Discussion PSU recommendations and power supply discussion thread - Tom's hardware

Page 75 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Yes. ANY RMx unit, when purchased new, is generally a really solid buy. So are the HX and AX variants. Even the TX units aren't too bad compared to a lot of what else is out there. But the RMx units, not to be confused with the regular RM units, are pretty darn decent.
 
Aug 10, 2021
24
3
15
Yes. ANY RMx unit, when purchased new, is generally a really solid buy. So are the HX and AX variants. Even the TX units aren't too bad compared to a lot of what else is out there. But the RMx units, not to be confused with the regular RM units, are pretty darn decent.

sounds good. I was originally going to go with an HX or HXI but they're no where in sight that aren't being scalped so as long as the RM1000x is a solid unit for a basic gaming/office machine I'll snag that. ... It seems that the market has shifted in the last 5 years when the old G2 and G3 Supernovas were king and PC Power and Cooling were putting out monster PSUs (I just retired my last PC Power and Cooling 950 watt beast that has been running 11 years straight)
 
Aug 10, 2021
24
3
15
There are Super Flower Leadex III units which are essentially the same thing but quieter, they're most of the time relatively cheap too, just pay attention where's the seller located, they sometimes ship all the way from China.

I've considered them since they made the 850 watt Supernova G2 I have now, but they don't have the Leadex III's in 1,000 watts. Only the SE's and V Pros which are a not AS nice.
 
Even if MEIC is responsible for the issue with the units, Gigabyte is still mostly responsible in my opinion.

Gigabyte QC tests units before they sell them, so there has to be a problem with Gigabyes per unit QC testing if it allows for so many DOA units etc.

I could be wrong, but I recall hearing the P-GM is a modified platform from MEIC, not just off the shelf.
 
That's a good question, and one that I think around here probably mostly only @jonnyguru or Aris would be really qualified to answer for the most part, but just at a guess I'd say that it is BOTH of their ends. The OEM is responsible for making sure the platform is safe and protections work properly BUT the party contracting the work has just as much of a responsibility to make sure that what they have contracted out works the way it is supposed to and is safe for their customers and clients. If either of them fails to do their job properly, then both are responsible if that product, en masse, is problematic.
 
Considering Gigabyte claims to tests all the units before they sell them, id say the quality control issue is definately on Gigabyte.

There has to be a problem with Gigabyes per unit QC testing if it allows for so many DOA units.

I could be wrong, but I recall hearing the P-GM is a modified platform from MEIC, not just off the shelf.
No idea though if Gigabytes "quality control testing" is thorough, or even adequate for that matter. Their QA process might be restricted to a far more limited battery of tests than what even some of the lesser PSU reviewers are known to do.
 

TK31

Distinguished
Nov 4, 2011
67
9
18,545
with my old EVGA G2 850 being left in my old rig for a gaming machine for my wife (building a new gaming rig with 5900x and and 3080 in it as soon as my turn in the EVGA waitlist comes up) and I was thinking of going with a 1000 watt psu just for a little headroom (I know the 850 should work fine but I wanted the 1000 for spikes and just in cases over the years since it's not THAT much more cash).... it seems like most of the major 1000 watt supplies are being scalped or ooos, but is the RM1000x (2021 or 2018, not sure which is better) a solid buy? The HX series is oos everywhere and not sure about the seasonic ones, especially after seasonic focus series had all the issues a few years with the FM serial series.
I've purchased this exact PSU (2018) about 4 days ago and it is great. the only gripe i have is the included type 4 cables have massive inflexible heatshrink sections where they hide the in-line capacitor. This really wreaks havoc trying to wire it up with tight tolerances. Would recommend a nice braided extension for the 24-pin ATX if you have issues with trying to get this into tight spaces.
 
