desktop pc, cd-roms, hdd, and many more are "prone" to this accident. anyone could explain what is really happen there?

thanks
Posted on 2003-11-27 02:04:51 by dion
I would have thought as an electronic enthusiast that you would understand what a SPIKE is, and what it does, and how to filter it... a better question would be why spike filtering is not inherent in all power supplies as a minimum standard !!
Posted on 2003-11-27 02:18:22 by Homer
uhm... when that spike occurs? at power up or power down? ahh... i guess power down...
spikes in power down? i never heard about it. just heard at power up only
Posted on 2003-11-27 02:25:42 by dion

I would have thought as an electronic enthusiast that you would understand what a SPIKE is, and what it does, and how to filter it... a better question would be why spike filtering is not inherent in all power supplies as a minimum standard !!


my sentiments exactly, damn cheap world :mad:
Posted on 2003-11-27 05:16:16 by Hiroshimator
I will try to explain it in simple words:

Because powerlines have an implicit inductance, and while the power flow is established the poweline itself accumulates an magnetic energy. Depending on the values of implied inductance(s) and the value of current flow this energy can be quite big.... and it just sits there waiting for....

The time when the power is suddenly cut off: the magnetic enery will attempt to keep the current flow constant (the law of electromagentic induction) and because the connection is severed it can do so ONLY by growing the tension very fast and very high.

Of course it will fail ultimattely (aka when the energy accumulated into its implied inductance will be wasted), but by then 1-2KV or even more higher voltage pulses might have reached your sensitive electronic devices...even for a very short time this can be very harmfull.

Why are those things not filtered and sources not protected against it?

The question is redundant, the answer is simple: money,competition and capitalismus

Mostly ALL things we hummans so have 2 reasons:
1) make it possible at first
2) make is cheap latter

the seccond reason takes is tribute very fast in capitalismus
Posted on 2003-11-27 10:19:10 by BogdanOntanu
uhm... is it at the power supply, the place where i should put the filter on? any detail on that? is this called "surge"?
Posted on 2003-11-27 20:29:59 by dion
Yes theoretically you should place a fast Zenner diode with a reverse surge tension slightly greater than the normal voltage input.

The Zenner diode will abruptly start conducting when a tension spike appers and limit voltage to only a few Volts above the normal value... sometimes with the cost of its integrity. Varistors could also be used.

Against current spikes one can use a torroidal core with two antiphased inductors that will anihilate eachothers spikes...

etc
Posted on 2003-11-27 21:32:40 by BogdanOntanu
BogdanOntanu pretty well hit the nail on the head.

Heres the math: V = L * dI/dt

The voltage developed is such that the current it produces OPPOSES the reduction of current flow.

Here is Canada, eh, we have a 120v distribution system (AC at 60hz). If im drawing say 30amps of current into my home at 120v and the inductive values in the lines are say 30mH, and all of a suddent the power is cut out at the source mid cycle (90 deg in the sine wave) i have peak current being interupted (30amps) and falling to 0 amps within 1uS (assuming low capasitance in the lines).

Do the math -> V = 30mH * (30 A)/(1uS) = 0.03 * 30 * 1000000 = 900,000V

This is one hell of a kick into my fuse box. The only reason it doesnt fly off the wall is becaue the POWER, or energy stored in the lines, is very low. At this voltage the current is low (much like a static shock you get sometime from your clothes). But when it comes to computer related equipment the chips and components are often sized to handle 2 times the operative voltages. If my power supply gives me 12Vdc, capasitors are in the range of 30v dc. Getting a hit form thousands of volts for a very short period of time allows current to *tunnel* through the devices insulation or dielectric component, effectively crippling the device. This is why your equipment *fails* to operate.


How to fix? Best option is to put a crow bar circuit in to shunt excess voltage to ground. Zener diodes are also a cheap fix, but can get tosted from too much current...

BTW, for similar reasons you should always put reverse biased diodes across stepper motor coils, such that thier induced current kickback when the coil is denergized is alow to shunt through the diode (in its forward biased direction). The higher the rate of current from the kickback, the more energy will be released per second and depleats magnetic fields causing the kickback voltage in the first place.

Hope this helps..
Regards,
:NaN:
Posted on 2003-11-27 21:44:44 by NaN

How to fix? Best option is to put a crow bar circuit in to shunt excess voltage to ground. Zener diodes are also a cheap fix, but can get tosted from too much current...


crow bar? what is it?? yes, i dont want to use zener, because if it is burn, the rest of the circuit still get the impact.
Posted on 2003-12-01 04:56:45 by dion
Another good protection would be highly insulated steep down magnetic transformers
Unfortunately they are heavy and costly

Most switching power supply used in computers are cheap
(because they elimminate the above transformers)
but are do not provide enough electrical separation from the main power lines...so such pulses propagate easyly into PC power supply and from there to the very sensible ICs on mainboard

As always on this planet: we know exactly how to do things right, but it is NOT economically viable because of competition... and we like to lie about "efectiveness"
Posted on 2003-12-01 06:29:35 by BogdanOntanu

Another good protection would be highly insulated steep down magnetic transformers
Unfortunately they are heavy and costly


do you mean isolation transformer? each side ,primer and secondary is separated winded?
Posted on 2003-12-02 05:41:19 by dion
Right! what Nan said. That is very important to put diodes in reverse bias accross all selenoids to block back EMF. 1N914's are usually suitable.
Posted on 2003-12-02 07:34:04 by mrgone
Search digikey.com for Varistor and ZNR. These are crow bar devices. Once a voltage is reached, the device's resistance drops extremely quick, thereby clamping any spikes.

