[mike s]

On Sunday, April 17, 2011 4:19:12 AM UTC+1, Big Bad Bob wrote:
On 04/15/11 03:34, mike s so wittily quipped:
On Thursday, April 14, 2011 11:13:26 PM UTC+1, mick wrote:



I just grabbed that book. Good isn't it?

Let's see...
opto-isolator fed from bias supply: input into PIC
No further action without bias
PIC ramps up heater(s)
Allow some warm-up time
Close a relay to allow B+ via a resistor
Another delay
Close second relay to short out B+ resistor& allow full supply

owzat?


Nice, a "finite state machine". So now all that's needed is something similar for safe power down, oh and the nasty cases for when the person switching it on, or off, changes their mind part way through a sequence and throws the switch the other way.

properly handling rapid on/off sequences would be a 'nice touch'. I'd
start the sequence from the beginning again each time. Probably the
safest way. As for power off, easiest method, just "power off".


Sometimes it's not just a 'nice touch'. I know of someone destroying tens of thousands of pounds worth of computer kit by doing just this. Heck I've even heard of people being killed by gas purge systems that didn't go through the cycle properly. Sometimes the phone ring when you throw the switch and there seems to be an in-built reaction to throw it back.


My PSU is very simple, but as it has a fairly hefty directly heated rectifier I added a thermistor to protect against power switch jitter - which would tend to blow the mains fuse. Almost certainly because the rectifier was warm, but the smoothing cap discharged, so the rectifier behaved more like silicon than vacuum on these occasions.

thermistors and similar components seem to be the most common way of
handling this 'old school' style.

Yet I'll bet that nearly every fancy electronic system drawing more than a few tens of watts also has at least one NTC thermistor to handle some part of the switch on surge. Most likely need one - to protect the fancy start-up system at least.