This week on Ebay: 1 gig MMC-2. A bit expensive but it would be fun to try. The lister says it is a genuine intel part. Two available. I was told that 850mhz was the max. This CPU looks slightly different than all of the other MMC-2s.

I saw this referred to over on Wim's BIOS forum. I've never heard of this 1000MHz MMC-2 module. It does not appear in any official Intel lists of their MMC-2 module specifications. But some people seem to think that it is a "real" Intel MMC-2 module that was custom-built specifically for some version of the Panasonic Toughbook. It also seems that it is slightly taller than the standard MMC-2 modules, but I can't confirm that, nor can I tell if this would be an issue installing it into a 600E/600X or 770X/770Z machine.

I too am curious about it...but not enough to lay out $150.

Phil.

I had conversed with Laptoplucknow via email a couple of days ago and during that exchange, found this in the archives:

http://forum.thinkpads.com/viewtopic.php?t=62945

It does look like somebody has extracted the original CPU off the MMC-2 board and installed a socket.

Thanks to Ray's advice i joined the forum.
Well it looks like someone could not resist, one is gone.
The person listing does not give a lot of info, and because i don't Ebay i can not
even read the full text of the Q/A about the CPU.
This processor has an extra notch on one side. If it is custom, what cpu card did they start with.
They usually post a screen capture of CPUZ with their listings but not for this one.
I checked out the link Ray sent me and i noticed that in the CPUZ screen shot they provide
it shows a BGA/microPGA and not an MMC-2.
Now a question: is there any use for the socket that the 56k modem was pulled from,
or a way to wire an extra usb port or lan connection to fill the hole left from the RJ11 jack? (600E)

This is the eBay auction: http://cgi.ebay.com/ws/eBayISAPI.dll?Vi ... 0314142013

And the following Q&A is currently posted:

This 1ghz MMC-2 isn't p'n'p... how to fasten the cooler? Ok, with the old plate and special glue... and modified cooler - or glue without plate - hoping the vent never dies...

It looks like normal MMC-2-board, reballed with 1ghz processor. Does anyone know the model/serialno. from this boards?


Edit:
I've got some pictures.

Great surprise! It's not a reballed MMC-2 board!

It has no Intel PBA-number.

The board has not the MMC-2-standard-layout. This ghz-board is larger (~5mm). The underside looks very different (other layout) to other MMC-2-boards. On top there is a 6-ways-DIP-switch, and it's a socket-cpu!!!... looks changeable! Both wasn't seen at 'real' MMC-2-boards.

Because it's larger, I think, it will not match in some machines (like 600x).

Uwe.

Itä's entirely possible that Intel manufactured some non-standard CPU's on special orders for OEMs and also that official Intel sSpec listing doesn't have many CPU's that exist in reality.

For example, try finding 900 MHZ Pentium III Xeon CPU's with 2MB cache. They are not listed but I have several of them working in servers (sSpecs SL4XY and SL4XZ).

Just in case that they will reappear on ebay and someone wants to try it:
I bought one of these and I can confirm that they work at 1 Ghz if you apply the speestep-mod (for mainboards supporting speedstep, e.g. 600X 650 Mhz, this is not needed, of course). The pictures of the speedstep-mod for the normal MMC-2-processors do not fit, but I have made pictures for this particular MMC-2-PCB, if someone ever needs them. The switches are for the voltage setting in the maximum performance mode. By default (if you set the switches as shown on the sticker) , the 1 Ghz CPU is undervolted to 1,6 Volt (rated for 1,7 Volt by Intel at 1 Ghz). But it is difficult to attach a cooler to the processor because there are no mounting points and it is ~2mm higher than normal MMC-2-processors.
With the new processor, the AC adapter and the battery have to be disconnected from the notebook once it was powered on because otherwise it would not POST properly when you power it on again after shutting it down but rebooting and standby always works (but in standby, the fan is not switched off any more but rotates at half speed).

Thanks so much for the follow-up and explanation regarding this special MMC-2 board.

Can I just double-check, are you doing all your work on your 770X? I wonder if the power adapter issue might be avoided by using a 600X, especially a 600X PIII 650...?

And can you re-explain what happens during the POST? It will boot up okay after various key presses or something, but before you reboot, you have to disconnect both the battery and the AC adapter? And these disconnections aren't necessary if you go to standby instead of shutting down. Is that right?

Phil.

The processor is used in my 770X using a selfmade heatpipe cooler. I already had a selfmade cooler for the 850 Mhz (there are photos in the 770 upgrade thread) but for the 1,08 Ghz I made a new one with a bigger surface. Using the IBM fan is impossible with the 1 Ghz µPGA2-socket processor (no mounting points, 2mm higher as usual MMC-2-PCBs, heat).

If I plug in the battery and/or the AC adapter and switch on, everything works as usual. I can reboot and go into standby + wake up again from standby and shut down with no problems.

But if I leave the battery and/or the AC adapter connected and power the unit on again (after it has shut down) it will never POST properly. Sometimes it will not switch on the fan and complain about it with error 192 (the fan works in general but is not switched on by the BIOS), most of the time it will not even show an error message but just try to power on again and again as if the BIOS is in a loop. The indicators on the LCD are blinking in an interval.

This does never (!) happen if I just unplug the battery and the AC adapter after it has shut down and after some seconds plug them in again and switch on. Then it will boot up fine everytime.

This is not a big issue to me because it is very easy to avoid the boot-up problem. And usually you do not shut down and power on again immediately. I think that there are some circuits on the new MMC-2-board which are not properly switched off after shutdown (so they stay on) and if the BIOS wants to initialize them again during POST, it fails because they are still on. This is only a guess but it would explain why it always works if there has not been any power source connected to the Thinkpad before switching it on.

It would definitely be interesting to know if this problem would also occur on the speedstep enabled 600X 650 Mhz.

Just a follow-up:
On a 770z mainboard, the socket adapter does not work, it has to be a mainboard from a 770x. It looks like there are some more differences between the 770x and the 770z mainboard besides the asseet-id function.