# How T=R*C tolerance of capacitor?

Discussion in 'The Projects Forum' started by mfr, Apr 24, 2011.

1. ### mfr Thread Starter New Member

Apr 24, 2011
19
0
Hello,

I have a circut that I found on this forum, that I don't want to keep powered up, rather I'd like the power to be switched on via the same switch that triggers the timer delay. I thought that if I introduced a 50ms delay to powering up the 555-2 and that if the Frequency of the 555-1 clock is high enough, then the circuit would still generate a 'random' delay anyays due to the hopefully insconsistent (tolerance) behaviour of a capacitor in serises with a resistor used to delay the power of 555-2 by 50msec or so. I was thinking at 3.9K resistor and a 4.47uF capacitor to ground the LM555-2 for 50-60msec.

Is that a possible route?

The circuit:
https://cid-9c1723531baec47d.skydri...sid=9C1723531BAEC47D!204&authkey=R1KjyQY!CJs\$

Basically I'd like the ON/Off switch to trigger the random delay without a momentary-on switch, hence only one shot at it until powered off manually.

Thanks,

Mihai

Dec 26, 2010
2,147
300
Try it and see, but I would think that the RC timing would not be all that random.

3. ### mfr Thread Starter New Member

Apr 24, 2011
19
0
Well if the 4017 counter clock would be at 1MHz, then would you not expect some level of inconsistency on the part of the RC network bringing the second 555 online 50 ms later ? I'd expect the counter to hopefully not be on the same count at that moment...

Dec 26, 2010
2,147
300
As I said, try it and see. Some inconsistency is probable. Whether it would be enough to get a really random result is hard to say. My gut reaction is probably not, but I do not have any maths to support that - do you? Unless you can model or otherwise calculate the statistics for the RC timer, I don't see much alternative to trying it.

That's not the usual sense of "tolerance" as applied to resistors and capacitors, by the way. Different examples of a 10% capacitor may vary by that amount, but an individual capacitor would not be expected to drift by that amount, at least not in the short term. If you want to achieve true randomness (which is actually very difficult) you might try looking at a noise generator.

5. ### designnut Member

Apr 21, 2011
33
1
If you want to randomize something look into a PRNG using a shiftt register and gate feedback. I used it for many uses.