• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

1.33 GHz Mac Mini

MBongo

Active member
A question for those that may have one of these specific, stock G4 Mac minis.

1.33 GHz Mac Mini ONLY

Q: Is there a resistor present at the R52 location (under the heatsink)?

BetterView.png
Resistor location red-dotted here on a 1.25 GHz Mac mini.


And if you’ve ever removed (or tried to remove) your G4 Mac mini’s heatsink, you might try this:

How.png
Easier heatsink removal, with a nylon spacer and steel pin from the backside.

 

herd

Well-known member
Wasn't the "1.33GHz Mini" the silent upgrade version of the low end model (1.25GHz)? R52 is just one of the PLL resistor inputs to the CPU. If you're trying to change the CPU speed, you can check the datasheet or one of the overclocking guides around. Would any Mini set for 1.33GHz be a "1.33GHz Mini?"

There was also a bump in the Model number from 10,1 to 10,2 but I don't know if there was much difference in hardware. I think the higher clocked versions may have used a different fan and/or heatsink. At least I've seen different versions and with a limited sample this seems to be a pattern, but it could be down to different suppliers. I would say that the 64MB (vs 32MB) of VRAM would be something to look for, because you can't change this by moving a resistor or two.

I can't see any pictures on that forum but I don't think measuring the air temperature is really telling you much. Warm air coming out the back just means that the fan and heatsink are working. For example, you could potentially jam the fan and the CPU would burn up with no air coming out the back. The CPU temperature is what you want to know. There is a built in temperature sensor, but it's not available to tools like Temperature Monitor. However, the fan is variable speed and the controller uses the sensor for feedback. So a simple thing to check is how fast the fan runs and how quickly the fan speeds up when the CPU is loaded. I got one Mini that would always run with a very fast fan. I opened it and found the heatsink and fan filled with cat hair. Once cleaned, the fan ran much slower (while the CPU may be be around the same temperature and the air coming out the back is likely hotter).
 
Last edited:

MBongo

Active member
Thanks @herd. Appreciate your insights and opinions - however I did ask specifically about “stock” 1.33 GHz G4 minis and not for any G4 Minis “set” or overclocked to 1.33 GHz. Perhaps additional clarification was necessary?

The 1.33 GHz model AND the 1.5 GHz are members of the 10,2 Family - while the 1.25 & 1.42 GHz were from the earlier 10,1 Family. The 1.5 GHz model featured the increased VRAM of 64 MB (over all others @ 32 MB VRAM).

TenOne&Two.png

And my inquiry here (and elsewhere) pertains primarily to many of those oveclocking guides that you mention and as to whether or not the stock 1.33 GHz version was originally produced without the R52 resistor... that those guides note need be removed to attain 1.33 GHz. Now, that might be because the guides apply primarily to the 10,1 family… but that is partially what I am also attempting to clarify. *One might note too that the settings for many of the overclock speeds above 1.58 GHz also note the removal of R52?

(Perhaps in a related vein, in order to downclock a 1.42 GHz mini to 1.33 GHz - would one need to remove the R52, R351 and R355?)

Inset images below are from a 1.25 GHz G4 Mac mini.

ReDone.png

Unfortunate that you can’t view any pictures on MacOS9Lives.com - but that is possibly, easily remedied by registering with the site much as one registers here or over on MacRumors, Tinker Different, etc. There is no fee.

In recent testing here, a distinct difference in air exhaust temperatures was discovered with TWO of seven minis. (All stock, none overclocked and including 1.25 GHz, 1.42 and 1.5 GHz. Sadly, no 1.33 GHz.) Keeping in mind that the TWO minis (a 1.42 GHz and a 1.5 Ghz) both still had the paper thin, black, original heatsink thermal pad in place. Eighteen (nineteen?) year-old machines that had never had those pads renewed, removed or thermal pasted. The five other minis here (all upon cold boot) registered initial exhaust temps of 84˚F in the first minute or so. After ten to 15 minutes those temps rose to and remained steady @ right around 100˚F (at idle). The TWO minis that still had their original black thermal “pads” quickly rose to and measured 107˚F to 112˚F …within the first 10 minutes, after dead cold boots.

Then, after removal of the original thermal pads and application of fresh thermal paste (Arctic Silver) those TWO minis registered exhaust temps just like the other five. Perhaps a valid question here is: what computer is expected to continue to run for 18 or 19 years without ever having had fresh thermal paste applied and/or thermal pads replaced?

Now granted, some may take a look at those spring-loaded heatsink retainer pins and opt out of proceeding any further. [ Pin removal is also covered here: http://macos9lives.com/smforum/index.php/topic,6875.0.html ] Alas, registration required to view images.

Top.png

“I got one Mini that would always run with a very fast fan. I opened it and found the heatsink and fan filled with cat hair. Once cleaned, the fan ran much slower (while the CPU may be be around the same temperature and the air coming out the back is likely hotter).” - herd

Was a new thermal heatsink pad or thermal paste applied?
Do you still have that machine and if so, what is its ”idle” exhaust temperature after a cold boot and 10-15 minutes, now?
And also granted, if your testing environment’s temperature is very high, temperature ranges may indeed vary somewhat.


