Why no Crowbar in PC PSUs?
Phil Allison
it seems the 5 volt rail on most ATX and other PSUs has no safety crowbar
circuit. An SCR and small zener would do the job and save a lot of
expensive damage if (when) the PSU goes out of reg.
Is there a genuine reason or is it just penny pinching not to fit one?
Regards, Phil
D.Castles
Dave
Rod Speed
Richard Freeman <nos...@atps.net> wrote in message
news:3c22c118$1...@news.iprimus.com.au...
> D.Castles <d.ca...@latrobe.edu.au> wrote
>> Wow! a crowbar on a switchnmode the mind boggles.
> in fact thinking about it you will only need one crowbar on the 5V
> rail as it is only the 5V rail that is regulated in PC power supply
Utterly mangled. If a fuse was used on the 5V rail and the crowbar takes
that out, that wont do a damned thing about the overvoltage on the other rails.
And it aint a trivial matter to have a fuse on the
5V rail with modern PC power supplys anyway.
And there must be a fuse on the 5V rail because there is a very significant fire
risk if the 5V rail is just crowbarred on overvoltage. 200-300W is gunna produce
a pretty spectacular result if you just crowbar the 5V output without a fuse.
> in fact you can buy a chip which does the whole job
Pigs arse it does on that dissipation of 200-300W or a fuse question.
> or looking on the internet it appears Phil is correct a typical
> crowbar circuit consists of a zener an SCR and Resistor.
Pity its more complicated than that with a PC power supply.
You're either gunna have to have a fuse that the crowbar takes out, and
thats not trivial, and wont do a damned thing about the other rails if you go
that route, or you're gunna have to do something about the 200-300W the
supply will continue to dump into the crowbar when a fuse isnt used.
There's a reason that crowbars are very uncommon indeed with PC power supplys.
Then there's the tiny matter of the UL approval etc.
w_tom
Guidelines require a crowbar on all power supply DC outputs.
> The overvoltage sense circuitry and reference shall reside in
> packages that are separate and distinct from the regulator
> control circuitry and reference. No single point fault shall be
> able to cause a sustained overvoltage condition on any and
> all outputs. The supply shall provide latch-mode overvoltage
> protection ....
They couldn't be more specific without using saying overvoltage crowbar. It is
required on all ATX power supplies.
Show me a clone power supply that meets these requirements. So far, I have seen
none. This requirement has been standard on switching power supplies long before
there was an IBM PC. Why are they not in so many clone computers? Clone
computers are sold upon price - not value. Maybe that answers the question?
"D.Castles" wrote:
> Wow! a crowbar on a switchnmode the mind boggles.
> Dave
>
> Phil Allison wrote:
>
> > ...
Rod Speed
w_tom <w_t...@usa.net> wrote in message news:3C23CA02...@usa.net...
> Without directly requiring a crowbar by name, the ATX Power Supply
> Design Guidelines require a crowbar on all power supply DC outputs.
Bullshit. A crowbar is an entirely separate over voltage protection.
There is no requirement what so ever for that in the design guidelines.
AND a crowbar SHORTS the output, taking out the fuse deliberately, and
there is no requirement what so ever for that either in the design guidelines.
>> The overvoltage sense circuitry and reference shall reside in packages
>> that are separate and distinct from the regulator control circuitry and
>> reference. No single point fault shall be able to cause a sustained
>> overvoltage condition on any and all outputs. The supply shall provide
>> latch-mode overvoltage protection ....
> They couldn't be more specific without using saying overvoltage crowbar.
Wrong again. Thats nothing like what crowbar protection involves.
> It is required on all ATX power supplies.
Pity it aint crowbar protection.
> Show me a clone power supply that meets these requirements. So far, I have seen none.
And even the fancy expensive PC power supplys dont either. There's a message there.
> This requirement has been standard on switching
> power supplies long before there was an IBM PC.
Pigs arse it ever was with CROWBAR protection.
> Why are they not in so many clone computers?
It aint even seen with IBM branded PC power supplys.
None of the IBM branded PCs ever had crowbar protection.
> Clone computers are sold upon price - not value.
Pity about the IBM branded PCs that didnt have crowbar protection either.
> Maybe that answers the question?
Nope, just flaunts your complete pig ignorance of what crowbar protection
involves. And how IBM branded PC power supplys were done too.
> D.Castles wrote:
>> Wow! a crowbar on a switchnmode the mind boggles.
Rod Speed
>>> Switch mode power supplies are very Common in Midranges
>>> and Mainframes along with the associated Crowbar
>> Pity they are clearly protecting much more
>> expensive stuff than with a PC power supply.
> Indeed they are 10's of thousands of dollars more expensive stuff in fact
>> I cant think of a single supplier of PC power supplys that routinely supplys crowbar
>> protected supplys, so its very unlikely to actually be due to 'penny pinching'
> Likewise I cannot say I have seen any P.C power supply with
> Crowbar protection - a future silicon chip project perhaps????
