Reminder not to open your own power supply unless you are a trained technician who understands the dangers. We're really excited about this content. Patrick Stone has been working for weeks to learn more about PSU testing & reviews, and in this one, he teaches about how power supplies work by using the Dell G5 PSU as an example. It's surprisingly good! This sort of gets back more to our roots of doing some educational content. I learned a lot by watching back the section that Stone hosted, and we hope you all find it interesting also. We still have a lot to learn with PSUs, but give us some time and we'll keep advancing. We are new to PSU reviews and are still in the stages of figuring out presentation, style, format, etc., but we've had fun working on this one and look forward to doing more. Thanks for all the interest in PSU content! You can see the full saga of the Dell G5 5000 - Part one on component choice & support: th-cam.com/video/4DMg6hUudHE/w-d-xo.html Part two on benchmarks & bloatware: th-cam.com/video/5N7aYtkzKJc/w-d-xo.html We even reviewed the included GPU in our standardized GPU test bench: th-cam.com/video/HFaJZq13tr8/w-d-xo.html Back-order a GN Anti-Static Modmat: store.gamersnexus.net/products/modmat Buy GN's Red & Black Mouse Pad: store.gamersnexus.net/products/gn-charge-redblack-mousepad Blue & Black Mouse Pad: store.gamersnexus.net/products/gn-component-blueblack-mousepad Wireframe Desk-Sized Mouse Mat: store.gamersnexus.net/products/gn-wireframe-mouse-mat
Never seen stone before. Give him some space on the channel to explain things like this, he's untapped potential for the media side of things, if I'm not mistaken about his absence in the videos.
Really liked this. Especially explaining each part. I have a corsair 1500w psu platinum certification and in the icue options it let's me choose between single and multirail. Never seemed to make a difference (I'm not using near the capability of the psu even with the 3090 or my crazy water cooling). Now I have an idea on what to research more. Maybe a video on why I would select one over the other? When I should? What are the consequences? Etc. Great job as always guys
@@jamieh4086 each rail on your psu will have a rating of X watts. Unless you need more than X watts on one connector, stick with multi-rail, as this is safer if somthing fails short-circuit.
To prevent any potential confusion for newcomers to the electronics design field please call it a square wave. Rectangular is technically correct, but there is no electronics engineering textbook that calls it that. It is always a square wave with some amplitude at a specific frequency.
@@GamersNexus also please make sure you explicitly mention in the video not to disassemble a PSU if you don't know what you're doing since the built up charges in the capacitors can potentially kill you. As great as this content is, I really wouldn't want some young kid getting hurt tinkering with them.
It's called square not because it's a polygon with four sides, but because the intersection between parts of the waveform are square (i.e. at 90 degrees to one another), in the same way you'd call a room "square" when the walls and ceiling meet at 90 degrees as they are supposed to. You don't need to cede ground on the use of "rectangular" because that's just wrong. There's no such term to describe an intersection of 90 degrees.
These types of supplies are also used in Dell's poweredge server line, hence the weird form factor. It makes them swappable. And servers is something you cant cheap out on as a brand.
Can verify the PCB design is very similar to that of a Flex form factor PSU. Appears to be potentially a custom housing though. Just took apart a Dell PowerEdge 495 Watt 80+ Platinum PSU Model: F495E-S0. The one in the video is very similar with some component differences. Clearly based on the same platform though.
@@GamersNexus looks like a PE server PSU with wires soldered on instead of a hot swap pin set. I have a PE R710 PSU up in my bin of parts, and can take a look at it, it is also made by liteon im sure. I can send one to you if you want, but they have some other juju on them that allows them to both hotswap, and also can communicate with the server BCM, for full offline control and monitoring.
Marjie S - Thats just not true. Expecially on this particular "retail consumer" product line (PSUs from Dell). The Poweredge Servers have different PSUs made from Delta. Your comment is like looking at well-built car tires and saying just cause theyre round and made of rubber, they can be used on Semi trucks. Not even close. Theyre only swap-able within their own product category lines - Vostros, Optiplex, Latitudes, etc. The only reason they have a "weird form factor" is for Proprietary sales. Form-factors and swap-ability are mutually exclusive by choice from Dell. The server lines are different products and have their own requirements for components (lights-out, roll-overs, etc) - especially with Xeon, Epyc, and Nvidia AI units in those boxes.
Just some minor corrections: The chokes and caps at the start of the primary side are mainly for filtering out the noise from the power supply from getting back out on the power lines, rather than smoothing the power coming into the power supply (that's done mostly by the bulk capacitor, but also more finely on the secondary side). Also the "double-coil" inductor next to the bulk cap isn't mainly for current-limiting, it's known as a common-mode transformer (or common-mode choke, both terms can be googled), and is also for keeping certain forms of noise inside the PSU, in this case common-mode noise (hence the name).
@@sajivsingh5601 yes, but efficiency doesn't matter if the power output is garbage in quality. That quality is very important for sensitive electronics
Actually, one of the most important things that a power supply can do, is to NOT blow up your system if there is a failure. I had a friend come over a few weeks back with a PC that was no longer turning on (PC was about 10 years old). It had an Antec power supply, and one of the 12V rails was briefly spiking up to almost 13V on powerup, but the built in overvoltage protection would immediately shut down the supply. He had been having problems with phantom shutdowns for a few months, but it had become so common as to be unusable a few days before he brought it down for repair. No other hardware was damaged, it was just a case of replacing the power supply (and reinstalling Windows, as the image had been corrupted to the point of bluescreening with the random early startup power losses.) PSU's need to have overvoltage fault protection like this. Back around '96, for my first new PC I purchased a generic AT pentium prebuilt (from a local shop, everything was off-brand in that system except for the processor and hard disk); several years later it failed, and the power supply decided to commit suicide and take the motherboard, ram, and video card along for the ride. After that point I decided I would never use anything but Antec or Enermax power supplies for anything else. I subsequently have had some of these power supplies die, but they have never taken out any other components along with them. I would guess that every power supply has this protection built in these days (since its part of the ATX PSU spec), but it would be nice to test it to see if its working as intended.
Thanks! We're really excited about it too. It's a new review category for us. Patrick (Stone) has been working hard at this testing for PSUs and we're finally in a stage to start sharing. I'm personally excited about it because PSUs are a weak area in my knowledge, and I can learn from the work he's doing. We'll be reading comments closely to keep improving as we figure out how to present the data and how to analyze our findings.
@@Noah-lj2sg Very true, and it also keeps the market sharp and competitive. I am saying this while listening to music via a dac which has been measured by a youtuber.
"Here we have some screws. Here are some wires. Here we have the active power factor correction and the is the rectangular waveform and the lower voltage smoothing MOSFETs" That escalated fast.jpg
This is exactly how many university lectures start. "So, 1+1 is 2 .... (suddenly some partial differential equations) .... and the rest is trivial, so just do it at home if you want to know what's up with that."
You all definitely earned a Titanium 80+ certification for efficiency in explaining power supplies. Have 0% interest in this Dell PSU, but this was the most interesting video ive seen in while. Thanks, cant wait to you guys progress on this type of testing eviews.
@@GamersNexus In all honesty, buying these things standalone from dell is difficult. You need a dedicated dell business account rep, they'll ask for a service tag of the model you need a PSU for because they have >100 different PSUs in all form factors, and then they'll tell you multiple times they can't find it before suddenly finding it and shipping you the wrong form-factor twice before getting the right one, all the while offering to sell you a new machine instead. However, if you have the same machine under a "ProSupport Plus" warranty, they have no problem finding the exact PSU and shipping it overnight.
@@futureb1ues Yeah, my company just scraps Dell systems when their power supplies blow out. It's nearly impossible to get replacement parts from them. They usually fail during power surges. Sometimes when someone accidentally ruins 48V through a serial or USB port.
@@futureb1ues That's the whole idea behind Dells proprietary parts. To keep government and business into Dell's little ecosystem. Sadly your district probably wont allow you to swap systems to standard parts. The one I worked for did that, bulk ordered parts and had junior high and high school students put them together with IT then slowly replaced the elementary schools PC's over a couple years.
It depends if you are referring to current or voltage. Most forward, half bridge or full bridge buck isolated stepdown switch-mode power supplies (apart from discontinuous types such as flybacks) have a PWM squareware voltage measured with respect to primary ground on the primary input to energy storage inductor and a contiunous sawtooth current in that inductor.
Having experienced two catastrophic PSU failures in my life (One starting a fire, the other killing every platter drive in my NAS to the tune of $7000 in hardware), I really appreciate that someone is now doing comprehensive PSU testing. Just FYI, the garbage PSU's I was using were branded (Gigabyte for the one that caught on fire, Rosewell for the one that died and over-volted/ or over-currented frying all my hard-drives). In both cases, it seems that the transformers broke down, from the limited autopsy that I did. Since then I've learned my lesson and I have been buying way-overspec'd EVGA power-supplies. Assuming EVGA knows what they are doing, but I guess we will soon find out...
