GA-7VRXP Memory problems

Hi, Simon.

Here's a link to a page with pictures of the fix...

http://mysite.wanadoo-members.co.uk/...es/cap_mod.htm

"Simon Elliott" wrote in message
t...
Homer J. Simpson wrote:

Hi, Simon.

Do you know if your Rev. 1.1 board has the voltage regulation problem
fixed?
A couple of years ago owners of the Rev. 1.1 boards RMA'd them and
Gigabyte
repaired them and sent it back to them. The repair, if I remember
correctly, was the addition of a 4.7uF 30V tantalum capacitor on one leg
of
a voltage regulator MOSFET next to the NB_FAN header.

Prior to the voltage regulation fix being available from Gigabyte some
were
raising the AGP Voltage, DDR Voltage or VCore Voltage to get enough
voltage
to the video card, memory modules or CPU respectively. This wasn't
always
successful.

"Simon Elliott" wrote in message
...


Good question Homer. Honest answer, don't know.

To my knowledge the board has never been RMA'd. I bought it new from a
fair a couple of years ago, as the basis of my current box. It's not
clocked but has worked quite happily with the optimum settings

I can find the NB_FAN socket but don't know what a voltage regulator
MOSFET looks like. Is it one of several chip-like things running down
the left hand side of the socket? I can't see anything that looks like
an extra capacitor. As you can guess, I'm working at the limits of my
knowledge here.

Thanks for taking the time to help me. BTW, hows Midge? :-)

--

Simon

Homer J. Simpson wrote:

Hi, Simon.

Here's a link to a page with pictures of the fix...

http://mysite.wanadoo-members.co.uk/...es/cap_mod.htm



I see what you're talking about and no, I don't have the fix. The board
is two years old, therefore too old to be RMA'd and this kind of work is
beyond me; so it sounds like something I'm going to have try and work
around.

Based on this and the information from your previous posting, I'm
guessing that my problem is a voltage issue. Presumably the three
different sticks I've tried have all been affected as a result. My
current hypothesis is that I might have some success using another 256MB
stick (complementing the one already in place), instead of 512MB and I'm
off in a minute to make the swap.

Your previous posting refers to "raising the .. DDR Voltage ... to get
enough voltage to the ... memory modules". If I find that the 256MB
doesn't work alongside my original (effectively meaning I can't add any
extra memory) then this increase is something I might consider. Can you
offer any advice about what sort of increments should I be trying and
anything else I should be aware of to avoid making a costly mistake?

Thank you again for your ongoing help.

--

Simon

Simon Elliott wrote:

I'm guessing that my problem is a voltage issue. Presumably the three
different sticks I've tried have all been affected as a result. My
current hypothesis is that I might have some success using another
256MB stick (complementing the one already in place), instead of 512MB
and I'm off in a minute to make the swap.

And what do you know? It works.

Swapped the 512MB for 256MB (also Kingston brand) and the computer boots
without a problem. In my case 2 x 256MB works where 1 x 512MB doesn't.

I can only conclude that Homer was on the right track, there is a
voltage issue with my board and that a 256MB stick is compatible, where
a 512MB stick isn't. I'm currently waiting on eBay, but I might be
getting a third 256MB stick and now I'm quietly optimistic that this one
will work as well.

Thanks again Homer for your help.

--

Simon

The default DDR Voltage is 2.5V. The increments are by 0.1V. The maximum
DDR Voltage that you can select is 2.8V. This means that you have 3
different voltages that you can try.

If none of the DDR Voltage changes works, try raising the VCore Voltage to
+5.0%. You may have to try various combinations of DDR and VCore voltages.

Check your system's voltage status in the Hardware Monitor & Misc Setup BIOS
setup page and post them here.

"Simon Elliott" wrote in message
...
Homer J. Simpson wrote:

Hi, Simon.

Here's a link to a page with pictures of the fix...

http://mysite.wanadoo-members.co.uk/...es/cap_mod.htm



I see what you're talking about and no, I don't have the fix. The board
is two years old, therefore too old to be RMA'd and this kind of work is
beyond me; so it sounds like something I'm going to have try and work
around.

Based on this and the information from your previous posting, I'm
guessing that my problem is a voltage issue. Presumably the three
different sticks I've tried have all been affected as a result. My
current hypothesis is that I might have some success using another 256MB
stick (complementing the one already in place), instead of 512MB and I'm
off in a minute to make the swap.

Your previous posting refers to "raising the .. DDR Voltage ... to get
enough voltage to the ... memory modules". If I find that the 256MB
doesn't work alongside my original (effectively meaning I can't add any
extra memory) then this increase is something I might consider. Can you
offer any advice about what sort of increments should I be trying and
anything else I should be aware of to avoid making a costly mistake?

Thank you again for your ongoing help.

--

Simon

Homer J. Simpson wrote:

The default DDR Voltage is 2.5V. The increments are by 0.1V. The maximum
DDR Voltage that you can select is 2.8V. This means that you have 3
different voltages that you can try.

If none of the DDR Voltage changes works, try raising the VCore Voltage to
+5.0%. You may have to try various combinations of DDR and VCore voltages.

Check your system's voltage status in the Hardware Monitor & Misc Setup BIOS
setup page and post them here.

"Simon Elliott" wrote in message
...


Thanks for the tips Homer.

I'd bought the extra memory because I was experiencing an occasional
'hiccup' in normal operation. If I had several applications running and
swapped between them quickly then the computer would hang for a couple
of seconds, CDs or radio would 'stutter' and then normal operation
would resume.

I attributed this to a memory bottleneck and time spent accessing the
swap file. Since I added the extra 256MB I've not experienced the
problem (fingers crossed).

So I'm happy with what I've now got and don't intend (in the near
future) to be adding any more memory; particularly if it's greater than
a 256MB stick.

So I've saved your previous posting for future reference and here's my
BIOS readings for academic interest.

AGP Voltage: 1.5V
DDR Voltage: 2.5V
Vcore Voltage: Normal
CPU Host Clock: (Mhz): 133
---
VCo +1.728V
VTT: +1.184V
+3.300V: +3.264V
+5.000V: +4.865V
+12.000V: +12.224V
5VSB +5.886V

Thanks again for your help.

--

Simon

Your Vtt (memory termination voltage) at +1.184 V is low by more than 5%.
It is suppose to be ½ of the DDR Voltage (i.e. 1.25V). The Vtt on my Rev. 2
board is +1.248 V, which is low by 0.16%.

Too bad you didn't post the Vtt for the 512MB stick(s) when you tried them.
It most likely would have been lower than +1.184 V. This would definitely
point to a voltage regulation problem that needs to be fixed.

If you would have raised the DDR Voltage, with the 512MB stick(s) you've
tried, you probably would have been able to use it.

If you know someone who is an electronics hobbyist and can obtain and solder
that capacitor to your board, that would be the ultimate solution.

"Simon Elliott" wrote in message
...
Homer J. Simpson wrote:

The default DDR Voltage is 2.5V. The increments are by 0.1V. The maximum
DDR Voltage that you can select is 2.8V. This means that you have 3
different voltages that you can try.

If none of the DDR Voltage changes works, try raising the VCore Voltage to
+5.0%. You may have to try various combinations of DDR and VCore
voltages.

Check your system's voltage status in the Hardware Monitor & Misc Setup
BIOS
setup page and post them here.

"Simon Elliott" wrote in message
...

Thanks for the tips Homer.

I'd bought the extra memory because I was experiencing an occasional
'hiccup' in normal operation. If I had several applications running and
swapped between them quickly then the computer would hang for a couple of
seconds, CDs or radio would 'stutter' and then normal operation would
resume.

I attributed this to a memory bottleneck and time spent accessing the swap
file. Since I added the extra 256MB I've not experienced the problem
(fingers crossed).

So I'm happy with what I've now got and don't intend (in the near future)
to be adding any more memory; particularly if it's greater than a 256MB
stick.

So I've saved your previous posting for future reference and here's my
BIOS readings for academic interest.

