Ultra-Simple H.F. Test Oscillator

Tony Deff · Joined: May 25, 2008 Posts: 51 Location: Suffolk, UK

(This posting is undergoing major editing)

You (more than) sometimes find that the timing of your 7555 is not quite as you'd expected, and by more than a trifle.
Frequently, you are told that this because of your component tolerances, yet you may look at your 2% resistor and 10% capacitor and wonder. What you are seldom told is that the 555 has its tolerances too, especially in its built-in resistor chain.

Resistors inside semiconductor networks are notoriously imprecise, but they are supposed to match each other closely, and obtaining a ratio of 2/3 supply at pin 5 is what it's all about. In National Semiconductor's original LM555, this is formed by a chain of three nominal 5K resistors (in the 7555, these are 100K and in the TS555, 200K).

The LM555's spec. sheet details the accuracy of these resistors in a curiously indirect way: it details the Control Voltage Level (Electrical Characteristics) min., typ. and max. for a supply of 15V and 5V. For the LM555C (Commercial variant) at 15V, this is 9,10 or 11V, i.e. ±10%.

Because of the exponential function, a 10% deviation in the target voltage is distorted by much more in the time domain.

If you look at Graph 1, a target reduction from Y=66.7% (red line) to 60% (a 10% reduction) reduces the time from X=1.1 CR (blue line) to about 0.9 CR — a reduction of some 18%.

A target increase of 20% (from 66.7% to 80%) increases the time from 1.1 CR (Time Constants) to 1.6 CR — an increase of some 40%!

Other tolerances that affect the timing are the Threshold input current (negligible in the CMOS versions) and leakage into pin 7.

A way of increasing the time period up to double thus presents itself by simply strapping pin 5 with a resistor to supply, increasing the threshold voltage to which the capacitor must charge.

Resistance is futile.jpg

Description:

Filesize:

15.48 KB

Viewed:

2009 Time(s)