ASUS XG Station Pro Review - Cool, Calm and Collected
 
Notifications
Clear all

ASUS XG Station Pro Review - Cool, Calm and Collected  

 of  11
  RSS

goalque
(@goalque)
Noble Member Admin
Joined: 4 years ago
 
Posted by: wimpzilla

Thanks you.

I was wrong on some aspect of the pcb, the picture greatly helped to sort out, i updated the picture you sent me.
-What i thought the VRM controller is in reality an ITE IC, often designated as I/O controller also managing the t°, voltage, fans, current monitoring.
-The main 3 phases VRM for the gpu 12v have doubled Hi side and Low side mosfets to help the current balance on the phase.
-All the other minor rails needed as the 5/3.3v or lower, are provided by power stages, being a set of Hi/Lo mosfets integrated into a single package.
-Not sure about, i suppose a simple LDO voltage regulator is maybe used to feed the voltage needed to drive the mosfets.
-A low current phase is maybe build using the internal mosfet of the IC tagged by the ?, pretty curious if this IC is not the VRM controller at the end.

Overall i'm surprised by the amount of VRM this pcb alone pack, without knowing the exact IC part number used it's hard to tell more.
I would suggest you to test both connectors on the back of the pcb while the enclosure is running, would be interesting to know on which power plane it's running, 12v or 19.5v.
If you could try to read the number on the IC marked with a ?, the one that is connected to the small inductor, i would suppose this one would be the VRM controller.
IF the back connectors support 12v and would feed either the gpu and board, it would be nice if ASUS include them into the finished build.
Would be great to have the ability to either run both a psu or a power brick, maybe the only enclosure that would offer this option.

asus xg station pro main tbt board

Here's the new picture, I couldn't test the IC marked section with ? but probed other points on the board.

blue = ~18.8V
orange = ~2.2V
purple = ~2.8V - 2.9V
green = ~4.7V
yellow = ~11.6V

I think there is a possibility to use a normal 12V PSU or HDPLEX 400W. Those two 8-pin sockets have direct 12V connection to the front of the board (yellow lines, multimeter's continuity test).

Or could we just plug the PSU directly to the other 8-pin?

The front panel is easily removable so you can drill mounting holes for an HDPLEX 400W, and slide back the panel, covering the screw heads.

automate-eGPU EFIapple_set_os.efi

Mid 2015 15-inch MacBook Pro eGPU Master Thread

 
2018 13" MacBook Pro [8th,4C,U] + Radeon VII @ 32Gbps-TB3 (ASUS XG Station Pro) + Win10 1809 [build link]  


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 4 years ago
 

So in order to read the output of a VRM you need to probe with the multimeter the output of the VRM choke, as you did with the green point.
One side of the choke is output, the other side is linked to the mosfets.
The decoupling caps after the choke are also a good probing point, but i don't know precisely which one are used for the decoupling.
Dunno if it use the SMD or the normal polymer caps for decoupling, you need to test the continuity between the components, you should have mosfets>choke>decoupling cap>load.

The other thing you need to check is if the 2 unused 8pin connectors link also elsewhere on the pcb or if these are only linked to the other 8pin connectors.
It become interesting only if these unused 8pin power up also the TB logic and disable the other unused VRM when so, allowing you to run the board out of a PSU.
Otherwise i simply conclude that the unused 8pin are just a pcb, power plane design, having the choice to mount vertical or horizontal connectors for the gpu output.
You could directly power up the gpu with the external PSU bypassing completely the board 8pin output, the wise choice.

Maybe it's still fine, but the voltages you read are quite low compared to the standard: 11.6v for 12v; 4.7v for 5v; 2.9v for the 3.3v.
Meaning the VRM compartment is somehow suffering, maybe check if you can rest your finger on the mosfets when the enclosure is under load.

This post was modified 3 years ago

2012 13-inch Dell Latitude E6320 + R9 270X@4Gbps-mPCIe (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo ThinkPad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


ReplyQuote
goalque
(@goalque)
Noble Member Admin
Joined: 4 years ago
 

Appreciated your input (I am not an electrical engineer). I chose the probing points randomly. The section near the main IO socket (18.8V) is covered by a shield and would not like to remove it.

I couldn't find other 12V points yet (except the x16 slot). There seems to be a capacitor between the 8pins because I heard only a short beep sound during the continuity test but the sound was continuous between the marked yellow lines.

