Power Supply Dual internal Xbox One S PSUs?

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
So I have been looking for a 12v internal power solution to incorporate with my pico PSU. The HDplex 400w AC/DC is close but it is 19v output which will not work with the Pico PSU and would require the purchase of the HDplex 400w power block... With that said, the Xbox One X and S both include internal 12v PSUs. The One X PSU is 245 watt but unfortunately is just a bit to large width wise to fit in my next case design. That leaves me with the One S PSU which is only 120 watt when I need at least 220.

So long story short, does anyone think it would be possible to incorporate two One S PSUs into a single case using a single wall plug? The idea would be to tie the two PSUs together through the power input port and have one power the GPU while the other powers the Pico PSU. The power output connector looks to be a standard 6 pin that could plug directly into the GPU. That said I need to find a pinout to confirm this. I have seen dual PSU setups before but have little experience on the matter... Would something like what I just proposed be a stable solution for internal 12v power?

Xbox One S PSU



 

Kmpkt

Innovation through Miniaturization
KMPKT
Feb 1, 2016
3,382
5,936
You can do this using the Y-PWR Adapter from Pico Box:

http://www.mini-box.com/Y-PWR-Hot-Swap-Load-Sharing-Controller

This device will take two power inputs between 5V and 30V at 10A (15A peak) and sum them into a single line of 20A (30A peak). You'd need to likely make a 6 pin to 4 pin cable adapter for each unit, but this should provide you with a 240W 12V power supply.
 

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
You can do this using the Y-PWR Adapter from Pico Box:

http://www.mini-box.com/Y-PWR-Hot-Swap-Load-Sharing-Controller

This device will take two power inputs between 5V and 30V at 10A (15A peak) and sum them into a single line of 20A (30A peak). You'd need to likely make a 6 pin to 4 pin cable adapter for each unit, but this should provide you with a 240W 12V power supply.

That looks to be an excellent solution to my issue. Thanks!
 
  • Like
Reactions: VegetableStu

CC Ricers

Shrink Ray Wielder
Bronze Supporter
Nov 1, 2015
2,234
2,557
When you get the power supplies, you may want to splice the 6-pin connectors from them. That is actually a fair bit smaller than the 6-pin connector for GPU power cables, so they won't fit inside the adapter.



Here is a size comparison. Standard ATX power connections use the Mini-Fit size. The Xbox One PSUs have the Micro-Fit size.
 

Kmpkt

Innovation through Miniaturization
KMPKT
Feb 1, 2016
3,382
5,936
Was just researching Mini Fit terminals (after looking at your post) and saw there is a Mini Fit Plus crimp type that handles 13.0A per circuit instead of 9.0A per circuit. Anyone have experience with these or have any idea why these aren't the industry standard with PCs? 13.0A per circuit at 12V would be pretty nice for things like PCIe no?
 

Therandomness

Cable-Tie Ninja
Nov 9, 2016
229
270
Was just researching Mini Fit terminals (after looking at your post) and saw there is a Mini Fit Plus crimp type that handles 13.0A per circuit instead of 9.0A per circuit. Anyone have experience with these or have any idea why these aren't the industry standard with PCs? 13.0A per circuit at 12V would be pretty nice for things like PCIe no?
Pretty sure it'd just be a pain to adopt a new standard at this point, considering the amount of things that use the larger size already.
 
  • Like
Reactions: Soul_Est

Kmpkt

Innovation through Miniaturization
KMPKT
Feb 1, 2016
3,382
5,936
Mini fit and Mini fit plus use the same housings, its just the Mini fit plus crimps are rated for a higher wattage.
 

blubblob

SFF Lingo Aficionado
Jul 26, 2016
104
127
Was just researching Mini Fit terminals (after looking at your post) and saw there is a Mini Fit Plus crimp type that handles 13.0A per circuit instead of 9.0A per circuit. Anyone have experience with these or have any idea why these aren't the industry standard with PCs? 13.0A per circuit at 12V would be pretty nice for things like PCIe no?
Mini fit and Mini fit plus use the same housings, its just the Mini fit plus crimps are rated for a higher wattage.
That Mini Fit Jr. and Mini Fit Plus share the same housing is exactly the reason why Plus will likely never replace the normal Jr. connectors.
Because there is no underlying bus system to negotiate the rating of a specific connector the system wouldn't know whether it is dealing with a Jr. or Plus connector. Combine that with there being hundreds of PSUs out there using those lower rated connectors you would always have to assume the worst.
There is no real upside to using those connectors as long as you have to deal with backwards compatibility.
 

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
Received everything Friday! So far I have the PSUs wired up in a test configuration with 4 pin connectors and successfully tested them with the Y-PWR. Later on this week I plan to conduct an actual load test using my current system.

Physical dimensions for the PSU are as follows...

128mm length
60mm width
44mm height

With a bit of modification the housing height can be dropped to 40mm. Removing the PSU from the housing drops about 4mm from all dimensions.
 

Kmpkt

Innovation through Miniaturization
KMPKT
Feb 1, 2016
3,382
5,936
Just to let you know, I ran the Y-PWR device through some stress tests a couple of days ago and it works quite nicely. The small component bank (MOFSETs I believe) can get warm and are the rate limiter on the function from what I can tell. I would recommend trying to get your hands on a small heatsink with thermally adhesive tape for these if you can.
 

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
Just to let you know, I ran the Y-PWR device through some stress tests a couple of days ago and it works quite nicely. The small component bank (MOFSETs I believe) can get warm and are the rate limiter on the function from what I can tell. I would recommend trying to get your hands on a small heatsink with thermally adhesive tape for these if you can.

I was wondering how well the Y-PWR handled the load. I should have a spare stick on heatsink lying around to use. Thanks for the tip! I’ll update this thread when I conduct my testing!

Also wondering if this thread shouldn’t be moved to the mod section...
 

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
Initial load testing complete!

A single PSU was capable of powering my system through boot up and during light tasks such as web browsing. Idle wattage recorded at the wall was as follows...





A single PSU was not capable of sustained power output during gaming loads of 140 watts or more. Connecting the Y-PWR and second PSU solved this issue of course. Average gaming load of my system (Ryzen 1400 and GTX1060) was about 190 Watts with 100% system load topping out at 215 watts.





One issue I did notice during the hour long gaming session was that one PSU was getting noticeablely warmer then the other. I can not find any specific documentation on exactly how the Y-PWR distributes the load but I am assuming it should be a 50/50 split?

If so then I don’t see how one should be getting warmer then the other. What I believe was happening was that the Y-PWR was constantly switching to the PSU that was warmer due to a voltage difference of more then 100 mv between the two. When measured without load one PSU was at 11.97 V while the other was at 11.88 V. I believe I need to work on bringing these voltages closer together and then re-run the test to see if the issue persists. How should I proceed with achieving this 100mv voltage drop?

Overall though the test was a success!
 
Last edited:

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
So more interesting observations about these little PSUs!

I currently have just one connected to my rig and just completed an on/off 30min gaming session in MW2. I have MW2 cranked to highest settings and 4k resolution. CPU utilization is about the same as battlefield 1 or other modern AAA titles but GPU utilization is around 70%. The system was rock solid during this gaming session and did not pull more the 130 watts during the entire test. So I am fairly confident that my rig with a 1050ti instead of a 1060 could get away with just one of these units! The PSU was very quiet (no coil whine) and the outer casing did not get warm at all! Very impressed and it has me considering dropping my GPU to a 1050ti Low Profile so that I can get my entire case smaller... If only I could find one at a reasonable price that is...
 

CC Ricers

Shrink Ray Wielder
Bronze Supporter
Nov 1, 2015
2,234
2,557
Oh hey, I remember that 3D printed case! Nice to see it again going forward with this setup.