Brings back memories as a tv repair technician. I've had to repair numerous tv power supplies that work just like the one described here. On another note: Dell computers are only useful for business applications where they will be thrown out and replaced with new ones as needed. They are not geared towards the general public as they can't be upgraded, which as you described, is a shame.
This look like a propietary version of server PSU. This is probably used on Dell server products which is why it's actually good. Dell business line is pretty good.
Sending this comment over to Stone so he can take note of it. I'm actually not sure on the answer for that -- he did a lot of testing, so maybe, but I'll ask him to look into this comment for future pieces if not!
As a PCB designer by day PSU design is mundane but incredibly hard to get right... If these are built to UL or CE there are decent creep age and clearance requirements...
Just want to mention it again how happy I'm to have found your channel after HardOCP folded. It really replaced everything I used to get from that site. Love your informative channel. Appreciate the hard work from Patrick Stone. Thanks and back to you, Steve. Never knew Liteon makes PSU too. I remember fondly though how great their optical drives were back in the days. I have always felt Dell systems are well built and hardly breaks. I often recommend Dell Optiplex line for my clients. Now it's great to know one reason why.
Yeah, i remember back in the early 2000's i had so many cd and dvd players that failed, most of them LG, Samsung, TEAC etc. But i had a LiteOn DVD-RW that lasted me for years, and i've written a lot of discs with it. This left me with a very good impression on this company, and heard many people having similar opinions.
Please please more in depth PSU reviews. Now that Jonnyguru is gone, it’s tough to find a real in-depth PSU review with actual quality testing. The skill and equipment required aren’t common. Great job guys!
When testing voltage regulation you also need to rapidly and randomly vary the load on the output - TiN explained it once when you were visiting him and KINGPIN
I normally listen to Steve to help me sleep, no joke, I sleep to TH-cam videos every night (I have really bad tinnitus). But this other guy, oh man, he knocked me out in seconds...
a few tests are missing: transient response, dynamic load and the time the psu can supply the components with power when the primary side has an interruption. Otherwise good first effort!
Dell has probably been using PSUs like this for awhile, probably costs too much (or they just have too much stock) to switch toe AVX12VO. At some point that should tip in favor of switching though.
I have a feeling the PSU& motherboard are used in dells' business/server options and it's just cheaper to make the home PC fit what the majority of there business is.
Nah, dell likes to lock their business customers into their ecosystem. They're also stingy about shipping replacements, they'll always try to sell you a new system.
Its more evil than that. People like to drone on about Apple locking you in but Dell really locks you in and it must be profitable because HP has gone the same way. I have an HP Z220 (i7 3770) that I got referbed and turned into a low end gamer/Hackintosh and it has all standard ATX parts. Fast forward to today and they are doing the same thing Dell does with custom PSUs, off size motherboards and proprietary connectors. The entire supply chain has a vested interest in preventing reuse of electronics. AS long as Im careful with it, my old HP will be usable for decades. PSU dies? Get any ATX standard as a replacement. CPU dies? Drop in another 3xxx series. Lightening fries everything? I can still put new guts into the case and call it a sleeper. Each time I do that I am denying some part of the supply chain a sale so they are all more than happy to go along with Dell, HP and others in making non replaceable parts. Im a hardcore Apple guy from back in the Apple ][ days and Apple is just as guilty but the difference is that most of the time Apple is doing it because the entire system is non-standard. Its not like they are intentionally going off ATX to make an iMac. No one is complaining about Dell or HP laptops having custom connectors either. Heck Ill even give them a break on Alienware and Omen systems in cool looking cases that can only happen with custom parts. But when you take a bog standard box and intentionally cripple it by going ever so slightly off ATX, you are just showing how evil you are.
@18:50, changing the vertical scale of the 'Power Watts' in your chart to a logarithmic scale might show you a more meaningful comparison between power in and power out across 0% load and up. Just something you may try and see what it looks like.
If it doesn't have a discharge resistor, sure. Since it's being sold by a huge company with liability concerns, odds are they didn't decide to skip that particular 5-cent component.
@@AndrewFremantle Did you watch the video about the rest of this PC? And the way they cheaped out on this PSU by making it proprietary instead of using the 12VO standard? Odd are they _DID_ decide to skip that particular 5-cent component.
@@AndrewFremantle It wouldn't surprise me that cheap PSUs skip that critical safety resistor, however I spotted it in this particular PSU, shame that the non standard format makes it useless outside of this line of systems.
@@AndrewFremantle "Small" correction, 0.01 cent component (yes they are really that cheap). And i have personally never seen any power supply, not even the cheapest backstreet chinese ones that arent anywhere close to being FCC or CE compliant that dont have discharge resistors on their bulk caps. The times where this was relevant was 40+ years ago, when SMD wasnt a thing and a resistor did cost 1. a couple cents to a couple dozen cents and 2. occupied some worthwhile physical space. But at that time this wasnt nearly as relevant, as switching power supplies werent really a thing, thus no high voltage caps in most devices.
This is what frustrates me the most about dell their engineers clearly know what they’re doing but it all gets ruined as usual by suits who have no idea what they’re doing making weird form-factors for arbitrary reasons and slapping predatory sales tactics for warranties on sales if they actually stuck to normal standards and didn’t send you notifications in your sleep asking you if you want a warranty it would actually be a pretty good buy honestly.
Actually there is a reason why these kinds of things exist, they seem to be re-used components from their enterprise hardware and server rack hardware is some of the most non-standsrd stuff you will ever encounter. Most brands moved away from desktop style PSU because single space servers would never fit them. The reason why is clearly shown here, the PSU is very efficient because when you're putting equipment at datacenter scale, even small percentages in efficiency can result in significant gains when you're talking about thousands of servers. The controversial decision is more along was it a good decision to bring this to their consumer devices.
I have one 250 Watt TFX Chieftec PSU and two 300 Watt TFX Chieftec PSUs. The 250 watt is not rated because it has low PFC but all three TFX PSUs are almost silver (90%) efficiency. Chieftec also sells a 500 Watt ATX PSU with terrible efficiency (less than 80 Plus white). I rarely see crappy TFX PSUs. I would love to get the Seasonic 300W mini-TFX PSU for a SFF PC.
I love the idea of you guys doing PS testing and reviews. They are IMO the most overlooked piece of hardware when specing out a PC build but often times one of the most important. Couple that with no one in the PC reviewer space doing any thing like this makes you guys unique. This is awesome guys! Good work!
Yeah, looks like a 2U psu. For them i think it is cheaper, server PSU specially from last generations are really cheap, they probably have a couple hunded thousands doing nothing in some werehouse.
I've been ready for this content, I'm excited to have you guy's teach me about power supplies. I've honestly learned a lot from your channel and a couple others it's cool to see how things work.
Thanks Steve. This is when I whish we could "like" videos more than once. The value we get from this channel is awesome, the work you guys do is quality and appreciated, I'll say it again easily the best PC hardware channel on the youtubes. As a fellow electronics tinkerer Patrick's presentations are great and it's refreshing to see a tech youtuber speaking my language, well done. Your videos are a break of sanity for us in an insane world, and you have me rolling with the Computex references. Thanks guys. Back to you Steve.
Patrick please read this! Please do research on how APFC works. To put it simply it is a boost converter, consisting of the yellow ferrite inductor, diode and mosfet. The rectified AC is only slightly smoothed by a ~1uF capacitor and then the boost converter raises and lowers its PWM to consume more and less current based on the voltage waveform of the AC, that way it can fake a unity power factor as best as possible and also step up the voltage to ~400v to allow higher efficiency and universal voltage capability. The inductor is not a simple "current limiter" its the inductor that the boost circuit is using to store the power.
For future ATX PSU reviews, I would additionally like to see measurements of noise under certain loads. And maybe comments on cable quality, length and stiffness - important factors when building and cable managing.
This is an interesting new category of content, I hope to see you do SFX PSU testing too just to see what kind of compromises must be done for the smaller form factor.
@@le_travie7724 I hope they get rid of it at the moment intel produced another weird keynote with more and better meme-potential snips. You can literally see it.
This was really refreshing to watch. I used to spend all my time in mid 2000s looking at PSU reviews, just to find that golden cheap nugget with Seasonic on the inside.
We need real power supply testing videos. This is one of the most overlooked parts to building a pc and nobody really does this on youtube. Everywhere on the internet people just try to scare you into buying the most expensive power supply you can buy. It would be really nice to know what is the best power supply for the value and why. Right now everyone just recommends the corsair cx and if you read comments on reddit or other places every other 80+ bronze power supply is going to blow up your system.
There's quite a few other bronze units I've seen recommended on reddit though. XPG Pylon, EVGA BQ, some CM MWE variants, Corsair CXM and CXF, to name a few. And it depends on the budget. The CX is possibly the best bronze unit but it costs more as well.
