinary 555 because its output voltage swing is not symmetrical, it does not go up high enough, but the Cmos 555 does when its load is a high resistance.

 

[tcmtech]

It's close enough.

If not exact it can easily be compensated for through pin 5.

 

[dr pepper]

Around 5 watts.
A ir2153 would be good, non of my suppliers have any, the only place I found them is ebay for too much.
I founf this:
http://www.discovercircuits.com/PDF-FILES/dfliposc2.pdf
I could use one as the oscillator and the other in the same chip as a divider, which also gives me a q and not q and do away with a gdt by winding the primary with centre tap and grounding either end in push pull, dead time mgith be an issue (but maybe not with a jk flip flop) I'd have to breadboard it and see if it works.

 

[MrAl]

Hi,

Oh ok, you can get away with more with only 5 watts.

But the 555 oscillator (as audioguru pointed out) will not be exactly symmetrical because of the internal transistor drop unless that drop is very small. It must be very small too, because the sensitivity analysis shows that for a 1 percent decrease in output voltage from the supply voltage will cause a 2.2 percent change in the ratio of the charge time to the discharge time, and that means an on time of about 51 percent and off time of about 49 percent of the total time, and that is for only a small drop in output voltage measured from the Vcc supply line. In other words, if the supply is +10v then 1 percent of that is 0.1v, and so if the transistor drops just 0.1v then the duty cycle is going to be 51 percent on and 49 percent off. This again is because the 2/3 point is only symmetrical with the 1/3 point when the supply itself is charging the capacitor. That might be acceptable for a small 5 watt unit however.

With a drop due to the transistor of VT volts the charge time is:
RC*log((3*VT-Vs)/(3*VT-2*Vs))

while the discharge time is:
log(2)*RC

Now if we make VT=Vs (Vs the supply voltage) then we do see the 50 percent duty cycle, but that does not include the transistor voltage drop then.

Im sure we could find a way to improve this though.

 

[ericgibbs]

hi dr,
Look at this link, its for the Tools section of this site.
https://www.electro-tech-online.com/tools/IC555V2T2.php
E

 

[spec]

 

[jpanhalt]

That circuit is close, because of the >100:1 ratio of (R10 + RV1) :R9, but it is not considered to be 50% duty cycle. It also suffers from having R9 quite small. According to the datasheet:

NB:Ra in the datasheet = R9 in your schematic.

John

 

[spec]

 

[tcmtech]

Personally I don't follow the logic of saying

I want to put togther a cheap nasty smps to power a project.

using a 555 timer IC as the driver for the design then wanting high accuracy duty cycle control using common complaints that are likely to have fairly loose working tolerances to begin with.

Apparently you want to use a USB power port which is good for ~5 watts @ 5 VDC but have not yet said what you are wanting to convert that power into at the end point which is ultimately what will determine the design and choice of components used in the power conversion circuit.

So what are you wanting to power and what are it's power requirements you think it needs?

 

[spec]

That circuit is close, because of the >100:1 ratio of (R10 + RV1) :R9, but it is not considered to be 50% duty cycle. It also suffers from having R9 quite small. According to the datasheet:
View attachment 97272
NB:Ra in the datasheet = R9 in your schematic.

John

Nonsense

 

[jpanhalt]

I guess we will just have to accept your unsubstantiated word for it over Philips.

John

 

[alec_t]

Here's a conventional way of getting an adjustable mark-space ratio, so that 50% can be set simply.