  • Like
Reactions: MikeyMarin
I've purchased this exact PSU (2018) about 4 days ago and it is great. the only gripe i have is the included type 4 cables have massive inflexible heatshrink sections where they hide the in-line capacitor. This really wreaks havoc trying to wire it up with tight tolerances. Would recommend a nice braided extension for the 24-pin ATX if you have issues with trying to get this into tight spaces.
I have the G2 750 myself. I bought this with the intention of it being THE core unit for my main system but when I realized after the fact that the inline capacitors were a huge problem for custom sleeving the cables, it got moved to a secondary system and I purchased a Seasonic Prime titanium which doesn't have them. But then I have had various time related and health issues and never got around yet to sleeving the cables even though I have all the paracord, shrink tube, terminals and release tools to do it. I might never do it. Just doesn't seem as important as it used to in light of other "life" issues. But, I fully agree that the bulky bundles where the inline caps and cabling are shrink tubed together do make for a little bit of grief when trying to accomplish decent cable management. All things considered, they are still pretty damn good units though. I think the G2 units are better than the G3 units personally. I believe there were some fundamental issues with the G3 units. Not necessarily in peformance or protections testings, but something. I can't remember exactly what the issues outlined by Aris were but will have to revisit that so that I'm not talking out of the side of my mouth.
 
  • Like
Reactions: TK31

4745454b

Titan
Moderator
I would be inclined to say Gigabytes fault. In theory the oem makes xyz to the specs provided. If Gigabyte wanted cheap caps and protections set too high, that's on them. The oem is just doing what they were told. Now if they were told to set the OPP to 120% and they picked 133%, or if they used whatever caps they had laying around instead of the proper ones that's obviously their issue. I'm sure Gigabyte would have sued.
 
Companies don't "sue" their OEMs. They simply make them make changes. But they only make them make changes IF they discover that changes need to be made. If they don't do adequate testing, then, once again, it's on them both. It's on the OEM because Gigabyte ASSUMED they would supply a product that complied with the design that was submitted, but since it was, presumably, an existing platform, they apparently felt they didn't need to do a lot of QA testing in that regard. Apparently they were wrong. So, it IS seemingly both of their faults.

A good OEM would make sure the product was adequate for the outline. A good company would MAKE SURE that what was being supplied with their name on it was actually capable of the expectations of the design they had requested. I don't see much way that either of them can shuck responsibility in that regard. But, then again, I'm assuming common sense which is something that seems to be sorely lacking in the world these days on a variety of levels.
 

Aeacus

Titan
Ambassador
In theory the oem makes xyz to the specs provided.

From GamersNexus vid, we can see that there are up to 3 different components used (e.g filtering caps) within the same PSU series and even the same wattage 750W units have different components in them. <- Is that OEM (MEIC) who changed the components on their own? Or is it Gygabyte who requested different components within the same PSU model?

If it's the former, Gigabyte could sue the MEIC. If it is latter, Gigabyte should not be selling PSUs, if the can't make up their mind which components to use within PSU.
 
Last edited:

Eximo

Titan
Ambassador
Don't have to sue if you have a well written contract, they either make it right, or then the suing starts. In this case, they probably had a clause about part source being up to the OEM as long as they met the technical requirements. That they turned out to be subpar is really on the component manufacturers. (Though one would have thought significant testing would have been done, or at least sampling, to check for failure rates on the finished product)

Business sense > common sense when money is involved. They don't want to take the loss, so they are pushing their vendors to sell the things despite the negative press. That Newegg, in particular, goes along with it is pretty stupid, but then again they have become a marketplace of crappy products. I knew it was bad news when they started selling appliances...

My guess would be Gigabyte set a price target and left them to it. Gamer's Nexus also mentioned that part availability may have been a driving factor. Shortages everywhere, if they had to use three or more sources for parts to meet the delivery schedule, that is what they had to do.
 

Rogue Leader

It's a trap!
Moderator
Companies don't "sue" their OEMs. They simply make them make changes. But they only make them make changes IF they discover that changes need to be made. If they don't do adequate testing, then, once again, it's on them both. It's on the OEM because Gigabyte ASSUMED they would supply a product that complied with the design that was submitted, but since it was, presumably, an existing platform, they apparently felt they didn't need to do a lot of QA testing in that regard. Apparently they were wrong. So, it IS seemingly both of their faults.

Thats not true. Having worked in Manufacturing I can tell you this type of dispute happens. The OEM says they made it to spec, the company says they didn't, there are fights. Of course you prefer to resolve it without lawyers, but it happens more than you'd expect.
 
