CUBX + 370GU
Bob Byrne
I've looked thru Google for info on updating to a
Celeron FC-PGA2 (Tualatin) CPU and found everything I need
except what BIOS to use in a CUBX board.
I currently have V1007 final and have found several flavours
of V1008 beta 4, the latest being
1008b4_CMD_1.9.16.zip (7/18/02)
It has a readme -
"ASUS CUBX (CMD-648) BIOS"
"BIOS v. 1008 BETA 4 (Tulatin mod) - CMD BIOS v. 1.9.16"
I assume this would be the one to use if I need to update at all.
Can anyone remember back that far and advise me on what I
need to do?
--
Bob
BTW - why am I doing this you ask?
Found a Celeron 1.3Ghz 370GU Pack _very_ cheap.
Stephan Grossklass
>
> Hello.
> I've looked thru Google for info on updating to a
> Celeron FC-PGA2 (Tualatin) CPU and found everything I need
> except what BIOS to use in a CUBX board.
> I currently have V1007 final and have found several flavours
> of V1008 beta 4, the latest being
> 1008b4_CMD_1.9.16.zip (7/18/02)
> It has a readme -
> "ASUS CUBX (CMD-648) BIOS"
> "BIOS v. 1008 BETA 4 (Tulatin mod) - CMD BIOS v. 1.9.16"
>
> I assume this would be the one to use if I need to update at all.
> Can anyone remember back that far and advise me on what I
> need to do?
It certainly can't hurt to update to this BIOS (not only because of
Tualatin microcodes, but also for the probably newer CMD 648 BIOS). Of
course you should really only use it with the original CUBX, not with
the -L or -E.
Stephan
--
Home: http://jgrossklass.bei.t-online.de/ | Webm.: http://www.i24.com/
PC#6: i440LX, 2xCel300A, 128 MB, 9 GB, ATI AGP 32 MB; WinNT4.0SP6a
This is a SCSI-inside, Legacy-plus, TCPA-free computer :)
Bob Byrne
"Stephan Grossklass" <sgros...@yahoo.de> wrote:
> Bob Byrne schrieb:
> > I currently have V1007 final and have found several flavours
> > of V1008 beta 4, the latest being
> > 1008b4_CMD_1.9.16.zip (7/18/02)
> It certainly can't hurt to update to this BIOS (not only because of
> Tualatin microcodes, but also for the probably newer CMD 648 BIOS). Of
> course you should really only use it with the original CUBX, not with
> the -L or -E.
Thanks, my system is very stable so I don't like to change things
unless absolutely necessary. I thought the microcode was only
so a true reading of the CPU being used was displayed.
Of course if it is required for a higher multiplier, which I suspect,
than I'll flash it. I'll need a multiple of 13.
I do have a CUBX, non L or E.
--
Bob
Paul
<bby...@bigpond.NOCAPSnet.au> wrote:
You can use CTMC to add a microcode for your new processor. To use
CTMC, you have to install the new processor, before doing the microcode
update. CTMC writes the microcode in a "volatile" portion of the
flash, next to the area used to hold DMI and ESCD information.
The microcode is used to patch any hardware bugs that might exist
in a given stepping of microprocessor. So, the steps you'd need
would be:
1) Flash your motherboard BIOS to the latest. If the board survives,
this proves that the flash is still writeable.
2) Look up the details of the new processor here - you can get the
CPUID by searching for the SSPEC on this site:
http://processorfinder.intel.com
3) Find a BIOS from another board know to contain the stepping of
Tualatin you plan on using. This BIOS file will be the source of
your new microcode patch. (At retail now, this is quite likely
to be CPUID 06B4, as the 06B1's are most likely all sold.) Bernd
found the latest microcodes in two files, which he mentioned
in this posting:
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=3E686B97.F187EC41%40fh-nuernberg.de
4) Get a copy of CTMC from CT magazine here:
ftp://ftp.heise.de/pub/ct/ctsi/ctmc10.zip
5) I posted the rest of my experience here. I did final steps of
the upgrade by booting from a DOS boot floppy, so I could flash
back to the same BIOS as step (1) if some error message arose. See:
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&selm=nospam-1104030856270001%40192.168.1.177
The only other thing you should be warned about, is the Vcore voltage
regulator. I'm sure you know about the fact that the voltage regulator
has to work at 1.5 volts, for a Tualatin upgrade to work. What isn't
stressed enough with an upgrade like this, is the amount of current
which is consumed by the processor at 1.5 volts. It is my speculation
that the current is higher than the max current design point of the
regulator design. (I used an Upgradeware Slot-T, but I think
the 370GU is similar in concept, only it is a S370 to S370 adapter.)
In my experiments, a Tualatin 1100/100 barely makes the MOSFETs in the
Vcore circuit warm. When I then tried a Tualatin 1400/100 and ran a
copy of BURNMMX as a load test, the MOSFET temp went up to 60C and
I stopped the test. The thing about MOSFETs, is they have a "safe
operating area", and if they get too warm, their internal resistance
goes up. Since power dissipation is equal to I**2 * Ron, as the
on-resistance Ron increases, the device gets hotter and so on. This
is thermal runaway. At this point, I don't know how to predict what
the risk is, as I cannot figure out what current the Vcore circuit
was designed for (it could be anywhere between 14 and 20 amps).
Here is my current performance table - using a P2B-S motherboard,
GF3 Ti200 video card, and retail CPU HSF:
Coppermine 1000/100/128KB Memory running 100MHz CAS2 Generic_Slocket 1.8V
Sensor ----> Processor Fet/Toroid Motherboard
Idle 29C 28C 26C
3DMark2001SE 42C 35C 28C (3DMarks = 4636)
Tualatin 1100/100/256KB Memory running 100MHz CAS2 Upgradeware Slot-T 1.5V
Sensor ----> Processor Fet/Toroid Motherboard
Idle 32C 35C 27C
3DMark2001SE 34C 36C 27C (3DMarks = 5176)
Tualatin 1400/100/256KB Memory running 100MHz CAS2 Upgradeware Slot-T 1.5V
Sensor ----> Processor Fet/Toroid Motherboard
Idle 37C 40C 26C
3DMark2001SE 41C 45C 28C (3DMarks = 5579)
BURNMMX 54C 60C 29C
So, if you do the upgrade, there is little danger of thermal runaway
during POST or while booting the board. But, when the time comes to
run PRIME95 or CPUBURN or BURNMMX or the like, keep a finger handy
to feel the surface of the MOSFETs in the Vcore regulator circuit.
If you cannot keep a fingertip on them for more than two seconds,
the surface temp of the MOSFET is 55C or so. I personally don't feel
that comfortable running them at that temperature -- I'm conservative
by nature (and it isn't summer around here yet). Since we have
different motherboard models, it is always possible that your Vcore
is rated for more current than mine is, in which case the MOSFETs
will only be warm to the touch.
HTH,
Paul
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> "Bob Byrne" wrote:
> > "Stephan Grossklass" <sgros...@yahoo.de> wrote:
> > > It certainly can't hurt to update to this BIOS (not only because of
> > > Tualatin microcodes, .......
> > ...... I thought the microcode was only
> > so a true reading of the CPU being used was displayed.
> You can use CTMC to add a microcode for your new processor. ....
Thanks but too much info (eyes glazing over) :-)
Here's where I stand .......
Don't want to flash BIOS unless absolutely necessary, you always take
the risk of something going wrong.
The new BIOS has the Tualatin microcode already so I don't see
the need for CTMC.
The CPU I'm installing is 1.3GHz and I think I'll have to update the
BIOS to get a high enough multiplier (13).
As far as heat goes I'm lead to believe that 1.5v Celerons (.13u)
run cooler than 1.75v P3's (.18u) given the same Mhz clock.
--
Bob
Paul
<bby...@bigpond.NOCAPSnet.au> wrote:
The issue is not the heat in the processor. It is heat created
in the voltage regulator circuit. Thermal loss in this circuit
is proportional to current drawn and to a first order approximation
is not dependent on the output voltage. All I'm suggesting is
to put a finger on the components of that circuit (next to the processor
socket), to make sure that when your new processor is running 100%
load, it isn't overheating something. Since there is unit to unit
variation between processors, you might get one that doesn't draw
the maximum current. Otherwise, it sounds like flashing a new BIOS
is all that is holding you back (as in some cases the multiplier
fix was added late in the life cycle of the BIOS).
Paul
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> The issue is not the heat in the processor. It is heat created
> in the voltage regulator circuit. Thermal loss in this circuit
> is proportional .........
> ......... Since there is unit to unit
> variation between processors, you might get one that doesn't draw
> the maximum current. .......
And what then?
Buy another CPU?
Forget everything, cut my losses and go back to scratch?
I'm trying to understand the relevance of this except as an
academic exercise.
--
Bob
Paul
<bby...@bigpond.NOCAPSnet.au> wrote:
I'm trying to point out one of the exposures of using high performance
processors not listed in the cpusupport webpage. Asus is not guaranteeing
that they work, so it is up to us, the users, to figure it out for
ourselves. The Powerleap/Upgradeware and the like companies are not
going to shed any tears if our motherboards suddenly stop working,
because they have our money.
To do this job properly, you need to know which version of the VRM spec
was used to design your motherboard. My list of collected specs is
below, with the comments next to the spec number being my opinion alone.
You can read the specs and extract your own parameters. Note that Asus
never tells us what the maximum Vcore current is for the board, neither
do they list the VRM spec which applies to the board. The only evidence
we get, is the processor maximum frequency stops at a certain point on
the cpusupport webpage, which _implies_ an engineering limit. (An
example of this can be seen in the case of the P4S8X, which had power
problems early in its life - the cpusupport webpage does not list a
P4 3.06GHz processor as an option for that board, which I interpret
as an admission of guilt.)
As near as I can tell, my motherboard (P2B-S) is either VRM 8.1 or
VRM 8.2. It really depends on which revision of the board you own, as to
which of the specs apply. (The HIP6004B regulator supports 1.30v-3.5v
so when that chip is used, the board is VRM 8.2). I don't believe that
VRM 8.5 applies, unless a board is specifically designed for Tualatin
processors. So, if the processor you use draws 22.8 amps, and the
VRM spec used to design the motherboard calls up 14.2 or 16.1 amps,
I think you can see the _potential_ for trouble. Now, Asus could always
have designed for a little more than the VRM current requirement, but
we'll _never know_ .
If you are a gambler like the rest of us crazy upgraders, you can
try a processor not listed by Asus. Personally, if this P2B-S dies
tommorrow, I'll be pissed, but it is not the end of the world. If
this is your only computer, DON'T DO IT.
You don't really need to download all of these, but they are there
for you if you are curious.
VRM 8.1 1.80-2.05, 2.1-3.5 volts 14.2 amps (early rev P2B-xx family?)
http://developer.intel.ru/download/design/Celeron/APPLNOTS/24340801.pdf
VRM 8.2 1.30-2.05, 2.1-3.5 volts 16.1 amps (late rev P2B-xx family?)
http://developer.intel.ru/download/design/pentiumii/xeon/designgd/24377302.pdf
VRM 8.3 (Xeon only)
ftp://download.intel.com/design/PentiumII/xeon/designgd/24387002.pdf
VRM 8.4 1.30-2.05 volts 22.8 amps (Coppermine uses 1.75V...)
http://www.intel.com/design/pentiumiii/designgd/24533505.pdf
VRM 8.5 1.050-1.825 volts 28.0 amps (Tualatin uses 1.475V...)
http://www.intel.com/design/pentiumiii/designgd/24965902.pdf
***********
Now, I'll try working an example with your board. The cpusupport webpage
at Asus lists the fastest processors for the CUBX as Celeron 1100 or
P3-1.0 GHz. Here are all the processors that I think they mean:
Pentium® III processors Filter Results
sSpec# CPU Speed Box/OEM Bus Speed Package Type P(W),V,P/V=I
SL5QV 1.00 GHZ BOTH 100 MHz 370 pin PPGA FC-PGA1 29.0,1.75,16.6A
SL5GR 1.00 GHZ BOTH 133 MHz 370 pin PPGA FC-PGA2 29.9,1.475,NA
SL5FQ 1.00 GHZ BOX 133 MHz 370 pin PPGA FC-PGA1 29.0,1.75,16.6A
SL5DV 1.00 GHZ BOX 133 MHz 370 pin PPGA FC-PGA1 29.0,1.75,16.6A
SL5B3 1.00 GHZ BOTH 133 MHz 370 pin PPGA FC-PGA2 29.0,1.75,16.6A
SL52R 1.00 GHZ BOX 133 MHz 370 pin PPGA FC-PGA1 29.0,1.75,16.6A
SL4MF 1.00 GHZ BOX 133 MHz 370 pin PPGA FC-PGA1 26.1,1.7,15.3A
SL4C8 1.00 GHZ BOTH 133 MHz 370 pin PPGA FC-PGA1 26.1,1.7,15.3A
The second item is a Tualatin, which is not what Asus intended. Whether
your board works properly at 133MHz or not, the listed power seems to be
the same anyway, and the computed maximum current is 16.6 amps.
Intel® Celeron® processors Filter Results
Click on the sSpec number to see the complete set of processor specs.
sSpec# CPU Speed Box/OEM Bus Speed Package Type P(W),V,P/V=I
SL6JR 1.10 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 29.5,1.5,NA
SL6CA 1.10 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 30.8,1.5,NA
SL5ZE 1.10 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 29.8,1.475,NA
SL5XU 1.10 GHz BOTH 100 MHz 370 pin PPGA FC-PGA1 33.0,1.75,18.9
SL5XR 1.10 GHz BOX 100 MHz 370 pin PPGA FC-PGA1 33.0,1.75,18.9
SL5VQ 1.10 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 28.9,1.475,NA
The four "NA" items are Tualatin, which is not what Asus intended. So, Asus is
saying that the max of 18.9 or 16.6 amps is possible from your motherboard. Now,
we use the 18.9 amp limit to vet the possible Tualatins. First, I try
looking at 1.3GHz Tualatins (which run at 1.5 volts).
Intel® Celeron® processors Filter Results
Click on the sSpec number to see the complete set of processor specs.
sSpec# CPU Speed Box/OEM Bus Speed Package Type P(W),V,P/V=I
SL6JT 1.30 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 32.0,1.5,21.3
SL6C7 1.30 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 32.0,1.5,21.3
SL5ZJ 1.30 GHz BOX 100 MHz FC-PGA2 33.4,1.5,22.3
SL5VR 1.30 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 33.4,1.5,22.3
So, the computed current of 21.3 or 22.3 amps for this choice of processor
is 13% or 18% more than the fastest processor that Asus quotes. So, what
is safe?
A rough range of values can be found in:
http://support.intel.com/design/celeron/datashts/29859604.pdf (page 26)
1.4 GHz 22.6
1.30 GHz 22.5
1.20 GHz 21.5
1.20 GHz 20.6
1.10A GHz 19.9
1A GHz 19.1
900 MHz 18.1
or the processorfinder web page returns these, and I compute an
approximate maximum current by dividing the thermal power by the Vcore:
Intel® Celeron® processors Filter Results (FSB 100, FCPGA2)
Click on the sSpec number to see the complete set of processor specs.
Click on the sSpec number to see the complete set of processor specs.
sSpec# CPU Speed Box/OEM Bus Speed Package Type P(W),V,P/V=I
SL6JV 1.40 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 33.2,1.5,22.1
SL6JU 1.40 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 ?
SL6C6 1.40 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 ?
SL5VR 1.30 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 33.0,1.5,22.0
SL6JT 1.30 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 32.0,1.5,21.3
SL6C7 1.30 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 32.0,1.5,21.3
SL6JS 1.20 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 32.0,1.5,21.3
SL6C8 1.20 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 32.0,1.5,21.3
SL5Y5 1.20 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 29.9,1.475,20.3
SL5XS 1.20 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 29.9,1.475,20.3
SL6CA 1.10 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 30.8,1.5,20.5
SL5ZE 1.10 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 28.9,1.475,19.6
SL5VQ 1.10 GHz BOTH 100 MHz 370 pin PPGA FC-PGA2 28.9,1.5,19.3
SL6JR 1.10 GHz BOX 100 MHz 370 pin PPGA FC-PGA2 29.0,1.5,19.3
SL6JQ 1.00 GHZ BOX 100 MHz 370 pin PPGA FC-PGA2 29.5,1.5,19.7
SL6CB 1.00 GHZ BOTH 100 MHz 370 pin PPGA FC-PGA2 29.5,1.5,19.7
SL5ZF 1.00 GHZ BOX 100 MHz 370 pin PPGA FC-PGA2 27.8,1.475,18.8
SL5VP 1.00 GHZ BOTH 100 MHz 370 pin PPGA FC-PGA2 27.8,1.5,18.5
So, there are a couple of 1.1 GHz (with 100MHz FSB) Tualatins that are close
to this estimated limit. So, you can now judge for yourself, as to what the
risk is on your motherboard. (And, it is still possible that the Asus design
is for more than the 18.9 amp number, but I don't know a way to figure it
out.)
***********
Good luck with whatever you decide,
Paul
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> ......... So, you can now judge for yourself, as to what the
> risk is on your motherboard. .......
And that's the bottom line, it's a judgement call.
Thanks for all that info, can't say I understand it fully but I get
the gist. To be honest I wouldn't know a MOSFET from a
toothbrush and I'm not about to go sticking my fingers on
a m/board while it's running ("danger Will Robinson") without
at least a pic of what I'm supposed to touch.
Like anyone who overclocks, and I never have before, it's a
risk. I've been trying to find the fastest legitimate CPU for a
CUBX but haven't had any luck, everything for sale is 133FSB,
I need 100. I just happen to stumble onto a dealer here in
Oz selling the 370GU and it appears from the info I've found
that they work. I get a 370GU and new Celeron 1.3GHz for
AU$170, can't pass it up.
I have a spare CUSL2-C BP I'm using to build another PC and
I'd imagine I could pick up a CUBX somewhere if I have to.
--
Bob
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> ........
> The only other thing you should be warned about, is the Vcore voltage
> regulator. .......
> .....
> In my experiments, a Tualatin 1100/100 barely makes the MOSFETs in the
> Vcore circuit warm. ......
Ok Paul I'm researching all the possibilities and think I know where the
MOSFETs are. There seems to be two of them between the ATX power
supply and the CPU. I found an article showing MOSFET cooling:
http://www.vr-zone.com/guides/kt7/
If I get a heat sink at least I'll be in front even without the CPU upgrade.
Can't hurt to fit one can it?
--
Bob
Paul
<bby...@bigpond.NOCAPSnet.au> wrote:
Use the finger test first :-) Get a copy of burnmmx (part of cpuburn4
package) and start it up, then touch the two MOSFETs with your fingertip.
If it is getting hard to hold a finger on them, then it is time to
<cntrl> C the burnmmx program and fit your heatsinking device. My
Tualatin 1100 makes them barely warm, while my Tualatin 1400 makes them
a lot warmer.
I looked at the link you gave above, and unfortunately they didn't have
a picture of the heatsink fitted to the MOSFET. If it is a U channel
aluminum HS with thermal tape on the back and it rests on top of the MOSFET,
I see nothing wrong with it. Just be careful that no metal from the
heatsink touches the pins on the MOSFET, as nasty things might happen
(including taking out your power supply).
I currently have a small fan blowing over the surface of the MOSFETs,
toroids, and caps. On my motherboard, that area is a dead zone in terms
of case air flow, so the combination of heatsink plus just a little
air will work wonders for the surface temperature.
Motherboards use the copper foil as a heatsink for the MOSFETs, which
causes the surface of the motherboard to heat up. The backside of the
board can get warm as well - it is too bad there is no airflow on
the backs of motherboards, as this would help as well.
I've already purchased some replacement MOSFETs for my motherboard,
but my replacement experiment is a "go slow" kind of project right
now. I have to buy some "chipquik" (low temperature desoldering alloy)
to remove them, and I'm still tossing around some possibilities as
to how best to heatsink the new MOSFETs. I'm thinking of changing
the MOSFET package from D2PAK (that just lays on the board), to a
TO-220 package (that stands up off the board). The TO-220 package
has a screwhole in it, so a bolt-on heatsink can be used. The
reason I'm not rushing to do this, is it is very easy to damage
the copper foil on these motherboards, so my mobo might not survive
the attempt. (I removed one pad by accident, while changing the
voltage regulator.)
Something else that occurred to me, after this series of posts, is
how the overcurrent protection works on my regulator. The datasheet
for the regulator, says that it monitors the current through the
MOSFET, by measuring the voltage across it while it is switched
on. This is V=Ion*Ron. The idea is, if there is a short on the output,
the higher Ion causes a threshold to be passed, and the regulator
shuts down, to protect the MOSFET. Now, in a thermal runaway, the
Ion is a constant, but the Ron is gradually rising, as the device
gets too hot to touch. In theory, this too should cause the thing
to shut down. If it doesn't shut down, this could mean Asus just
didn't implement this feature, or the threshold is set really high
to prevent nuisance tripping. If this feature really works, it
would protect only one of the two MOSFETs.
Enough theories, have fun :-)
Paul
Paul
(Paul) wrote:
The http://www.vr-zone.com/guides/kt7/ link, mentions the heatsinks
on the Epox 8KTA3+ board. This link shows a clear picture of them -
http://www.ocinside.de/index_e.html?/html/workshop/nmc8ttx_epox8kta3_vcore.html
It looks like these are standup MOSFETs (TO220 package) and the
heat sinks are some very nice extruded bolt-on HS. So, the style
of heatsink you use should match whatever style is used on your board.
Unfortunately, the D2PAK MOSFETs that lay on the board are harder to
cool, as a heatsink that has a reasonable fit is usually smaller then
the ones pictured on the "ocinside" site. One style of heatsink actually
solders to the board, and comes very close to touching the sides of the
MOSFET as it lays on the board - this style is actually cooling the PCB
as it gets heated by the MOSFET. Just sticking a heatsink to the plastic
body of the MOSFET will help a bit, but is not as good as a metal to
metal contact.
This vendor has an online catalog - you can download the full catalog
(at a hefty 40MB currently) or download individual pages. The pictures
in these catalog pages might give you some ideas as to what you might
use. I don't know of any other online electronics retailer who have
as detailed a catalog.
http://dkc1.digikey.com/us/PDF/T032/CatalogPages.html (see pg.476-481)
HTH,
Paul
Paul
Here is another idea. This is from a tomshardware review. These heatsinks
need clearance around them, which many motherboards don't have.
http://www6.tomshardware.com/mainboard/20030414/images/vrm_msi.jpg (picture)
http://www6.tomshardware.com/mainboard/20030414/i875p-17.html (review)
Paul
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> "Bob Byrne" wrote:
> > Ok Paul I'm researching all the possibilities and think I know where the
> > MOSFETs are. There seems to be two of them between the ATX power
> > supply and the CPU. ......
> Use the finger test first :-) Get a copy of burnmmx (part of cpuburn4
> package) and start it up, ......
I have MBM 5 and it comes with a little program called Heat Up.
> ...... If it is a U channel
> aluminum HS with thermal tape on the back and it rests on top of the MOSFET,
> I see nothing wrong with it. ......
I'm trying to find something like that locally.
> ......... I'm thinking of changing
> the MOSFET package from D2PAK (that just lays on the board), to a
> TO-220 package (that stands up off the board). The TO-220 package
> has a screwhole in it, so a bolt-on heatsink can be used. ...
Yes, I noticed the difference too. I assume the lay-down type was a
way of using the m/board as a heatsink.
--
Bob
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> ........
> Unfortunately, the D2PAK MOSFETs that lay on the board are harder to
> cool, ...........
> ........ Just sticking a heatsink to the plastic
> body of the MOSFET will help a bit, but is not as good as a metal to
> metal contact.
Yeah, I've been having a closer look at my CUSL2-C, it's similar to my
CUBX and they do feel very plastic as well. Hope they are not thermal
insulators. Like you said, the finger test is the best way to find out and
go from there.
> This vendor has an online catalog ......
It's not a matter of finding a heat sink to suit as much as finding
one locally. I will probably end up cannibalising something else to
suit my needs. Because of the small size it will have to be copper to be
efficient and high rather than wide to fit.
One of these could perhaps be mounted on its' end:
http://www.pcca.com.au/ramsink.html
--
Bob
JK
<bby...@bigpond.NOCAPSnet.au> wrote:
>Hello.
> I've looked thru Google for info on updating to a
>Celeron FC-PGA2 (Tualatin) CPU and found everything I need
>except what BIOS to use in a CUBX board.
>I currently have V1007 final and have found several flavours
>of V1008 beta 4, the latest being
>1008b4_CMD_1.9.16.zip (7/18/02)
>It has a readme -
>"ASUS CUBX (CMD-648) BIOS"
>"BIOS v. 1008 BETA 4 (Tulatin mod) - CMD BIOS v. 1.9.16"
>
>I assume this would be the one to use if I need to update at all.
>Can anyone remember back that far and advise me on what I
>need to do?
I have used the 370GU with CUBX and with both a celeron 1200 and 1300.
The bios cannot find out to post with 1300, but 1200. If I set 13x112
it posts correctly again. But this is of no importance really what it
writes during boot.
For the cpu codes:
I used just cbrom and loaned the cpucode.exe from the CUV4X that I saw
had cpucodes 06B0, 06B1, 06B4 where the 06B4 was rare at that time.
I saw no issue with voltage regulator heat.
best regards
John
Bob Byrne
"JK" <johnkn...@mail.dk> wrote:
> I have used the 370GU with CUBX and with both a celeron 1200 and 1300.
> The bios cannot find out to post with 1300, but 1200. If I set 13x112
> it posts correctly again. But this is of no importance really what it
> writes during boot.
Are you saying that a 1300 will be seen as a 1200 but in fact run
at full speed?
> For the cpu codes:
>
> I used just cbrom and loaned the cpucode.exe from the CUV4X that I saw
> had cpucodes 06B0, 06B1, 06B4 where the 06B4 was rare at that time.
Haven't a clue what you just said :-)
> I saw no issue with voltage regulator heat.
Now that's good news.
--
Bob
Paul
<bby...@bigpond.NOCAPSnet.au> wrote:
> "JK" <johnkn...@mail.dk> wrote:
>
> > I have used the 370GU with CUBX and with both a celeron 1200 and 1300.
> > The bios cannot find out to post with 1300, but 1200. If I set 13x112
> > it posts correctly again. But this is of no importance really what it
> > writes during boot.
>
> Are you saying that a 1300 will be seen as a 1200 but in fact run
> at full speed?
>
> > For the cpu codes:
> >
> > I used just cbrom and loaned the cpucode.exe from the CUV4X that I saw
> > had cpucodes 06B0, 06B1, 06B4 where the 06B4 was rare at that time.
>
> Haven't a clue what you just said :-)
He is referring to the Microcode Update stored in the BIOS. An Asus
BIOS has support for a certain number of "steppings" of Intel
processor, and each one of these versions of processors has a
CPUID register, with a hexidecimal value like 06B4. Intel has a
feature in each processor, where code can be stored inside a small
RAM on the processor. The code is used to patch broken features on
the processor, but as to what the bugs might be, you would have to
check the "Spec Update" document for the processor from the Intel
site.
If the proper 2KByte Microcode patch is not stored in the BIOS, then
the BIOS will issue an error message about it. I didn't have any
visible problems running without the Microcode patch under Win98se.
To fix the problem, there are a couple of recipes, with the CBROM
BIOS editing program being just one of the solutions. I managed to
use another method (via CTMC, which stands for CT MicroCode,
from CT magazine), and it fixed the BIOS without breaking it.
If you are running WinXP, WinXP can detect that the Microcode is
missing, and load it after the system is running. This would leave
the interval during which the computer is booting unprotected, but
would ensure "computing correctness" after the desktop appears.
Bob Byrne
"Paul" <nos...@needed.com> wrote:
> He is referring to the Microcode Update stored in the BIOS. An Asus
> BIOS has support for a certain number of "steppings" of Intel
> processor, .....
> ..........
> If you are running WinXP, WinXP can detect that the Microcode is
> missing, and load it after the system is running. This would leave
> the interval during which the computer is booting unprotected, but
> would ensure "computing correctness" after the desktop appears.
I'll take your word for it, I still don't understand this stuff, too technical
for me. I'll flash the BIOS with the latest version, which supposed to
have the Tualatin code and see what happens.
I am running XP Pro.
--
Bob
Bob Byrne
"Bob Byrne" wrote:
> "Paul" <nos...@needed.com> wrote:
> > ..........
> > If you are running WinXP, WinXP can detect that the Microcode is
> > missing, and load it after the system is running. This would leave
> > the interval during which the computer is booting unprotected, but
> > would ensure "computing correctness" after the desktop appears.
> ... I'll flash the BIOS with the latest version, which supposed to
> have the Tualatin code and see what happens.
> I am running XP Pro.
Just flashed the BIOS and it seems ok. Hope to have the new
Tualatin CPU in the next few days.
--
Bob