The Seasonic SSP-300SUG fully modular 150mm FlexATX PSU

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
Just received this unit today and thought I would take a look at it.
It's the only 150mm long FlexATX PSU with full modularity, and at 300W it has enough power for an R9 Nano and a decent i7.
The awesome thing about that is that you can make your own cable harness with perfect cable lengths without having to solder or losing your warranty, which is normally a huge issue with FlexATX PSUs as almost none of them even have PEG connectors for the GPU.

With 80Plus Gold efficiency there's potential for quiet operation during idle and moderate load.

I won't actually turn it on until I get the prototype parts for Brevis S, but I got pretty excited just looking at it.

Here are my first impressions:


Thanks for reading!
 
Last edited:

Boil

[H]ard|Gawd
Joined
Sep 19, 2015
Messages
1,439
Interesting, especially when thinking about the Hutzy XS…

So, could one simply remove the wires that were not needed (SATA power, IDE & Floppy power). Could one also 'strip down' the CPU wires if, say, one only needed a 4-pin CPU assembly…?

If so, would make for a very tidy cable system; with a M.2 boot/system drive & no other drives, one could have a basic three cable system; MB,CPU, & GPU…!

Pull the 'extra' wires, shorten the ones you are keeping to the appropriate length, sleeve same; one custom sleeved Flex ATX PSU…!!!
 

Boil

[H]ard|Gawd
Joined
Sep 19, 2015
Messages
1,439
Damnit, just looked at this the Seasonic website, it does not have a dedicated GPU connection…
 

Boil

[H]ard|Gawd
Joined
Sep 19, 2015
Messages
1,439
Okay, so I am guessing one would use the pins/connections for the SATA/IDE/Floppy power to make a 8-pin connector for the GPU…?!?

I am definitely interested in this, if it is possible to mod the harness to include a 8-pin GPU connector…!
 

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
Interesting, especially when thinking about the Hutzy XS…

So, could one simply remove the wires that were not needed (SATA power, IDE & Floppy power). Could one also 'strip down' the CPU wires if, say, one only needed a 4-pin CPU assembly…?

If so, would make for a very tidy cable system; with a M.2 boot/system drive & no other drives, one could have a basic three cable system; MB,CPU, & GPU…!

Pull the 'extra' wires, shorten the ones you are keeping to the appropriate length, sleeve same; one custom sleeved Flex ATX PSU…!!!

Yes, that's the idea.

Unfortunately, because so many pins on the cable harness have two wires crimped to them, there is no way to remove all pins for SATA/IDE connectors and be done with it. The only completely separate part are the ATX12V connectors for the CPU. Everything else is connected together in some way.
For a case manufacturer this would mean that they had to supply a custom harness with the PSU, but that's much more affordable than supplying a completely custom PSU.

Okay, so I am guessing one would use the pins/connections for the SATA/IDE/Floppy power to make a 8-pin connector for the GPU…?!?

I am definitely interested in this, if it is possible to mod the harness to include a 8-pin GPU connector…!

Yes it is, but you'll have to redo the whole harness. You can reuse the connector housings, but you'll have to get new pins. I guess you can leave all the cables in the ATX24pin connector and just shorten them, so it's not too complicated.
 

Boil

[H]ard|Gawd
Joined
Sep 19, 2015
Messages
1,439
I did mean that one would reuse the wires from the SATA/IDE/Floppy, sorry if I wrote it to read that I would actually reuse the existing pins… I know those would need replaced… So, any idea of the proper procedure for creating a viable 8-pin GPU power cable from the existing SATA/IDE/Floppy cables…? Like, the proper pin-out & such…?
 
Last edited:

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
I did mean that one would reuse the wires from the SATA/IDE/Floppy, sorry if I wrote it to read that I would actually reuse the existing pins… I know those would need replaced… So, any idea of the proper procedure for creating a viable 8-pin GPU power cable from the existing SATA/IDE/Floppy cables…? Like, the proper pin-out & such…?