I could try powering the board carefully from those unused 8pins to see if voltage conversion works inversely.

automate-eGPU EFIapple_set_os.efi

Mid 2015 15-inch MacBook Pro eGPU Master Thread

 
2018 13" MacBook Pro [8th,4C,U] + Radeon VII @ 32Gbps-TB3 (ASUS XG Station Pro) + Win10 1809 [build link]  


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 4 years ago
 

No please don't, it's NOT a good idea at all, if there is not a decent circuit that manage to cut the other VRM compartment.
You will for sure send 12v to the gpu but also send 12v to then entire output VRM power plane, not sure it will be nice. ^^
As said above i just conclude these are another power plane design to where to mount the 8 pin gpu output.
So don't plug any PSU anywhere on this board. ^^

You can't find the 12v because the probe point are under the pcb, the big 3 choke legs are under not visible here.
But you should find which caps have continuity with the leg choke to have an easier probe point.

This post was modified 3 years ago

2012 13-inch Dell Latitude E6320 + R9 270X@4Gbps-mPCIe (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo ThinkPad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


goalque liked
ReplyQuote
goalque
(@goalque)
Noble Member Admin
Joined: 4 years ago
 

Okay, I won't do it Smile Even though we were able to "back power" the old Thunder2 board, this is not the same as we are dealing with higher voltages, not only 12V.

automate-eGPU EFIapple_set_os.efi

Mid 2015 15-inch MacBook Pro eGPU Master Thread

 
2018 13" MacBook Pro [8th,4C,U] + Radeon VII @ 32Gbps-TB3 (ASUS XG Station Pro) + Win10 1809 [build link]  


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 4 years ago
 

Exactly, here we start with higher voltage to compensate for the current, higher voltage but lower current before the 12v conversion, like what we found on laptops.
If you can't find some fancy other continuity link between these unused 8 pin and some other parts of the board, i would say there no control switch over 12v/18v supply.

This post was modified 3 years ago

2012 13-inch Dell Latitude E6320 + R9 270X@4Gbps-mPCIe (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo ThinkPad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


ReplyQuote
goalque
(@goalque)
Noble Member Admin
Joined: 4 years ago
 
Posted by: wimpzilla

The other thing you need to check is if the 2 unused 8pin connectors link also elsewhere on the pcb or if these are only linked to the other 8pin connectors.
It become interesting only if these unused 8pin power up also the TB logic and disable the other unused VRM when so, allowing you to run the board out of a PSU.

They are linked to the 12V points of the x16 slot (a short beep sound, had to swap the red/black probes of the multimeter to hear the sound again, a capacitor between?).

EDIT: I also found a couple of more short beep connections from the unused 8pins, some on the purple area (2.9V) and the yellow ones (11.6V):

automate-eGPU EFIapple_set_os.efi

Mid 2015 15-inch MacBook Pro eGPU Master Thread

 
2018 13" MacBook Pro [8th,4C,U] + Radeon VII @ 32Gbps-TB3 (ASUS XG Station Pro) + Win10 1809 [build link]  


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 4 years ago
 

It mean that there is no control over the voltage imo and the board can't be powered using the read unused 8 pin.
It confirm what i was saying before, the 12v power plane is shared by the major 3 VRM assembly, the one with the big squared grey inductor/choke.
In fact you found one output of the 12v VRM i was speaking before, the top right big soldering point is the leg of the choke, the next bellow the decoupling cap.
It feed then the pci-e socket and some of the components of the TB chip.

Dunno why you got some intermittent beeping, maybe the resistance over the power plane increase so the multimeter have a hard time telling what is connected.
The continuity works measuring a resistance along a path, if the resistance is bellow a defined value, the multimeter will beep signaling continuity along the path.
But if the resistance value increase too much over the distance, the multimeter will have a hard time telling if it's just a open path or if there is some components along.

This post was modified 3 years ago

2012 13-inch Dell Latitude E6320 + R9 270X@4Gbps-mPCIe (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo ThinkPad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


goalque liked
ReplyQuote
goalque
(@goalque)
Noble Member Admin
Joined: 4 years ago
 

Ok, then it's double confirmed. I was just hopeful. Thanks!

automate-eGPU EFIapple_set_os.efi

Mid 2015 15-inch MacBook Pro eGPU Master Thread

 
2018 13" MacBook Pro [8th,4C,U] + Radeon VII @ 32Gbps-TB3 (ASUS XG Station Pro) + Win10 1809 [build link]  


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 4 years ago
 

I was hopeful too tbh, i thought too maybe i would be possible to power the board using these unused 8 pin with an external psu.
But these are just a power plane design, to have the output connectors arranged in another way.
Thank you for your testing, you cleared this question that i had long time ago.  😉 

Maybe you can reach Asus and get the power supply slitter to use 2 brick psu to power the enclosure, then check again if there is any change when you read the voltages.
Also check if the mosftes get hot, if you can't rest your finger on top of them it mean that a heatsink on top will greatly improve their current and voltage capability.

This post was modified 3 years ago

2012 13-inch Dell Latitude E6320 + R9 270X@4Gbps-mPCIe (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo ThinkPad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


ReplyQuote
 of  11