Stopping power glitches

william chan

My SX processor is powered by a transformerless circuit from 240v AC.
What I am afraid is that spikes from the AC line due to sparks and switching of loads may cause the SX to reset or go haywire.

What is the best way block these spikes?

Thanks.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my

VIRAND

If you're using a diode, a zener, and two capacitors for a power supply, that blows everything sooner or later.
The least you can do is use large, high voltage rated parts.

A wallwart will sacrifice itself and save your chips.

The next best thing is probably adding 2 or 3 big MOV's.
But then the first big surge will smoke them
and the second one will smoke your chip,
unless there is also a fuse before the MOV (which will definitely "reset" your chip...but hopefully not blow it).

Mike Green

Adding an RC or LC filter to the power supply output can help dampen transients. You can also add LC filtering to the power supply input. Often these are designed to suppress noise from a switching supply from getting into the power source, but they also work to suppress transients from getting into the power supply. Often they come built into a power cord socket / switch / filter assembly as seen in desktop PC power supplies and they're fairly cheap because of the volume of business.

william chan

What values of the R and C would you recommend after the zener diodes?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my

Mike Green

You'll need to describe your power supply including some idea of the current drain and voltages involved.

Beau Schwabe

william chan,
·
If this is an ESD event that you are trying to block, then this circuit might work for you.
·
It's basically a derivative of the type of ESD circuit that you would see in an IC, except that the PNP/NPN combo would be replaced with a single MOSFET.· The principle idea however is the same whether it's an actual ESD event, or there is an extremely noisy power supply.
·
The basic theory of operation is that you charge a capacitor through an RC·circuit·(22 Ohm + 470 Ohm + 100uF)· ...·this·in combination with the Zener creates a voltage reference (or low pass filter) between the capacitor and the 470 Ohm resistor that remains steady, even with a spike... we can’t really use this though to do much real work, because the 470 Ohm resistor is fairly high.· It limites the available current down to about 10mA.· If we lower the resistor value, then the available current increases, but the voltage reference becomes more susceptible to the noise spikes that we want to filter and stay away from, so·here in this circuit we are only using it as a voltage reference.· Next, notice the·PNP.· Under normal conditions, the PNP transistor is OFF.· Since the Base of the PNP is connected to our reference, if the Emitter experiences a positive voltage 0.6V above the reference (5.1 Volts in this case) due to a voltage spike on the input,·the PNP will turn on causing·the NPN transistor to also turn on.· Because of the configuration of the NPN, it is designed to snub or short out the voltage spike.· Depending on the amount of energy in your spikes, will·determine how the NPN transistor should be rated.·· That takes care of positive voltage spikes.· If the voltage spike goes the other direction, the reverse biased diode is all that's necessary to do the trick.· It too should be rated to handle the amount of energy in your spikes.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 8/22/2008 3:55:46 AM GMT

william chan

Beau,

Where you get this circuit? It looks ingenious!
But I think the zener should be at least 7v so that there is enough voltage drop for the 5v regulator.

I'll implement it in my board right away.

Thanks!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my

Post Edited (william chan) : 8/22/2008 5:10:22 AM GMT

Beau Schwabe

william chan,
·
This circuit is typical of what is used in the IC industry for capturing an ESD event.... In most cases however a MOSFET transistor is used.
·
I converted it, so that Bi-polar transistors could be used also.··
·
"But I think the zener should be at least 7v so that there is enough voltage drop for the 5v regulator"
·
... The intent was that 'Vsupply' was somewhat regulated or close to the Zener voltage to begin with·... The output that was denoted as saying 'To Regulator',·should probably just say 'regulated' because in a sense it is a 5.1V regulator in it's own right.
·

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Beau Schwabe

IC Layout Engineer
Parallax, Inc.

Post Edited (Beau Schwabe (Parallax)) : 8/22/2008 6:17:21 AM GMT

John Bond

Thanks Beau - That circuit is a gem

John Bond

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

metron9

I did a simulation but modified the circuit a bit. Using a 12V AC source and some extra caps to stabelize the voltage. I used a LED instead of a zener with a 6.5V forward voltage. i also put another LED in there to simulate a load. The simulator gets confused if you don't use resistors on the base of transistors so they are there as well.


·Select File Import on the simulator once it opens cut and paste code below and click import.
·This link will open the simulator http://www.falstad.com/circuit/

$ 1 5.0E-6 12.050203812241895 76 5.0 50
g 176 336 128 336 0
t 272 176 336 176 0 -1 4.502357099764968 -0.5262228271953484 100.0
t 336 240 384 240 0 1 -5.0287170484208055 0.6413520818545151 100.0
r 176 96 176 176 0 470.0
c 176 176 176 320 0 9.999999999999999E-6 5.143846303079973
w 176 320 176 336 0
w 176 336 272 336 0
w 336 96 336 160 0
w 336 96 272 96 0
w 272 96 176 96 0
w 176 176 272 176 0
w 384 224 384 128 0
w 384 128 384 96 0
w 384 96 336 96 0
w 384 256 384 336 0
w 384 336 272 336 0
d 432 336 432 96 1 0.805904783
w 432 336 384 336 0
r 176 96 96 96 0 10.0
s 96 96 96 176 0 0 false
w 96 208 96 176 0
d 96 240 96 208 1 0.805904783
R 96 240 96 272 0 1 100.0 12.0 0.0 0.0 0.5
c 496 144 496 288 0 0.0030 5.031430330094042
d 384 96 432 96 1 0.805904783
w 496 288 496 336 0
w 496 336 432 336 0
w 496 144 496 96 0
w 496 96 432 96 0
162 576 144 576 336 1 2.1024259 1.0 0.0 0.0
w 576 336 496 336 0
r 576 144 576 96 0 220.0
w 576 96 496 96 0
O 576 96 576 48 1
g 48 96 48 48 0
r 336 192 336 240 0 1.0
r 272 336 272 304 0 1.0
162 272 176 272 224 1 6.5 1.0 0.0 0.0
c 96 96 48 96 0 9.999999999999999E-5 5.8321966297834
w 272 224 272 304 0
o 4 64 0 38 10.0 1.953125E-4 0 -1
o 22 64 0 38 20.0 0.4 0 -1

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!

william chan

What was the result of your simulation test?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
www.fd.com.my
www.mercedes.com.my

metron9

Well the circuit maintained 5V output with an input signal or 100 HZ AC 12V simulating 12V spikes. Go ahead and give it a try.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!