Here are some temp readings from a very recently deceased (not mine) 1.42 GHz G4 Mac mini:
• 1 min. bottom: 30ºC / 86ºF
• 4 min. bottom: 38ºC / 100ºF
• 15 min. bottom: 41-42ºC, hottest spot under processor: 45ºC / 113ºF
• Hottest spot at exhaust: 47ºC / 117ºF

Ideally, knowing the actual CPU temp would be absolutely grand. But without that, exhaust air temperature readings are quite possibly the next best thing (and maybe the easiest to obtain). And I didn't base my temperature findings on just one mini. Seven various G4 minis were monitored. Seventeen degrees higher at idle, after 10-15 minutes runtime? That should tell you something on a bone-stock G4 mini.

Now again, anyone have a stock 1.33 GHz G4 Mac mini?

If anyone has a G4 Mac mini and they have never had the heatsink off to replace the thermal pad or re-paste it…
they may need to “take its' temperature” and remove, clean, re-pad or re-paste that heatsink?
 
Last edited:

Byrd

Well-known member
Do you usually post in a condescending tone? I’m presuming you need to double/triple post that information to other sites as people from one forum wouldn’t want to help you out after reading that.
 
Last edited:

max1zzz

Well-known member
I happen to have a 1.33Ghz mini sitting on my desk awaiting reassembly and can confirm R52 is not fitted

As a odd aside this mini is badged as a 1.25Ghz unit but is actually a 1.33Ghz one, I wonder if this was a case of apple using old stock shells or of this had a in warranty MLB replacement and got the faster board (It 100% hasn't been modified, I know where this specific mini came from and the original owner definitely did not overclock it)
 

Phipli

Well-known member
As a odd aside this mini is badged as a 1.25Ghz unit but is actually a 1.33Ghz one, I wonder if this was a case of apple using old stock shells or of this had a in warranty MLB replacement and got the faster board (It 100% hasn't been modified, I know where this specific mini came from and the original owner definitely did not overclock it)
All the 1.33 and 1.5GHz Mac Minis are like this aren't they? That's why we say silent upgrade. You only find out by looking in about this Mac.

Sadly never owned one, but my dad has a 1.5GHz.
 

herd

Well-known member
The G4 Mini uses a 7447 CPU and has a 167MHz bus speed. The CPU speed is a multiple of the bus speed, so any Mini with a 1.33GHz CPU will have the PLL resistors configured a certain way. Check the 7447 datasheet for possible PLL settings and the overclocking guides for which resistors correspond to the PLL inputs.

As I mentioned, the air coming out the back is not really a direct guide for the CPU temperature because the fan speed is variable (among other things). Try this simple experiment: measure the "exhaust" temperature with your setup. Open the Mini and cover roughly half of the CPU heatsink with a piece of tape. Close the Mini and repeat the measurement with NO other changes. What do you find?
 

Snial

Well-known member
It's odd it's not possible to read the PLL multiplier from software.

Hmm, it was possible to read the PLL multiplier on the 603e:

1.1.2.2.2 Clock Configuration Available in HID1 Register
Bits 0–3 in the new HID1 register (SPR 1009) provides software read-only access to the configuration of the PLL_CFG signals. The HID1 register is not implemented in the 603.

OK, HID1 also exists on the G4, page 2-7.

"
Hardware implementation-dependent registers. Control various functions, such as the power management features, and locking, enabling, and invalidating the instruction and data caches. The HID1 includes bits that reflects the state of PLL_CFG[0:4] (PLL_CFG[0:5] in the MPC7448) clock signals and control other bus-related functions.
See Table 2-46 for specific synchronization requirements on this register."


Yes, so mfspr 1009 can do it (is that privileged?)


mfspr HID1

I don't think you'd need sync instructions just to read it.

This tells you how to write a simple kext:


I think it prints out a message on boot somehow?

So, I guess from the example, doing

#include <sys/systm.h>
#include <mach/mach_types.h>

int32_t Hid1Get(void)
{
register int32_t hid1;
__asm("misprint 1009,%0" : "=r" hid1: : hid1); or whatever the read reg asm parameter in gcc is.
return hid1;
}

kern_return_t MyKext_start (kmod_info_t * ki, void * d)
{
printf("MyKext has started. Hid1=%d\n",Hid1Get());
return KERN_SUCCESS;
}


would do it.
 
Last edited:

herd

Well-known member
The chart posted above only shows half of the resistor positions for the Mini. For example, I think it would be technically more accurate to move R52, instead of disconnecting it.

The HID1 register can also be viewed with Reggie:

HID1.png
 

Snial

Well-known member
The chart posted above only shows half of the resistor positions for the Mini. For example, I think it would be technically more accurate to move R52, instead of disconnecting it.

The HID1 register can also be viewed with Reggie:
I've never heard of "Reggie" is it a 3rd party utility?
 
Top