Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

EddieK420

macrumors member
Jan 28, 2009
86
62
Hey guys. Thank you for all your help. I just went ahead and did the Pixlas mod using the open ended cable from moddiy and some splice taps. Tapped into the main power cable from the PSU. Working great now does not shut down in Cuda-Z heavy test mode anymore. And doesn't shut down in Resident Evil 7 or For Honor in Windows 10 anymore either. So let's see. Hope this is safe!
 

itdk92

macrumors 6502a
Nov 14, 2016
504
180
Copenhagen, Denmark
Just bumped into this thread by mistake, but I was so surprised by how people are recommending a possibly not safe way of doing stuff (read: abuse the 2 x mini 6 pins), over other proven and safer ways (draw additional power from SATA ports or through Pixlas mod), just to prove their point(?)

I must be missing something :confused:
 

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
Just bumped into this thread by mistake, but I was so surprised by how people are recommending a possibly not safe way of doing stuff (read: abuse the 2 x mini 6 pins), over other proven and safer ways (draw additional power from SATA ports or through Pixlas mod), just to prove their point(?)

I must be missing something :confused:

Drawing a bit more then 75W from each mini 6pin is also proven a safe way. There are lots of users doing that for years already. Which is enough data to shows that the actual risk is low.

Also, I don't think its risk is higher than Pixlas mod, which is a hardware mod, no one know the actual limit, and only exist for a relatively short period of time.

The safest way never change. Get a card that fit in the 225W limit.

Use the SATA ports should be a safer option, but some people may really need all SATA ports for HDD, and really prefer a high TDP card without extra PSU.

So, drawing a bit more from the mini 6pin is always an option.

If draw >75W from the mini 6pin is potentially not safe. Pixlas mod also clearly potentially not safe. And using 2xSATA port to feed a 6pin is just theoretically correct, but none of the official documents from Apple, or any GPU manufacture says thats safe to do so. Which also means potentially not safe.

In fact, all 3 methods are not the official way to power a GPU. All of them subject to some kind of risk. It depends on how the users apply these work around.

e.g. use mini 6pin only to power a heavily OCed FuryX may shut down the Mac Pro straight away. Which is actually a protection. But using Pixels mod to power the same FuryX, the Mac won't shutdown because of no protection avail. But may end up burn something(e.g. some wire), or overload the PSU in some extreme condition. Which way is safer?

The SATA port can deliver ~55W from 3 of the power line, but they may not good for GPU draw because GPU often have load spikes, which may significant enough to kill a SATA ports. And the 6pin is designed to handle fluctuate power demand (including load spikes), which way is safer if all we need is just extra few watts from the 6pin? (e.g. a 235W card).

The Pixels mod able to power the most demanding card, which should fully release the 980W PSU's potential, but it's never "safe". It's a hardware mod, possible tap to the wrong line during mod. Or short circuit etc.

Drawing >75W from the mini 6pin also never safe. Which is clearly outside the official limit. But no hardware mod require, and no need to occupy any SATA port.

Using a SATA port to power a GPU also never safe. That's a port for SATA device, GPU simply not belongs to that group. But technically all power draw are within limit.

So, I will say, not matter which way, all of them are unsafe (officially). But all 3 methods are proven work as expected if the user do it correctly. All of them has their own advantage and risk. It's just the users choice.
 
Last edited:

itdk92

macrumors 6502a
Nov 14, 2016
504
180
Copenhagen, Denmark
I really do not understand.

Before you say
Drawing a bit more then 75W from each mini 6pin is also proven a safe way. ...

but then you (correctly) state
Drawing >75W from the mini 6pin also never safe. Which is clearly outside the official limit.

So which one is it? :)

If Pixlas' mod is carried out correctly, with the correct cables, exc it is a MUCH safer way than try to overload the mini6 pins of the motherboard until one finds the power limit.

If then we take into account extreme situations (e.g. a Fury X running furmark), well then obviously either way might still not be 100% safe. But Pixlas Mod would still be safer (if one made some basic math about the power consumption of his cMP.)



Funnily enough, I read a thread not long ago, of a person having a Titan Xp shutting down his/her mac.

And then it turns out he/she was running it with the mini 6 pins(!)

So obviously it shut down.


I don't think I really understand how people can't see the risks of powering a TDP>225W graphic card with the mini 6 pins - and on top of that, there are other people actually starting threads misleading others into thinking it's OKAY.

Let me repeat (for the others, I know you know), it's NOT safe, and not okay. The fact some cMP user haven't experienced shutdowns does not make something unsafe, safe.


As I see it, only Pixlas mod or an external PSU are the solutions to 1-2 GPUs with TDP>225W.
 
  • Like
Reactions: Dr. Stealth

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
I really do not understand.

Before you say

but then you (correctly) state

So which one is it? :)

If Pixlas' mod is carried out correctly, with the correct cables, exc it is a MUCH safer way than try to overload the mini6 pins of the motherboard until one finds the power limit.

If then we take into account extreme situations (e.g. a Fury X running furmark), well then obviously either way might still not be 100% safe. But Pixlas Mod would still be safer (if one made some basic math about the power consumption of his cMP.)



Funnily enough, I read a thread not long ago, of a person having a Titan Xp shutting down his/her mac.

And then it turns out he/she was running it with the mini 6 pins(!)

So obviously it shut down.


I don't think I really understand how people can't see the risks of powering a TDP>225W graphic card with the mini 6 pins - and on top of that, there are other people actually starting threads misleading others into thinking it's OKAY.

Let me repeat (for the others, I know you know), it's NOT safe, and not okay. The fact some cMP user haven't experienced shutdowns does not make something unsafe, safe.


As I see it, only Pixlas mod or an external PSU are the solutions to 1-2 GPUs with TDP>225W.

What is the official correct way to carry out the Pixlas mod? NONE!

Draw more from the 6 pin if OFFICALLY outside limit. But REAL WORLD proven can work flawlessly. (Same as Pixlas mod, if carry out correctly! e.g. Only draw <100W on any 6pin, but not go >120W)

Overload directly from the 6pin port is unsafe? But overload the wire behind the 6pin port is safe? They are still the same wires! Apple never ever OFFICALLY says where is the 75W bottleneck coming from. How do you know it's not the wire?

A TitanXP can shutdown the Mac, because there is a shutdown PROTECTION! Running the GPU without protection is safer? The shutdown proved 2 things.

1) the GPU draw too much
2) the protection works

Think this way, Intel CPU has thermal shutdown protection. If the CPU overhear, the computer shutdown. Is it safer to disable this protection (assume we can) and let the CPU keep running even overheat? Please think about what is "protection". Plxlas mod effectively bypass the shutdown protection only gives out more flexibility, but not make it safer. It's the other side indeed.

You put a GPU that outside the officials limit is NEVER safe. Bypass the protection won't make it safer.

My point is, NONE of those 3 methods are safe OFFICALLY, but all the them are proven safe to use in REAL WORLD (fact). You can pick the Pixlas mod and believe that's the safest. But that subjective (opinion). There is simply zero evidence to prove that any of these 3 work around is safer than the others objectively (fact).

May I make it clear. I am not arguing that overload the mini 6 pin is safer than Pixlas mod. But it work (if do it correctly). However. Pixlas mod is also not safer than overload the mini 6pin. But it also work (if do it correctly).

Either options has their own limit and concern. And ALL of them are not safe officially. NONE of them are OK officially.

Is delid safer than use washer when upgrade the dual processors 4,1 CPU. For the socket, delid is safer. For the CPU, washer is safer. That's is, none of them are officially safe.

We discuss lots of unofficial hardware upgrade / mod in this forum. And most of them are proven OK in real world. Is that means safe? Obviously NO (officially). But it's up to the user to accept the risk of not.

Is dual X5690 safe? NO (officially), but we know it works (real world proven).

Overload the 6pin is not good to the port, or to the logic board trace. But won't physically damage the wire (from PSU). And no need to touch the PSU.

Pixlas mod will physically damage the wire (just under control, it actually cuts the wire a bit). Or required soldering on the PSU. But safe for the logic board trace (simply don't use it).

Protential physical damage to the wire or PSU is safer than protential overload the port or trace? They simply cannot be compared directly!