@@downnnnnn Tripping issues mean that the protection circuitry works. That’s a plus in my book. Current spikes are supposed to be handled on the card side. That’s a card fault rather than a PSU one. The prices can be a bit steep but that is the cost of quality components. Depends on your priorities.
Love this addition to your test suite. As many have said, looking forward to more P/S testing/reviews. Just a FYI, when I was learning how circuits work it always helped me to have a block diagram with an overall view of the stages of the circuit so that I could go from general theory to detailed. This may help others to better understand the overall view.
As an electronic professional I thought Patrick did a great job explaining the basic functionality of all the circuits needed to get a computer, or pretty much anything, working these days. I had teachers that knew less than that and couldn't explain a psu to save their lives. 😂 Next stop: budget firecrackers, extinguished with LN2!
Hey Gamers Nexus Team, for power supply testing videos where you open the unit up - a general "do not try this at home" warning for viewers not to open their own power supply due to the possibility of injury or death from electrocution I think would go a long way and be the ethical thing to do, even if all your normal viewers would already know to do this
I agree, there should be a warning in the video. I had no idea how dangerous a psu could be until around a month ago when I stumbled upon that information. Previously I just thought you should not tinker with them since the risk of breaking something that could become dangerous once power got plugged in again.
oh yes - show me more PSU testing! I am quite fascinated with the operation of a PSU so this video helps reveal more of what happens inside that component. Patrick Stone is well spoken and his explanations were easy to follow. Your testing seems well done as always. Looking forward to the next video on this!
Haven't seen much of Patrick Stone to be honest but really appreciate his work here. He has a great way of talking and a likeable personality which makes viewing a pleasure. Great job my man, keep up the good work and looking forward to learning more from you ☺️
Yep, they clearly missed it. And so did I. Based on their testing results I’d say it barely passes if you take into account tolerances and potential measurement errors. Remember that only a proper lab with calibrated equipment, trained staff and vetted procedures can assert that with any level of certainty.
@@rdoursenaud 1. Maybe you missed it, but you were watching a video. It's not your job to check every claim they make. They worked for hours with that device, I am still surprised.... 2. I think, the rating is missing. Steve only told us "It is close to 80Plus Titanium", but did it reach "Platinum", "Gold" or whatever? I think, that would be helpfull. But this video is to practice the procedure -> next time!
@@manfredbans3975 1. It’s always my/your job to check any claim made! How else would I prevent bullshit from entering my reasoning? "Trust but verify". 2. The answer (and Gamer’s Nexus mistake about the unit not being 80 Plus branded) is at 18:33. The measurements meet (exceed) Platinum level and are not quite Titanium. So the unit exceeds every 80 Plus supersets except the most stringent one. On the other hand, I find this metallurgy analogy confusing and unnecessary. This benefits more Ecos Consulting’s bottom line than the consumer. Hence my flawed guess in my previous answer (I confused the levels).
Your ads for the EVGA Z20 and then when Linus had that video where EVGA was doing 50% off on them convinced me to buy one. I can say, I'm not disappointed.
Very important thing not mentioned in the video: PSU capacitors store a LOT of charge. Don't take your PSU apart for any reason unless you are sure you have safely discharged them. They could literally kill you.
@@ufoisback5088 Well it probably wouldn't hurt if nature and fate were able to hand out a few more Darwin awards these days. If you don't cull the stupid from society then they propagate more stupid which then adds even more stupid and you soon end up with current society where you have people who need to be explicitly informed that "dangerous thing is dangerous."
Great video. I am an engineer and understand electronics but still really enjoyed his description and testing. Looking forward to more of these on one of my favourite channels.
It would be better to have a video dedicated to explaining how a PSU works prior to posting a PSU review rather than putting both in one video. The explanation part of how a PSU works here is quick and dirty and would benefit having a longer video focusing solely on this topic as it is interesting enough to be featured on its own.
I know how difficult (and expensive!) PSU testing is, but there's simply not enough reliable PSU tests out there so I'm pretty excited for this stuff. Great work.
This is actually a new thing for me, since all the prebuilts I've ever seen always carried the crappiest PSU that were humanly possible to be manufactured. Shame that a good thing, for a change, is spoiled by non standard specs.
@@JohnWiku most if not all oem pc's comes with high quality psu's. Sure they may look like generic junk on the outside but those badboys are on par with your regular 80+ bronze at minimum. This is not a new trend, its standard industry practice.
this was quite good. i'd recommend a viewer warning to be wary of disassembling power supplies, since those caps are rather massive, and could certainly cause some serious injury. i'm looking very forward to new content presented with the GN clarity, especially since PSU reviews can be particularly dry and difficult to parse. it might be worth considering a basic "how PSUs work" that can always be referred back to as well, with a few graphs showing how the power waveform gets affected throughout a simplified PSU, and what is and is not desirable.
Thanks! We'll probably do one of those once we get more experience with what sort of stuff needs more clarification. We'll base that on common questions in comments!
@@GamersNexus also it’s worth including a block diagram in the PSU video (as suggested by @neobleidd) and explain some of the terminology (e.g. PFC) in there, as they are music to my ear (a trained electronic engineer) but as far as I believe not to most people.
@@GamersNexus welp looks like LTT might have beaten you to the punch (^: if they've got "turbo nerd edition", i fully expect a gn style "ludicrous nerd edition"
Great content. The rundown on the function of the internal components was awesome. Part of my job is designing flyback power supplies for industrial control modules. Computer power supplies are also typically flyback topologies. Flybacks are great because they provide isolation and the ability to easily generate multiple voltage rails. The filtering on the AC input side of the power supply does prevent noise from making its way into the power supply (conducted susceptibility), but just as important is preventing harmonics of the switching frequency from making it OUT of the power supply and back on to the grid. This is part of the conducted emissions compliance that the power supply must meet before getting the CE mark or other regulatory approval like FCC Part 15. Switching power supplies are extremely noisy, especially at 100's of watts. When the MOSFET on the primary side of the transformer turns on and off based on commands from the PWM controller it generates noise and that noise is suppressed by the inductors, chokes and caps. Those MOSFETs on the secondary side of the transformers are interesting. They must be in an "ideal diode" configuration. Diodes have a typical forward voltage drop of 0.3 to 0.7 volts. This is a big deal for efficiency because this voltage drop multiplied by the secondary current (500W / 12V = ~42 amps) is the power lost in the diodes. An ideal diode is a circuit using a MOSFET and other components to act like a diode, but with 0 forward voltage drop. This increases the efficiency of the power supply by reducing losses in the secondary rectifier diodes to near 0. Less losses mean less power in for the same power out. I think your testing is good. I would like to see a step load test to measure transient response. I don't know if the Intel spec covers this, but basically you load a power supply to some low-ish % or rated load (33% lets say) and then switch the load rapidly to a high current (say 66%) and observe the response of the voltage. It will initially dip, and then the PWM controller will increase the duty cycle to the primary MOSFET to output more power. The voltage typically overshoots the setpoint (12V) and then the PWM controller reduces the duty cycle and the output voltage settles. You want the power supply to respond quickly (minimize the dip) with minimal overshoot. Along with this testing it would be interesting to know the switching frequency. You could measure this on the output side of the transformer with your oscilloscope.
Not true! Their workstations before Sandy Bridge (Core 2nd generation +) used standard ATX. Take any Precision workstation prior to Sandy Bridge - t3500, t5500, t7500 , t7400s etc could be swapped with PSUs off-the-shelf. HP and Lenovo were the same. Dell wised-up.
@@IIGrayfoxII ??? Color coding wires uniquely doesnt make it NOT standard ATX. If that was the case, all third party aftermarket PSU makers (EVGA, ASUS, PowerColor, Corsair, Enermax, Seasonic, etc) would NOT meet "your" definition of "standard ATX". C'mon stop trolling. Go learn a thing-or-two about what makes a PSU meet standard ATX power requirements. LOL
@@staryoke The ATX color coding is apart of the standard. Yellow is +12v Red is +5v Orange is +3.3v White was -5v(no longer apart of the ATX2.x standard) Blue is -12v Green is PWR_ON Gray is PWR_OK Black is ground Purple is +5vSB High end PSU that we gamers get technically do not meet the standard due to this. But when it is not a high end PSU but a basic one. having orange be +5 and purple be +12v will throw you off if you're unaware of dell doing this
@@IIGrayfoxII "High end PSU that we gamers get technically do not meet the standard due to this." LOL! C'mon! I think you should put out a trade paper stating this and listing the PSU makers by name. First you would receive a cease-and-desist for libel and slander and then your ass would get sued into the ground - and rightfully so. Clearly youre not an electrical engineer, because you wouldnt have made such a stupid statement with a straight face - unless youre just trolling! The colors of the wires are an industry suggestion, not a requirement. More important is the gauge and quality on hand (at the time of manufacturing). Color is secondary to those requirements. Contact any of the aforementioned manufacturers, they will tell you that.