I cant see its feasible. Its just too hard to even add the fuses in the output
lines, let alone the SCRs to blow them etc with modern ATX supplys. Even
just a modules that you plug the ATX connector from the power supply into,
with its own cable and ATX connector to the motherboard isnt likely to be
viable given that ATX supplys dont have remote sensing etc.
And I bet thats why even the relatively expensive replacement supplys like
those from PC Power & Cooling dont have crowbar protection for PC supplys.
Its feasible for minis and mainframes because of the much higher
value hardware being protected from a supply gone mad, but not with
PCs where the hardware being protected isnt worth a hell of a lot.
Richard Freeman
>
> > Likewise I cannot say I have seen any P.C power supply with
> > Crowbar protection - a future silicon chip project perhaps????
>
> I cant see its feasible. Its just too hard to even add the fuses in the
output
> lines, let alone the SCRs to blow them etc with modern ATX supplys. Even
> just a modules that you plug the ATX connector from the power supply into,
> with its own cable and ATX connector to the motherboard isnt likely to be
> viable given that ATX supplys dont have remote sensing etc.
The safer bet is to blow the whole Power supply (presuming it doesnt go in
to shut down from over current as they seem to do whenever I have
inadvertantly shorted one out ) at $40-60 for a replacement AT/ATX power
supply it is still cheaper than several hundred dollars for a replacement
Motherboard and a couple of hundred dollars for the Hard drive (not to
mention the data that everyone carefully Backs up :-)
In addition the SCR to do it does not have to sink as much power as
suspected in the event of the PSU continuing to operate. assuming a
pessimistic 1.2V forward drop at 30 Amp the SCR gives us 36 Watts (with most
of the rest of the 300W being dissipated by the PSU) in the event the PSU
continues to deliver power this will have the added affect of dragging the
other Rails down anyway.
well thats my Theory anyway the next and obvious step is to try it I have a
couple of AT supplies sitting around doing SFA if it works with them the
next step would be to find a Sacrificial ATX supply
> And I bet thats why even the relatively expensive replacement supplys like
> those from PC Power & Cooling dont have crowbar protection for PC supplys.
and also the rarity (as you have already mentioned of the feedback circuit
failing and the PSU going overvoltage).
> Its feasible for minis and mainframes because of the much higher
> value hardware being protected from a supply gone mad, but not with
> PCs where the hardware being protected isnt worth a hell of a lot.
Regards
Richard Freeman
Rod Speed
>>> Likewise I cannot say I have seen any P.C power supply with
>>> Crowbar protection - a future silicon chip project perhaps????
>> I cant see its feasible. Its just too hard to even add the fuses in the output
>> lines, let alone the SCRs to blow them etc with modern ATX supplys. Even
>> just a module that you plug the ATX connector from the power supply into,
>> with its own cable and ATX connector to the motherboard isnt likely to be
>> viable given that ATX supplys dont have remote sensing etc.
> The safer bet is to blow the whole Power supply
Thats much harder to do, particular when the power supply has already gone
mad and you cant assume that its shorted output protection is working anymore.
The SRCs would have to be able to dissipate well over 300W with
no fuse on the output rails to blow and thats a completely non trivial
thing to do within the confines of a PC case without producing a fire.
> (presuming it doesnt go in to shut down from over current as
> they seem to do whenever I have inadvertantly shorted one out )
Sure, but thats a non failed power supply. You
cant assume that that protection is still working.
> at $40-60 for a replacement AT/ATX power supply it is still
> cheaper than several hundred dollars for a replacement
> Motherboard and a couple of hundred dollars for the Hard drive
> (not to mention the data that everyone carefully Backs up :-)
Yeah, no argument there if there was some sure way of killing
the power supply when you have decided its gone mad.
> In addition the SCR to do it does not have to sink as much power
> as suspected in the event of the PSU continuing to operate.
> assuming a pessimistic 1.2V forward drop at 30 Amp
You cant assume that 30A. Thats what its designed to deliver
without the voltage sagging. Thats not what it can deliver to
a shorted output rail when the shorted output has failed.
And you cant assume the 1.2V either when you have deliberately
shorted the output using the existing output leads.
> the SCR gives us 36 Watts (with most of the
> rest of the 300W being dissipated by the PSU)
> in the event the PSU continues to deliver power
Thats not right either. 300W is what the power supply can DELIVER
and has nothing to do with what its dissipates internally. What it
dissipates internally is determined by its efficiency, a separate matter
entirely. It'd normally dissipate something of the order of 30-60W
internally when delivering its rated 300W to the outputs.
> this will have the added affect of dragging the other Rails down anyway.
Yes, until the protection circuitry sets fire to the contents
of the case. Which it will do very quickly with the supply
continuing to deliver something like 500W to the outputs.
Thats the reason real crowbar protection has fused
supply outputs. The crowbar deliberately blows the fuse.
And thats a non trivial exercise with a big supply like
that, blowing that 40A output fuse isnt trivial at all to do.
> well thats my Theory anyway the next and obvious step is to try it
> I have a couple of AT supplies sitting around doing SFA if it works
> with them the next step would be to find a Sacrificial ATX supply
I cant see its even worth trying. You must have a fuse for real crowbar
protection and its a distinctly no trivial exercise to design an addon crowbar
to a PC power supply. Even just the fuse alone is a real difficulty, let alone
the fact that you need one for each of the major rails, 3 with an ATX supply.
