How is this IC powered?

From the T input, through the ESD diodes inside the chip.

Which leads to this zero-crossing detector circuit:

Nasty. An external diode from T to Vdd would make that a lot safer.

James

There's not a lot to admire about that circuit. An analog PUT-type pulser, or one of those trigac things, would be cheaper and more reliable.

Well, that's inspired. Earl Muntz lives!

The square wave on the clock gets rectified by the chip's on-board protection diode, which charges the cap, which powers the chip.

It probably also fries the chip over time -- but hey, the thing only needs to work long enough for someone to buy it, right?

You beat me to it!

No, it probably is just fine, under normal conditions. But.... Fence chargers are hooked up outside the house, and the first time a lightning bolt strikes within half a mile, it probably induces a huge jolt into the line input, and poof!

So, I think some protective circuitry would be a good idea.

Jon

And so it worked :)

Jamie

Thanks. I knew there must be an internal path, but had never seen this arrangement and didn't suspect that it was so direct. I must say that even as a very naive circuit-eer it looks really hacky to me.

Another question, about component selection. The 100u charge cap is charged through 2 75R 5w resistors. Giving a time constant of 1 ms. But the cap is discharged only every second or so & a much larger time constant would work. Say 100ms, using 7.5k resistors. Then the resistors wouldn't have to be 5 watt-ers. Am I missing something?

Thanks again, Bob

My thought is, the voltage on the anode of the SCR must go to 0 volts for it to shut off. Mikek

yes, that is exactly what it does. The primary coil will saturate and magnetize, mean while, the charge cap in the supply depletes itself to the point where it can not sustain the minimum holding current. The SCR then opens and the main supply then slowly charges up the main cap via the line R.

Years ago, I used to live in a lot of farm country, I repaired many fence pulsers. I have also built many for people using uC and serial networks so that one can detect when a line got touched and report back to the server PC. This would alert the farm hand or who ever was incharge of watching over the animals to take a look or at least get a record of reports.

Some of the farms around me had a lot of small sectioned areas, each having their own controller.

Mother nature really loves those controllers!

Jamie

I see what you're saying about the SCR turning off, but I don't see how the use of the small (75R) resistors helps that. Wouldn't _larger_ input resistors help the SCR turn off (reduce source current)?

Bob

BTW - if there's a newsgroup more suited for my naive questions, I'll gladly take them there.

rs

=20

The output transformer in conjunction with the series line resistors looks = like a parallel RLC to the SCR. If there is no resistive load on the fence,= i.e. a body of some kind to discharge through and damp the reactance, the = circuit will ring in such a way that the transformer magnetizing inductance= current will keep the SCR turned on. In that case, smaller line resistors = will damp out the oscillations faster than large ones. There are UL, CSA, a= nd other international standards regulating the output of electric fence co= ntrollers, and they do set limits on the allowable duration of transients c= oupled onto the fence. This is why they're using smallish 75R line resistor= s.

like a parallel RLC to the SCR. ... This is why they're using smallish 75R line resistors.

Thanks. I knew they wouldn't be using 5W resistors without a reason.

Bob

The 75 ohm resistors are there for the MOV. That way it won't blow up on the first surge that hits the unit.

?-)

The resistors don't do anything to protect the MOV against the 8x20u current wave transient which is the most common test for protection against lightning induced surge. The differential 150R looks more like a line match from that perspective.