So this is the pinout for the PSU side connector:

KUaQTH9.png


The relevant pins for PEG (GPU power) are 12V (yellow) and GND (dark grey).

12V:
Pin 2 is connected to ATX24pin and SATA.
Pin 3 is connected to ATX24pin and IDE/Floppy.
Pin 14 is connected to the first half of the ATX12V connector.
Pin 15 is connected to the second half of the ATX12V connector.

GND:
Pin 16 is connected to the first half of the ATX12V connector.
Pin 17 is connected to the second half of the ATX12V connector.
Pin 18 is connected to SATA.
Pin 19 is connected to IDE/Floppy.

So the easiest way to add a PEG connector is to remove the ATX12V connectors and use those wires.
When you want to remove IDE/Floppy or SATA, then 5V and 3.3V are relevant as well:

5V:
Pin 20 is connected to ATX24pin and SATA.
Pin 21 is connected to IDE/Floppy.

3.3V
Pin 22 is connected to SATA.

So, to get rid of those, you'll have to remove the pins 2, 3, 18, 19, 20, 21 and 22, re-crimp the 5V wire of ATX24pin to pin 20.
Then you can use pin 2 and 3 to connect to the 12V contacts of the PEG connector and ATX24pin and use pin 18 and 19 to connect to the GND contacts of the PEG connector.

I hope that's detailed enough.


The stock fan is 2 pin, so wouldn't it run at full speed constantly?

No. Pretty much all PSU and GPU fans are 2 pin from what I know. They're voltage controlled, not PWM controlled. The third pin on normal fans is for the RPM readout, the fourth is for RPM.
The problem with non-RPM fans as case fans is that the mainboard can't know the fan curve (some can't even use voltage control), which might not be linear, so fine-tuning is very hard. With PSUs, the manufacturer is exactly aware of the fan an can thus trim the voltage response to the desired result. Or they cheap out and just use a thermistor.

If this unit is decent in terms of noise, it would be a great fit for both of our cases. And pretty much all FlexATX cases in general.
 

Boil

[H]ard|Gawd
Joined
Sep 19, 2015
Messages
1,439
TY very much for the pin-out breakdown…! I do believe I could fumble through the rest (maybe with a little bit of help from the Internet)…!
 

andgo

Weaksauce
Joined
May 15, 2014
Messages
99
Did you test out the PSU at full load?

How is the fan curve / noise?

Would it be possible to power an R9 Nano + i7 @ 4ghz on 300watt?
 

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
Boil is correct. The prototype parts are scheduled to arrive on the 3rd of May, so until then I won't do any testing.

Would it be possible to power an R9 Nano + i7 @ 4ghz on 300watt?

I say it is possible. The i7-6700 has a turbo clock of 4GHz and a TDP of 65W, the R9 Nano has a TDP of 175W. While TDP is not equivalent to power consumption, it's a good estimate.
Give each of them 15W overhead (just to be sure) and you arrive at 270W, still way below the maximum 300W. You can easily put an SSD or two in there as well, still won't reach the 300W.
Additionally, you'll never have GPU and CPU and all drives under 100% load at the same time, that's just not a realistic scenario.

So yes, very much doable.
 
Last edited:

andgo

Weaksauce
Joined
May 15, 2014
Messages
99
Boil is correct. The prototype parts will arrive on the 3rd of May, so until then I won't do any testing.



I say it is possible. The i7-6700 has a turbo clock of 4GHz and a TDP of 65W, the R9 Nano has a TDP of 175W. While TDP is not equivalent to power consumption, it's a good estimate.
Give each of them 15W overhead (just to be sure) and you arrive at 270W, still way below the maximum 300W. You can easily put an SSD or two in there as well, still won't reach the 300W.
Additionally, you'll never have GPU and CPU and all drives under 100% load at the same time, that's just not a realistic scenario.

So yes, very much doable.