>> And I bet thats why even the relatively expensive replacement supplys like
>> those from PC Power & Cooling dont have crowbar protection for PC supplys.
> and also the rarity (as you have already mentioned of the
> feedback circuit failing and the PSU going overvoltage).
Sure, tho you'd expect that there would be a market with people like Phil.
Phil Allison
> The safer bet is to blow the whole Power supply (presuming it doesnt go in
> to shut down from over current as they seem to do whenever I have
> inadvertantly shorted one out ) at $40-60 for a replacement AT/ATX power
> supply it is still cheaper than several hundred dollars for a replacement
> Motherboard and a couple of hundred dollars for the Hard drive (not to
> mention the data that everyone carefully Backs up :-)
> In addition the SCR to do it does not have to sink as much power as
> suspected in the event of the PSU continuing to operate. assuming a
> pessimistic 1.2V forward drop at 30 Amp the SCR gives us 36 Watts (with
most
> of the rest of the 300W being dissipated by the PSU) in the event the PSU
> continues to deliver power this will have the added affect of dragging the
> other Rails down anyway.
> well thats my Theory anyway the next and obvious step is to try it I have
a
> couple of AT supplies sitting around doing SFA if it works with them the
> next step would be to find a Sacrificial ATX supply
> Regards
> Richard Freeman
I think you have to consider a number of scenarios in the case of an ATX
PSU with a crowbar on the 5 volt rail. (The 5 volt rail is normaly the one
sensed for feedback.)
Suppose the supply control goes out of regulation due to an internal fault
making the switching transistor conduct at a high duty cycle so all the
rails go up. The SCR crowbar will trigger instantly clamping the 5 volt rail
to about 1.2 volts. A very large current will now flow severely overloading
the supply and limiting the over voltage appearing on the other rails until
either:
1. The supply senses the short on the 5 volt rail and shuts down in a
fraction of a second or if that cct is also faulty,
2. The supply blows its mains fuse.
3. The switching transistor fails and blows the mains fuse.
4. The supply gets hot enough to go into thermal shutdown.
In case 4 the SCR may fail due to overheating and go short. This is not a
concern as you are trying to save a lot of other hardware. Also, the ATX
supply is not usually economic to repair anyhow.
It seems that ATX supplies should have an overvoltage shutdown and latch
system on the 12, 5 and 3.3 volt rails but not all do and how fast it
operates is not clear.
If this system is absent then a crowbar (or two) should be capable of
limiting the consequential damage in case of a runaway PSU.
I await the results of your tests!
Regards, Phil
grog
Would it not be too slow? By the time the Zenner does it's zennering and the
SCR goes about blowing a rather large fuse, the xGHz computer has probably
enough time to send an email informing of the pending PSU failure.
greg
Rod Speed
Phil Allison <bi...@bigpond.com> wrote in message
news:PevV7.14358$HW3....@newsfeeds.bigpond.com...
> Richard Freeman <nos...@atps.net> wrote
> > The safer bet is to blow the whole Power supply (presuming it doesnt go in
> > to shut down from over current as they seem to do whenever I have
> > inadvertantly shorted one out ) at $40-60 for a replacement AT/ATX power
> > supply it is still cheaper than several hundred dollars for a replacement
> > Motherboard and a couple of hundred dollars for the Hard drive (not to
> > mention the data that everyone carefully Backs up :-)
> > In addition the SCR to do it does not have to sink as much power as
> > suspected in the event of the PSU continuing to operate. assuming a
> > pessimistic 1.2V forward drop at 30 Amp the SCR gives us 36 Watts (with most
> > of the rest of the 300W being dissipated by the PSU) in the event the PSU
> > continues to deliver power this will have the added affect of dragging the
> > other Rails down anyway.
> > well thats my Theory anyway the next and obvious step is to try it I have a
> > couple of AT supplies sitting around doing SFA if it works with them the
> > next step would be to find a Sacrificial ATX supply
> I think you have to consider a number of scenarios in
> the case of an ATX PSU with a crowbar on the 5 volt rail.
> (The 5 volt rail is normaly the one sensed for feedback.)
> Suppose the supply control goes out of regulation due to an internal fault
> making the switching transistor conduct at a high duty cycle so all the
> rails go up. The SCR crowbar will trigger instantly clamping the 5 volt rail
> to about 1.2 volts. A very large current will now flow severely overloading
> the supply and limiting the over voltage appearing on the other rails until either:
> 1. The supply senses the short on the 5 volt rail and shuts
> down in a fraction of a second or if that cct is also faulty,
> 2. The supply blows its mains fuse.
Thats not likely to happen with an addon crowbar protection.
Just because its gunna be hard to clamp the output that
severely given the wiring involved in the high current supplys.
Dont forget that its no longer a single wire outside the supply. So
you'd have to have an SCR on all the wires for the rail being crowbarred
and even then, you're unlikely to be able to blow the mains fuse.
> 3. The switching transistor fails and blows the mains fuse.
