Nice, where did you get the power?
Short info: A is B and B is A. The rear is CPU A.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 1,1-5,1 Cold, Hot, or Just Right?
- Thread starter GiveYourOpinion
- Start date
- Sort by reaction score
Short info: A is B and B is A. The rear is CPU A.
Its not the heatsink that deforms, its what the N/B chip its self sits on, its not ali.
The power comes from SATA to a fan controler on side of my heatsink A i didnt lay it out as per Apple it was just my own layout to explain what happens.
Any increase in strength due to work hardening disappears at temperatures above 250ºС
There is no aluminum anywhere in the nortbridge encapsulation, it's an exposed die over a PCB substrate.
1. What would you say can be inferred if the temperature is exactly at the top, 75 Celsius, of the range you give?
2. Do you believe dry dust, including "dustbunnies", inside, say, a cMP, can be responsible for increasing such temperatures compared to a cMP that does not have the dust? (I don't mean "caked on" dust, or debris, that adheres; I mean "dry" dust that freely moves if sufficient "air flow" is directed onto it.)
3. Is there some manner of + or - margin of error value that can be assumed for the temperature readings from senosrs inside a cMP; particularly, considering they may be many years old?
Do you want to give the URL(s) to open source BOOTROM dumping tools you recommend that can be downloaded from, say, GITHUB?
I know there is not a shortage of such tools, but I tend to prefer using the ones that can be dowloaded from github, as opposed to only being available for download from the author's personal or business website.
Not really meaningful if you don't know the average ambient temperature, if it's measured while complete idle or after load, etc. 75ºC over Finland winter is overheating, while the same temperature in a tropical country summer is normal and expected.
Btw, I've taken a second look at your first post screenshots and you have a MacPro1,1/2,1? We are all talking about early-2009 to mid-2012 Mac Pros.
Yes, sure. Anything that blocks the heat transfer will affect the temperature. Sometimes it's so dusty inside that the air don't flow anymore.
I not a metrology nut, but I try to have consistent readings, my setup is a Fluke 189 + 80BK-A K-type thermocouple probe + IR189USB data logging cable. Infrared or low end laser thermometers usually don't work very well with reflective metals and readings of heat sinks are usually all over the place, specially for some of the MacPro5,1 northbridge heatsinks that have a almost chrome-like mirror surface finish treatment.
Mac Pro sensors are fairly precise, very little deviation in relation of my set-up, for all Mac Pros that I tested, MacPro4,1 on wards.
No need to worry about a MacPro1,1/2,1 BootROM, it's too old to have a SPI flash memory.
Waste of time and effort, but you can always compile flashrom 0.99 from source and use the internal programmer - flashrom 0.99 works with all MacPro's since at least the early-2008 Mac Pro - I've tested with MacPro3,1 to MacPro6,1.
GitHub - flashrom/flashrom: Send patches to https://review.coreboot.org: https://www.flashrom.org/Development_Guidelines#GitHub
Send patches to https://review.coreboot.org: https://www.flashrom.org/Development_Guidelines#GitHub - flashrom/flashrom
github.com
Last edited:
Ambient temperature 30 celsius, as in the image i posted when I started this thread.Not really meaningful if you don't know the ambient temperature, if it's measured while idle or after load, etc. 75ºC while on Finland winter is overheating, while the same temperature in a tropical country summer is normal and expected.
Btw, I've taken a second look at your first post screenshots and you have a MacPro1,1/2,1? We are all talling about early-2009 to mid-2012 Mac Pros.
Yes, sure. Anything that blocks the heat transfer will affect the temperature. Sometimes it's so dusty inside that the air don't flow anymore.
I not a metrology nut, but I try to have consistent readings, my setup is a Fluke 189 + 80BK-A K-type thermocouple probe + IR189USB data logging cable. Infrared or low end laser thermometers usually don't work very well with reflective metals and readings of heat sinks are usually all over the place, specially for some of the MacPro5,1 northbridge heatsinks that have a almost chrome-like mirror surface finish. Mac Pro sensors are fairly precise, very little deviation in relation of my set-up, for all Mac Pros that I tested, MacPro4,1 on wards.
Sorry, but why you are worried about a MP1,1/2,1 BootROM? It's too old to even have a SPI. Anyway, you can always compile flashrom 0.99 from source:
GitHub - flashrom/flashrom: Send patches to https://review.coreboot.org: https://www.flashrom.org/Development_Guidelines#GitHub
Send patches to https://review.coreboot.org: https://www.flashrom.org/Development_Guidelines#GitHub - flashrom/flashrom
tsialex said, "Sorry, but why you are worried about a MP1,1/2,1 BootROM?"
When I started this thread I tagged it MP 1,1 - 5,1.
I have, right now, within 20 meters of me, a number of Mac Pro(s) of different vintage, including 1,1 and 4,1/5,1 and even 5,1; not to mention 2e. Did I indicate I was referring, specifically, or exclusively, to a 1,1? I tagged the thread MP 1,1 - 5,1 when I started the thread.
Nontheless, thuough, I am concerend about lots of things someone else, including you, may not be concerned about; just as you may be concerned about many things I am may not even be aware of, or I am not concerned about..
I registerd this Macrumors forum account, and I started this thread to, hopefully, interact w/ people, including those with enthusiasm for maintaining cMP(s), among other things, that are also interested in sharing their knowledge. Some things I may already know a little bit about and some of them my knowledge may be at an enormous deficit compared to, for example, yours. Therfore, I typically ask questions on forums to, among other reasons, expand my knowledge base and interact w/ others who may be like minded.
I can not necessarily discern, at first glance, if someone, yourself for instance, may not be interested in sharing their knowledge w/ me, or answering questions I ask, anymore than I can read their minds. If I expressed interest in stuff you do not care about, or just do not want to discuss w/ me, then, I apologize for directing the associated questions in your direction, and, in that case, I will try to make an effort to avoid asking you anything else, directly.
Thx for responding, this time, and have a good year.
I looked at your screenshot while I was answering and noticed that it was from a MacPro1,1, anyway, you didn't wrote if it was at completely idle or after load, this matters greatly.
Seems you clearly didn't understood my answer, MacPro1,1 and MacPro2,1 have FWB memories for the BootROM storage and the supported OS X releases for these Macs use the NVRAM very sparingly, so it won't brick as a MacPro4,1 to 6,1 that have SPI flash memories that store both the BootROM and the NVRAM volume with just one memory and use the NVRAM constantly. MacPro3,1 also very rarely bricks since Apple used a different and one time only half way approach, with FWB memory for the BootROM and SPI for the NVRAM that can still boot even with a corrupted NVRAM volume.
Also, to avoid this type of thing in the future, make it clear about the MacPro model you are talking about - just tagging MP1,1 to 5,1 it's not enough to receive meaningful answers since MacPro1,1 to 3,1 are completely distinct internally from MacPro4,1 and MacPro5,1.
I don't understand why you posted this, I even edited my original post long before you replied (please check) to make it clear about not needing to worry with MacPro1,1/2,1 BootROMs.
You seem extraordinarily knowledgeable about Macs and there are, likely, many people on this site, and many people who may have never even visited this site, that are very appreciative for your contributions, for your help, and for your work.
Thx, again, for your responses.
I rarely have seen a thread in any forum where subsequent posts all adhere to the topic of the first post. This thread does not appear different, and that is not a problem, for me.
In the post where I quoted you and asked about open source BOOTROM dumping tools I did not say anything about a 1,1 Mac Pro; my original tag for this thread was MP 1,1-5,1. I was just asking if you were aware of open source BOOTROM dumping tools hosted on github.
You are not obligated to reply to me, at all, but it seemed as if you "read in" something I had not asked instead of, if you wanted to answer, just saying if you knew of bootrom dumping tools hosted on github, or not. In fact, I had my 4,1/5,1 Mac Pro's in mind, but that should not really matter; you seem very knowledgeable about Macs and I just thought you might be interested in answering the question I asked in that particular post.
Incidentally, I have some bootrom tools I have downloaded, but I have gotten malware alerts about them; just trying to find bootrom dumping tools that might function differently, better, or, and, not prompt malware alerts.
Ultimately, there is not a problem, though; thx, again, for your responses. I would not be surprised to hear you have helped lots of Mac owners who could not find any help, elsewhere.
Thx, again, for your responses.
I rarely have seen a thread in any forum where subsequent posts all adhere to the topic of the first post. This thread does not appear different, and that is not a problem, for me.
In the post where I quoted you and asked about open source BOOTROM dumping tools I did not say anything about a 1,1 Mac Pro; my original tag for this thread was MP 1,1-5,1. I was just asking if you were aware of open source BOOTROM dumping tools hosted on github.
You are not obligated to reply to me, at all, but it seemed as if you "read in" something I had not asked instead of, if you wanted to answer, just saying if you knew of bootrom dumping tools hosted on github, or not. In fact, I had my 4,1/5,1 Mac Pro's in mind, but that should not really matter; you seem very knowledgeable about Macs and I just thought you might be interested in answering the question I asked in that particular post.
Incidentally, I have some bootrom tools I have downloaded, but I have gotten malware alerts about them; just trying to find bootrom dumping tools that might function differently, better, or, and, not prompt malware alerts.
Ultimately, there is not a problem, though; thx, again, for your responses. I would not be surprised to hear you have helped lots of Mac owners who could not find any help, elsewhere.
You can use my dumper what does a combination of dumping / analyzing.
see https://forums.macrumors.com/thread...our-on-4-1-5-1-machines.2333460/post-31048702
Its not in Github but the AppleScript and ShellScript sources are not hidden.
it uses Flashrom and Syncretic Scanvss as external tools.
see https://forums.macrumors.com/thread...our-on-4-1-5-1-machines.2333460/post-31048702
Its not in Github but the AppleScript and ShellScript sources are not hidden.
it uses Flashrom and Syncretic Scanvss as external tools.
Last edited:
For those of you interested, this is my Northbridge chip after being subject to regular 82C for at least 3 years (that's how long I got this computer, perhaps it was longer.)
No warping, no deformation. Cleaned, applied Thermal Grizzly compound to both CPUs and NB, put back together.
No noticeable difference in temperatures on first boot, I know the thermal compound needs some time to properly settle though, so will check in a week.
BTW the spring rivets were in perfect shape. I assume someone was there before me, as this computer sports two X5690s.
No warping, no deformation. Cleaned, applied Thermal Grizzly compound to both CPUs and NB, put back together.
No noticeable difference in temperatures on first boot, I know the thermal compound needs some time to properly settle though, so will check in a week.
BTW the spring rivets were in perfect shape. I assume someone was there before me, as this computer sports two X5690s.
83°C is very hot. What thermal paste are used? How much did you used? how was the paste applied? What is the temperature in your room?
Temp difference between heatsink and IOH is also a bit to high. Normal is around 13 K your difference is 18 K.
Temp difference between heatsink and IOH is also a bit to high. Normal is around 13 K your difference is 18 K.
It's been always like this.
Thermal Grizzly Aeronaut.
Smooth spread technique for about 0.2 mm thickness.
The diff is 19 deg C now. Perhaps has something to do with the NB radiator temp probe not sticking to the radiator, it became separated on disassembly.
Thermal Grizzly Aeronaut.
Smooth spread technique for about 0.2 mm thickness.
The diff is 19 deg C now. Perhaps has something to do with the NB radiator temp probe not sticking to the radiator, it became separated on disassembly.
Aeronaut has only 8,5 W/mK. For the IOH I would use the best I can get. Kryonaut has almost twice the thermal conductivity of 14,2 W/mK.
I have made a small line (your picture) from left to right. The contact pressure does the rest.
I have also added a 1mm thermal pad (12,8 W/mK) on the ceramics of the chip.
I have made a small line (your picture) from left to right. The contact pressure does the rest.
I have also added a 1mm thermal pad (12,8 W/mK) on the ceramics of the chip.
Last edited:
Thermal Grizzly Aeronaut is not the best paste, it's name thermal grizzly sell's it. Kryonaut is much better but even that at extreme temps can separate in high temp tests and your N/B temps are high for some reason. Noctua NT-H2 is a very stable paste at high temps.
I would check the heatsink its self as they are not exactly well made, and yours might have a concave surface. the temp probes are just stuck on the heatsink, you could re apply thermal epoxy to re attach it. but IOH temps are really high 83c is very hot.
Thanks @TzunamiOSX & @Matty_TypeR this Kryonaut + thermal pad combo seems a good idea.
Using TGPro last night I turned on the boost A fan to the maximum and the NB diode temperature dropped to ~50C in no time. Thermal compound aside, I assume it may be some poor workmanship from my side.
Using TGPro last night I turned on the boost A fan to the maximum and the NB diode temperature dropped to ~50C in no time. Thermal compound aside, I assume it may be some poor workmanship from my side.
My IOH is around 70°C with a room temperature of 28°C (it is hot out there).
TGPro is set to
Intake/Exhaust around 1000 RPM on idle
Booster +8% (when IOH over 66°C = 11%)
TGPro is set to
Intake/Exhaust around 1000 RPM on idle
Booster +8% (when IOH over 66°C = 11%)
Last edited:
It's 27°C where I live and I set TGPro to keep the IOH diode temperature lower until my Kryonaut + stuff order arrives.
BTW if anybody has a spare PCI fan that is not equal in price to diamonds, I would gladly buy it (I'm in EU). My Noctua mod didn't went well, I kinda screwed up the connector.
BTW if anybody has a spare PCI fan that is not equal in price to diamonds, I would gladly buy it (I'm in EU). My Noctua mod didn't went well, I kinda screwed up the connector.
OK, an update is due.
I reinstalled the NB radiator getting rid of the original plastic fasteners and created a combination of M3 vinyl screws that I shortened a bit and original springs from the fasteners. From the bottom of the board the screws are fixed with vinyl nuts, with a non-conductive 2mm thick rubber washer between the nut and the actual circuit board.
The length of the screws was chosen by trial and error, I ended up with threads being 18 mm long (threads only, as the head dimensions may vary). This allows some vertical movement of the radiator, but the springs are about 50% less visible (50% more compressed) compared to the original setup with plastic rivets.
I chose Noctua NT-H1 that was readily available in my neighborhood hardware store and followed the Noctua application guide (single 3mm drop).
I used the thermal pads to follow the layout @TzunamiOSX pictured above. Here comes the warning: use thermal pads only if they are 1 mm thick - or less. I didn't do a detailed check while in store and bought 1.5mm thick pads instead. I warned you. Once I assembled everything back together (disassembly includes removing the board from the metal carrier, removing both CPU radiators, cleaning the old thermal compound from the CPUs, re-pasting the CPUs, the assembly is exactly as exciting) it only turned out the thermal pads I used were too thick and did not allow the desired contact between the NB chip and its radiator. When I started the computer, the X58 IOH diode rocketed to 95°C immediately.
Consumer advice break: use TGPro or any other fan control software that allows sensor-based fan speed policy to avoid a total disaster on your first startup. Yesterday before going to bed I set my BoostA to react to X58 IOH diode value, so fortunately once TGPro loaded it immediately started the fan on full speed. The temperature dropped to 85°C, so it was shutdown time and disassembly again.
Which is why you want to have the proper, thin thermal pads to not do things twice.
So I bought some 1mm pads, applied them and assembled everything together again. It's now running for 20 minutes (probably mostly idle, but my measurements of ~82°C from past years were with more or less same usage patterns) with following result:
The room temperature is currently 28.75°C (European heatwave, yikes!) so I believe this is a decent result. Also, as you see, the temperature indication difference between heatsink and diode also dropped to very acceptable 9 degrees.
Thanks guys for inspiring me to finally fix it.
I reinstalled the NB radiator getting rid of the original plastic fasteners and created a combination of M3 vinyl screws that I shortened a bit and original springs from the fasteners. From the bottom of the board the screws are fixed with vinyl nuts, with a non-conductive 2mm thick rubber washer between the nut and the actual circuit board.
The length of the screws was chosen by trial and error, I ended up with threads being 18 mm long (threads only, as the head dimensions may vary). This allows some vertical movement of the radiator, but the springs are about 50% less visible (50% more compressed) compared to the original setup with plastic rivets.
I chose Noctua NT-H1 that was readily available in my neighborhood hardware store and followed the Noctua application guide (single 3mm drop).
I used the thermal pads to follow the layout @TzunamiOSX pictured above. Here comes the warning: use thermal pads only if they are 1 mm thick - or less. I didn't do a detailed check while in store and bought 1.5mm thick pads instead. I warned you. Once I assembled everything back together (disassembly includes removing the board from the metal carrier, removing both CPU radiators, cleaning the old thermal compound from the CPUs, re-pasting the CPUs, the assembly is exactly as exciting) it only turned out the thermal pads I used were too thick and did not allow the desired contact between the NB chip and its radiator. When I started the computer, the X58 IOH diode rocketed to 95°C immediately.
Consumer advice break: use TGPro or any other fan control software that allows sensor-based fan speed policy to avoid a total disaster on your first startup. Yesterday before going to bed I set my BoostA to react to X58 IOH diode value, so fortunately once TGPro loaded it immediately started the fan on full speed. The temperature dropped to 85°C, so it was shutdown time and disassembly again.
Which is why you want to have the proper, thin thermal pads to not do things twice.
So I bought some 1mm pads, applied them and assembled everything together again. It's now running for 20 minutes (probably mostly idle, but my measurements of ~82°C from past years were with more or less same usage patterns) with following result:
The room temperature is currently 28.75°C (European heatwave, yikes!) so I believe this is a decent result. Also, as you see, the temperature indication difference between heatsink and diode also dropped to very acceptable 9 degrees.
Thanks guys for inspiring me to finally fix it.
At the moment my room temperature is 30°C an my IOH is over 74°C with TGPro it was possible to bring down the temp to around 72°C by speed up the booster to (I think it was) +19%.