Well, looking forward to see someone try it :)
 

tamas970

n00b
Joined
May 11, 2016
Messages
1
Great, looking forward to the results:) I am planning to set up a small NAS with a NUC board and 2x3.5" HDD-s (<30W needed), and looking for a capable small PSU. (the supplied NUC PSU would most likely fail to feed the big hard drives)

Do you have any chance to measure efficiency at idle?

Boil is correct. The prototype parts are scheduled to arrive on the 3rd of May, so until then I won't do any testing.



I say it is possible. The i7-6700 has a turbo clock of 4GHz and a TDP of 65W, the R9 Nano has a TDP of 175W. While TDP is not equivalent to power consumption, it's a good estimate.
Give each of them 15W overhead (just to be sure) and you arrive at 270W, still way below the maximum 300W. You can easily put an SSD or two in there as well, still won't reach the 300W.
Additionally, you'll never have GPU and CPU and all drives under 100% load at the same time, that's just not a realistic scenario.

So yes, very much doable.
 
Last edited:

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
Unfortunately I don't have the test-gear necessary for that. The prototype parts arrived, so hopefully I'll get a chance to use this PSU until Friday or Saturday, but don't expect too much of a technical review. I can mainly run it and tell you how loud it felt to me.
 

Hahutzy

Limp Gawd
Joined
Sep 4, 2015
Messages
169
I ordered one today, want to test 2 things:

1) Whether my case's C14 cutout fits

2) Noise @ load vs FSP 400W
 

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
For preliminary calculations, use these measurements (sorry, forgot to post them):

NnwC5ifl.jpg

yM4VArll.jpg

JSxGEHDl.jpg
VXx6Yuel.jpg


You'll have to calculate the center point position yourself, I'm in a hurry^^

Make sure to report your findings about noise, I'd be very interested in that!
 

Hahutzy

Limp Gawd
Joined
Sep 4, 2015
Messages
169
That's worrying.. I drew up something quick and it's showing me there's 1.9mm between the top of the C14 connector, and the left-most 6-32 tap on top of it.

Which means even if the cutout fits, you need to use a 6-32 screw with a screwhead diameter of <3.8mm to not be obstructed by the connector.

They didn't really think this through did they..

If I have to accommodate the measurements as they stand, the material between the cutout and the hole is somewhere along the lines of <0.3mm. No thanks.

I think the long term solution is to tell them to move the C14 connector down 1mm at time of assembly.

ikgFeVI.png
 
Last edited:

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
Yeah I think we'll just have to send them an email and ask whether that will be fixed. I can't imagine that being a knowing design decision, it just seems like a mistake that nobody has spotted yet.
 

Hahutzy

Limp Gawd
Joined
Sep 4, 2015
Messages
169
I received my SSP-300SUG today. Same deal. So even though it's a mistake, at least it's consistent.

I'm going to send an email to SeaSonic on monday.

Edit Quick note: Pin 21 is literally useless if the Molex/Floppy connectors are removed.
 
Last edited:

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
Good to know. I already sent them a mail but they didn't reply, I'll probably call them and ask what's up.

You just reminded me that I wanted to upload the wiring diagram I used to make a new cable harness. Here it is:

 

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
So after using the unit in my rig for over a month now, I did a bit of stress testing today.
My setup is a short GALAX GTX970 OC, an i7-4770T (45W TDP) and 16GB of G.Skill Value F3-1600C11S-8GNT RAM, so maximum power consumption should be well below the 300W of this PSU.

Until now, I was relatively happy with the unit, but I never exhibited massive loads on it. I played Rocket League for five hours straight, no problem, but that's not particularly hard on the GPU, it's a well optimised Unreal Engine 3 game.

During that time it was relatively quiet. As a matter of fact, it was about as loud as the the GPU fan during gaming, but with a somewhat higher pitch. Pretty reasonable in my eyes. The GPU load was at 40 to 55% , power consumption between 55 and 65%, as observed with HWMonitor.
During that time, the LiHeat riser was covering the top vents of the PSU, which didn't affect the noise at all.

Onto the tests: First, I made sure that the vent holes weren't covered to make sure the PSU didn't get suffocated.

