Re: Simple 555 PWM - disappointing performance

Terry Given <my_name@xxxxxxxx> wrote:

>Terry Pinnell wrote:
>> Terry Given <my_name@xxxxxxxx> wrote:
>>
>>
>>>Two comments re. your circuits.
>>>
>>>Firstly, you brute! 100nF slapped on the output of poor old CMOS gates.
>>>Still, rise time is slow, around 300ns IIRC, so I = C*dV/dt =
>>>0.1uF*13V/0.3us = 4A, IOW it'll take a month of sundays to rise, and
>>>frighten the pants off of the cmos gate. I would have slicer R3,R4 in
>>>half, and popped the caps in there.
>>
>>
>> Hadn't thought about those sort of implications - thanks.
>>
>>
>>>Secondly, the caps across the 2 switches. These will, at turn-on, pull
>>>4001-6 and 4001-13 high (actually it will be an exponential spike, time
>>>constant 10ms). Whether or not this is an issue, I leave as an exercise
>>>to the reader (IOW I didnt check) but its something to be aware of.
>>
>>
>> Not sure I follow you; could you amplify please?. The idea of both
>> those RC filters was twofold:
>>
>> - reduce switch bounce (although, in theory, it shouldn't really
>> matter, as the bistables should latch on only the very first pulse).
>>
>> - reduce noise, as the wiring to the main unit was likely to be quite
>> long.
>>
>
>that they do.
>
>Initially, these capacitors have no charge. that means Vcap = 0. At the
>"instant" of turn on, the +13V supply rises (however fast or slow) to
>+13V, but there is 0V across each of the debounce/filter caps. for the
>caps to 0V, this means a logic zero. for the caps to +13V, (open/stop
>and close/stop switches) the resistor is *initially* +13V and hence
>logic level 1. Then the cap charges (to +13V) thru the 100k resistor,
>time constant 10ms, resulting in an exponential decay to 0V, at which
>point (really depends on gate Vil) the logic level becomes the zero you
>expect..
>
>If the +13V supply rises much slower than the RC time constant, this
>effect can be masked. its quite common when using lab supplies,
>especially if you turn them "off" with the current limit dial, like I do :)
>
>I've shot myself in the foot just recently with exactly that - a
>TUSB3210 based system, the micro's oscillator *wont start* if the supply
>voltage ramps up too slow! I spent a few hours trying to figure out why
>my programmed micro lit up LEDs it shouldnt....doh.
>
>Cheers
>Terry

Thanks for the follow-up.

--
Terry, West Sussex, UK

.