I've been looking forward to these kind of videos for a long time now, no other youtuber does power supply testing, let alone this in depth and in such an informative way! This will go a long way to help show which power supply brands are trustworthy and what to look out for.
Best damn power supply video I've ever seen. While I know the words he used I am too far removed from my classes to remember how it all works. Great explanation, Patrik Stone.
So glad you guys are adding PSU's for content. Jonny guru and his website has been out for some time and I don't think we could ask for a more reliable stand-in.
This format of PSUs is common in server machines and, as such, they tend to be good quality and made by reputable OEMs such as Delta, for example...even if they look "uncool" for modern hardware beauty standards. Dell, also being in the server business, probably uses some of the hardware they comission for their servers in prebuilts too, like this PSU. That practice also has an ugly side, like weird layouts and proprietary hardware we see in their (and HP's too) prebuild desktops...
It's about double the width of the server line, probably to reduce noise. Also with no pcb card connector. As far as internals it looks completely redesigned vs those. e.g. th-cam.com/video/dnD-zaOznRo/w-d-xo.html again, probably for noise.
Server PSUs are normally rated for much higher power delivery than the nominal 500W output of this Dell desktop power supply -- for example a twin-EPYC 2U server will have multiple redundant 1400W PSUs. Modern manufacturing lines can produce small runs of "custom" units in a cost-effective manner based off existing circuit designs that simply need a bit of tweaking to meet the desired specification and this is probably the reason for the design of this particular PSU. Badly designed power supplies with iffy components that fail hard can cause system damage which leads to warranty repairs, product recalls and in the worst case fires, explosions and interruptions to TH-cam video recordings. The small cost of designing and manufacturing a decent and safe power supply for even a lower-end system is outweighed by the bad press and downstream costs when the makers get it wrong (see Gigabyte's rebadged explodey-bricks for a recent worked example).
Really appreciate the PSU terminology education. I have heard the term "rails" for a long time and never really understood exactly what was being referenced.
yeah seriously. don't just go around poking at bare PSU. touch the wrong thing and you might get zapped from a high voltage that might be left in some caps ot something.
@@EmergencyChannel bold of you to assume I've never been zapped by 220v AC or got a nasty burn from dropping a soldering iron. These things need to be handled with care because it hurts a lot when you don't
Work in power electronics (specifically power FETs), and will back this as a great review! I am sharing this with our summer interns because it's a great physical example of what our customers want to make! Two points of feedback: - Block diagram during the physical dimensions review of the staging would be nice, but that's probably too much for most people - OCP and OPP are both time dependent. They should trip in some small amount of time (like microsecond range or less). I would be interested in knowing when the fault trips and how long it takes. Since yall were slowly ramping up to the fault detection, I would be interested in a verification by jumping straight to that point or slightly below.
I don't know if I would call the Earth Dreams engine reliable at all considering all the issues people were having with severe oil dilution problems because the engine doesn't heat up enough on short trips. And then the fix Honda deployed to those people basically killing the amount of heat the heater could output in order to try to get the engine to proper operating temp faster.
Thanks for the detail on the circuit paths. I would have never expected that there were two transitions from AC to DC. I guess cleaning everything up, then synthesizing AC to pass through an isolation transformer makes sense. This would make a good 12 VDC supply for a test bench after one is done using it in the computer.
Showing the ripple when it begins to show up in the oscilloscope waveform would be a great addition when testing more PSUs in the future. Some people don't know how it looks or why it is bad. Those waveforms you guys showed up don't mean much.
Great tear down, the PSU is probably not the first components considered when buying a prebuilt, but it’s one of the most important. Can’t wait to see future PSU reviews.
Presentation was good on the info. I would say you can simplify the break down going over the psu internals to just the important parts and quality of parts used.
Naturally having a standards compliant component in a system is totally unacceptable. Everyone knows that (duh). As such they had to modify that slightly so it wouldn't meet the standard anymore, even if it costs extra.
Love this content. Keep up the great technical analysis. One thing not mentioned was the whole reason to move to 12v only PSUs is an increase in efficiency due to the simplified design of eliminating the 5v output. Google has been using 12v only power for long time in their data centers.
Most excellent job on reviewing the PSU from Dell (especially given the non standard nature of the plugs) I really appreciated Patrick's walkthrough of the powers route through the internals of the PSU and the presentation as whole was good and well balanced. Not too high level nor so technical that a regular "Joe" like myself, can't keep up. Great job 👍
Reminder not to open your own power supply unless you are a trained technician who understands the dangers. We're really excited about this content. Patrick Stone has been working for weeks to learn more about PSU testing & reviews, and in this one, he teaches about how power supplies work by using the Dell G5 PSU as an example. It's surprisingly good! This sort of gets back more to our roots of doing some educational content. I learned a lot by watching back the section that Stone hosted, and we hope you all find it interesting also. We still have a lot to learn with PSUs, but give us some time and we'll keep advancing. We are new to PSU reviews and are still in the stages of figuring out presentation, style, format, etc., but we've had fun working on this one and look forward to doing more. Thanks for all the interest in PSU content!
You can see the full saga of the Dell G5 5000 - Part one on component choice & support: th-cam.com/video/4DMg6hUudHE/w-d-xo.html
Part two on benchmarks & bloatware: th-cam.com/video/5N7aYtkzKJc/w-d-xo.html
We even reviewed the included GPU in our standardized GPU test bench: th-cam.com/video/HFaJZq13tr8/w-d-xo.html
Back-order a GN Anti-Static Modmat: store.gamersnexus.net/products/modmat
Buy GN's Red & Black Mouse Pad: store.gamersnexus.net/products/gn-charge-redblack-mousepad
Blue & Black Mouse Pad: store.gamersnexus.net/products/gn-component-blueblack-mousepad
Wireframe Desk-Sized Mouse Mat: store.gamersnexus.net/products/gn-wireframe-mouse-mat
It would be interesting to see a comparison between consumer and server PSUs, if possible.
That PSU looks like the ones in their business conputers, which are usually made by Delta, who are top quality. Interesting that dell chose that PSU.
Excellent to see because the HP Pavillion I've got is totally identical to the Dell you broke down. 😎
@@Hogdriva This is made by LiteOn. We said that in the video.
I loved the rundown of the psu circuit and what are their respective jobs, very informative!
Good stuff you keep releasing!
Back to you Steve!
.
That is NEVER going to get old
This is the first time I've not glossed over while having electronic circuitry explained to me. Looking forward to more.
Awesome to hear!
Check out bigclive for more.
Never seen stone before. Give him some space on the channel to explain things like this, he's untapped potential for the media side of things, if I'm not mistaken about his absence in the videos.
Really liked this. Especially explaining each part. I have a corsair 1500w psu platinum certification and in the icue options it let's me choose between single and multirail. Never seemed to make a difference (I'm not using near the capability of the psu even with the 3090 or my crazy water cooling). Now I have an idea on what to research more.
Maybe a video on why I would select one over the other? When I should? What are the consequences? Etc.
Great job as always guys
@@jamieh4086 each rail on your psu will have a rating of X watts. Unless you need more than X watts on one connector, stick with multi-rail, as this is safer if somthing fails short-circuit.
To prevent any potential confusion for newcomers to the electronics design field please call it a square wave. Rectangular is technically correct, but there is no electronics engineering textbook that calls it that. It is always a square wave with some amplitude at a specific frequency.
Sent the comment over to Patrick for his consideration. Thanks for the input.
@@GamersNexus also please make sure you explicitly mention in the video not to disassemble a PSU if you don't know what you're doing since the built up charges in the capacitors can potentially kill you. As great as this content is, I really wouldn't want some young kid getting hurt tinkering with them.
It's called square not because it's a polygon with four sides, but because the intersection between parts of the waveform are square (i.e. at 90 degrees to one another), in the same way you'd call a room "square" when the walls and ceiling meet at 90 degrees as they are supposed to.
You don't need to cede ground on the use of "rectangular" because that's just wrong. There's no such term to describe an intersection of 90 degrees.
@@PeterMikeSolomon can you tell me how to discharge those capacitors?
(I change fans on PSU for DIY)
@@bk-sl8ee Either a Small resistor to drain it as heat touched to both contacts, or a screwdriver if ya don't mind potental damage.
This "Back to you, Steve." kills me every single time. Thank you so much Intel for the Memes!
AT 18:15 you could add the graph of 80gold or Platinum certification on the graph to be able to compare better
Excellent idea!
@@GamersNexus thank you. you are doing an excellent job for the reviews and the industry cheers from France
@@GamersNexus Yep, just colored background zones would do the trick nicely!
These types of supplies are also used in Dell's poweredge server line, hence the weird form factor. It makes them swappable.
And servers is something you cant cheap out on as a brand.
That's great to learn about the Poweredge line. Thanks for letting us know!
@@gg-gn3re There are used in some of them. Not all of them.
Can verify the PCB design is very similar to that of a Flex form factor PSU. Appears to be potentially a custom housing though. Just took apart a Dell PowerEdge 495 Watt 80+ Platinum PSU Model: F495E-S0. The one in the video is very similar with some component differences. Clearly based on the same platform though.
@@GamersNexus looks like a PE server PSU with wires soldered on instead of a hot swap pin set. I have a PE R710 PSU up in my bin of parts, and can take a look at it, it is also made by liteon im sure. I can send one to you if you want, but they have some other juju on them that allows them to both hotswap, and also can communicate with the server BCM, for full offline control and monitoring.
Marjie S - Thats just not true. Expecially on this particular "retail consumer" product line (PSUs from Dell). The Poweredge Servers have different PSUs made from Delta. Your comment is like looking at well-built car tires and saying just cause theyre round and made of rubber, they can be used on Semi trucks. Not even close. Theyre only swap-able within their own product category lines - Vostros, Optiplex, Latitudes, etc. The only reason they have a "weird form factor" is for Proprietary sales. Form-factors and swap-ability are mutually exclusive by choice from Dell. The server lines are different products and have their own requirements for components (lights-out, roll-overs, etc) - especially with Xeon, Epyc, and Nvidia AI units in those boxes.
Just some minor corrections:
The chokes and caps at the start of the primary side are mainly for filtering out the noise from the power supply from getting back out on the power lines, rather than smoothing the power coming into the power supply (that's done mostly by the bulk capacitor, but also more finely on the secondary side).
Also the "double-coil" inductor next to the bulk cap isn't mainly for current-limiting, it's known as a common-mode transformer (or common-mode choke, both terms can be googled), and is also for keeping certain forms of noise inside the PSU, in this case common-mode noise (hence the name).
@COFASA Uhhhh...That is the main purpose the PSU is there for. If it is good at that then I guess it is good.....
@@sajivsingh5601 yes, but efficiency doesn't matter if the power output is garbage in quality. That quality is very important for sensitive electronics
Actually, one of the most important things that a power supply can do, is to NOT blow up your system if there is a failure. I had a friend come over a few weeks back with a PC that was no longer turning on (PC was about 10 years old). It had an Antec power supply, and one of the 12V rails was briefly spiking up to almost 13V on powerup, but the built in overvoltage protection would immediately shut down the supply. He had been having problems with phantom shutdowns for a few months, but it had become so common as to be unusable a few days before he brought it down for repair. No other hardware was damaged, it was just a case of replacing the power supply (and reinstalling Windows, as the image had been corrupted to the point of bluescreening with the random early startup power losses.)
PSU's need to have overvoltage fault protection like this. Back around '96, for my first new PC I purchased a generic AT pentium prebuilt (from a local shop, everything was off-brand in that system except for the processor and hard disk); several years later it failed, and the power supply decided to commit suicide and take the motherboard, ram, and video card along for the ride. After that point I decided I would never use anything but Antec or Enermax power supplies for anything else. I subsequently have had some of these power supplies die, but they have never taken out any other components along with them.
I would guess that every power supply has this protection built in these days (since its part of the ATX PSU spec), but it would be nice to test it to see if its working as intended.
And they said power cleanup for the PC is done at the input X-Y caps and Inductors 😂😂
Thanks! Noted and passed on to Stone!
I’ve been so excited for you guys to start showing power supply testing. Keep up the great work!
Thanks! We're really excited about it too. It's a new review category for us. Patrick (Stone) has been working hard at this testing for PSUs and we're finally in a stage to start sharing. I'm personally excited about it because PSUs are a weak area in my knowledge, and I can learn from the work he's doing. We'll be reading comments closely to keep improving as we figure out how to present the data and how to analyze our findings.
@@GamersNexus I'm definitely interested in those Gigabyte power supplies. I'm also interested in what a low quality PSU can do to a system over time
@@Noah-lj2sg Very true, and it also keeps the market sharp and competitive. I am saying this while listening to music via a dac which has been measured by a youtuber.
@@GamersNexus looks like future intel motherboards like likely become like laptops mobo
Any video with MOSFET discussion gets an upvote from this Electrical Engineer. Well done, Patrick!
I'm glad to see the "Thanks Steve" transitions from the Intel video in use. Absolutely perfect.
"Here we have some screws.
Here are some wires.
Here we have the active power factor correction and the is the rectangular waveform and the lower voltage smoothing MOSFETs"
That escalated fast.jpg
😂 I thought the same
It went from English to Chinese pretty fast
@@caio5987 Electronese*
This is exactly how many university lectures start. "So, 1+1 is 2 .... (suddenly some partial differential equations) .... and the rest is trivial, so just do it at home if you want to know what's up with that."
You guys are going to take those intel transitions to the grave, aren’t you.
I'll make the Intel transitions a meme if it's the last thing I do.
@@GamersNexus Thanks, Steve.
It's what the people want.
intel is trans? based.
@@archimedies9369 quite the opposite if you actually knew what based meant.
Turns out Intel's keynote was more beneficial to you than they could have imagined. Back to you, Steve.
You all definitely earned a Titanium 80+ certification for efficiency in explaining power supplies. Have 0% interest in this Dell PSU, but this was the most interesting video ive seen in while. Thanks, cant wait to you guys progress on this type of testing
eviews.
Love seeing the other team members in videos
"We know that no one's actually buying this thing standalone..."
::Stares in K-12 IT::
Ha, yeah, that's about the target audience!
@@GamersNexus In all honesty, buying these things standalone from dell is difficult. You need a dedicated dell business account rep, they'll ask for a service tag of the model you need a PSU for because they have >100 different PSUs in all form factors, and then they'll tell you multiple times they can't find it before suddenly finding it and shipping you the wrong form-factor twice before getting the right one, all the while offering to sell you a new machine instead. However, if you have the same machine under a "ProSupport Plus" warranty, they have no problem finding the exact PSU and shipping it overnight.
@@futureb1ues Yeah, my company just scraps Dell systems when their power supplies blow out. It's nearly impossible to get replacement parts from them.
They usually fail during power surges. Sometimes when someone accidentally ruins 48V through a serial or USB port.
@@futureb1ues That's the whole idea behind Dells proprietary parts. To keep government and business into Dell's little ecosystem. Sadly your district probably wont allow you to swap systems to standard parts. The one I worked for did that, bulk ordered parts and had junior high and high school students put them together with IT then slowly replaced the elementary schools PC's over a couple years.
@@MikeStavola they don't get replacements on amazon? Although by the time a psu goes it's been a few years probably
8:58
the wave form is called "square wave" not rectangular.
happy to help.
Passed the feedback along to Stone. Thanks.
It depends if you are referring to current or voltage. Most forward, half bridge or full bridge buck isolated stepdown switch-mode power supplies (apart from discontinuous types such as flybacks) have a PWM squareware voltage measured with respect to primary ground on the primary input to energy storage inductor and a contiunous sawtooth current in that inductor.
Thanks Steve!
That has to be the best PSU description / explanation I have ever seen , I love when I actually learn something. Thank You Patrick, and Steve..
Having experienced two catastrophic PSU failures in my life (One starting a fire, the other killing every platter drive in my NAS to the tune of $7000 in hardware), I really appreciate that someone is now doing comprehensive PSU testing. Just FYI, the garbage PSU's I was using were branded (Gigabyte for the one that caught on fire, Rosewell for the one that died and over-volted/ or over-currented frying all my hard-drives). In both cases, it seems that the transformers broke down, from the limited autopsy that I did. Since then I've learned my lesson and I have been buying way-overspec'd EVGA power-supplies. Assuming EVGA knows what they are doing, but I guess we will soon find out...
Probably one of their rack power supplies, tweaked a bit to hush it up
Exactly what I thought. No way Dell spent money to get a PSU of this quality designed just to get something proprietary.
Likely with cheaper parts substituted where possible.
@@YeOldeTraveller Yeah the "Elite" bulk cap in their server PSU variation would be probably a Nichicon or Nippon CC.
Ain't those psu made by delta? I havent seen the video but if is made by them Its bulletproof..if they mounted quality components
@@fabiofoltran4361 Liteon
Brings back memories as a tv repair technician. I've had to repair numerous tv power supplies that work just like the one described here. On another note: Dell computers are only useful for business applications where they will be thrown out and replaced with new ones as needed. They are not geared towards the general public as they can't be upgraded, which as you described, is a shame.
This look like a propietary version of server PSU. This is probably used on Dell server products which is why it's actually good. Dell business line is pretty good.
Server PSUs are way above this. T30 might use something similar, not the proper servers.
Are you guys checking for creepage and clearance guidelines for safety? I know some PSUs fail it hard.
