Jump to content

US$189 AKiTiO Thunder2 PCIe Box (16Gbps-TB2)


Recommended Posts

I just soldered the J6 cable like the photos and tape A123 and B123 and disconnect J4 cable (yellow/black).

After connected the Akitio default PSU (for test) nothing came up no led no fan, did I do anything wrong? Or do I need to use the ATX PSU in order to use this method?

post-30795-14494999736512_thumb.jpg

post-30795-14494999735983_thumb.jpg

post-30795-14494999736233_thumb.jpg

Link to comment
Share on other sites

I just soldered the J6 cable like the photos and tape A123 and B123 and disconnect J4 cable (yellow/black).

After connected the Akitio default PSU (for test) nothing came up no led no fan, did I do anything wrong? Or do I need to use the ATX PSU in order to use this method?

Nando's instructions don't say anything about disconnecting that yellow/black cable. If you like to try it back powered as I did, remove the tapes as well. And please always use only the ATX PSU with a molex-to-barrel adapter, no other power sources.

  • Thumbs Up 1
Link to comment
Share on other sites

So it means there are two difference methods.

1. Taped the A123 B123 and connect J4 yellow/black cable

2. Disconnect J4 cable and untaped the A123 B123

Both need power from Molex to Barrel PSU

I understand correctly,right?

Oh and one thing, this photo from you show that you still connect J4 cable. Did you remove it later?

http://forum.techinferno.com/attachments/enclosures-adapters/14171d1424708172-c3_capacitor.jpg

Link to comment
Share on other sites

So it means there are two difference methods.

1. Taped the A123 B123 and connect J4 yellow/black cable

2. Disconnect J4 cable and untaped the A123 B123

Both need power from Molex to Barrel PSU

I understand correctly,right?

Oh and one thing, this photo from you show that you still connect J4 cable. Did you remove it later?

http://forum.techinferno.com/attachments/enclosures-adapters/14171d1424708172-c3_capacitor.jpg

Correct, and the GPU on the x16 slot. The method 2 seems to be as stable as the taping trick with the Asus R9s, but still crashes in low power conditions.

The photo describes the taping trick situation.

  • Thumbs Up 1
Link to comment
Share on other sites

Is there anything I can do to increase power stability?

Yes, solder in the missing C3 (100UF/16V) and C4 (10uF/25V) capacitors as is provided by the higher-rated but nearly identical 120W OWC Mercury Helios product to provide some additional smoothing/filtering of the 12V_ATX input power to the slot. See http://forum.techinferno.com/enclosures-adapters/8848-%24299-owc-mercury-helios-16gbps-tb2-discussion.html#post121898

Could we have more infos on the type of capacitor. I'm having a really hard time to find them online. Thank you

Edit: Found them:

Part codes are EEE-TP1C101AP and EEE-FP1E100AR in eBay

Link to comment
Share on other sites

Hello,

I am having an issue where my MacBook Pro hangs if I boot the card before the mac and also if I boot the mac holding alt/option and then turn the eGPU on. Anyone got any ideas as to why my MacBook Pro refuses to boot with the eGPU connected? The Thunder2 is detected fine if I boot into OS X then turn on the eGPU but obviously the card is not detected. Windows doesn't see the card at all if I boot with it connected after initial boot (AKA at login screen)

My setup is as follows :-

1. MacBook Pro Retina Late 2013 with GT750M

2. MSI GTX 960 Gaming

3. Powered PCIe x16 raiser with the Akitio PSU disconnected (back powered by raiser)

4. XFX 850w PSU (overkill i know but its what I have laying around)

5. Two different thunderbolt cables

6. Windows 8.1 Pro 64bit (EFI) & OS X 10.10.3

Any help would be much appreciated.

Thanks

Link to comment
Share on other sites

@macman: But the AKiTiO is working with only beeing powerde by the riser?

The AKiTiO has a green power light, a blue light to say thunderbolt Is connected and the fan is spinning so I assume it's being successfully powered by the raiser

Sent from my iPhone using Tapatalk

Link to comment
Share on other sites

Well, when I saw the flicker, the molex from the PCI-e riser was resting on that little spot, but the plastic was only touching it. But I'm pretty sure that's what caused the small fire.

I just put up everything so I don't burn anything else lol. But I plan on buying another riser like goalque said.

But I do know that green and blue light still pop up. Just hoping that it still works.

I still don't understand. Either some piece of metal touched something it shouldn't have or there is some other flaw in your setup.

Last night I spent another 2-3 hours with Far Cry 4 running full 4K 60 Hz, this time with GTX980. I was amazed to discover that the game is 90+% GPU. My 2014 Mini spent most time at 2.4 Ghz and the CPU WAS NEVER ABOVE 12 Watts and 60C while the GPU was running around 70-90% of TDP.

This is why eGPU so effective, the game barely used CPU. (Or Intel has an error in the monitoring app)

Anyhow, I digress. Getting it booted can be tricky but once the eGPU is running the 4K, the machine is 100% stable. As in NO CRASHES in last 2 nights. I walked away and left the game running while I ate dinner and watched TV. Came back and it was still ready to go.

I feel strongly that people are far better off WITHOUT A RISER. I have a couple, sitting in a box. May be different using AMD hardware as I have not tested but with 680, 780, titan Black, 980 and Titan-X I have found perfect function and no power issues by just bending end of case and using standard power.

If I was putting this together to sell I would probably switch the 22AWG wires but I have been unable to reproduce any electrical issues by running this way.

I think it is likely that some folks have mistaken driver instability for power issues and thus all this riser talk.

Link to comment
Share on other sites

Hello,

I am having an issue where my MacBook Pro hangs if I boot the card before the mac and also if I boot the mac holding alt/option and then turn the eGPU on. Anyone got any ideas as to why my MacBook Pro refuses to boot with the eGPU connected? The Thunder2 is detected fine if I boot into OS X then turn on the eGPU but obviously the card is not detected. Windows doesn't see the card at all if I boot with it connected after initial boot (AKA at login screen)

My setup is as follows :-

1. MacBook Pro Retina Late 2013 with GT750M

2. MSI GTX 960 Gaming

3. Powered PCIe x16 raiser with the Akitio PSU disconnected (back powered by raiser)

4. XFX 850w PSU (overkill i know but its what I have laying around)

5. Two different thunderbolt cables

6. Windows 8.1 Pro 64bit (EFI) & OS X 10.10.3

Any help would be much appreciated.

Thanks

Just a follow up to this, i think i have ruled out the AKiTiO being the problem here as I unplugged the GPU and left the raiser connected (still back powering) and had no issues with hanging when I tried the same test scenarios.

Does this mean the GPU is at fault or could the raiser still be the cause? The raiser (appears) to be high quality with two +12 wires and two GND wires connecting to front and back side with a capacitor on one side.

Thanks in advance for any ideas people can put forward.

Just a note however I'm not willing to explore any of the ideas which involve hardware modification (soldering etc) as I don't want to void the warranty or be left with a device I can't return if I can't get it working.

Link to comment
Share on other sites

I still don't understand. Either some piece of metal touched something it shouldn't have or there is some other flaw in your setup.

Well it somewhat seems to be working now but I still can't get Windows or Yosemite to recognize the GPU.

I've been following this method but using powering method 2 which is molex-to-barrel mod with my own PSU.

SOFTWARE INSTALLING

  1. Install Windows 8.1 64 bit from a USB stick by using ISO file. This is a standard Boot Camp installation.
  2. After installing Windows, download and install all Windows updates.
  3. Update Boot Camp drivers (Boot Camp Support Software 5.1.5640) and shut down
  4. Do the hardware preparing part 1-8
  5. Install Nvidia driver 340.52 (Drivers | GeForce)
  6. Shut down MacBook Pro and continue from the hardware preparing step 8

HARDWARE PREPARING

  1. Open AKiTiO's box by a normal screwdriver, take out the PCIe board and attach the GPU to its x16 slot
  2. Attach PSU's power cable (blue) to the GPU and do the "paperclip trick" as shown in picture 1.
  3. Attach DVI/HDMI cable to the GPU (if using a back bracket, I wasn't able to fit DVI-cable because the plastic part was too wide) and use any external monitor you like
  4. Place something under the GPU so that it will be steady (a removed front panel is perfect for this)
  5. Attach AKiTiO's 60W power plug to the DC jack of the PCIe board
  6. Connect TB cable
  7. Power up AKiTiO's box (green light should appear) and the PSU at the same time (I use a power strip with a switch on the floor)
  8. Turn on MacBook Pro (blue light should appear). If it is booting to OSX partition, change startup disk to Boot Camp partition from the OSX preferences.

I've been tracing my steps back and still nothing.

When I do steps 6-8 (Hardware Preparing) the GPU is struggling to get the fans spinning and it makes this weird grind noise that is constant whenever the Thunderbolt cable is plugged in, but when I take out the Thunderbolt cable the noise stops. I then plug it back in it the noise continues.

(EDIT: Found out that the noise is the coil whining from the GPU which the GTX 970 is known for).

Hardware that I'm using:

500B EVGA PSU

GTX970 SC

AKitio Thunder 2

molex-to-barrel

(I stopped using the PCI-e raiser because it started getting hot on one of the sides)

Link to comment
Share on other sites

I did both methods the led came up the fan on GPU spin at first I thought it's working. But in DIY eGPU Setup it can't detect the GPU no matter what.

I wonder did I break anything on the board?

Link to comment
Share on other sites

I did both methods the led came up the fan on GPU spin at first I thought it's working. But in DIY eGPU Setup it can't detect the GPU no matter what.

I wonder did I break anything on the board?

I did exactly the same thing.

I did both methods and it powered on and everything, but now the fan is struggling to move.

The first method I did was with my own PSU plugged in and also AkiTio's PSU plugged in(not recommended at all) and the whole thing booted up and the fans started spinning. Only con was after about 5-10 secs the fans stopped spinning and Windows couldn't detect the eGPU.

Link to comment
Share on other sites

I did exactly the same thing.

I did both methods and it powered on and everything, but now the fan is struggling to move.

The first method I did was with my own PSU plugged in and also AkiTio's PSU plugged in(not recommended at all) and the whole thing booted up and the fans started spinning. Only con was after about 5-10 secs the fans stopped spinning and Windows couldn't detect the eGPU.

Consider buying a replacement board: http://forum.techinferno.com/computer-components/9474-akitio-thunderbolt-2-pcie-expansion-box-pcie-board-only.html#post128271

Link to comment
Share on other sites

Just a follow up to this, i think i have ruled out the AKiTiO being the problem here as I unplugged the GPU and left the raiser connected (still back powering) and had no issues with hanging when I tried the same test scenarios.

Does this mean the GPU is at fault or could the raiser still be the cause? The raiser (appears) to be high quality with two +12 wires and two GND wires connecting to front and back side with a capacitor on one side.

Thanks in advance for any ideas people can put forward.

Just a note however I'm not willing to explore any of the ideas which involve hardware modification (soldering etc) as I don't want to void the warranty or be left with a device I can't return if I can't get it working.

Apologies for constantly bumping my question but I have more info to add and am still looking for a solution or an idea as to why this is happening.

I have done more testing today with a Nvidia GTX 280 and a AMD HD4670 (I think) and noted the following results.