Then I did the first test, GPU-only running Heaven Benchmark 4.0 on extreme presets. GPU load and power consumption were at 100%, power consumtion even got to 105% at some point.
And boy did the PSU suddenly get loud. Not quite a mini-jet engine, but reasonably close, at least with my rig sitting directly in front of me. Still perceivable with headphones on, though my headphones are very open to external noise.

Not really chuffed by that, but it did run well, and after half an hour I decided that the test was passed.

For the second test I wanted to go all-in, so Prime95 on the CPU for a few minutes until temps and noise stabilised and then Heaven on top of that. The CPU got into up to 65W of Package power consumption. Unsurprisingly, this additional load caused the PSU to spin it's teeny fan even louder, definitely over the acceptable level, but what did suprise me was the fact that after 30 seconds of combined Heaven and Prime, the whole PC shut off in an instant.

Not at all happy with that result. I know Prime puts an unrealistic load on the CPU, but I would've never thought that the SSP-300SUG couldn't handle that. I guess my system does draw more than 300W in this case. Unfortunately, I have no way of testing that.
 

Hahutzy

Limp Gawd
Joined
Sep 4, 2015
Messages
169
I just got done testing the 300SUG, and... it ran without crashing at all.

Tested using:
i5-4440s + Noctua L9i
Gigabyte 970 ITX
Mushkin 240GB SSD
2 sticks of 4GB ram

Tests I did:

Unigen Heaven, extreme settings x1
3DMark Firestrike x1

30-min torture test:
Start with prime-95, Inplace Large FFT 4 threads
After 10-minutes, Start furmark on default settings (8x MSAA 1280x720)



It all ran fine. In fact the fan noise was less whiny than the FSP 400W.

Color me confused :\
 

Zap

[H]ard|Gawd
Joined
Apr 1, 2004
Messages
1,421
For the second test I wanted to go all-in, so Prime95 on the CPU for a few minutes until temps and noise stabilised and then Heaven on top of that. The CPU got into up to 65W of Package power consumption. Unsurprisingly, this additional load caused the PSU to spin it's teeny fan even louder, definitely over the acceptable level, but what did suprise me was the fact that after 30 seconds of combined Heaven and Prime, the whole PC shut off in an instant.

Not at all happy with that result. I know Prime puts an unrealistic load on the CPU, but I would've never thought that the SSP-300SUG couldn't handle that. I guess my system does draw more than 300W in this case. Unfortunately, I have no way of testing that.

Maybe time to invest in a cheap Kill-A-Watt to measure power draw? In theory your setup shouldn't hit the limit but maybe your factory overclocked GPU (and software reporting 105% power draw) put it close?

What temperatures were the CPU and GPU running at the time of shutdown? Is this PC inside a case (re-read thread but couldn't find case info)? Maybe over-temp protection kicked in if the heat from CPU/GPU was getting sucked into PSU?
 

iFreilicht

[H]ard|Gawd
Joined
Sep 23, 2014
Messages
1,348
My case is the prototype of Brevis S, so about as open-air as it gets. You are right that I should probably get my hands on a power meter, though. The temperatures of CPU and GPU were fine, the GPU was at 82°C and the CPU at 65°C, but yes, the heat from the GPU could've been a problem, as it pretty much directly exhausts onto the side of the PSU.
 

Hahutzy

Limp Gawd
Joined
Sep 4, 2015
Messages
169
Just an update on my test unit.

I ended up using a file to grind down 1mm of plastic at the top of the C14 connector.
Also grinded down 0.15mm of plastic on each side of the C14 connector.

It now properly fits into the cutout I have for Hutzy XS.

I emailed SeaSonic earlier this week and have sent them a picture of the problem with the screw head hitting the C14 connector because it is too far up.

I told them they should do the following:

1) Move entire connector plastic housing down 1mm
2) Reduce the width of the plastic housing by 0.5mm from each side.

This will make it conform to FSP's offerings.

They acknowledge that they received my email have and sent it to Engineering.

We'll see where this goes.
 
Last edited:
Top