Sending this comment over to Stone so he can take note of it. I'm actually not sure on the answer for that -- he did a lot of testing, so maybe, but I'll ask him to look into this comment for future pieces if not!
As a PCB designer by day PSU design is mundane but incredibly hard to get right... If these are built to UL or CE there are decent creep age and clearance requirements...
Just want to mention it again how happy I'm to have found your channel after HardOCP folded. It really replaced everything I used to get from that site. Love your informative channel. Appreciate the hard work from Patrick Stone. Thanks and back to you, Steve.
Never knew Liteon makes PSU too. I remember fondly though how great their optical drives were back in the days.
I have always felt Dell systems are well built and hardly breaks. I often recommend Dell Optiplex line for my clients. Now it's great to know one reason why.
Yeah, i remember back in the early 2000's i had so many cd and dvd players that failed, most of them LG, Samsung, TEAC etc. But i had a LiteOn DVD-RW that lasted me for years, and i've written a lot of discs with it. This left me with a very good impression on this company, and heard many people having similar opinions.
Definitely excited for more PSU reviews, will be great to have something to watch when I'm looking to buy my next one.
The electrical engineer in me loved this! Would love more videos like this of other PC components.
18:30 "...Lack of any 80Plus branding." Goes back in time to 4:58 and you see 80-Plus-Platinum.
And 21:53
Please please more in depth PSU reviews. Now that Jonnyguru is gone, it’s tough to find a real in-depth PSU review with actual quality testing. The skill and equipment required aren’t common. Great job guys!
When testing voltage regulation you also need to rapidly and randomly vary the load on the output - TiN explained it once when you were visiting him and KINGPIN
I normally listen to Steve to help me sleep, no joke, I sleep to TH-cam videos every night (I have really bad tinnitus). But this other guy, oh man, he knocked me out in seconds...
a few tests are missing: transient response, dynamic load and the time the psu can supply the components with power when the primary side has an interruption. Otherwise good first effort!
Dell has probably been using PSUs like this for awhile, probably costs too much (or they just have too much stock) to switch toe AVX12VO. At some point that should tip in favor of switching though.
I have a feeling the PSU& motherboard are used in dells' business/server options and it's just cheaper to make the home PC fit what the majority of there business is.
Nah, Dells been using custom proprietary stuff including PSU's since at least the Pentium II days.
Nah, dell likes to lock their business customers into their ecosystem. They're also stingy about shipping replacements, they'll always try to sell you a new system.
Its more evil than that. People like to drone on about Apple locking you in but Dell really locks you in and it must be profitable because HP has gone the same way. I have an HP Z220 (i7 3770) that I got referbed and turned into a low end gamer/Hackintosh and it has all standard ATX parts. Fast forward to today and they are doing the same thing Dell does with custom PSUs, off size motherboards and proprietary connectors. The entire supply chain has a vested interest in preventing reuse of electronics. AS long as Im careful with it, my old HP will be usable for decades. PSU dies? Get any ATX standard as a replacement. CPU dies? Drop in another 3xxx series. Lightening fries everything? I can still put new guts into the case and call it a sleeper. Each time I do that I am denying some part of the supply chain a sale so they are all more than happy to go along with Dell, HP and others in making non replaceable parts. Im a hardcore Apple guy from back in the Apple ][ days and Apple is just as guilty but the difference is that most of the time Apple is doing it because the entire system is non-standard. Its not like they are intentionally going off ATX to make an iMac. No one is complaining about Dell or HP laptops having custom connectors either. Heck Ill even give them a break on Alienware and Omen systems in cool looking cases that can only happen with custom parts. But when you take a bog standard box and intentionally cripple it by going ever so slightly off ATX, you are just showing how evil you are.
@18:50, changing the vertical scale of the 'Power Watts' in your chart to a logarithmic scale might show you a more meaningful comparison between power in and power out across 0% load and up. Just something you may try and see what it looks like.
Just got an email from Newegg saying that Gigabyte is now taking returns or exchanges for PSUs in a certain range of serial numbers. Great job GN!
So im going to study electrical engineering in august, i will watch this video again after im done, keep up the good work!
I feel like Patrick Stone and Buildzoid could go to a scrapyard together and build a budget supercomputer in a couple of days
"Thanks Steve" made me giggle way too much
The first PSU review was the reason I subscribed to GN! Hoping to see more of these more technical component reviews.
Maybe a small tip, but if people ever do this in the future. Please remember that the capacitor can keep charge for quite a while!
If it doesn't have a discharge resistor, sure.
Since it's being sold by a huge company with liability concerns, odds are they didn't decide to skip that particular 5-cent component.
@@AndrewFremantle Did you watch the video about the rest of this PC? And the way they cheaped out on this PSU by making it proprietary instead of using the 12VO standard? Odd are they _DID_ decide to skip that particular 5-cent component.
In any case, better safe than dead.
@@AndrewFremantle It wouldn't surprise me that cheap PSUs skip that critical safety resistor, however I spotted it in this particular PSU, shame that the non standard format makes it useless outside of this line of systems.
@@AndrewFremantle "Small" correction, 0.01 cent component (yes they are really that cheap).
And i have personally never seen any power supply, not even the cheapest backstreet chinese ones that arent anywhere close to being FCC or CE compliant that dont have discharge resistors on their bulk caps.
The times where this was relevant was 40+ years ago, when SMD wasnt a thing and a resistor did cost 1. a couple cents to a couple dozen cents and 2. occupied some worthwhile physical space. But at that time this wasnt nearly as relevant, as switching power supplies werent really a thing, thus no high voltage caps in most devices.
These power supply test are exactly what I want to see more of. Particularly the current stability with ripple and regulation testing. Love the work!
This is what frustrates me the most about dell their engineers clearly know what they’re doing but it all gets ruined as usual by suits who have no idea what they’re doing making weird form-factors for arbitrary reasons and slapping predatory sales tactics for warranties on sales if they actually stuck to normal standards and didn’t send you notifications in your sleep asking you if you want a warranty it would actually be a pretty good buy honestly.
Actually there is a reason why these kinds of things exist, they seem to be re-used components from their enterprise hardware and server rack hardware is some of the most non-standsrd stuff you will ever encounter. Most brands moved away from desktop style PSU because single space servers would never fit them.
The reason why is clearly shown here, the PSU is very efficient because when you're putting equipment at datacenter scale, even small percentages in efficiency can result in significant gains when you're talking about thousands of servers.
The controversial decision is more along was it a good decision to bring this to their consumer devices.
I have one 250 Watt TFX Chieftec PSU and two 300 Watt TFX Chieftec PSUs.
The 250 watt is not rated because it has low PFC but all three TFX PSUs are almost silver (90%) efficiency.
Chieftec also sells a 500 Watt ATX PSU with terrible efficiency (less than 80 Plus white).
I rarely see crappy TFX PSUs. I would love to get the Seasonic 300W mini-TFX PSU for a SFF PC.
I love the idea of you guys doing PS testing and reviews. They are IMO the most overlooked piece of hardware when specing out a PC build but often times one of the most important. Couple that with no one in the PC reviewer space doing any thing like this makes you guys unique. This is awesome guys! Good work!
It's probably an adapted server power supply
Yeah, looks like a 2U psu. For them i think it is cheaper, server PSU specially from last generations are really cheap, they probably have a couple hunded thousands doing nothing in some werehouse.
The Connector Extractor is what i needed for Years, thanks dude
Honestly, that was really cool, and also surpising lol
Go crazy on these PSUs teardowns and reviews and love to see more of them! Nobody reviews PSUs you've got absolutely ZERO competition there.
I've been ready for this content, I'm excited to have you guy's teach me about power supplies. I've honestly learned a lot from your channel and a couple others it's cool to see how things work.
Thanks Steve.
This is when I whish we could "like" videos more than once. The value we get from this channel is awesome, the work you guys do is quality and appreciated, I'll say it again easily the best PC hardware channel on the youtubes. As a fellow electronics tinkerer Patrick's presentations are great and it's refreshing to see a tech youtuber speaking my language, well done.
Your videos are a break of sanity for us in an insane world, and you have me rolling with the Computex references. Thanks guys.
Back to you Steve.
Patrick please read this! Please do research on how APFC works. To put it simply it is a boost converter, consisting of the yellow ferrite inductor, diode and mosfet. The rectified AC is only slightly smoothed by a ~1uF capacitor and then the boost converter raises and lowers its PWM to consume more and less current based on the voltage waveform of the AC, that way it can fake a unity power factor as best as possible and also step up the voltage to ~400v to allow higher efficiency and universal voltage capability. The inductor is not a simple "current limiter" its the inductor that the boost circuit is using to store the power.
For future ATX PSU reviews, I would additionally like to see measurements of noise under certain loads. And maybe comments on cable quality, length and stiffness - important factors when building and cable managing.
This is an interesting new category of content, I hope to see you do SFX PSU testing too just to see what kind of compromises must be done for the smaller form factor.