Neither of these cards hang at boot like the GTX 960 was causing but the GTX 280 sometimes caused windows to freeze at the little swirl. Couldn't get the driver to install though and I noticed that the FaceTime HD camera showed as code 12 in device manager (not enough system resources) so I gave up with that card. The AMD HD4670 (which requires no external power where as the other two cards required PCIE power) but I left it powered by the raiser only (AKiTiO PSU is still left in the box) and that didn't cause any issues like the previous two cards and Windows even installed a driver for it and I had external display, haven't tested the gaming side of things though (although I bet the GT750m in the laptop has more power).

Tested the GTX 960 in a PC and it displayed video fine so I think the card is ok.

Should I give up with the GTX 960 and order something else? Or is there a solution anyone can suggest.

Thanks

  • Thumbs Up 1
Link to comment
Share on other sites

Apologies for constantly bumping my question but I have more info to add and am still looking for a solution or an idea as to why this is happening.

I have done more testing today with a Nvidia GTX 280 and a AMD HD4670 (I think) and noted the following results.

Neither of these cards hang at boot like the GTX 960 was causing but the GTX 280 sometimes caused windows to freeze at the little swirl. Couldn't get the driver to install though and I noticed that the FaceTime HD camera showed as code 12 in device manager (not enough system resources) so I gave up with that card. The AMD HD4670 (which requires no external power where as the other two cards required PCIE power) but I left it powered by the raiser only (AKiTiO PSU is still left in the box) and that didn't cause any issues like the previous two cards and Windows even installed a driver for it and I had external display, haven't tested the gaming side of things though (although I bet the GT750m in the laptop has more power).

Tested the GTX 960 in a PC and it displayed video fine so I think the card is ok.

Should I give up with the GTX 960 and order something else? Or is there a solution anyone can suggest.

Thanks

This might give an answer why your GTX 960 is not working properly with the 750M equipped MBP:

http://forum.techinferno.com/diy-e-gpu-projects/9378-installing-nvidia-web-drivers-osx-betas.html#post127372

And thanks about AMD HD4670 note, interesting to see new AMD experiences as they are rarely used with the AKiTiO.

  • Thumbs Up 1
Link to comment
Share on other sites

This might give an answer why your GTX 960 is not working properly with the 750M equipped MBP:

http://forum.techinferno.com/diy-e-gpu-projects/9378-installing-nvidia-web-drivers-osx-betas.html#post127372

And thanks about AMD HD4670 note, interesting to see new AMD experiences as they are rarely used with the AKiTiO.

Ah thank you so much for that link, I'll return the 960 then and get a GTX 7 series card. Just double checked and it is a ASUS ATI/AMD HD4670 512MB.

Which GTX 7 series card do you recommend?

Link to comment
Share on other sites

My experience with an evga 770 sc was a total disaster. Then i got a 970 and everything was much better but still did not manage to get it 100% stable under osx for gaming. Only my experience but choose carefully. If you take a 770 try to take a reference design not overlocked like i had.

Link to comment
Share on other sites

Just a follow up to this, i think i have ruled out the AKiTiO being the problem here as I unplugged the GPU and left the raiser connected (still back powering) and had no issues with hanging when I tried the same test scenarios.

Does this mean the GPU is at fault or could the raiser still be the cause? The raiser (appears) to be high quality with two +12 wires and two GND wires connecting to front and back side with a capacitor on one side.

Thanks in advance for any ideas people can put forward.

Just a note however I'm not willing to explore any of the ideas which involve hardware modification (soldering etc) as I don't want to void the warranty or be left with a device I can't return if I can't get it working.

@macman - I have the same issue, as do many of us who are trying to use a Maxwell (GTX 750ti, GTX 960/970/980, etc.) GPU with a rMBP with a dedicated graphics card (such as the 750m). I'm running basically the same hardware you are (late 2013 750m rMBP, AKiTiO w. ATX PSU, GTX 970). I find that, if I keep trying to boot my machine over and over, it will eventually start. It's unpredictable, but I would say it successfully boots about 20% of the time.

Link to comment
Share on other sites

@macman - I have the same issue, as do many of us who are trying to use a Maxwell (GTX 750ti, GTX 960/970/980, etc.) GPU with a rMBP with a dedicated graphics card (such as the 750m). I'm running basically the same hardware you are (late 2013 750m rMBP, AKiTiO w. ATX PSU, GTX 970). I find that, if I keep trying to boot my machine over and over, it will eventually start. It's unpredictable, but I would say it successfully boots about 20% of the time.

Its a very strange problem, does anyone know what is so special about this config which causes this issue? Do they conflict in some way or something?

Link to comment
Share on other sites

@macman, we don't know for sure, but the consensus seems to be that something about having an onboard NVIDIA dGPU and a Maxwell eGPU together prevents the firmware from consistently recognizing the eGPU. The problem only appears to manifest with Maxwell eGPUs and Nvidia dGPUs. The 15'' rMBP with intel Iris only is not affected, and non-Maxwell GPUs (e.g. the GTX 760) are not affected.

  • Thumbs Up 1
Link to comment
Share on other sites

Ah thank you so much for that link, I'll return the 960 then and get a GTX 7 series card. Just double checked and it is a ASUS ATI/AMD HD4670 512MB.

Which GTX 7 series card do you recommend?

Unfortunately I don’t have any dGPU Mac, but you can take a look at implementations hub 750M section, often the manufacturer and the model of the GPU is mentioned:

http://forum.techinferno.com/diy-e-gpu-projects/6578-implementations-hub-tb-ec-mpcie.html

If there are 3DMark tests done, it usually means a stable system - but not always since my Asus R9 270X and 280X crash when the GPU is not stressed, such as after a couple of rounds of 3DMark11 physics.

  • Thumbs Up 1
Link to comment
Share on other sites

How do AMD cards fair in these setups? Are they largely unstable or do people have success?

@Oscar J informed me that he got the his second Gigabyte 7970 OC working perfectly stable without a powered riser + 650M equipped MBP on OS X. The first Gigabyte 7970 was identical but crashed after 5 minutes with the Indigo renderer. I suspect the reason could be different firmware or BIOS chip. I’ve been tracking down the reason for both the black screen and booting issues of the different architecture GPUs, and I bet that the common factor is EFI version:

http://forum.techinferno.com/diy-e-gpu-projects/8619-tb2-macbooks-allow-monitors-used-nvidia-egpus-osx.html#post120215

That’s why some users report black screens with the new 2015 model:

http://forum.techinferno.com/diy-e-gpu-projects/7910-diy-egpu-macbook-experiences-57.html#post128390

In order the GPU is able to talk the same language with the host Mac, they need the common protocol. I have tried to hack the Mac firmware, no success. The logic board had to change because there did not exist any tool to rewrite 2014 15” MBP firmware back to original according to discussion with a senior Apple specialist. Also the local Mac service was confused how on earth I was able to make my MBP a brick :D The only sign of life was a spinning fan, and luckily I had a warranty.

Apple’s systems are pretty closed. I have a hunch that they are restricting eGPU use with specific Mac models. MVC may rescue us in the future with those custom flashed Nvidia GPUs, or we could consider AMD’s window of opportunity…

AMD Gaming: AMD enables incredible DirectX®... | AMD Blogs

And my API Overhead Feature Test under Windows 10 Pro Technical Preview (build 10041), DirectX12 works flawlessly with the AMD engineering sample driver (installed automatically):

post-28870-14494999739669_thumb.png

AMD Radeon R9 280X video card benchmark result - Intel Core i7 4702HQ,Apple Inc. Mac-3CBD00234E554E41

DX12 shows 3X more draw calls per second than DX11! Wow!

For me the eGPU screen output is important, I don’t need the full OpenGL performance on OS X because Windows is for gaming, and the OpenCL performance per price is undoubtedly better with the AMDs. And yes, out of the box Crossfire is nice. It’s a pity that some AMDs are unstable with the AKiTiO.

  • Thumbs Up 2
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.