Thats pretty unlikely too.
> 4. The supply gets hot enough to go into thermal shutdown.
Or the external protection catches fire as the insulation on those those
shorted rail wires all go up up in flames very spectacularly indeed and
you're fucked whats in the case much more comprehensively due to the
fire than you would have done with the over voltage. Basically ensuring
that nothing in the case is any use at all due to the fire.
And you could easily produce a fire that sets fire to the house as well.
Turning what was a severe bad day into an unspeakable catastrophe.
With the insurance company deciding that it was your addon
crowbar that did it and so its your problem, not theirs.
And thats very likely what would happen, because its very rare indeed
that a PC sets fire to a house and that would be very closely investigated
by the regulatory authoritys to ensure it doesnt happen again.
> In case 4 the SCR may fail due to overheating and go short. This
> is not a concern as you are trying to save a lot of other hardware.
Sure, but it could well set fire to the rail wires insulation.
> Also, the ATX supply is not usually economic to repair anyhow.
Sure, it is certainly viable to deliberately kill the ATX supply.
Just not feasible with an addon crowbar with absolute certainty.
> It seems that ATX supplies should have an overvoltage shutdown
> and latch system on the 12, 5 and 3.3 volt rails but not all do
Yes, but that is at least something that can be checked with the branded supplys.
> and how fast it operates is not clear.
And that measured. Much more feasible to do that with one
of the branded supplys than to design a viable addon crowbar.
It might even be feasible to get some operation like Choice to do it.
> If this system is absent then a crowbar (or two) should be capable
> of limiting the consequential damage in case of a runaway PSU.
Cant see it with an addon.
> I await the results of your tests!
Trouble is that no test can ever really test what matters, what happens
in the situation the addon crowbar is attempting to protect, where what
fault protection the power supply has is not sufficient to prevent a severe
over voltage of what is powered from the supply. Its never really gunna
be feasible to test that with a couple of sacrificial AT supplys.
The only way is to design the crowbar from scratch as a completely independent
protection, and ensure that the design covers ALL the eventualitys, because by
definition an addon crowbar is almost never gunna be used with a PC.
I cant see that its ever gunna be feasible to do it without an output fuse on each
rail being protected and very non trivial even with a fuse on the most important rails.
It is certainly possible to design a power supply from scratch with full crowbar protection,
with the voltage sensing being moved to downstream of the fuses, but its never gunna be
economic with the value of what its protecting in a PC. And you'd never be able to flog
more than a tiny number of those supplys to those who understand the desirability of
full crowbar protection. And you'd have to go the whole hog as well, not only full crowbar
protection, but since they will have to pay substantial money for that, you'd have to fix
the other irritations with PC supplys like the fan noise as well to have any hope of enough
sales to make the retail price viable.
And even if you did all that, you'd still go bust trying IMO.
Which is why you cant just buy one now.
Many insurance policys cover 'fusion' anyway so it would make a lot more sense
to just ensure you have one of those policys and in the very unlikely situation that
the PC power supply does fry whats powered from it, just celebrate and go and
buy a new PC with your 'new for old' claim on your insurance. With obviously a
decent backup of the data you will slash your wrists if you lose.
Rod Speed
grog <shem_an_na@SPAM_ME_NOTyahoo.com> wrote in message
news:JPvV7.8736$wD1....@news-server.bigpond.net.au...
> Richard Freeman <nos...@atps.net> wrote
> Would it not be too slow? By the time the Zenner does it's zennering and the
> SCR goes about blowing a rather large fuse, the xGHz computer has probably
> enough time to send an email informing of the pending PSU failure.
Yeah, I dont think Richard has grasped the difficulty
in that area with those very high current supplys.
Rod Speed
>> Would it not be too slow? By the time the Zenner does it's zennering and the
>> SCR goes about blowing a rather large fuse, the xGHz computer has probably
>> enough time to send an email informing of the pending PSU failure.
> Crowbar Circuits work in many situations by blowing Fuses/Powers
> supplies before any damage is done to other circuitry
Yes, but with those very high current rails, thats a very non trivial exercise.
Its pretty trivial with the lower current linear supplys, which is why most of those
did have crowbar protection and because thats their main failure mode, over voltage.
> the question in this instance is wether the Crowbar will work on the 5V rail alone of a
> PC power supply or wether an effective crowbar involves more than just the 5V rail
The problem is actually blowing the fuse, in spades with an addon crowbar.
Even the fact that all the high current rails have multiple wires to the
ATX connector because they are high current rails is a major difficulty.
Phil Allison
"grog" <shem_an_na@SPAM_ME_NOTyahoo.com> wrote in message
news:JPvV7.8736$wD1....@news-server.bigpond.net.au...
>
> Would it not be too slow? By the time the Zenner does it's zennering and
the
> SCR goes about blowing a rather large fuse, the xGHz computer has probably
> enough time to send an email informing of the pending PSU failure.
>
> greg
Once the threshold voltage is reached at say 6 volts or so the zener
conducts and fires the SCR in microseconds. The rail voltage will then drop
to around two diode drops or 1.2 to 1.5 volts. If the PSU then shuts down,
fine. If it persists the load is still protected.