Thank you for doing power supply reviews. Not enough people are doing detailed high quality video reviews of them like you are trying to now.
I was fully expecting "you can literally see it" when he took the outer casing off. Imagine my disappointment.
This jumps from explaining an AC outlet to talking about “in rush current” or whatever really quick.
I guess "thanks Steve" and "back to you Steve" memes are here to stay
It hasn't gotten old yet. I'm sure they'll get rid of it eventually
Good, good.
@@le_travie7724 I hope they get rid of it at the moment intel produced another weird keynote with more and better meme-potential snips.
You can literally see it.
This was really refreshing to watch. I used to spend all my time in mid 2000s looking at PSU reviews, just to find that golden cheap nugget with Seasonic on the inside.
We need real power supply testing videos. This is one of the most overlooked parts to building a pc and nobody really does this on youtube. Everywhere on the internet people just try to scare you into buying the most expensive power supply you can buy. It would be really nice to know what is the best power supply for the value and why. Right now everyone just recommends the corsair cx and if you read comments on reddit or other places every other 80+ bronze power supply is going to blow up your system.
There's quite a few other bronze units I've seen recommended on reddit though. XPG Pylon, EVGA BQ, some CM MWE variants, Corsair CXM and CXF, to name a few. And it depends on the budget. The CX is possibly the best bronze unit but it costs more as well.
The EVGA B5 is pretty decent as well. It’s not great, but it’s still better than the BR and BQ series
I’d personally trust Seasonic at any efficiency. They tend to use higher end components compared to other brands and are put together very well.
I wouldn’t, seasonic has pretty overpriced units, also there focus plus series has some tripping issues with ampere
@@downnnnnn Tripping issues mean that the protection circuitry works. That’s a plus in my book. Current spikes are supposed to be handled on the card side. That’s a card fault rather than a PSU one. The prices can be a bit steep but that is the cost of quality components. Depends on your priorities.
Love this addition to your test suite. As many have said, looking forward to more P/S testing/reviews.
Just a FYI, when I was learning how circuits work it always helped me to have a block diagram with an overall view of the stages of the circuit so that I could go from general theory to detailed. This may help others to better understand the overall view.
As an electronic professional I thought Patrick did a great job explaining the basic functionality of all the circuits needed to get a computer, or pretty much anything, working these days. I had teachers that knew less than that and couldn't explain a psu to save their lives. 😂
Next stop: budget firecrackers, extinguished with LN2!
Back to you, Steve. Never gets old.
Hey Gamers Nexus Team, for power supply testing videos where you open the unit up - a general "do not try this at home" warning for viewers not to open their own power supply due to the possibility of injury or death from electrocution I think would go a long way and be the ethical thing to do, even if all your normal viewers would already know to do this
Hi Dad
I agree, there should be a warning in the video. I had no idea how dangerous a psu could be until around a month ago when I stumbled upon that information. Previously I just thought you should not tinker with them since the risk of breaking something that could become dangerous once power got plugged in again.
oh yes - show me more PSU testing!
I am quite fascinated with the operation of a PSU so this video helps reveal more of what happens inside that component.
Patrick Stone is well spoken and his explanations were easy to follow.
Your testing seems well done as always.
Looking forward to the next video on this!
Dude you’re getting a Dell! I’m sorry for your loss...
Haven't seen much of Patrick Stone to be honest but really appreciate his work here. He has a great way of talking and a likeable personality which makes viewing a pleasure. Great job my man, keep up the good work and looking forward to learning more from you ☺️
Did you miss the "80 Plus Platinum" Label at 4:53? Does it meet the requirements for that?
yes he did
Yep, they clearly missed it. And so did I. Based on their testing results I’d say it barely passes if you take into account tolerances and potential measurement errors. Remember that only a proper lab with calibrated equipment, trained staff and vetted procedures can assert that with any level of certainty.
@@rdoursenaud 1. Maybe you missed it, but you were watching a video. It's not your job to check every claim they make. They worked for hours with that device, I am still surprised.... 2. I think, the rating is missing. Steve only told us "It is close to 80Plus Titanium", but did it reach "Platinum", "Gold" or whatever? I think, that would be helpfull. But this video is to practice the procedure -> next time!
@@manfredbans3975 1. It’s always my/your job to check any claim made! How else would I prevent bullshit from entering my reasoning? "Trust but verify".
2. The answer (and Gamer’s Nexus mistake about the unit not being 80 Plus branded) is at 18:33. The measurements meet (exceed) Platinum level and are not quite Titanium. So the unit exceeds every 80 Plus supersets except the most stringent one.
On the other hand, I find this metallurgy analogy confusing and unnecessary. This benefits more Ecos Consulting’s bottom line than the consumer. Hence my flawed guess in my previous answer (I confused the levels).
Your ads for the EVGA Z20 and then when Linus had that video where EVGA was doing 50% off on them convinced me to buy one.
I can say, I'm not disappointed.
Very important thing not mentioned in the video: PSU capacitors store a LOT of charge. Don't take your PSU apart for any reason unless you are sure you have safely discharged them. They could literally kill you.
Definitely a great piece to know.
That really should be explained more explicitly than some legal notice "don't do that at home".
@@ufoisback5088 Well it probably wouldn't hurt if nature and fate were able to hand out a few more Darwin awards these days. If you don't cull the stupid from society then they propagate more stupid which then adds even more stupid and you soon end up with current society where you have people who need to be explicitly informed that "dangerous thing is dangerous."
Great video. I am an engineer and understand electronics but still really enjoyed his description and testing. Looking forward to more of these on one of my favourite channels.
It would be better to have a video dedicated to explaining how a PSU works prior to posting a PSU review rather than putting both in one video. The explanation part of how a PSU works here is quick and dirty and would benefit having a longer video focusing solely on this topic as it is interesting enough to be featured on its own.
I feel like I’m being forced to learn about the things you guys are learning about but it’s fun because we’re doing it together.
Thanks GN
I'm waiting for the gigabyte PSU review
I guess that they delayed it so they could use the new testing equipment
It's been in endurance testing for a few weeks now! Should be ready soon.
@@cyberceel seems that way. GN do great work. I can wait
@@GamersNexus can't wait man
@@GamersNexus You're implying it hasn't exploded.
Where' the kaboom? There was supposed to be an earth-shattering Kaboom!!
I know how difficult (and expensive!) PSU testing is, but there's simply not enough reliable PSU tests out there so I'm pretty excited for this stuff. Great work.
It’s a pity that the only powerful thing about Dell PCs is the power supply itself.
a testament to the psu being the most likely cause of fire if made improperly. as long as its not in a nzzxt case... i mean uh... no i got nothing.
And that's seems to be the case with OEM PSU, drawbacks is proprietary PSU
This is actually a new thing for me, since all the prebuilts I've ever seen always carried the crappiest PSU that were humanly possible to be manufactured. Shame that a good thing, for a change, is spoiled by non standard specs.
NZOC - ??? Im pretty sure the CPUs and RAM are good (made to a decent standard) too.
@@JohnWiku most if not all oem pc's comes with high quality psu's. Sure they may look like generic junk on the outside but those badboys are on par with your regular 80+ bronze at minimum. This is not a new trend, its standard industry practice.
The new opening sound effects sound so good on my phone. When the bass hits my phone rumbles. Love it!
this was quite good. i'd recommend a viewer warning to be wary of disassembling power supplies, since those caps are rather massive, and could certainly cause some serious injury.
i'm looking very forward to new content presented with the GN clarity, especially since PSU reviews can be particularly dry and difficult to parse.
it might be worth considering a basic "how PSUs work" that can always be referred back to as well, with a few graphs showing how the power waveform gets affected throughout a simplified PSU, and what is and is not desirable.
Thanks! We'll probably do one of those once we get more experience with what sort of stuff needs more clarification. We'll base that on common questions in comments!
@@GamersNexus also it’s worth including a block diagram in the PSU video (as suggested by @neobleidd) and explain some of the terminology (e.g. PFC) in there, as they are music to my ear (a trained electronic engineer) but as far as I believe not to most people.
@@GamersNexus welp looks like LTT might have beaten you to the punch (^:
if they've got "turbo nerd edition", i fully expect a gn style "ludicrous nerd edition"
Great content. The rundown on the function of the internal components was awesome. Part of my job is designing flyback power supplies for industrial control modules. Computer power supplies are also typically flyback topologies. Flybacks are great because they provide isolation and the ability to easily generate multiple voltage rails.
The filtering on the AC input side of the power supply does prevent noise from making its way into the power supply (conducted susceptibility), but just as important is preventing harmonics of the switching frequency from making it OUT of the power supply and back on to the grid. This is part of the conducted emissions compliance that the power supply must meet before getting the CE mark or other regulatory approval like FCC Part 15. Switching power supplies are extremely noisy, especially at 100's of watts. When the MOSFET on the primary side of the transformer turns on and off based on commands from the PWM controller it generates noise and that noise is suppressed by the inductors, chokes and caps.
