1.2 Tualitan- how many volts is too many?

Uncle Vinnie wrote:

Thanks David..

Actually, this PC is a Compaq- it's coming up on 5 years old. I've taken it
about as far as it can go...

I was thinking of boosting speed a little by upping the voltage (per
Fishface) but as it is, it is very stable, unless I keep it running without
AC!
(It's in my office, a converted attic- Without the AC on, the attic probably
hits 110/120F! PLUS if I forget to keep the desk cabinet open that it's
in!)

That would do it. Especially the closed desk turning it onto a mini oven
from lack of airflow, and being just about already there starting at 120F.


My first idea was to replace the power supply since, per the part number,
it's only 145 watts (how can that be?). I was going to take the 230w
Powerman from my daughter's Intel (being replaced with a 380), and use it.

BUT, I am not sure of 2 things- first, how can the Compaq p/s be only 145w?

Because it's sized for 'what it needs' rather than "whatever you end up
building" as is the case with a generic PSU where they have no idea what
you'll put in there so it has 'more than enough of everything' (for the
'size', e.g. 250 Watts). Or, so one hopes.

That's also why not all generic 250 watt PSUs are the same. They still have
to make *some* kind of assumptions as to how much power to allocate on
the 3.3, 5, and 12 volt rails so one might provide more to the 5, vs the
others, while another favors the 3.3 and yet another is more generous on
the 12. But they all still 'add up' to 250, assuming they didn't cheat on
the spec like many do. And those assumptions change over time as the
'typical' PC available at the time changes. So the allocations in a PSU
built back 'in the beginning' of the ATX days won't necessarily have the
same power allocation as one made today.

Point being, even a 'million watt' PSU won't help if all that power is on
the wrong power rail for your particular system.

And that's before you even get to things like the noise and regulation specs.

Second, is it a standard ATX p/s?

Most likely a Micro or Flex ATX PSU. In which case, a standard ATX won't fit.


The board, btw, is a Mitac 6513wu, with Compaq's part number 191767. The p/s
is Compaq p/n 124848.




"David Maynard" wrote in message
...

Uncle Vinnie wrote:

Yes, and yes!

I am running it over the next few days set at 1.55 (vcore at 1.527).

If it was the PS all along (which it might be, and since my first attempt
was to replace a 230 with 250- which probably meant my PS wasn't bad but
that 230 or even 250 just wasn't enough), I will then crank it up a
little and see what happens!

However, if anyone needs a cranked up PC, that would be me....

That will wait a week... (I have a Compaq 5000T/ 5BW. Motherboard
believed to be actually Mitac 6513WU.

Mitac did a number of boards for Compaq. I've got one I picked up as a
replacement for an HP motherboard.

I bet you noticed that Compaq isn't HP. So did I after it got here but the
fly by night *&%$ who sold it didn't.

I bought it back when the tualatin mod was a hand wiring job and was going
to use it to test for a customer's HP and it was supposed to BE the HP
mobo. I had my own tualatin to test with and he agreed to pay for the mobo
(it was very cheap) to run it on so if it worked I would swap boards and
the customer got a tualatin plus I got his board for a 'spare' (and paid
for the work). If it didn't I ended up with the new board as a 'spare'.

As my perfect 'all bases covered' plan turned out I had to eat it because
I couldn't run the test. One of those "best laid plans of mice and men..."
things.



Here, I might lead into the same problem... The stock PS is well 300...
could be around 230... I need to check, and feeding more voltage might
put more of a drain wouldn't it??

Power consumption increase linearly with speed but with the square of the
voltage, which is why increasing voltage is the last thing you want to do.
When overclocking you increase speed first and voltage only if needed and
only as little as needed because it's working both for you (volts for
stability) and against you (heat generation). And the 'against you' part
gets bad faster than the good it's doing real quick.



Thank you both for all your help... As soon as all settles in, I will
post a huge thank you... you''ve helped alot....


"David Maynard" wrote in message
...


Fishface wrote:


David Maynard wrote:




Might as well try the stock voltage again too.


I say try it at 1.7v and the 133 MHz bus setting! Might run 1.6 Ghz...
http://www.hardwarecooling.com/produ.../reviews_id/15



Since he hadn't been able to get it stable at ALL it's best to first
establish an operational baseline.

Uncle Vinnie wrote:
BTW, that is the PS I am replacing, 230 Powerman....

I just noticed that it is in fact a 235w Power Man. I heard once
that they were made by Fortron. I think there is a brand called
Powman-- whether it is the same, I do not know. It specs at
22A. on the 5v, 12A on 3.3v, and 8A on 12V. Depending upon
from which line the CPU voltage is regulated by the motherboard,
I'm sure that individual results could differ.