CoolerMaster G650M 650W PSU (RS650-AMAAB1-xx)

Testing Methodology & Results

Old school testing of power supplies is pretty straight forward, install the PSU into a system that can make a good load on the Power supply break out the MultiMeter, KillAWatt, infra-red thermometer and dBA Meter. What we look for is as close to the rated full load of the PSU as we can take the machine without popping the power supply or expensive computer components. In the case of the CoolerMaster G650M 650W Power supply we are looking for between 550 and 560W for as heavy a load as we want to put on it. Why 550 or 560W instead of 650 you ask, 550W is approximately 90% load for the +12V 624W rated single large Rail. Spitball math 624 X .90 = 561 so at 560W we are a hair below 90% of the +12V rating which is actually an insane load to consider running for any length of time other than for testing. You should figure the anticipated maximum load of your machine and say it’s our projected 560W total draw then you will probably want a 750W PSU because at maximum load (560w) a 750W PSU is about 75% loaded. Follow that wisp of wisdom or don’t but remember components pushed to the brink often take a dip in the lake that is your wallet,

Once we have the rig delivering the desired load at peak we test the PSU with the rig at an idle, then ramp it up in increments and take readings at different voltage draws until we finally have the full load draw we want then we take the final readings, thermals and noise.

To reach our load target we will be using out X99 test rig but with the ultra modern Palit 4870×2 and a reference 4870 in triple X-Fire. A single 4870×2 in our test rig at full load runs 483W at the wall. so adding the 4870 in triple X-fire should put us in the butter Zone.

Well be using Cinebench and ROG Bench to drive both the GPU and the CPU as hard as they will go, whichever provides the greatest load is the winner or we use both while we enjoy an IV of please don’t pop.

Test Rig

Test Rig “Hexzilla”
Case Type	SilverStone Raven SST-RV01B-W
CPU	Intel Haswell E 5820k 3.3GHz
Motherboard	Asus X99 Deluxe
Ram	Kingston Fury 32GB 2400MHz (15 – 15 -15)
CPU Cooler	Noctua NH-D14 (W/2011v3 Hold down kit) TT Water 2.0 Pro TT Water 2.0 Performer SilverStone Tundra TD03-E
Hard Drives	Kingston M.2. SM2280S3 120GB M.2. SSD 2 x Crucial MX 100 SSD 1 x Two TB Seagate Platter Storage Drive
Optical	Asus Quiet Trac BR Liteon DVD Burner
GPU	Palit 4870×2 Refrence 4870 Triple X-Fire
Case Fans	2 X 180mm 700 RPM 18dBA 1 x 120mm 950 RPM 18dBA
Docking Stations	None
Testing PSU	CoolerMaster G650M
Legacy	None
Mouse	Razer Lachesis
Keyboard	Razer BlackWidow Chroma
Any Attempt Copy This System Configuration May Lead to Bankruptcy

Testing Results

Power Output Results(V)
Power Consumption (W)	193W	267W	326W	400W	550W
3.3v	3.41	3.38	3.37	3.35	3.31
5v	5.17	5.15	5.09	5.07	5.04
12v	12.41	12.34	12.19	12.13	12.06

It looks like the CoolerMaster G650M just missed the under 3% mark at an idle, a PSU that holds under 3% isn’t industry standard. The Industry standard is 5% or under, under 3% gives a PSU a shot at a Gold award if all other things, noise, thermal readings and the whole ball of wax are good. Anything over 3% Gold isn’t an option. At 193W we were just over the 3% line but as we increased the load the voltage regulation tightened up and got well under 3%. The voltage regulation on the CoolerMaster G650M is good and exceeds the industry standard.

Thermally speaking at full load we measured the exhaust at 47°C and which isn’t bad for thermal exhaust so there’s no problem there.

Noise wise the fan was clearly audible during heavy load periods and at two feet away our Decibel meter, a Wensn FZ-032 rated at 30dBA – 130dBA we recorded a 37dBA level with the PSU outside the chassis sitting on a neoprene pad. Buttoned up inside the Raven RV01 measured from 2 foot away from the side of the chassis we registered a 35dBA and the fan was still audible but not overly annoying. We prefer silent operation on PSU so the noise level will cost the CoolerMaster G650M a few points.