It's just the users opinion to decide which one is safer (that's the users personal opinion, but not a fact).
 
Last edited:

itdk92

macrumors 6502a
Nov 14, 2016
504
180
Copenhagen, Denmark
I never talked about "official"? Obviously nothing we have done with these cMP has been official in a long time :D

Draw more from the 6 pin if OFFICALLY outside limit. But REAL WORLD proven can work flawlessly
..

Well, beside for those for which it didn't work (which are becoming more and more)?

Have you ever heard a report of a user who correctly installed the mod (read: following the macrumors forums' thread)?



My conclusion (<-- yes, it's an opinion):

I would never, ever, recommend to draw 150W (8 Pin) from a cMP mini 6 pin, because (we can all agree) it's certain that it's bad, even though for some users worked fine <-- This does not make it neither flawless nor safe

I will always recommend Pixlas mod or an external PSU, because even if the first requires a hardware mod, and tapping power directly from the PSU, it bypasses the logic board, and taps into unused power - hence if used thoughtfully (thinking about what other components one has, and what GPU(s) to mount) it will provide power enough.


ps. - I was obviously referring to the thread starter when I wrote about spreading misinformation
ps. 2 - I thought you were a fan of Pixlas mod, as a better way to power manage these cMPs? :)
 

Prince134

macrumors 6502
Aug 17, 2010
338
153
I really do not understand.

Before you say

but then you (correctly) state

So which one is it? :)

If Pixlas' mod is carried out correctly, with the correct cables, exc it is a MUCH safer way than try to overload the mini6 pins of the motherboard until one finds the power limit.

If then we take into account extreme situations (e.g. a Fury X running furmark), well then obviously either way might still not be 100% safe. But Pixlas Mod would still be safer (if one made some basic math about the power consumption of his cMP.)



Funnily enough, I read a thread not long ago, of a person having a Titan Xp shutting down his/her mac.

And then it turns out he/she was running it with the mini 6 pins(!)

So obviously it shut down.


I don't think I really understand how people can't see the risks of powering a TDP>225W graphic card with the mini 6 pins - and on top of that, there are other people actually starting threads misleading others into thinking it's OKAY.

Let me repeat (for the others, I know you know), it's NOT safe, and not okay. The fact some cMP user haven't experienced shutdowns does not make something unsafe, safe.


As I see it, only Pixlas mod or an external PSU are the solutions to 1-2 GPUs with TDP>225W.


All of our folks here are simply sharing their opinions, which are only trying to help some one. Pixlas mod works because it by pass the backplane's protection. Don't forget why backplane needs protection? Well because each of 6 pin trace on PCB has its limit . However, why both 6 pin also can work? Answer is if you know what balanced drawing means. Look at the product EVGA Powerlink, it can help you balance the drawing from either 6 pin over loading and therefore works in most of case.

No one in their mind would run furmark overnight without thinking that can shortening whole system's lifespan. I wouldn't run furmark since why I need to? if I can play BF one for as much longer as I want.

Pixlas mod not necessary the safest, even though it draws current from the 12V source bypassing the mainboard. The contacts may not be clean and if consider in some cases that 1080 Ti shot off still can occur since not all GPU are born equal and GPU PCBs have their own protections as well. If Pixlas contacts are too thin, it still can't handle sudden demand.

Since you mentioned about Titan xP shot off, let me tell you what I found: The GP102 die (titan x, xP and 1080 Ti) is larger, and the memory being 11-12GB, so of course draw more currents comparing to other Pascals. Since all dies are different, plus the variation of memory chips and the parts on pcb of the GPU, some of these GP102 cards would shot off suddenly even before it starts load. In my experience I see this happened when running Unigene Superposition while running 8K. In this case because it allocates more memory than 4k, 2.5K, etc, voltage dropped so quick and really you haven't seen its on load. This has something to do with the GTX circuit design and also requires some (bad) luck, ie, either Nvidia change design or cherry picking can avoid this. (if you knew CPU overclocking you know what it is). But a soldering of a pair of 12V and GND from the PSU ultimately solve all these issues that happened to me.

Again, all these discussion all based on the fact that the 980W PSU is very reliable. With a pair of 12V wire, you can run two Titan X. From my wattage meter reading with two Titan X, 2 X5690 with 96GB Ram running Octane render, total consumption is around 500W. Running Battlefield one is the same since one GPU has to be idle given no SLI. Individually, running Battlefield One draws more current than octane render due to more CPU involved. But all in all, the PSU certainly gives you all you wanted.
 
Last edited:
  • Like
Reactions: itdk92

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
I never talked about "official"? Obviously nothing we have done with these cMP has been official in a long time :D



Well, beside for those for which it didn't work (which are becoming more and more)?

Have you ever heard a report of a user who correctly installed the mod (read: following the macrumors forums' thread)?



My conclusion (<-- yes, it's an opinion):

I would never, ever, recommend to draw 150W (8 Pin) from a cMP mini 6 pin, because (we can all agree) it's certain that it's bad, even though for some users worked fine <-- This does not make it neither flawless nor safe

I will always recommend Pixlas mod or an external PSU, because even if the first requires a hardware mod, and tapping power directly from the PSU, it bypasses the logic board, and taps into unused power - hence if used thoughtfully (thinking about what other components one has, and what GPU(s) to mount) it will provide power enough.


ps. - I was obviously referring to the thread starter when I wrote about spreading misinformation
ps. 2 - I thought you were a fan of Pixlas mod, as a better way to power manage these cMPs? :)

I also never ever recommend anyone to draw 150W from the mini 6pin. In fact, I highly against this idea. My real world test shows the shutdown protection kicks in a bit above 120W (fact). I personally treat it's "safe" to draw about 100W (opinion).

Well, if you don't care about the "official" limit, why care about that 75W? 75 is just the official limit, not the real world limit of the mini 6pin. If consider drawing more than 75 is unsafe, the only reason is because that exceed the official limit. Draw more than 75 won't shutdown the Mac straight away, if you treat this shutdown protection is "unsafe", then the "unsafe threshold" is 120W, not 75W.

If you care that 75, you are talking about the official limit.

So, if you only care about the real world limit, 75 is nothing, and drawing 100W from the mini 6pin is "safe". So do the Pixlas mod.

If you care that 75, Pixlas mod is also not safe, because that's not safe officially.

Yes, I am a fan of Pixlas mod. But I told myself that I must use the same standard to compare it to overload 6pin. If the standard is real world limited, then both options are safe (if do it correctly). If the standard is official limited, both are unsafe (no matter how we do it).

For any high power GPU, e.g. 275W and above. I will recommend the Pixlas mod more then any other method. But for anything below 275W, I personally tends to recommend overload the 6pin. HOWEVER, MUST balance the load between the 6pins (e.g. Via a EVGA power link). The idea is to keep the power draw below 100W on each 6pin, and never ever let it touch 120W. In this case, the cMP will never shutdown. In other word, it's also a good way to power the card (same as Pixlas mod), but require less work.
 
Last edited:

l008com

macrumors regular
Jan 20, 2004
112
4
Lots of useful info here, but it's all on macpro4,1 and macpro5,1. I'm trying to find specs on first gen macpro1,1s? I'm trying to run a pair of 8 pin cards that draw right around 200 watts each. Over the 75w plus 75w limit, but well under if first Gen Mac Pros also have 150 watt mini-6pin adapters.

Also as a failsafe/plan B, what about combining a mini-6pin plus one of the optical molex connectors to get an 8-pin connector? I don't know what the limits are on any of these connectors :/
 

Frixo Cool

macrumors regular
Jun 10, 2004
123
25
Croatia
Just a short note that you can run Sapphire Radeon RX 580 Nitro+ Special Edition (the overclocked one) on a regular unmodified Mac Pro.

For reliable operation you need to completely ignore 6-pin input on the card and use only 8-pin with a splitter to 2 x 6-pin on the motherboard.

I measured the power consumption with Marcel Bresink's Hardware Monitor and the card is idling at only 7.4W on PCIe slot and around 10 - 15W per PCIe AUX lines A and B.

Then, I pushed the card very hard with the FurMark test and it passed the benchmark flawlessly. The consumption was 85-95W at AUX lines and 50-65W on PCIe slot. So, everything seems within specifications that we expect and I guess it's reasonably safe.
 

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
Just a short note that you can run Sapphire Radeon RX 580 Nitro+ Special Edition (the overclocked one) on a regular unmodified Mac Pro.

For reliable operation you need to completely ignore 6-pin input on the card and use only 8-pin with a splitter to 2 x 6-pin on the motherboard.

I measured the power consumption with Marcel Bresink's Hardware Monitor and the card is idling at only 7.4W on PCIe slot and around 10 - 15W per PCIe AUX lines A and B.

Then, I pushed the card very hard with the FurMark test and it passed the benchmark flawlessly. The consumption was 85-95W at AUX lines and 50-65W on PCIe slot. So, everything seems within specifications that we expect and I guess it's reasonably safe.

Sounds good, but why not further split the 8pin to dual 6+2? So that no need to leave the 6pin input useless.
E1A9551D-87F8-4937-8E8D-1BB5525FC14D.jpeg


The card won’t draw extra 75w. With still at least 55W “safe” power available from the mini 6pins. I personally believe it’s better to connect the other 6pin as well.
 

MIKX

macrumors 68000
Dec 16, 2004
1,815
691
Japan
Wow ! all of the post make m think that 4,1 & 5,1 owners will have to compromise between 6 core CPU & newer GPU's TOTAL power requirements under stress.

I may be playing it on the safe side but I'm going for 2 X 6 core X5670's @ 2.93Ghz and maybe a HD 7970 O/C ( or R9 280X ) at 5.0 GT/s link speed. Enough for my current to semi-future needs ( I'm not as young as I used to be ).

I've read that anything higher than X5670 really ups the power requirements + overall stress on the machine
 

Frixo Cool

macrumors regular
Jun 10, 2004
123
25
Croatia
Wow ! all of the post make m think that 4,1 & 5,1 owners will have to compromise between 6 core CPU & newer GPU's TOTAL power requirements under stress.

I may be playing it on the safe side but I'm going for 2 X 6 core X5670's @ 2.93Ghz and maybe a HD 7970 O/C ( or R9 280X ) at 5.0 GT/s link speed. Enough for my current to semi-future needs ( I'm not as young as I used to be ).

I've read that anything higher than X5670 really ups the power requirements + overall stress on the machine

No, not true. Go all the way with the fastest processor you can get. For long time I use 12-core 3.46 Xeon 5690 setup and I have 4 HDDs + SSD. I pushed it as hard as I could and the maximum power usage was 498W. That's very far from 980W that power supply can handle.
 

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
Wow ! all of the post make m think that 4,1 & 5,1 owners will have to compromise between 6 core CPU & newer GPU's TOTAL power requirements under stress.

I may be playing it on the safe side but I'm going for 2 X 6 core X5670's @ 2.93Ghz and maybe a HD 7970 O/C ( or R9 280X ) at 5.0 GT/s link speed. Enough for my current to semi-future needs ( I'm not as young as I used to be ).

I've read that anything higher than X5670 really ups the power requirements + overall stress on the machine

Not really. AFAIK, the theory is "because Apple never ship dual 130W CPU 5,1. Therefore, going beyond dual X5675 (95W) is unsafe (or over stress the Mac)".

I can't say this is wrong. I have no evidence that 130W CPU won't cause any extra ware to the cMP. I don't even have any official document to indicate that the cMP's socket is 100% standard (especially about the power handling characteristic).

But I personally prefer to believe dual 130W CPU + single powerful GPU is fine. The fact is there are so many members at here that did this for years already. Not a single case have evidence that cause any damage.

So, let's do some calculation.

The PSU is rated 980W. Even though both CPU cost 260W. There are still 620W available.

If installed a "outside limit GPU" (like my 250W 1080Ti). There are still 370W remain.

A typical WD Black HDD only draw ~10W max. So, lets say 4 HDDs draw 50W. Then 320W remaining.

Since Apple said a single core 5,1's idle power consumption is 125W (this included the CPU, GPU, and HDD). So, let's say the logic board itself cost 100W. Then 220W remaining.

Let's assume each DIMM can draw extra 10W when under stress, and all 8 slots occupied. Then still has 140W available.

To further stress the cMP. We installed a 4 ports USB 3.0 PCIe card, and each port able to deliver 2A (10W). Then another 40W is used. Still has 100W remaining.

TBH, it's quite hard to use up all power available on the cMP.

The above estimations are quite on the safe side already. e.g. a CPU rarely really pull 130W even under full stress.

So, unless intentionally power the cMP in a hot room to reduce the PSU output, and then fully stress all CPU, GPU, HDD, DIMM, USB ports.... etc at the same time. I don't think we can really reach the cMP PSU's limit for most normal high demand usage.
 
  • Like
Reactions: Frixo Cool

Surrat

macrumors 6502
Jun 20, 2014
478
171
United States
You can go higher than 2.93ghz 6c and stay within the 95 watt original limits.
The x5675 3.06ghz 6c is 95 watts, as is the x5672 3.2ghz 4c processors.

I upgraded one of my flashed 2009 towers with a pair of x5672 that I delidded and they work great.
 
Last edited:

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
Just saw an interesting post on Facebook.

This guy power the PNY blower 1080Ti via a EVGA powerlink. And it can run CUDA-Z without shutdown.
31073341_10204503679388247_5668963729543266304_o.jpg


So, may be CUDA-Z isn't really that demanding, but can create a power spike that's strong enough to shut down the Mac.

May be it's because there are some capacitors inside the powerlink to "smoothen" the power draw.

If the theory is correct, I wonder if this kind of cables (with build in capacitor) can help as well.

https://www.btosinte.com/Seasonic-8...Cable-w-inline-capacitor-SS-IN-8-2x62PCIe.htm
 

MikkelAD

macrumors regular
Feb 17, 2018
188
33
Use the SATA ports should be a safer option, but some people may really need all SATA ports for HDD, and really prefer a high TDP card without extra PSU.

That's what I did... As you can see on the picture I finally got my hardware home

IMG_1583.JPG


The 2*mini 6-pin to 8-pin cable seems to be great and can power it (my GTX 970) so the lights and fans turn on but the 2*sata to 6-pin cable won't fit in my Mac Pro so did not get any image. I had to crack the "sides" on the sata connectors to get it connected to the motherbord (guess there is no other way since it's the power from the slots one needs and not the data part) ? ? ?

but......

That cable isn't long enough since it can't connect to the GPU despite being 20cm

I bought this one:

https://www.ebay.com/itm/20cm-Dual-SATA-15Pin-Male-M-to-PCI-e-6-Pin-Female-F-Video-Card-Power-Cable/162714342690?ssPageName=STRK:MEBIDX:IT&_trksid=p2057872.m2749.l2649

I am stuck :( :( :(

My plan was to power this card "the right way" by using two SATA ports but seems I need to take the road with mini 6-pin to 6-pin and mini 6-pin to 8-pin ? ? ?
 
  • Like
Reactions: MIKX

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
That's what I did... As you can see on the picture I finally got my hardware home

View attachment 760118

The 2*mini 6-pin to 8-pin cable seems to be great and can power it (my GTX 970) so the lights and fans turn on but the 2*sata to 6-pin cable won't fit in my Mac Pro so did not get any image. I had to crack the "sides" on the sata connectors to get it connected to the motherbord (guess there is no other way since it's the power from the slots one needs and not the data part) ? ? ?

but......

That cable isn't long enough since it can't connect to the GPU despite being 20cm

I bought this one:

https://www.ebay.com/itm/20cm-Dual-SATA-15Pin-Male-M-to-PCI-e-6-Pin-Female-F-Video-Card-Power-Cable/162714342690?ssPageName=STRK:MEBIDX:IT&_trksid=p2057872.m2749.l2649

I am stuck :( :( :(

My plan was to power this card "the right way" by using two SATA ports but seems I need to take the road with mini 6-pin to 6-pin and mini 6-pin to 8-pin ? ? ?

It's fine to use mini 6 -> 8pin to power a 6+8pin 970. The 970's TDP is quite low actually, nothing to worry about.

But if you want to do it "the right way", then why not get another cable? A longer dual SATA -> 6pin or a FEMALE 6pin to 6pin EXTENSION cable can also fix your problem.

In fact, for 970. I prefer to get another FEMALE 8pin -> Dual 6+2 pin. So that you don't need any SATA power.

I can using this "bridge method" to power a 1080Ti, for 970, this should be nothing.
 

MikkelAD

macrumors regular
Feb 17, 2018
188
33
Thanks for the answers!

I know GTX 970 (6-pin + 8-pin) isn't that demanding like some of the faster cards for the Mac Pro but I want to do it the best way possible and maybe be able to overclock it in windows (it's a fantastic card for that, many achieve stock GTX 980 speeds)

A longer dual SATA -> 6pin

But where? I have been looking all over. All cables are 15 to 20 cm long and all of them seems to have "clippers" on the sides so you have to break them of to get the cable connected on to the "wider" SATA ports on the Mac Pro motherboard...

FEMALE 6pin to 6pin EXTENSION cable can also fix your problem.

Should work too yes but it's about 3cm that is missing and all the extension cables I am able to find are like 40-50 cm and that's a whole lot of extra crap lying inside my beautiful baby :)

In fact, for 970. I prefer to get another FEMALE 8pin -> Dual 6+2 pin. So that you don't need any SATA power.

Please post links for all the cables you imagine I would have to use with this method. At this point I'm a bit confused...
[doublepost=1525110024][/doublepost]
Please post links for all the cables you imagine I would have to use with this method. At this point I'm a bit confused...

https://www.ebay.com/itm/PCI-E-8p-F...GPU-Video-Card-Power-Cable-30cm-/263422269544

So you want to connect this kind of cable to the 2*mini 6-pin --> 8-pin I already bought?

Will this method distribute power better than one mini 6-pin to 6-pin and one mini 6-pin to 8-pin ? ? ?
 

h9826790

macrumors P6
Apr 3, 2014
16,656
8,587
Hong Kong
Guess that will be my back-up plan!

No ideas for the other solutions? Am I completely retarded or isn't it really uphill to get the right cables like I described in the last post ? ? ?

It's your choice. I can't tell you where to get the longer cable. I can easily get those long enough cable from the computer shop which just 20min away from my place. So, no link can provide.

And if you don't want any "too long" cable inside the case as well. Then obviously it's impossible to buy a 3cm extension cable. However, it is very easy to mod the cable by yourselves to make it 3cm longer.

And yes, there is no "right connector" cable you can get on the SATA side. AFAIK, you have to remove the latch.

As I said, I personally prefer the "female 8pin -> dual 6+2pin" method. And you found the cable.
 

MikkelAD

macrumors regular
Feb 17, 2018
188
33
It's your choice. I can't tell you where to get the longer cable. I can easily get those long enough cable from the computer shop which just 20min away from my place. So, no link can provide.

Damn I live in the wrong country :) Just thought it would be easy to find on eBay/Amazon but it really isn't :(

And if you don't want any "too long" cable inside the case as well. Then obviously it's impossible to buy a 3cm extension cable. However, it is very easy to mod the cable by yourselves to make it 3cm longer.

Well a new kind of wire and that should be doable maybe...

And yes, there is no "right connector" cable you can get on the SATA side. AFAIK, you have to remove the latch.

By the latch you mean those two clippers (1 on each side of the SATA cable) right?

As I said, I personally prefer the "female 8pin -> dual 6+2pin" method. And you found the cable.

Yep looks like that will be the way to go most likely. Damn I hope my Gigabyte GTX 970 G1 will work probably now that it's not a reference GTX 970...

I read the post you linked earlier about your own GPU upgrade. Can you somehow explain to me why the power load will be more evenly distributed with the "bridge method ? ? ?

Again thanks for all the answers!


 
Last edited:

MikkelAD

macrumors regular
Feb 17, 2018
188
33
"h9826790" I always strive to learn a bit along the way but if it's too complicated to explain:

"Why the power load will be more evenly distributed with the "bridge method"

Maybe you could post me a link for some kind of explanation? :)
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.