Are you using a surge protecter?

I highly recommend a surge/brownout protecter.

A brownout is an under voltage condition (incandescent lights look brown). Had a copier damanged by brownout.
Posted on 2003-12-03 02:34:01 by eet_1024
does 78xx/other regulator will protect the circuit from those phenomenon?

btw, what unit used by surge protector device?
Posted on 2003-12-03 23:29:50 by dion
A 78xx is a bad choice for a surge supressor, it will breakdown at the voltages concerned and may just end up being no more than an expensive peice of wire. Zener diodes are a better choice, you can also use Metal Oxide Varistors (MOVs) though they are generally slower to switch than Zeners. If you look for data sheets on different MOVs you will find schematics for various surge suppressors. In general the cuircuit has to provide low impedance at voltages higher than the desired voltage and high impedance at the desired voltage or lower. They usually follow this type of configuration:
Posted on 2003-12-03 23:43:31 by donkey
What are you trying to protect? What do you want to protect it from?

Is it a consumer appliance you want to protect from blackouts - where spikes are generated by the removal of current from inductive loads (like motors)

If it is your own ciruit, use something like the schematic below. The varistor clamps the voltage, the inductor isolates the varistor, and the caps smooth out the supply. The reason for isolating the varistor is so that it just protects your circuit, and doesn't have to crowbare everything on the supply; which can be very noisy in an automotive application.
Posted on 2003-12-04 00:25:53 by eet_1024
thanks for all the replies, before.

i do not know exactly what i'm gonna protect, say in specifically, my desktop pc, from burned out because of the reason you guys said, if i am not wrong concludes, a surge phenomenon.

yes, because no other devices except pc's smps, that powering all the desktop stuff, i guess, i need to do something with it.

NaN told the theory, and some other suggest crowbar devices, such as znr and varistor. in this point, i dont want to use zener, because if it is failed, it wont protect the rest circuit anymore, i mean no such mechanism that if its broken, it just something like blown up fuses and disconnect the powersupply line (as i know that znr was put in pararel manner to the line, not serial).

while you, eet, said somewhat brownout?, i dont know if it can be like that.

donkey, if 78xx just broken when it faces the surge, do you ever found a broken one with out voltage become the same like input voltage, say, in=12, out =12 for 7805?

i never found such, instead a lower output. in this case, if it was good enough that it wont be bad in such manner(in=out), i can concude i can use these conventional regulator only. ahh... of course i cant use 78xx to supply my pc. i have to find another one.
Posted on 2003-12-04 05:01:58 by dion

donkey, if 78xx just broken when it faces the surge, do you ever found a broken one with out voltage become the same like input voltage, say, in=12, out =12 for 7805?


Yes.

A 7805 would not be the best choice for a 12 volt supply though, there can be alot of heat generated by dropping 7 volts depending on the current draw of the load. As with all devices the key is how much power (P=VI) is passing through the device and will be converted to heat. I think the maximum voltage input for the 7800 series is something like 18 volts, you would have to check the sheets to be sure though. So say that you were drawing 1 amp and had an input to a 7805 of 12 volts (-1.4 for the bridge)

5.6V * 1amp = 5.6watts

5.6 watts is alot of power for the device to absorb, you'll need a heatsink :)
Posted on 2003-12-06 14:39:29 by donkey
If you want to protect a consumer appliance, like a PC, use a surge protector. Get one that comes with a warranty that covers the equiment that is connected to it; so even if you PC is fried, the company will pay to repair or replace it.

Note that the warranties don't cover acts of god, such as lightening. If you are at risk of lightening damage, install a lightening arrester.

If you are designing a product that uses >500mA, you should consider using a switching supply. If you do go this route, you will have to do a lot of design. Even the component placement on the PCB is important. And watch out for caps that explode and/or catch fire.

Donkey,
If you're using a solid 12V AC input with a large cap, the regulator could see a continous 16-17V.

The schematic I post is for running a PIC and some LED's off of an automotive supply.

A good rule of thumb for heat is: If you smell flesh burning when you touch it, you need a heat sink.
Posted on 2003-12-08 23:32:56 by eet_1024

Donkey,
If you're using a solid 12V AC input with a large cap, the regulator could see a continous 16-17V.

The schematic I post is for running a PIC and some LED's off of an automotive supply.

A good rule of thumb for heat is: If you smell flesh burning when you touch it, you need a heat sink.


That's why you have eet in your nick :)

I just sort of did it in my head from what I remember when I used to play with electronics as a hobby. Built a few power supplies but nothing really complicated, LRC stuff only no switching but I only needed it for the projects in Radio Electronics. Does that magazine still exist ?
Posted on 2003-12-08 23:37:53 by donkey