Seasonic PRIME 650 Titanium Power Supply

PRODUCT: PRIME 650 Titanium
PROD LINK: PRIME Titanium Product Page
PRICE: $159.99 MSRP
Price is at time of testing!

PRIME 650 Titanium – Overshoot Transient Tests
VSB On VSB to 100%, 12V Off to 100%, 12V
O-Scope Shot O-Scope Shot O-Scope Shot

Ok, it’s time for the scoring change since this unit does not improve on the bigger units here. You may remember that up until now, I only score these units on obvious violations of the ATX spec. Some have argued to me that technically, these up-down-up again spikes in the shots do represent spec violations, but because of the way things are worded in the design guide I always felt it could be argued the other way as well. And even if these do represent technical spec violations, you have to also consider that they are not going to damage things because they are well under mean value. So how do I score on that?

Well, here’s how: Intel does say there has to be a “smooth and continuous” ramp to the nominal voltage, but there’s sort of some room for interpretation there, I think, that has some engineers stop and say, “well, that should be adequate to do the job.” I’m not playing that game anymore. For me, “smooth and continuous” means no stopping in the ramp and going all up-down-up again. Some units, I have noticed, have no trouble with that. Others, like this PRIME, have it a little harder.

Which units do a good job with the smooth and continuous part of it? The Seasonic Snow Silent 1050W comes to mind. Let’s see those scope shots again for comparison:

Seasonic SS-1050XP3 – Overshoot Transient Tests
VSB On VSB to 100%, 12V Off to 100%, 12V
O-Scope Shot O-Scope Shot O-Scope Shot

Much better. The standby rail is perfect, and the 12V rail only has the tiniest of spikes. Let’s have another example:

Rosewill Quark 1000 – Overshoot Transient Tests
VSB On VSB to 100%, 12V Off to 100%, 12V
O-Scope Shot O-Scope Shot O-Scope Shot

Good God, that’s amazing. Then again, Enhance is one of the few companies that really know how to nail these. CWT is another, as is Great Wall; as evidenced by my; Corsair SF450. But I’ve seen CWT blow it as well, as we saw with the; Thermaltake 1250W Titanium.

Now, I think a lot of the reason we’re seeing this from the PRIMEs comes from the fact that Seasonic likely couldn’t manage to get this design looking perfect and still get that efficiency or regulation. And if that’s the tradeoff, I’ll gladly take it. But I want to score in such a way that it encourages all OEMs to start improving on results like this because what I’m seeing here is just not my idea of “smooth and continuous.” There are Titanium units out there that look just like the Rosewill’s test results. Need a reminder? Here you go:

EVGA T2 850 – Overshoot Transient Tests
VSB On VSB to 100%, 12V Off to 100%, 12V
O-Scope Shot O-Scope Shot O-Scope Shot

Here’s the bottom line… there is now an S&C point in the performance scoring. That stands for “smooth and continuous,” not that white soul group you listened to in the 70s… Smoggins and Crumbley. If you get results like the EVGA and Rosewill up there, you get the whole point. If you nail it on one or two shots but blow the third, you lose half a point. All three? There goes the whole point. And I think we’re going to see this point going away quite often in the future because a whole lot of units don’t pan out like the EVGA and Rosewill. Oh, and I’m still scoring obvious and egregious spec violations on top of this, a half point at a time.

And I am not unreasonable. I might let units showing a tiny spike in the middle of the ramp go. That Snow Silent is almost a pass for me. But these PRIME units? Nope, sorry, I think the whole point is leaving on this unit. It’s arguably in spec, but it just does not hold up to the same level of awesomeness we’ve seen from the rest of these test results. And no, I will not score retroactively. The two other PRIMEs can keep their perfect 10 scores.

Whew… long page already. Let’s see if the hot tests look as amazing as the cold tests.

PRIME 650 Titanium – Hot Load Tests
Test # +3.3V +5V +12V DC Watts/
AC Watts
Eff. P.F. Intake/
Progressive Load Tests
1 1.02A 1.02A 4.61A 66.0W/
117.7V 92.1% 0.955 29°C/
3.311V 4.98V 12.14V
2 1A 1A 10A 134.5W/
118.7V 94.3% 0.982 32°C/
3.310V 4.98V 12.14V
3 3A 3A 24A 326.0W/
115.6V 94.1% 0.993 36°C/
3.310V 4.98V 12.14V
4 4.5A 4.5A 36A 487.0W/
115.2V 93.2% 0.992 40°C/
3.310V 4.98V 12.15V
5 6A 6A 48A 650.1W/
115.0V 92.1% 0.994 45°C/
3.310V 4.98V 12.15V
Crossload Tests
CL1 12A 12A 0A 101.0W/
116.4V 89.0% 0.978 31°C/
3.306V 4.98V 12.16V
CL2 0A 0A 54A 657.2W/
114.2V 92.3% 0.995 42°C/
3.319V 4.99V 12.14V

* Fanless operation.

LOL!!! Ok, that’s just getting silly, now. The 3.3V rail has improved to the point I almost needed four decimal places to show any instability. It’s at 0.03% right now… a number so good given the precision of my measurements I’m not sure my brain is capable of processing it. The 5V rail has gone nowhere again. The 12V rail is slightly less stable but is still doing 0.08%. All told, we have an average of 0.037%, which is somehow an improvement on the cold tests.

And efficiency is still amazing.

And the fan is still content to spin up a minute into test three.

And I think I might faint if these here scope shots are as awesome as the numbers in this table.

PRIME 650 Titanium – Oscilloscope Tests
Test # +3.3V +5V +12V
1 O-Scope Shot O-Scope Shot O-Scope Shot
2 O-Scope Shot O-Scope Shot O-Scope Shot
3 O-Scope Shot O-Scope Shot O-Scope Shot
4 O-Scope Shot O-Scope Shot O-Scope Shot
5 O-Scope Shot O-Scope Shot O-Scope Shot
CL1 O-Scope Shot O-Scope Shot O-Scope Shot
CL2 O-Scope Shot O-Scope Shot O-Scope Shot

Hmm… looks like this unit has succumbed to the random spike thing that used to plague Seasonic units of the past. And yet, those little spikes are so far inside spec they do not matter a bit. But there are enough of them that I do have to count them.

Even so, that only means we’re at 12mV at best for the minor rails. The 12V is sitting there at 15mV. While I’d like the spikes cleaned up, this unit still gets excellent scoring privileges here.

Time to take this unit apart and see if it’s made of electronics or pure magic. Or both.