AGP Voltage: 1.5V
DDR Voltage: 2.5V
Vcore Voltage: Normal
CPU Host Clock: (Mhz): 133
---
VCo +1.728V
VTT: +1.184V
+3.300V: +3.264V
+5.000V: +4.865V
+12.000V: +12.224V
5VSB +5.886V

Thanks again for your help.

--

Simon

Homer J. Simpson wrote:

Your Vtt (memory termination voltage) at +1.184 V is low by more than 5%.
It is suppose to be ½ of the DDR Voltage (i.e. 1.25V). The Vtt on my Rev. 2
board is +1.248 V, which is low by 0.16%.


That would explain a couple of blue screens I've had since I put the new
stick in, one on startup and one on shutdown (didn't bother recording
them, want to see if they become a regular occurrence first). Presumably
that's when the greatest load is placed on the machine.

Are those figures based solely on experience or do you have a reference
you could share?

Too bad you didn't post the Vtt for the 512MB stick(s) when you tried them.
It most likely would have been lower than +1.184 V. This would definitely
point to a voltage regulation problem that needs to be fixed.


Unfortunately, at the time I didn't know that sort of information would
have been useful.

If you would have raised the DDR Voltage, with the 512MB stick(s) you've
tried, you probably would have been able to use it.


I'd wondered if that might have been the case. I won't be looking for
any extra memory but if some comes my way cheaply then I might
experiment. Are there any risks of blowing components if the DDR voltage
is increased?

If you know someone who is an electronics hobbyist and can obtain and solder

that capacitor to your board, that would be the ultimate solution.

I'd like to see if those blue screens I mentioned become a frequent
problem. If so, then I might consider it. This box is my main computer
and I use it for my business, so I can't afford a long period of downtime.

--

Simon

DDR Memory Power specifications say that Vtt (memory bus tracking
termination voltage) is defined as ½ of the DDR reference voltage, which is
2.5 Volts. If you don't believe me you can do a Google on "Vtt memory
termination voltage" and read the Electrical Engineering specifications for
DDR memory power requirements. I have also taken Electrical Engineering at
university and I also have a degree in Computer Science.

You should be able to safely increase your DDR Voltage, even up to the
maximum of 2.8V, in the BIOS without damaging your memory modules. This
voltage increase doesn't have any effect on any of the other components on
your system board, just the memory bus. You currently have it set at 2.5V
but that is not what your system board is currently delivering to your
memory, my guess is that your memory is really getting about 2.37 Volts. I
would try setting DDR Voltage to 2.7V and see if your Vtt is any closer to
the preferred 1.25V. This should get you more stability.

"Simon Elliott" wrote in message
...
Homer J. Simpson wrote:

Your Vtt (memory termination voltage) at +1.184 V is low by more than 5%.
It is suppose to be ½ of the DDR Voltage (i.e. 1.25V). The Vtt on my Rev.
2 board is +1.248 V, which is low by 0.16%.

That would explain a couple of blue screens I've had since I put the new
stick in, one on startup and one on shutdown (didn't bother recording
them, want to see if they become a regular occurrence first). Presumably
that's when the greatest load is placed on the machine.

Are those figures based solely on experience or do you have a reference
you could share?

Too bad you didn't post the Vtt for the 512MB stick(s) when you tried
them. It most likely would have been lower than +1.184 V. This would
definitely point to a voltage regulation problem that needs to be fixed.

Unfortunately, at the time I didn't know that sort of information would
have been useful.

If you would have raised the DDR Voltage, with the 512MB stick(s) you've
tried, you probably would have been able to use it.

I'd wondered if that might have been the case. I won't be looking for any
extra memory but if some comes my way cheaply then I might experiment.
Are there any risks of blowing components if the DDR voltage is increased?
If you know someone who is an electronics hobbyist and can obtain and
solder
that capacitor to your board, that would be the ultimate solution.

I'd like to see if those blue screens I mentioned become a frequent
problem. If so, then I might consider it. This box is my main computer and
I use it for my business, so I can't afford a long period of downtime.

--

Simon

Homer J. Simpson wrote:

DDR Memory Power specifications say that Vtt (memory bus tracking
termination voltage) is defined as ½ of the DDR reference voltage, which is
2.5 Volts. If you don't believe me you can do a Google on "Vtt memory
termination voltage" and read the Electrical Engineering specifications for
DDR memory power requirements. I have also taken Electrical Engineering at
university and I also have a degree in Computer Science.


Took your advice and googled on 'Vtt memory termination voltage' but
couldn't find a page I understood. Most of the places I found were
talking about various manufacturers products and required more
background knowledge than I have. What little I could understand seemed
to confirm what you're saying though.

You should be able to safely increase your DDR Voltage, even up to the
maximum of 2.8V, in the BIOS without damaging your memory modules. This
voltage increase doesn't have any effect on any of the other components on
your system board, just the memory bus. You currently have it set at 2.5V
but that is not what your system board is currently delivering to your
memory, my guess is that your memory is really getting about 2.37 Volts. I
would try setting DDR Voltage to 2.7V and see if your Vtt is any closer to
the preferred 1.25V. This should get you more stability.




Had another blue screen at startup yesterday, a Stop 4E,
PFN_LIST_CORRUPT. Googling led me to -
http://support.microsoft.com/?kbid=291806
which said "This behavior may occur if you have bad RAM.".

I don't believe this to be the case, although I'll run a memtest86
sometime, just to check. If I'm understanding everything correctly then
my DDR voltage is close to the point of instability and as you suggest,
increasing it should move my board away from this point.

I plan to monitor the frequency of these BSODs and if they become too
frequent to try increasing the DDR voltage and see if they go away.

Your advice that this is unlikely to damage my memory is reassuring.

--

Simon Elliott

I would expect that you are going to get a lot of memory errors or complete
system freezeup or spontaneous reboot when you run memtest86. You cannot
rely on the results from memtest86 when you are not supplying the correct
power requirements to the memory modules. You need to test the memory on a
known stable system board to determine if your memory modules are defective.

The DDR Voltages in the BIOS, that are above the standard 2.5V, are there
for the overclockers so that they can attempt to achieve system stability
when running their memory modules faster than its original design
specifications.

"Simon Elliott" wrote in message
...
Homer J. Simpson wrote:

DDR Memory Power specifications say that Vtt (memory bus tracking
termination voltage) is defined as ½ of the DDR reference voltage, which
is 2.5 Volts. If you don't believe me you can do a Google on "Vtt memory
termination voltage" and read the Electrical Engineering specifications
for DDR memory power requirements. I have also taken Electrical
Engineering at university and I also have a degree in Computer Science.

Took your advice and googled on 'Vtt memory termination voltage' but
couldn't find a page I understood. Most of the places I found were talking
about various manufacturers products and required more background
knowledge than I have. What little I could understand seemed to confirm
what you're saying though.

You should be able to safely increase your DDR Voltage, even up to the
maximum of 2.8V, in the BIOS without damaging your memory modules. This
voltage increase doesn't have any effect on any of the other components on
your system board, just the memory bus. You currently have it set at 2.5V
but that is not what your system board is currently delivering to your
memory, my guess is that your memory is really getting about 2.37 Volts.
I would try setting DDR Voltage to 2.7V and see if your Vtt is any closer
to the preferred 1.25V. This should get you more stability.



Had another blue screen at startup yesterday, a Stop 4E, PFN_LIST_CORRUPT.
Googling led me to -
http://support.microsoft.com/?kbid=291806
which said "This behavior may occur if you have bad RAM.".

I don't believe this to be the case, although I'll run a memtest86
sometime, just to check. If I'm understanding everything correctly then my
DDR voltage is close to the point of instability and as you suggest,
increasing it should move my board away from this point.

I plan to monitor the frequency of these BSODs and if they become too
frequent to try increasing the DDR voltage and see if they go away.

Your advice that this is unlikely to damage my memory is reassuring.

--

Simon Elliott