Those MOSFETs on the secondary side of the transformers are interesting. They must be in an "ideal diode" configuration. Diodes have a typical forward voltage drop of 0.3 to 0.7 volts. This is a big deal for efficiency because this voltage drop multiplied by the secondary current (500W / 12V = ~42 amps) is the power lost in the diodes. An ideal diode is a circuit using a MOSFET and other components to act like a diode, but with 0 forward voltage drop. This increases the efficiency of the power supply by reducing losses in the secondary rectifier diodes to near 0. Less losses mean less power in for the same power out.
I think your testing is good. I would like to see a step load test to measure transient response. I don't know if the Intel spec covers this, but basically you load a power supply to some low-ish % or rated load (33% lets say) and then switch the load rapidly to a high current (say 66%) and observe the response of the voltage. It will initially dip, and then the PWM controller will increase the duty cycle to the primary MOSFET to output more power. The voltage typically overshoots the setpoint (12V) and then the PWM controller reduces the duty cycle and the output voltage settles. You want the power supply to respond quickly (minimize the dip) with minimal overshoot. Along with this testing it would be interesting to know the switching frequency. You could measure this on the output side of the transformer with your oscilloscope.
Dell have never stuck to the ATX standard.
Even their ATX ones dont use the correct colors, and at times their pinout is not the same.
Not true! Their workstations before Sandy Bridge (Core 2nd generation +) used standard ATX. Take any Precision workstation prior to Sandy Bridge - t3500, t5500, t7500 , t7400s etc could be swapped with PSUs off-the-shelf. HP and Lenovo were the same. Dell wised-up.
@@staryoke Dell used their own color code though
Yellow was not +12v, red was not +5v, orange was not +3.3v
@@IIGrayfoxII ??? Color coding wires uniquely doesnt make it NOT standard ATX. If that was the case, all third party aftermarket PSU makers (EVGA, ASUS, PowerColor, Corsair, Enermax, Seasonic, etc) would NOT meet "your" definition of "standard ATX". C'mon stop trolling. Go learn a thing-or-two about what makes a PSU meet standard ATX power requirements. LOL
@@staryoke
The ATX color coding is apart of the standard.
Yellow is +12v
Red is +5v
Orange is +3.3v
White was -5v(no longer apart of the ATX2.x standard)
Blue is -12v
Green is PWR_ON
Gray is PWR_OK
Black is ground
Purple is +5vSB
High end PSU that we gamers get technically do not meet the standard due to this.
But when it is not a high end PSU but a basic one.
having orange be +5 and purple be +12v will throw you off if you're unaware of dell doing this
@@IIGrayfoxII "High end PSU that we gamers get technically do not meet the standard due to this."
LOL! C'mon! I think you should put out a trade paper stating this and listing the PSU makers by name. First you would receive a cease-and-desist for libel and slander and then your ass would get sued into the ground - and rightfully so. Clearly youre not an electrical engineer, because you wouldnt have made such a stupid statement with a straight face - unless youre just trolling!
The colors of the wires are an industry suggestion, not a requirement. More important is the gauge and quality on hand (at the time of manufacturing). Color is secondary to those requirements.
Contact any of the aforementioned manufacturers, they will tell you that.
I've been looking forward to these kind of videos for a long time now, no other youtuber does power supply testing, let alone this in depth and in such an informative way! This will go a long way to help show which power supply brands are trustworthy and what to look out for.
Linus must feel threatened. Now HE'S done a "power supply basics" video too.
Best damn power supply video I've ever seen. While I know the words he used I am too far removed from my classes to remember how it all works.
Great explanation, Patrik Stone.
The memes from intel's presentation should be used every video until their next one
So glad you guys are adding PSU's for content. Jonny guru and his website has been out for some time and I don't think we could ask for a more reliable stand-in.
This format of PSUs is common in server machines and, as such, they tend to be good quality and made by reputable OEMs such as Delta, for example...even if they look "uncool" for modern hardware beauty standards.
Dell, also being in the server business, probably uses some of the hardware they comission for their servers in prebuilts too, like this PSU. That practice also has an ugly side, like weird layouts and proprietary hardware we see in their (and HP's too) prebuild desktops...
Great feedback and information. Thanks!
It's about double the width of the server line, probably to reduce noise. Also with no pcb card connector. As far as internals it looks completely redesigned vs those. e.g. th-cam.com/video/dnD-zaOznRo/w-d-xo.html again, probably for noise.
Server PSUs are normally rated for much higher power delivery than the nominal 500W output of this Dell desktop power supply -- for example a twin-EPYC 2U server will have multiple redundant 1400W PSUs. Modern manufacturing lines can produce small runs of "custom" units in a cost-effective manner based off existing circuit designs that simply need a bit of tweaking to meet the desired specification and this is probably the reason for the design of this particular PSU.
Badly designed power supplies with iffy components that fail hard can cause system damage which leads to warranty repairs, product recalls and in the worst case fires, explosions and interruptions to TH-cam video recordings. The small cost of designing and manufacturing a decent and safe power supply for even a lower-end system is outweighed by the bad press and downstream costs when the makers get it wrong (see Gigabyte's rebadged explodey-bricks for a recent worked example).
Really appreciate the PSU terminology education. I have heard the term "rails" for a long time and never really understood exactly what was being referenced.
*THE FOLLOWING PROCEDURE WAS DONE BY A TRAINED PROFESSIONAL, DO NOT TRY THIS AT HOME*
There's no way I'm the only person who's asshole puckered as he was pointing w/ that screwdriver
@@Arudis4x not gonna lie, my ass was clenching during that, it was terrifying to see him poke around with a screwdriver.
yeah seriously. don't just go around poking at bare PSU. touch the wrong thing and you might get zapped from a high voltage that might be left in some caps ot something.
If you haven't been zapped by AC, you haven't done enough tinkering.
@@EmergencyChannel bold of you to assume I've never been zapped by 220v AC or got a nasty burn from dropping a soldering iron.
These things need to be handled with care because it hurts a lot when you don't
Work in power electronics (specifically power FETs), and will back this as a great review! I am sharing this with our summer interns because it's a great physical example of what our customers want to make!
Two points of feedback:
- Block diagram during the physical dimensions review of the staging would be nice, but that's probably too much for most people
- OCP and OPP are both time dependent. They should trip in some small amount of time (like microsecond range or less). I would be interested in knowing when the fault trips and how long it takes. Since yall were slowly ramping up to the fault detection, I would be interested in a verification by jumping straight to that point or slightly below.
This is like putting a Honda Earth Dreams engine into a Yugo. Efficient, reliable power, just too bad everything else is garbage.
I don't know if I would call the Earth Dreams engine reliable at all considering all the issues people were having with severe oil dilution problems because the engine doesn't heat up enough on short trips. And then the fix Honda deployed to those people basically killing the amount of heat the heater could output in order to try to get the engine to proper operating temp faster.
Thanks for the detail on the circuit paths. I would have never expected that there were two transitions from AC to DC. I guess cleaning everything up, then synthesizing AC to pass through an isolation transformer makes sense.
This would make a good 12 VDC supply for a test bench after one is done using it in the computer.
Showing the ripple when it begins to show up in the oscilloscope waveform would be a great addition when testing more PSUs in the future.
Some people don't know how it looks or why it is bad.
Those waveforms you guys showed up don't mean much.
Great tear down, the PSU is probably not the first components considered when buying a prebuilt, but it’s one of the most important. Can’t wait to see future PSU reviews.
I love that you had to tear apart the power supply to find something redeeming lol
Presentation was good on the info. I would say you can simplify the break down going over the psu internals to just the important parts and quality of parts used.
Naturally having a standards compliant component in a system is totally unacceptable. Everyone knows that (duh). As such they had to modify that slightly so it wouldn't meet the standard anymore, even if it costs extra.
1 standards compliant component is too much goodness no mortal consumer should be allowed to enjoy
I feel like I just watched the opener to season three, and they decided to mix in a new character. But I love it. This is excellent content.
"Thanks Steve."
"Thanks Steve."
Back to you Steve!
You can literally see it!
Love this content. Keep up the great technical analysis. One thing not mentioned was the whole reason to move to 12v only PSUs is an increase in efficiency due to the simplified design of eliminating the 5v output. Google has been using 12v only power for long time in their data centers.
Gigabyte psu have a bad reputation, will be cool if you can review someone of those
They're in the process of doing it, one blew up on camera lol
Most excellent job on reviewing the PSU from Dell (especially given the non standard nature of the plugs) I really appreciated Patrick's walkthrough of the powers route through the internals of the PSU and the presentation as whole was good and well balanced. Not too high level
nor so technical that a regular "Joe" like myself, can't keep up.
Great job 👍