The rate of rise of the the PSU output voltage will not be very fast due
to all the filter electros on each rail. I think the crowbar is the fastest
method available.
Regards, Phil
Rod Speed
> grog <shem_an_na@SPAM_ME_NOTyahoo.com> wrote
>> Would it not be too slow? By the time the Zenner does it's zennering and the
>> SCR goes about blowing a rather large fuse, the xGHz computer has probably
>> enough time to send an email informing of the pending PSU failure.
> Once the threshold voltage is reached at say 6 volts or so the zener
> conducts and fires the SCR in microseconds. The rail voltage will then
> drop to around two diode drops or 1.2 to 1.5 volts. If the PSU then
> shuts down, fine. If it persists the load is still protected.
Only until the wiring starts to go up in flames very spectacularly indeed
as the power supply delivers 500W into your addon crowbar.
That fire could well end up sending the house up in flames.
Just a tad worse than the stuff supplied by the power supply getting over voltaged.
And likely not covered by your house insurance too.
> The rate of rise of the the PSU output voltage will not
> be very fast due to all the filter electros on each rail.
> I think the crowbar is the fastest method available.
Fastest way to set fire to your house, too.
Rod Speed
> Actually Trialling a PSU (not having made a crowbar yet) running the 5V
> rail into 20 * 5R6 10 Watt Resistors in parallel (0.28 ohms drawing 18A)
Thats not a crowbar. You're gunna have to draw a lot more than that to prevent
a power supply thats gone mad from over voltaging whats supplied from it.
> the first PSU destroyed itself within 15 minutes nope no fire no smoke the
> wires got warm to touch but not hot (which was the point in the exercise) .
Thats a completely useless test. Of course the wires didnt get more than warm,
the supply is SUPPOSED to be able to deliver 18A without the wires getting
distressed. It would be one hell of an inditement of the design if they had got hot.
Says nothing useful whatever about what happens with a true crowbar.
And even if the wire insulation doesnt melt, you're still gunna be dissipating something
of the order of 500W in the addon crowbar when you dont have fuses on the output
rails, that thats gunna set fire to the contents of the case quick smart if the protection
cuts in when there is no one around to unplug it when they smell it going up in flames.
> When I used the disk drive power connector to connect to the
> Resistor the connector did get hot so it appears any crowbar
> will have to be hard wired on to the wires from the PSU.
Still doesnt do anything about the very fundamental problem with a
crowbar with no fuses on the rails, what that 500W dissipated in the
crowbar will do to the other cables in the case in a very short order.
> On initial investigation it appears the mains fuse blew after the switching
> mosfet went short. so it appears the specification on this power supply
> (and I suspect similar claims on many/all other Chinese/Taiwanese PSUs)
> which claims 22A at 5V is far from a continuos power rating
Yeah, the reality in that era was that it was mostly used to
startup the big old hard drives and so wasnt a continuous load.
Things have changed now with modern ATX supplys tho, the total delivered to the
5V and 3.3V rails has very little startup surge involved and is relatively continuous
now. That why those damned Athlons are so picky about the power supply, they
are a pretty crude power hungry design and many power supplys dont do that well.
w_tom
power supplies, but he also does not understand a most basic concepts of switching
power supplies. Its called fold back current limiting. All properly designed
switching power supplies feature the concept. Overvoltage crowbars, as required by the
Intel ATX power supply spec, are more than fast enough and sufficient to shutdown a
power supply without motherboard damage - today as it was done even 25+ years ago.
More quotes from Intel specs - Rod Speed will use more insults and foul language to
deny these specs as well:
> An output short circuit is defined as any output impedance
> of less than 0.1 ohms. ... The power supply shall be capable
> of withstanding a continuous short circuit to the output without
> damage or over-stress to the unit ... The maximum short circuit
> energy in any output shall not exceed 240 VA.
As previously quoted from the spec, all properly designed ATX power supplies contain
some form of an overvoltage crowbar. No power supply, properly designed, will ever
output excessive voltage. However many cheap power supplies have found another way to
sell on price. For example, those Deer supplies contain no Intel ATX required
overvoltage protection - required as posted earlier.
Overvoltage crowbars are more than fast enough and easily implemented. If any
crowbar is triggered, then the supply, as required by the spec, will shutdown all
output voltages. Overvoltage damage from a power supply must never occur. Power
supply outputs must be shorted and create no power supply damage - create no mythical
"fire could well end up sending the house up in flames." It was standard long before
there was an IBM PC and yet still one posts in direct contradiction to basic technical
facts. The expression is called fold back current limiting - understood even by
novices in a power supply shop - and still not comprehended by a poster here.
If selling a power supply for $25 or $40, and since overvoltage crowbar parts can
cost $1 .... well many clone computer builders will never know the difference since
they only purchase based upon price alone. Therefore many clone computer supplies
don't have the required overvoltage crowbars. Some supplies (in direct violation of
specs) even self destruct when shorted. Brand name supplies contain the required
overvoltage protection.
