Analysis of the modification technique in the Remarked CPUs

We clarified the technique of the modification in the Remarked Sempron.
Moreover, we received the report of similar CPU from the United Kingdom. The United Kingdom#1 The United Kingdom#2
We got a mail from Poland. Poland

The shipment from AMD of the SocketA Sempron has already stopped. We confirmed the Remarked Sempron 2800 existed in the market today.
Therefore, please note "Sempron2800 sold now has the possibility of the fake". If you discovered it in the market, we will recommend the important information to be offered to AMD. And, let's request that it wants you to expel it from the market.

Then, please look at the photograph of the genuine Sempron 2800 first.
Reference : Image of the genuine AMD Sempron2800, 720x720px

This photograph is L11 bridge of genuine Sempron2800 +. The L11 sets the core voltage of CPU, and 1.60V is supplied to Sempron by this pattern. This time, please watch the L11 bridges. You will be able to find doubtful Sempron by this comparison.

  • 1.6V : Bridge of L11[1] and L11[3] is open.
  • CPU for Desktop : All L8 bridges are closed.

Analysis of the modification technique in the Remarked Sempron 2800+

We obtained this high resolution photograph. And, it succeeded in the clarification of an illegal technique in the fake Sempron.

Modification to the close in the L11.

Please watch this photograph.
You will find two small round dents. This dent is short-circuited of the bridge and internal GND plane. As a result, the recognition of the bridge is changed from the open to the close.

Modification to the open in the L11.

You will find the cut bridge wiring. In the photograph, it looks white. The recognition of the cut bridge changes from the close to open.

These L11 bridges are detected as 1.65V.


Modification of the multiplier in the L3.

In L3, an illegal modification by the dent was found.

These L3 bridges are detected as 12x.

Modification of the FSB in the L12.

In L12, an illegal modification by the dent and wound was found.

These L12 bridges are detected as 166MHz.


About these photographs

These photographs are Fake Sempron 2800 sold in Akihabara in Tokyo. We were going to exclude this from the market. Then, we sent the photograph of this Sempron and information on the shop to the Authorized distributor of AMD Japan. In addition, this information was sent to the call center in AMD Japan. We hear that AMD Japan contacted the AMD United States. As a result, only genuine Sempron was sold in Japan on June, 2006.

Remarked Sempron 2800 from Tokyo Japan [May 02, 2006] : Enlarge of image

Reference : Genuine AMD GeodeNX Enlarge image of GeodeNX1750


Remarked Sempron 2800+ from United Kingdom #1 [Septembar 04, 2006]

We received two reports of Sempron2800+ that caused the problem from the United Kingdom. Thank you.
The first report is Mr. Damian.

Hi, thanks for a great site.
I bought a SDA2800DUT3D sempron about 3 weeks ago here in the UK from a very big online store and have been having problems with it, overheating and crashing. It looks very much like some of the pictures on your site, particularly the bridges L6, L8 and L11

I sent a mail to AMD over a week ago along with the attached photo, but have not heard anything back from them. It seems that they are not bothered about remarked cpu's. I've since sent another mail and am waiting to see if they ignore that as well.

Here's my sempron, any ideas on whether it is fake or not. As you can see I tried adding more heatsink paste after it failed in an attempt to keep the temperature down.
thanks
Damian

Enlarge of image

I wrote:
You are correct. I certainly think that the Sempron is a fake.
However, it is difficult to clarify disagreeing of whether it is a fake or the truth.

When contact is done to AMD directly, AMD seems to be passive about the Remarked Sempron. It seems to be more effective to do contact to the distributing agent of AMD. We will recommend speaking at that time, 'I want to offer AMD important information about the Remarked Sempron'.

-----
There is a good method.
I expect that the core voltage becomes 1.65V by your photograph. The core voltage of correct Sempron SDA2800DUT3D is 1.60V. Could you detect the voltage of your Sempron?

It can be judged that the Sempron is abnormal if it is 1.65V. If it is defective CPU, it might be able to be returned to the shop.

notes

Even if the investigation of the remark begins, AMD seems to be passive to opening the details to the public. Moreover, the system that neatly accepts the remark information seems not to function.

However, the power of AMD is necessary to exclude doubtful CPU from the market. We pulled one's coat to AMD. And, it succeeded two times in Japan in the exclusion of those CPU from the market. In addition, we heard that AMD had dealt with this problem at once in Germany.

It is 1.65V according to an illegal modification.

Mr. Damian wrote:
The core voltage is 1.65 and not 1.60 (measured using Everest Home Edition) I've contacted AMD twice over the last 10 days and not had a reply yet. It would seem that they are not interested. People can make their own mind up as to why they don't seem to be bothered. Thanks again for your website otherwise I would not have known that the cpu is most likely fake.



Remarked Sempron 2800+ from United Kingdom #2 [Septembar 05, 2006]

The report from another United Kingdom is Mr. Banerjee. We confirmed this Sempron2800+ was the Fake.

