Hello!! Today I managed to raise the Mac Mini G4 10.2 from 1.5GHz to 2GHz WITHOUT CPU SWAP.
This Mac Mini model comes with a more modern Freescale CPU. It is capable of reaching much higher speeds, but as I already explained when I performed the CPU SWAP on a Mini 1.25GHz, the impediment is the maximum current limit of the motherboard.
In my blog I explained how to increase the voltage and current limit of the Mac Mini. A 330k resistor must be added IN PARALLEL to R5 to increase +0.05V the Vcore, and another 330k resistor to R3 to slightly increase the maximum current allowed to the CPU. These resistors act on the MAX1993 chip. I would not have discovered this limitation without @herd's instructions. You have the details here.
Once this is done, you just have to change the PLLs that adjust the speed of the CPU. Like everyone else, until now I did it by adding or removing resistors from R52, R362, R355, R351 and R358, but @herd told me that this is not the correct way, as can be read in this fragment of the datasheet, it must be indicated correctly the pull-up or pull-down, and in this way, the adjustment was "in the air" and any variation in tension could cause instability.
![PLLpullUpDn.png PLLpullUpDn.png](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72306-334cff46c9b9e7d177e90b2b38f553a5.jpg)
The correct table is this one, which Herd provided me, where it indicates that when removing one resistor, another must be added to its opposite one, and vice versa.
![Mac Mini G4 overclock table Herd.png Mac Mini G4 overclock table Herd.png](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72307-044b7e0dd07e85915141d8a4deb61851.jpg)
Once set to 2GHz speed, the Mac Mini started up and ran stable. I have passed PowerFractal and GeekBench tests, which give a result of 1043 points under 10.4.11 (My Mac Mini 10.1 @2GHz under 10.5.9 gives 1143 points)
![IMG_20240408_124523.jpg IMG_20240408_124523.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72308-7af4c82b1f4cec51f9fb832b3d0bd31e.jpg)
![IMG_20240408_125953.jpg IMG_20240408_125953.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72314-b3c4641a4347111159fbd8daffe4f757.jpg)
![IMG_20240408_130002.jpg IMG_20240408_130002.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72315-2c18191cbc35989e396b77394361be09.jpg)
![IMG_20240408_130047.jpg IMG_20240408_130047.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72312-d7fbd2555fcd4d4102c465dfd365e6ad.jpg)
![IMG_20240408_130706.jpg IMG_20240408_130706.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72313-d731ba95848f212d7fe404479c9596e4.jpg)
As a curiosity, in R50, R49, R45, R44 and R47 instead of 1K, Apple has used 10K, but I have tried both 1k and 0k and the result is the same.
I have been a bit brief with the entire explanation, I am sure that @herd can add more specific data, since he knows these processors much better.
.
This Mac Mini model comes with a more modern Freescale CPU. It is capable of reaching much higher speeds, but as I already explained when I performed the CPU SWAP on a Mini 1.25GHz, the impediment is the maximum current limit of the motherboard.
In my blog I explained how to increase the voltage and current limit of the Mac Mini. A 330k resistor must be added IN PARALLEL to R5 to increase +0.05V the Vcore, and another 330k resistor to R3 to slightly increase the maximum current allowed to the CPU. These resistors act on the MAX1993 chip. I would not have discovered this limitation without @herd's instructions. You have the details here.
![MAX1993.png](/bb/proxy.php?image=https%3A%2F%2Fblogger.googleusercontent.com%2Fimg%2Fb%2FR29vZ2xl%2FAVvXsEhduIDzWxelu5s1fapjuc4aWGxFeL6djr5KtLcTJShvPHXDWMnEwD4JskiXy072b77FzwzdOjOVEsc3qU4LHckkeX-EN_MQRxCC3fumoSVBaUw_Jrj8oFfhNsN-l2pGryP_ifXHQhB3vBzVCF3xZmPJ3qIpn3aXgw2m1F-JDt6uJrEWed183yoU_D9uDg%2Fw640-h624%2FMAX1993.png&hash=69e2b82931a6ba0edca881120a0fcb36)
Once this is done, you just have to change the PLLs that adjust the speed of the CPU. Like everyone else, until now I did it by adding or removing resistors from R52, R362, R355, R351 and R358, but @herd told me that this is not the correct way, as can be read in this fragment of the datasheet, it must be indicated correctly the pull-up or pull-down, and in this way, the adjustment was "in the air" and any variation in tension could cause instability.
![PLLpullUpDn.png PLLpullUpDn.png](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72306-334cff46c9b9e7d177e90b2b38f553a5.jpg)
The correct table is this one, which Herd provided me, where it indicates that when removing one resistor, another must be added to its opposite one, and vice versa.
![Mac Mini G4 overclock table Herd.png Mac Mini G4 overclock table Herd.png](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72307-044b7e0dd07e85915141d8a4deb61851.jpg)
Once set to 2GHz speed, the Mac Mini started up and ran stable. I have passed PowerFractal and GeekBench tests, which give a result of 1043 points under 10.4.11 (My Mac Mini 10.1 @2GHz under 10.5.9 gives 1143 points)
![IMG_20240408_124523.jpg IMG_20240408_124523.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72308-7af4c82b1f4cec51f9fb832b3d0bd31e.jpg)
![IMG_20240408_125953.jpg IMG_20240408_125953.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72314-b3c4641a4347111159fbd8daffe4f757.jpg)
![IMG_20240408_130002.jpg IMG_20240408_130002.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72315-2c18191cbc35989e396b77394361be09.jpg)
![IMG_20240408_130047.jpg IMG_20240408_130047.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72312-d7fbd2555fcd4d4102c465dfd365e6ad.jpg)
![IMG_20240408_130706.jpg IMG_20240408_130706.jpg](https://forumbucket.us-southeast-1.linodeobjects.com/data/attachments/72/72313-d731ba95848f212d7fe404479c9596e4.jpg)
As a curiosity, in R50, R49, R45, R44 and R47 instead of 1K, Apple has used 10K, but I have tried both 1k and 0k and the result is the same.
I have been a bit brief with the entire explanation, I am sure that @herd can add more specific data, since he knows these processors much better.
.
Attachments
Last edited: