Need to press restart to boot up.

  • Thread starter Thread starter pjdd
  • Start date Start date
P

pjdd

I've seen a few computers that won't start up unless the restart button
is pressed, but otherwise work normally. At least one of them was an AT
Pentium, and I think the rest were more recent ATX-based systems. I do
know that they were all stand-alone desktops.

I never had a chance to closely examine any of them but I'd like to
have some idea of possible causes. Slow spin-up HDD, RAM, mobo ? Has
anyone worked on and cured computers with similar symptoms ?
 
I've seen a few computers that won't start up unless the restart button
is pressed, but otherwise work normally. At least one of them was an AT
Pentium, and I think the rest were more recent ATX-based systems. I do
know that they were all stand-alone desktops.

I never had a chance to closely examine any of them but I'd like to
have some idea of possible causes. Slow spin-up HDD, RAM, mobo ? Has
anyone worked on and cured computers with similar symptoms ?
Click to expand...

Usually it's a sign of a failing or insufficient power
supply.
 
I've seen a few computers that won't start up unless the restart button
is pressed, but otherwise work normally. At least one of them was an AT
Pentium, and I think the rest were more recent ATX-based systems. I do
know that they were all stand-alone desktops.

I never had a chance to closely examine any of them but I'd like to
have some idea of possible causes. Slow spin-up HDD, RAM, mobo ? Has
anyone worked on and cured computers with similar symptoms ?
Click to expand...

This sounds sorta dumb, but it's real-world and I have it going in my
garage:

AT power supplies have an AC on-off switch on a cord that's generally
mounted in the front of the machine. ATX power supplies -- and a few late
AT ones -- are turned on and off by the motherboard in response to a
momentary pushbutton switch and a "soft power" circuit. The front panel
power switch and front panel reset switch in an ATX case are identical.
They're each connected to a jumper-style pair of pins (a "header") on the
motherboard by a twisted pair of wires. So:

(1) If the main power button itself fails, it's easy to just connect the
reset button to the "power on" header. The reset button becomes the
power-on button. The individual who did it knows this, but future users get
to figure it out.

(2) If an idiot assembles the computer and switches the two little plugs,
the buttons switch roles. That's a two-minute fix.

(3) If you mount an ATX power supply and compatible board in a proprietary
AT case -- certain combinations do work -- then the on/off switch becomes
meaningless and you connect the reset switch to the "power-on" header. The
reset button becomes the power-on button.

There could be other, less intentional, causes, but these are worth checking
out. None of them is really much of a problem if you're used to it.
 
Thanks for the replies, kony and Pelysma, but I'm afraid it doesn't
seem to be any of the reasons suggested.

1. Today I got hold of the P1 from my brother's printing press where
it's been used for DTP work for the past 10 years. It's a Pentium 166,
started out with 16MB, now 64MB, 1GB HDD, CD-ROM drive, 2MB PCI gfx
card, and nothing else. It has a 200W PSU that's been working all these
years.

2. One of the ATX types is an Athlon XP 2600+, onboard video and single
HDD, owned by an online acquaintance in a distant city. He seems to be
fairly experienced in assembling his own computer, so I assume that he
wired the switches correctly (I'll ask him again). Following my
suggestion, he replaced his PSU with a name-brand 450W unit last week,
but he still has to press the restart button.

3. I don't have access to any of the others ATM, but I recall that with
at least one of them, power comes on and the CPU and case fans turn
when the power button is pressed. It just refuses to boot up unless the
restart button is pressed. I can't say for sure if the POST screen
comes up, but I don't think it does.

Any other ideas please ?
 
Thanks for the replies, kony and Pelysma, but I'm afraid it doesn't
seem to be any of the reasons suggested.

1. Today I got hold of the P1 from my brother's printing press where
it's been used for DTP work for the past 10 years. It's a Pentium 166,
started out with 16MB, now 64MB, 1GB HDD, CD-ROM drive, 2MB PCI gfx
card, and nothing else. It has a 200W PSU that's been working all these
years.
Click to expand...

Presumably this is one of the systems having the
need-press-reset problem?

What evidence do you have that the power supply isn't the
problem? Understand something clearly- Just because a
system can eventually run from a PSU, that doesn't mean it's
ok, sufficient. Because in actuality, it didn't run ok if
you had to press the reset button. Evidence that one thing
works ok is not evidence that everything is ok relating to
power. Even so, it would be prudent to use a multimeter to
take voltage readings on all supply pins during, and after
the initial power-on event.

2. One of the ATX types is an Athlon XP 2600+, onboard video and single
HDD, owned by an online acquaintance in a distant city. He seems to be
fairly experienced in assembling his own computer, so I assume that he
wired the switches correctly (I'll ask him again). Following my
suggestion, he replaced his PSU with a name-brand 450W unit last week,
but he still has to press the restart button.
Click to expand...

Why do you mean "name-brand" but not the name? Details are
you friend, specifics matter. Never substitute a vagueness
for specific details when they can be used concisely. This
is not a complaint, merely that it is in your interest to do
so.

When all else fails one might strip the system down to
minimal components and retry it. Again a multimeter might
be used to measure voltages, but also before power-on event
since it is an ATX PSU. Come to think of it there were ATX
P1 systems so if system #1 is ATX, also measure it before
power on.