Mr. Banerjee wrote:
Hi, I've just bought two Socket A devices, a "Sempron 2800" and a "Duron 1800" which I thought were from a reputable supplier. The motherboard set the core voltage for the "Sempron" to 1.79 V, which I thought was a little suspicious. I did a search and came across your site. Thanks for making this information available.

It's clear to me that the Sempron is fake, but the Duron looks so similar it also seems suspicious. Can you tell me whether the Duron is also a fake?

Enlarge of image
Enlarge image of genuine Duron


I wrote:
I think that the Sempron 2800+ is a fake. I expect that the core voltage becomes 1.775V by your photograph. The core voltage of correct Sempron SDA2800DUT3D is 1.60V. If it is defective, it might be able to be returned to the shop.

>Duron is also a fake?
There is no problem. Duron is the truth. The fake Duron has not been discovered by present.

It is 1.775V according to an illegal modification.

Though it doesn't understand who is doing . . . They work into the L11, and have changed the core voltage into another value.

Close to open:
The bridge has been opened by newly making the pit.

Open to close:
An internal pattern is short-circuited by the dent, and it does in equivalence to the close.



Remarked Sempron 2800+ from Poland [Septembar 18, 2006]

We received the report of the remarked Sempron2800+ from Poland. Thank you.

Mr. Leszek wrote:
Hello, I`ve bought yesterday Sempron2800 processor. I saw that offered processors in one shop are all Remarked. So I went to another shop to see if there are original processor. However there weren't original, only Remarked versions. I talked to the guy at the computer shop about Remarked processor and he said that if my sempron work at 1.7? or something similar Vcore he will give my money back.

I went home, mounted processor, fited thermaltake silent boost cooling and turned pc on. Chandeg the FSB do 166MHz and... nothing wrong happend next, temperature was fine, about 40 degrees C. When I was stressing CPU maximum temp was about 51 degrees. my cpu works at 1.725V by the way. Next when I`ve compared performance with sisoft sandra with other processors my cpu was at the same performance as another similiar/same amd processors.

So, is it sure, that processors with pits at L8 and some other small changes are all remarked?

Enlarge of image


I wrote:
We watched a lot of SocketA processors by present. There was no product that there is an irregular bridge in Sempron. Therefore, it is a remarked. Even if it is illegal CPU, it operates in many cases correctly. However, it is overclocking of another product. You will think that GeodeNX1750 is just like your Sempron.

Please act carefully.

Of course, it is necessary to blot out the Remark from the market. However, SocketACPU is hardly supplied now. We have worry that the supply of SocketA might decrease further if this problem is exaggeratedly notified. It is the best for our user to solve it by what method.

Please note the following when you return the CPU to the shop.
The shop might have sold the processor without knowing it was Remarked.
New CPU might not be sold in the market.



Appendix L11 : VID Code for startup Core voltage

The following data is AMD GeodeNX1750. You will notice this GeodeNX1750 is the same as the Remarked Sempron2800 +.
In a word, Remarked Sempron2800 + was made from the illegal modification of GeodeNX1750.

Enlarge image of GeodeNX1750

SocketA VID Code
BridgesV_CORE (V)
L11
[4:0]
DesktopMobile
SempronXP-M
CCCCC1.850 2.000
CCCC:1.825 1.950
CCC:C1.800(N)1.900
CCC::1.775 1.850
CC:CC1.750(M)1.800
CC:C:1.725 1.750
CC::C1.700(P)1.700
CC:::1.675 1.650
C:CCC1.650(K)1.600
C:CC:1.625 1.550
C:C:C1.600(U)1.500(L)
C:C::1.575 1.450(Q)
C::CC1.550(H)1.400(V)
C::C:1.525 1.350(J)
C:::C1.500(L)1.300(W)
C::::1.475 -
SocketA VID Code
BridgesV_CORE (V)
L11
[4:0]
DesktopMobile
-Geode NX
:CCCC1.4501.275
:CCC:1.4251.250(X) Geode NX1750
:CC:C1.4001.225
:CC::1.3751.200(T)
:C:CC1.3501.175
:C:C:1.3251.150(C)
:C::C1.3001.125
:C:::1.2751.100(Y) Geode NX1250
::CCC1.2501.075
::CC:1.2251.050
::C:C1.2001.025
::C::1.1751.000(G) Geode NX1500
:::CC1.1500.975
:::C:1.1250.950
::::C1.1000.925
::::: - -

C is a closed bridge. The logic is 0.
: is an open bridge.The logic is 1.


Thank you for your report

We got the permission of opening mail to the public.

  • Mr. Damian from the United Kingdom.
  • Mr. Banerjee from the United Kingdom.
  • Mr. Leszek form Poland.
  • Mr. Tatataku form Tokyo Japan

Related to this article

We think that this is urgent and important. We permit the photograph and information on these page to be used in case of being to inform many people of this article. Please link of fab51.com.