Notice a silly statement from Rod Speed:
> Only until the wiring starts to go up in flames very
> spectacularly indeed as the power supply delivers
> 500W into your addon crowbar.
His statement is in direct violation to the quoted ATX power supply specification.
The number was 240 VA. But a foul language critic does not even understand a basic
power supply concept called fold back current limiting. How then can he know anything
about how much current would appear in an overvoltage crowbar. It required him to
first read specs before posting.
Rod Speed
w_tom <w_t...@usa.net> wrote in message news:3C27DAF7...@usa.net...
> Not only does this poster outrightly deny an exact
> quote from Intel's specs for ATX power supplies,
Lie. I just rubbed your nose in the FACT that Intel's
specs do NOT require the use of crowbar protection.
> but he also does not understand a most basic concepts of switching power supplies.
Lie. You wouldnt know what crowbar protection was if it bit you on the arse.
> Its called fold back current limiting.
Wrong. As always.
There are various ways of providing protection with any power supply.
Foldback protection has NOTHING to do with crowbar protection.
They are completely different approaches to protection.
The term crowbar protection comes from the much older approach of using a
crowbar literally to short the output of the power supply. The modern approach
is to use an SCR as an electronic crowbar which deliberately blows the output
fuse when the crowbar protection has detected that the supply output has gone
too high, due to some failure with the power supply. Usually a failure with a
linear supply. Linear supplys often do fail that way, the full unregulated
input is delivered to the output when the regulation fails shorted.
> All properly designed switching power supplies feature the concept.
Wrong again. All properly designed switching power supply certainly
have failure protection, but that is NOT crowbar protection.
> Overvoltage crowbars, as required by the Intel ATX power supply spec,
Lie. The intel specs dont even mention crowbar protection at all.
> are more than fast enough and sufficient to shutdown a power supply
> without motherboard damage - today as it was done even 25+ years ago.
Pity there is no crowbar protection in any mass market
ATX supply, even those produced by IBM etc.
> More quotes from Intel specs
Not one of which says a damned thing about crowbar protection.
>> An output short circuit is defined as any output impedance
>> of less than 0.1 ohms. ... The power supply shall be capable
>> of withstanding a continuous short circuit to the output without
>> damage or over-stress to the unit ... The maximum short circuit
>> energy in any output shall not exceed 240 VA.
Thats got absolutely NOTHING to do with crowbar protection
and is talking about what the supply has to do BEFORE its failed.
Even someone as stupid as you should be able to grasp that once the supply
HAS FAILED, all bets are off on that behaviour with a shorted output.
The WHOLE POINT of crowbar protection is entirely independant
protection ONCE THE POWER SUPPLY HAS FAILED AND IS
PRODUCING AN OVERVOLTAGE ON THE OUTPUT OF THE SUPPLY.
> As previously quoted from the spec, all properly designed ATX
> power supplies contain some form of an overvoltage crowbar.
Yet another lie. No mention what so ever in the spec you quoted of CROWBAR protection.
> No power supply, properly designed, will ever output excessive voltage.
Nice theory, pity about the supplys that have done that.
Including brand name supplys like one of the Macase supplys too.
> However many cheap power supplies have found another way to sell on price.
Pity the Macase aint cheap power supplys sold on price.
Pity the IBM branded supplys arent cheap power supplys sold on price either.
NOT ONE IBM BRANDED PC POWER SUPPLY HAS EVER HAD CROWBAR PROTECTION.
> For example, those Deer supplies contain no Intel ATX
> required overvoltage protection - required as posted earlier.
Wrong again. All supplys have over voltage protection.
NONE have CROWBAR overvoltage protection. And the Intel
spec DOES NOT require CROWBAR overvoltage protection.
> Overvoltage crowbars are more than fast enough and easily implemented.
Have fun explaining why NOT ONE IBM branded PC
power supply has ever had crowbar over voltage protection.
> If any crowbar is triggered, then the supply, as required by the spec, will shutdown
> all output voltages. Overvoltage damage from a power supply must never occur.
Pity that aint CROWBAR OVERVOLTAGE PROTECTION. THERE IS NO CROWBAR.
> Power supply outputs must be shorted and create no power supply damage
> - create no mythical "fire could well end up sending the house up in flames."
Thats ONLY when the power supply HAS NOT FAILED you fool.
> It was standard long before there was an IBM PC
Yes. Pity that NOT ONE IBM BRANDED PC POWER SUPPLY
HAS EVER HAD CROWBAR OVER VOLTAGE PROTECTION.
> and yet still one posts in direct contradiction to basic technical facts.
You wouldnt know what a basic technical fact was if it bit you on the arse.
You cant even manage to work out the difference between crowbar over voltage
protection and foldback current limiting protection which is something else entirely.
They are two COMPLETELY DIFFERENT types of protection.
Crowbar over voltage protection PROTECTS AGAINST OVER VOLTAGE, by
shorting the output of the power supply and deliberately blowing the output fuse.
Foldback current limiting protects when THERE IS EXCESSIVE CURRENT
drawn from the supply, due to a failure in whats being powered by the power
supply. The current is limited to minimise the damage done to whats powered
by the supply by reducing the current delivered to that.