I'd figure you'd be cutting it very close with a single PSU if you're using a GTX 1060. If you were using a 35W CPU you may be able to use just one but depends on the efficiency of the power transfer of these units, I suppose.

Nice to see you found a hidden gem out of these power supplies. They're like the smaller MeanWell units but with the bonus of having a built-in enclosure and not needing to wire an AC outlet to power it up.
 

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
Yea I never expected to run my current rig on one of these PSUs but I am very impressed with how well it was able to run my 1060 at 70% power! 140-150 watts seems to be about the absolute max these PSUs allow one on one without stability issues. So a 1050ti with a 65 watt processor should be just able to get by. A 35 watt processor would ensure rock solid operation under full artificial load.

Now I just need to figure out this voltage difference so I can re-run my load sharing test and see if I can bring the temperature down on the slightly higher voltage PSU.
 
Last edited:

Kmpkt

Innovation through Miniaturization
KMPKT
Feb 1, 2016
3,382
5,936
One issue I did notice during the hour long gaming session was that one PSU was getting noticeablely warmer then the other. I can not find any specific documentation on exactly how the Y-PWR distributes the load but I am assuming it should be a 50/50 split?

If so then I don’t see how one should be getting warmer then the other. What I believe was happening was that the Y-PWR was constantly switching to the PSU that was warmer due to a voltage difference of more then 100 mv between the two. When measured without load one PSU was at 11.97 V while the other was at 11.88 V. I believe I need to work on bringing these voltages closer together and then re-run the test to see if the issue persists. How should I proceed with achieving this 100mv voltage drop?

Yeah I believe this is the core fault in the Y-PWR and why I am developing the KMPKT Merge. The primary purpose of the Y-PWR is to enable a battery based UPS that will kick in when your AC-DC kick out. The design is such that you deliberately have the UPS output a higher/lower wattage (I forget which) so as to provide this functionality. I had two HDPlex 160W running the other night, each with their own Kill-A-Watt and you would periodically see the load dump to just one unit. Will be interesting to see how this works with a 1080 and 7700K. Hopefully I don't get a 400W draw momentarily on a 160W power supply (lol)
 

Zackmd1

Airflow Optimizer
Original poster
Jun 3, 2016
347
561
Yeah I believe this is the core fault in the Y-PWR and why I am developing the KMPKT Merge. The primary purpose of the Y-PWR is to enable a battery based UPS that will kick in when your AC-DC kick out. The design is such that you deliberately have the UPS output a higher/lower wattage (I forget which) so as to provide this functionality. I had two HDPlex 160W running the other night, each with their own Kill-A-Watt and you would periodically see the load dump to just one unit. Will be interesting to see how this works with a 1080 and 7700K. Hopefully I don't get a 400W draw momentarily on a 160W power supply (lol)

Yea 100mv is not a lot at all especially with these PSUs and load conditions... I know it was sharing some load though as the system would have rebooted due to lack of power since a single PSU was not capable of loading battlefield 1. 400w through a 160 watt brick should be very interesting! Haha

PS: anyone have experience with a 1050ti and 65watt cpu system? Am I close in my assumption that total system power shouldn’t be much more then 140 Watts?

I am seriously considering a 1050ti LP now with only one of these PSUs since a 1050ti would still be more then adequate for my gaming needs.
 

Therandomness

Cable-Tie Ninja
Nov 9, 2016
229
270
Received everything Friday! So far I have the PSUs wired up in a test configuration with 4 pin connectors and successfully tested them with the Y-PWR. Later on this week I plan to conduct an actual load test using my current system.

Physical dimensions for the PSU are as follows...

128mm length
60mm width
44mm height

With a bit of modification the housing height can be dropped to 40mm. Removing the PSU from the housing drops about 4mm from all dimensions.
With the PSU out of the housing, is there literally no way to make it any smaller? What components of the PSU force it to be that high?