3. I don't have access to any of the others ATM, but I recall that with
at least one of them, power comes on and the CPU and case fans turn
when the power button is pressed. It just refuses to boot up unless the
restart button is pressed. I can't say for sure if the POST screen
comes up, but I don't think it does.

Any other ideas please ?
Click to expand...

BIOS update? Doubt that's it though, it's probably power.
 
kony said:
Presumably this is one of the systems having the
need-press-reset problem?

What evidence do you have that the power supply isn't the
problem? Understand something clearly- Just because a
system can eventually run from a PSU, that doesn't mean it's
ok, sufficient. Because in actuality, it didn't run ok if
you had to press the reset button. Evidence that one thing
works ok is not evidence that everything is ok relating to
power. Even so, it would be prudent to use a multimeter to
take voltage readings on all supply pins during, and after
the initial power-on event.
Click to expand...

It may well have to do with allowing the drives to spin up to
speed. This is the time when the power is likely to be critical.
One automatic delay for drive access is to allow the system to do
the memory test, which will probably consume several seconds. If
the bios has provision for a separate delay to fire up the drives,
take it.
 
Thanks for the replies, kony and Pelysma, but I'm afraid it doesn't
seem to be any of the reasons suggested.

1. Today I got hold of the P1 from my brother's printing press where
it's been used for DTP work for the past 10 years. It's a Pentium 166,
started out with 16MB, now 64MB, 1GB HDD, CD-ROM drive, 2MB PCI gfx
card, and nothing else. It has a 200W PSU that's been working all these
years.

2. One of the ATX types is an Athlon XP 2600+, onboard video and single
HDD, owned by an online acquaintance in a distant city. He seems to be
fairly experienced in assembling his own computer, so I assume that he
wired the switches correctly (I'll ask him again). Following my
suggestion, he replaced his PSU with a name-brand 450W unit last week,
but he still has to press the restart button.

3. I don't have access to any of the others ATM, but I recall that with
at least one of them, power comes on and the CPU and case fans turn
when the power button is pressed. It just refuses to boot up unless the
restart button is pressed. I can't say for sure if the POST screen
comes up, but I don't think it does.

Any other ideas please ?
Click to expand...

According to a couple of Intel motherboard reference schematics, three
things reset the CPU:
1. Power_OK from the power supply (ATX power connector pin 8)
2. PowerGood from the motherboard CPU core voltage regulator, and
3. Reset switch.
You can try shorting to ground the Power-OK line (Pin 8) on the ATX
power connector. If it does not reset the CPU, then there's a problem
with the motherboard circuitry. If it does, then it's possible the
power supply is asserting Power_OK too soon.
Otherwise, the problem is probably with the motherboard core voltage
regulator circuitry.
 
According to a couple of Intel motherboard reference schematics, three
things reset the CPU:
Click to expand...

It may be quite premature to assume that the CPU needs
reset.
1. Power_OK from the power supply (ATX power connector pin 8)
2. PowerGood from the motherboard CPU core voltage regulator, and
3. Reset switch.
You can try shorting to ground the Power-OK line (Pin 8) on the ATX
power connector. If it does not reset the CPU, then there's a problem
with the motherboard circuitry. If it does, then it's possible the
power supply is asserting Power_OK too soon.
Otherwise, the problem is probably with the motherboard core voltage
regulator circuitry.
Click to expand...


The problem is usually not the motherboard voltage
regulators... even less likely in this case as if they were
problematic, it would tend to resurface in other system
operations.

The typical troubleshooting steps such as stripping system
down to bare minimums should be tried, in addtion to
reducing clock rates as much as possible. These suggestions
are made with the assumption that a known viable power
supply (as proven, not assumed) has been tried.
 
kony said:
It may be quite premature to assume that the CPU needs
reset.



The problem is usually not the motherboard voltage
regulators... even less likely in this case as if they were
problematic, it would tend to resurface in other system
operations.

The typical troubleshooting steps such as stripping system
down to bare minimums should be tried, in addtion to
reducing clock rates as much as possible. These suggestions
are made with the assumption that a known viable power
supply (as proven, not assumed) has been tried.
Click to expand...

As I indicated from the beginning, this was not a problem with a
computer I have sitting with me and my interest was mainly
academic.

At present I don't have access to any of the computers mentioned
except the Pentium 166, and there's no other system of similar
configuration close at hand for swapping components with. I have
not had time to take the voltage readings yet. In any case, if we
're going to consider all theoretical possibilities, acceptable
voltage readings may not be 100% reliable as a basis for
eliminating the PSU as the culprit. The speed of initial build-up
may play a part. Oh, and the "name-brand" 450W PSU used as a
replacement by my online acquaintance is a Zebronics.
 
As I indicated from the beginning, this was not a problem with a
computer I have sitting with me and my interest was mainly
academic.
Click to expand...

Ok, but it's quite common that such questions can't be
answered in an academic context because many problems can
have multiple causes, even combinations of problems as the
cause.

At present I don't have access to any of the computers mentioned
except the Pentium 166, and there's no other system of similar
configuration close at hand for swapping components with. I have
not had time to take the voltage readings yet. In any case, if we
're going to consider all theoretical possibilities, acceptable
voltage readings may not be 100% reliable as a basis for
eliminating the PSU as the culprit. The speed of initial build-up
may play a part. Oh, and the "name-brand" 450W PSU used as a
replacement by my online acquaintance is a Zebronics.
Click to expand...

That's not a name-brand power supply, it's a relabled
generic.
 
Back
Top