Foldback current limiting isnt even seen with PC power supplys
which use a different approach, shutting down the power supply
instead until the short on the output is removed.
> The expression is called fold back current limiting
Wrong. As always. Foldback current limiting has NOTHING to do with crowbar
overvoltage protection. Its protecting against a completely different failure, stupid.
> - understood even by novices in a power supply
> shop - and still not comprehended by a poster here.
You obviously havent got the remotest concept of what a spectacular fool you are
making of yourself in the eyes of everyone who knows anything about power supplys.
But then you keep doing that time after time after time, so you must be very thick indeed.
> If selling a power supply for $25 or $40, and since overvoltage crowbar
> parts can cost $1 .... well many clone computer builders will never know
> the difference since they only purchase based upon price alone.
Have fun explaining why NOT ONE IBM branded PC
power supply has crowbar over voltage protection.
And its completely trivial to test any power supply and see if its got foldback
current limiting or not. PC power supplys almost never have that. They do however
have output short circuit protection and its completely trivial to prove that by testing.
> Therefore many clone computer supplies don't have the required overvoltage crowbars.
Have fun explaining why NOT ONE IBM branded PC
power supply has crowbar over voltage protection.
> Some supplies (in direct violation of specs) even self destruct when shorted.
Got absolutely NOTHING to do with what was being discussed, whether the
Intel spec says a damned thing about CROWBAR OVERVOLTAGE PROTECTION.
It doesnt.
> Brand name supplies contain the required overvoltage protection.
BUT NOT CROWBAR OVER VOLTAGE PROTECTION.
Even someone as stupid as you can inspect any PC power supply and you wont find
one thats got the OUTPUT FUSES that CROWBAR overvoltage protection requires.
Or the SCRs on each output rail either. Tho likely you
wouldnt know an SCR if it bit you on the arse either.
You might conceivably be able to work out what is a fuse tho.
> Notice a silly statement from Rod Speed:
>> Only until the wiring starts to go up in flames very spectacularly indeed
>> as the power supply delivers 500W into your addon crowbar.
> His statement is in direct violation to the quoted ATX power supply specification.
Yet another lie.
> The number was 240 VA.
Yet another lie. That isnt even POSSIBLE with the higher rated ATX supplys.
Even someone as stupid as you should be able to grasp that a 400W ATX supply
cant possibly be limited to 240VA. That would defy the laws of physics, stupid.
> But a foul language critic does not even understand a basic
> power supply concept called fold back current limiting.
Knew about it before you were even born thanks, fool.
PC power supplys DO NOT have foldback current limiting and that has NOTHING to
do with CROWBAR OVERVOLTAGE PROTECTION which is something else entirely.
> How then can he know anything about how much
> current would appear in an overvoltage crowbar.
Pity PC power supplys DO NOT have foldback current limiting, fool.
> It required him to first read specs before posting.
Reading them clearly didnt help you one iota, because you so technically pig ignorant
that you wouldnt know what crowbar overvoltage protection was if it bit you on the arse.
> Rod Speed wrote
>> Phil Allison <bi...@bigpond.com> wrote
Brian Goldsmith
"Rod Speed" <rod_...@yahoo.com> wrote in message
The term crowbar protection comes from the much older approach of using a
crowbar literally to short the output of the power supply.
For Humanities sake if not your own ,stop this puerile nonsense from spewing
out of your mind.NOBODY ever used a hexagonal cross sectioned piece of
steel weighing upwards of 15 Kg (commonly known as a crowbar) to short out a
power supply of the type mentioned in the original posting.
Brian Goldsmith.
Phil Allison
"Brian Goldsmith" <brian.g...@echo1.com.au> wrote in message
news:7cVV7.10808$wD1....@news-server.bigpond.net.au...
I have heard of a guy taking to his PC with an axe but not a crowbar,
maybe axes are more to hand on such occasions.
Regards, Phil
Phil Allison
> news:3c272343$1...@news.iprimus.com.au...
>
> > Actually Trialling a PSU (not having made a crowbar yet) running the 5V
> > rail into 20 * 5R6 10 Watt Resistors in parallel (0.28 ohms drawing
18A)
>
> > On initial investigation it appears the mains fuse blew after the
switching
> > mosfet went short. so it appears the specification on this power supply
> > (and I suspect similar claims on many/all other Chinese/Taiwanese PSUs)
> > which claims 22A at 5V is far from a continuos power rating
** I cannot see your original post of this Richard, it doesn't appear on
Google either - could you repost the message?
** Blowing up switching PSUs - what fun!
Regards, Phil
Brian Goldsmith
"Rod Speed" wrote:-
The term crowbar protection comes from the much older approach of using a
crowbar literally to short the output of the power supply.
Brian Goldsmith wrote:-
NOBODY ever used a hexagonal cross sectioned piece of
steel weighing upwards of 15 Kg (commonly known as a crowbar) to short out a
power supply
Rod Speed wrote:-
NOBODY ever said anyone did, pratchild.
You did Speed ,see above.Look up the meaning of the word
literally,"Following the exact words" OED.
Brian Goldsmith.
Rod Speed
Some lying pratchild claiming to be
Bwian Goldsmith <brian.g...@echo1.com.au> wrote in message
news:%fZV7.11641$wD1....@news-server.bigpond.net.au...
> Rod Speed <rod_...@yahoo.com> wrote
>> Bwian Goldsmith <brian.g...@echo1.com.au> wrote
>>> Rod Speed <rod_...@yahoo.com> wrote
>>>> The term crowbar protection comes from the much older approach
>>>> of using a crowbar literally to short the output of the power supply.
>>> For Humanities sake if not your own ,stop this
>>> puerile nonsense from spewing out of your mind.
>> Even you should be able to do a LOT better than that, pratchild.
>>> NOBODY ever used a hexagonal cross sectioned piece of steel
>>> weighing upwards of 15 Kg (commonly known as a crowbar) to
>>> short out a power supply of the type mentioned in the original posting.
>> NOBODY ever said anyone did, pratchild.
>> The name was around LONG before the PC had even been invented,
>> and before even computers had been invented too, stupid.
> You did Speed,
More of your childish lying, pratchild.
> see above.
Nowhere do I ever say that one was ever used on a PC power supply, pratchild.
> Look up the meaning of the word literally,
Completely and utterly irrelevant, pratchild. Electronics and computing
are absolutely RIDDLED with figurative use of normal words. Like
crowbar protection, booting, bootstrapping, feedback, etc etc etc.
> "Following the exact words" OED.
Completely and utterly irrelevant, pratchild.
Franc Zabkar
to keyboard and composed:
> Notice a silly statement from Rod Speed:
I notice them all the time.
>> Only until the wiring starts to go up in flames very
>> spectacularly indeed as the power supply delivers
>> 500W into your addon crowbar.
>
> His statement is in direct violation to the quoted ATX power supply specification.
>The number was 240 VA. But a foul language critic does not even understand a basic
>power supply concept called fold back current limiting...
That's because you don't get that kind of knowledge from Dick Smith
Funway kits.
-- Franc Zabkar
Please remove one 'g' from my address when replying by email.
Rod Speed
Some pathetic wogprat claiming to be
Franc Zabkar <franc...@dinggoblue.net.au> wrote in message
news:3c285fca...@news.dingoblue.net.au...
just the puerile wogshit that always pours from the back of it.
Richard Freeman
Rod has included my original posting intact in his message.
I was merely pointing out that blowing the F^@( out of an AT power supply
hardly required much effort in fact trying to get the PSU to provide its
supposedly rated Power was enough to blow the F^)( out of it and I suspect
an ATX powersupply will be pretty much the same as ATX supplies are pretty
much the same beast with a few changes (an additional low power permanant 5V
rail and a 3.3V rail.
As yet (and as I stated in my previous posting) I still have not had time to
lash up a Crowbar but now Christmas is over I might get a chance in the next
few days
Richard
Rod Speed
> I was merely pointing out that blowing the F^@( out
> of an AT power supply hardly required much effort
You only proved that it wasnt much effort with that particular supply.
Says nothing useful what so ever about a general crowbar design.
And still doesnt say anything about how you're gunna
prevent the power dissipated in the addon crowbar
protection from setting fire to the contents of the case.
15 mins of 500ish watts is plenty to do that.
> in fact trying to get the PSU to provide its supposedly
> rated Power was enough to blow the F^)( out of it
Doesnt say a damned thing about PC power supplys in general.
Particularly with modern supplys that do tend to get run at something
like their rated capacity with the most badly designed power hungry cpus.
> and I suspect an ATX powersupply will be pretty much the same
You're wrong on that. AT supplys were designed to supply significant short term
high current loads, to spin up the bloody great full height 5.25" hard drives. That
requirement is long gone now and an ATX supply sees far less startup current surge.
> as ATX supplies are pretty much the same beast with a few changes
Not on that startup surge current requirement they aint.
> (an additional low power permanant 5V rail and a 3.3V rail.
There's a lot more difference than just those two.
And just because a particular supply couldnt deliver its rated output current for
long before dying says absolutely NOTHING about PC power supplys in general.
If there aint someone around to turn the power off when something
is obviously burning, that 500ish watts being dumped in the addon
crowbar is absolutely guaranteed to set fire to the other cables in the
case eventually and could well end up burning the house down too.
Those monitor fires that did kill some individuals had much
lower rated power supplys than a modern ATX supply.
> As yet (and as I stated in my previous posting) I still have not had time to lash
> up a Crowbar but now Christmas is over I might get a chance in the next few days
Wont prove anything even when you do. You cant test what
happens with all the ATX supplys shipped with all failure modes.
Foley U. Matthews
applied digits to the keyboard and routed the information from some
kind of brain (presumably), thusly:
Foley U. Matthews. there | I believe in : Paying NO voluntary taxes, i.e.
are no e's in my true email | Lotteries, Gambling... The Executive Producer
Visit the Ellen Foley Info | is to blame!... and perhaps, Love (is/can be)
http://www.go.to/ellen-foley | "Fully expecting to be Hurt!"