Aspire 3100 CPU upgrade failing for unclear reasons

Hey, Charlie here, Comp TIA A+ Certified, so I do know what I’m doing for the most part.

Trying to upgrade an Acer Aspire 3100 from a “AMD 3500+ mobile Sempron” CPU (SMS3500HAX4CM Keene) to an “AMD Turion 64 x2” CPU (TMRM72DAM22GG Lion) and it's not wanting to take. It won't even POST after the transplant. Looking at the hardware the Aspire 3100 has an ATI Radeon Xpress 1100 chipset (read right off the die), which according to the documentation at the ATI web site is 100% fully compatible with the new CPU's so long as the CPU's HyperTransport clock can go down to 1Ghz. So silicon compatibility does not seem to be the issue as far as I can tell. CPU-world.com also tells a similar story. The biggest thing, and the most likely to be the problem is that the socket in the Aspire 3100 is a S1g1 socket, and the new CPU's are made for S1g2 sockets. Also just as true but of much less concern, the new CPU has significantly faster clocks than the older one (Which, the clocks should be able to underclock just fine you would think) and a slightly lower voltage (voltages actually overlap). Other than these small issues I can see no difference that would prevent it from at least clearing POST with one of the new CPU's. Despite this assumption, after several attempts with two different new CPU's I have yet to get the unit to beep and then POST like a good lappy should. Only after I reinstall the Sempron does it clear POST once again.

I may be grasping at straws here, since I'm really tired (it's 4:30 AM). But the only explanation I have for this so far is that either (1) the stupid difference between socket S1g1 vs the S1g2 is all it takes to make it not work. Or (2) some REALLY stupid small thing that would normally not effect any real system is preventing this Acer POS from working right. Or (3) Acer has intentionally blocked the CPU ID's of better processors from being installed for marketing reasons.

Any constructive thoughts?

:EDIT:

I made a small mistake. The one new CPU is certainly the "Lion" die for sure. However the other new CPU was actually a Turion 64 X2 TL-56, either a "Trinidad" or a "Tyler" die. Which according to Wikipidea both are 800Mhz Hyper Transport and S1g1 socket, just like the laptop. And of course it would not POST with that either, so it's most likely not the S1g2 socket issue. Since both the Tyler and Trinidad are S1g1 CPU's.

"It probabbly has to do with BIOS constraint than anything else"

That's about what I figured too. Does any one know if there already is a way to patch or hack the BIOS for laptops with problems like this? Or is it basically impossible? Or do you think maybe I could drop the Aspire 5100 BIOS in to it? I figure this has a chance of working since it is only the very next step up from the 3100 and has a dual core by default. I have seen both models up close side by side and they are virtually identical. I think the only difference between the 5100 and the 3100 is that the 5100 has more RAM, a DVD burner, a slightly better video card, and of course dual core CPU. So as long as there are no unit specific variables stored in the flash chip and it's 99.999% the same as the 3100 BIOS it should work IMO. Has any one ever herd of something like this working? Maybe I should do a hex comparison of the two BIOS ROM files....hummm....

Any way, thanks for the opinion.

I'm fully aware that the Lion die is S1g2. I mentioned the S1g1 vs S1g2 issue several times in fact and read up on all things I could find revolving around this issue... believe me. Also, I edited my first post a while ago showing that with a "Turion 64 X2 TL-56" installed it also did not want to POST (which is a S1g1 CPU). It's entirely possible that it was a bum CPU though, since it was pulled from a different laptop. If I can I'll try to get my hands on an other CPU that fits the bill and try that and see what happens.

On that note, you could be a bit more specific as to why it's incompatible. Saying "AS3100 only supports S1G1" doesn't really tell me anything I didn't already know. You could take a bit more time to explain your self. For example, from what I have recently gathered the different S1 sockets *might* have different pinout architectures? Is this why your saying it won't work? If so that would be useful information to know considering I would think it's blatantly obvious it wouldn't work. Though it would be nice to know the pinouts for sure. It's not exactly a for certain "they *DO* use different pinouts always." But that detail is not your fault.

Incidentally, I think it's really irresponsible of AMD and "friends" to make a system so ambiguous that also is intended to be upgrade friendly. They should do like Intel and change pins around just enough to not fit. Or they could do something like put a f%#$ing sticker on the socket and CPU that tells you it won't work. I mean, I'm reading that some people have actually had the CPU power section of there laptop blow up in a big flash of electricity just because they upgraded to a CPU that *PHYSICALLY FITS AND IS NOT MARKED AS INCOMPATIBLE*. That's heading for a lawsuit IMO. Some one could burn their house down. It's common knowledge in all PC repair literature as well as human nature in general that... "if it can physically fit, people will try it". I can't imagine AMD doesn't know this.

Anyway thanks for trying, I think I got the information I need already. I'm unsubscribed. See you all some other day maybe.