  • Like
Reactions: 4745454b
My guess is the reason the components are swapped out is due to the shortages and components chosen in the first place. Of course there could be various other reasons too.

From what I have heard and seen, even name brands are being forced to swap out components as component shortages are impacting capacitor selection etc. For example, the EVGA GT650 originally came with a Hitachi bulk cap, but due to shortages, they were forced to change to a different manufacturer for the newer units. Another example is that the EVGA BR500 used to use an Elite 105c bulk cap, but now uses a CapXon bulk cap unfortunately.

However, if whoever picked out the components in the first place went with reputable manufacturers instead of cut rate components from non reputable manufacturers, likely a LOT fewer components would have to be changed out.
 
Thats not true. Having worked in Manufacturing I can tell you this type of dispute happens. The OEM says they made it to spec, the company says they didn't, there are fights. Of course you prefer to resolve it without lawyers, but it happens more than you'd expect.
Yeah, no, I get that. I know that it "happens". I didn't mean it exactly the way I worded it. I just meant that usually they will want to work with their OEM to correct any problems, especially if it's an OEM that they work with a lot on many different products, rather than to just sue them offhandedly. Obviously if the company says "we have a problem and it needs fixed now" and the OEM doesn't play ball, they are going to sue. But they will almost always want to try a different strategy first UNLESS it's something where they themselves are being sued in which case, well, stuff rolls downhill.
 

Juular

Respectable
BANNED
Mar 14, 2020
1,061
258
1,940
There's such thing as Qualified Vendor List. You get one part in design but it's unavailable rn or the lead time is too long, so the OEM replaces it with the next best thing. It's very widespread, you would not find a single PSU model with manufacturing volume higher than a dozen per month with the very same components across all samples. The part switching is not the problem here (although usually the part switching is limited to capacitors and smaller less important components), the problem is the failure of MEIC and Gigabyte to properly validate the changes to the design. And then, after it because obvious that there are problems with the design after components swap, failure or unwillingness to fix the problem, i.e recall affected batches (at the very least try to do that, it's not as simple as it sounds). And do that before all this became a public knowledge, so essentially, they're valuing money more than their reputation and the convenience of their clients.
PS: Gotta thank them tho for proving the point that a random 80+ Gold PSU isn't going to be good just because it passed some arbitrary efficiency levels once.
 
Last edited:
PS: Gotta thank them tho for proving the point that a random 80+ Gold PSU isn't going to be good just because it passed some arbitrary efficiency levels once.
This is maybe one of the most important factors here. I can't count how many times I've heard people here and elsewhere say that any unit with a Gold, Platinum or Titanium efficiency rating HAVE to be good, because the only way to achieve those ratings is by using quality parts and using a good quality platform. Um, right.
 
Aug 10, 2021
24
3
15
oooo, one other thought. Does the 2021 version of the RM1000x come with the optional 4 pin CPU power connector or is it absent like the 2018 version?...I noticed in the visual I could see 2 of the 8 pin CPU 4+4 connectors, but not a dedicated 4 pin connector to add on to that "optional" port................................... or am I just being dumb and the 4+4 8 pin connector can separate and I can just use one of the 4+4/8 pin connectors as well as a second one and split the second 4+4/8 pin connector in half and use one of those in the optional 4 pin spot..

since I've never had to use a board that had that option and haven't bought a PSU in the last 5 years I didn't know if the mobos required a single dedicated cable coming from the PSU plugged into a SINGLE slot on the modular power supply with an 8 pin AND an optional 4 pin connector on it (like 6+2 PCIE cables) or whether you would just use 2 CPU cables, both plugged into their own individual slots on the PSU itself, with one going to the regular 8 pin connection, and the second CPU cable having the 4+4/8 pin connector split in half and using one of them to plug into that optional 4 pin slot

basically I just wanted to know if there was a special dedicated cable that came with the 8 pin and a 4 pin dangling off the side that is needed to use the optional 4 pin connector on these newer boarods, or whether you would just use 2 CPU cables and split the second cables 4+4 pin connector in half and use that way as this is the first time having one of these here new fangled mobos with the extra connector (yes, I know it's not necessary except for extreme over clocking, just like to have all the info at hand in case I ever do decide to go nuts with over clocking my 5900x)
 
Last edited: