Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

hello_there_123

Member
  • Content Count

    999
  • Joined


Reputation Activity

  1. Funny
    hello_there_123 got a reaction from AviaLaws in Best Ryzen 3600 Air Cooler   
    The stock cooler.
  2. Like
    hello_there_123 got a reaction from Parideboy in Common PSU Misconceptions   
    Some common PSU misconceptions
     
    “80+ rating is the best indicator of PSU quality”
     
    Not wholly true. You also have to look at ripple, voltage regulation, transient response, noise, fan bearing, topology, load regulation (including 12V crossloads), temperatures (includes rated temperature, 30C 40C 50C etc.). That’s why even though Raider II spinoffs like EVGA GQ and Be Quiet! Pure Power 10/11 are Gold, they are still worse than Corsair CX in quality (ACRF vs Half Bridge LLC). 
     
    “You should get a higher wattage PSU because of efficiency curve”
     
    First, let’s look at what an efficiency curve on a modern PSU looks like. 

    (Left percentage is 115VAC, right is 230VAC) Efficiency peaks at 200/650 = 31% of the rated wattage at ~91.5/93%, then decreases until it hits ~88/92%  at max load and ~87.5/91.5% above 700W. The difference between max load and peak load efficiency is 3.5/1%, which is extremely insignificant. 
    Even on older platforms, this trend of efficiency plateauing above ~20-40% load is present:

     
    Now for the real world calculations: let’s say that a person is debating on whether to get a 550W or 750W PSU for a build that consumes 400W when overclocked to its max. Also, he’s debating on whether to get the Corsair RMx 550 or 750. 
     
    These are the efficiency curves for both: 


    The 550W has ~90/92% efficiency at 400W, and the 750 also has ~90/92% efficiency at 400W. Conclusion? He should get the 550 since it’s probably much cheaper, he doesn’t need 750W anyway since his build consumes 400W max, and the efficiency is the same! Source 1 Source 2 Source 3 Source 4
     
    “Get more wattage than you need because it will last longer/running your PSU at max will significantly decrease its lifespan” 
     
    First off, most modern not trash PSUs are rated for 40C or higher, which is good for most cases. And the advertised wattage is usually CONTINUOUS, not PEAK. If a 550W continuous PSU at 40C were not able to deliver 550W constantly at 40C, then that would be false advertising and the manufacturer would get sued. Thus, the claim that running the PSU at max would significantly decrease its lifespan is generally false. Besides, most PSUs are overbuilt to ~120%+ of their rated wattage, and you can see that from where the protections are set. 
     
    More wattage = last longer isn’t necessarily true; a MOSFET or the fan can still fail under high wattage PSUs. In fact, higher wattage PSUs can be more dangerous; this is especially true for single rail PSUs, because highly set OCP (OPP’s even slower!) is much slower than lower set OCP. And in the event of a short within the PSU or within the motherboard, FET/VRM failure on the motherboard etc. the OCP/OPP on the PSU won’t be fast enough to trip before the short causes massive damage on the PSU and/or the motherboard. Source 1 Source 2
     
    “PSUs branded by [Brand name] are always good” 
     
    Every brand except Delta and Diablotek, etc have made good and bad PSUs (Delta have made only good PSUs if you don’t care about noise, and Diablotek only makes bombs). The most common brands I've seen people spout as “never having made a bad PSU” are Seasonic and EVGA. 
     
    First off, both have their fair share of bad PSUs. Look no further than Seasonic’s S12ii/M12ii/EVO series or their S12III series, and EVGA’s G3 (protections are set too high)/BQ/BT etc. 
     
    Secondly, EVGA doesn’t even make PSUs (that job goes to FSP, HEC, Super Flower, RSY, Andyson etc.), and Seasonic along with Super Flower often outsource to RSY because their manufacturing capacity is small. If Seasonic/Super Flower choose to manufacture a certain PSU themselves, the Quality Control will be worse on those because every part is outsourced, compared to FSP, CWT etc who do not outsource the PCB-A. 
     
    This misconception also falls under statements such as “just get a brand name PSU, they’re all fine”. 
     
    “Modularity is extremely important for the consumer” 
     
    Truth: unless the person in question is doing Cablemod, Mod vs Semi-Mod will make absolutely no difference, as the cables that are preattached on semimod PSUs will be necessary cables, such as the 24pin for motherboard. Also, in my and many other people’s experiences, on lower wattages (<=550W) nonmod is actually EASIER to work with than semi/full mod because there’s only a few cables left over from nonmod, which are easily ziptied and tucked under the PSU shroud (or left out in the open). Of course the issue of nonmod vs semi-mod is subjective, but at least I have found that the ~2 cables to ziptie after all the cables are attached on nonmod PSUs are easily dealt with. 
     
    “Japanese Caps are the most important factor of a PSU's quality” 
     
    A PSU is not just a box of caps. There is also the bridge rectifier, transformer, MOSFETs (particularly the 12V ones), fan quality, and topology to worry about when judging a PSU’s quality. The capacitor plague is over; for most modern PSUs, tier 2 Teapo/whatever else caps are as good as Japanese caps, and MOSFETs/the fan are usually the first to fail anyway (Stefan Payne). 
     
    “PSU calculators are good for estimating your system’s wattage” 
     
    Ever notice how PSU calculators have affiliate links for their recommended PSUs? Yeah, they’re not there for YOUR benefit, they’re there to make money. And in this case, it’s by overestimating the wattage a lot so that they make money off higher wattage PSUs, which cost more thus bringing better profit margins. Outervision is by far the worst offender here; when I put in an RX 5700 XT, 3600, 2x8gb RAM, 1 SSD, 1 HDD, I got ~550W for my power consumption during gaming. However, when I look at Tom’s Hardware power consumption measurements, I see that the 3600 draws about 70W when PBO is turned on during gaming, and RX 5700 XT full OC draws 250W during gaming. Add 30W for everything else (because everything else draws minimal wattage) and we get 350W, which is FAR from the 550W that Outervision proclaimed. 

    Be Quiet!’s PSU calculator doesn’t brashly overestimate as much as Outervision’s; when I put in the same specs, I got 387W, which is still overestimating by ~37W but not as much as 550W. However, Be Quiet! Calculator is still overestimating for their own benefit, not yours. Moral of the story? Use THW’s power consumption measurements for the CPU and GPU, then tack on 30W for your true power consumption when gaming and overclocking. Source 1 Source 2 Source 3 Source 4
     
    “Get a PSU with higher wattage than you need for headroom” 
     
    Nothing about the current trend of CPU and GPU power consumption says that their power consumption will go up in the future. In fact, with die shrinks, the opposite is likely to happen, as same core CPU from Zen+ to Zen 2 power consumption has dropped, similarly with GCN to RDNA. 
     
    For example, 450W is enough to run a Vega 56/64 and 8700k overclocked to 4.9ghz (both @Stefan Payne and @LienusLateTips did this). That shows that most builds really don’t need 550W+. 
  3. Like
    hello_there_123 got a reaction from Trentonxx in Best z390 Motherboard for i7 9700k   
    Zis Wai
  4. Like
    hello_there_123 got a reaction from CasJ in First gaming / workstation build (high-ish)   
    https://docs.google.com/spreadsheets/d/1B27_j9NDPU3cNlj2HKcrfpJKHkOf-Oi1DbuuQva2gT4/edit#gid=0
     
    As you can see, quite a lot of drives have Toshiba 64L 3D TLC NAND. 
     
    Instead you should be seeing the Rocket as a Phison E12 Reference drive
  5. Like
    hello_there_123 got a reaction from LukeSavenije in First gaming / workstation build (high-ish)   
    https://docs.google.com/spreadsheets/d/1B27_j9NDPU3cNlj2HKcrfpJKHkOf-Oi1DbuuQva2gT4/edit#gid=0
     
    As you can see, quite a lot of drives have Toshiba 64L 3D TLC NAND. 
     
    Instead you should be seeing the Rocket as a Phison E12 Reference drive
  6. Agree
    hello_there_123 got a reaction from Princess Luna in Cheapest Board for i9 9900?   
    That's funny, I thought Intel was much preferred over AMD for Adobe. 
  7. Agree
    hello_there_123 got a reaction from Mattias Edeslatt in how to choose a power supply   
    You're doing it wrong then.
  8. Like
    hello_there_123 got a reaction from Thready in 970 EVO vs Sabrent Rocket   
    Mostly because consumers are very uninformed and see Samsung as God. 
     
     
  9. Agree
    hello_there_123 reacted to dizmo in SSD's got even bigger? PLC nand   
    The durability rating is still awfully good. The 1TB is rated at what, 330TB? Who even writes that much? It'd take many, many years to reach that point, and by then you'll have likely upgraded or swapped it out for something else.
    It's not needed? It's not like it'd save costs or something.
  10. Agree
    hello_there_123 got a reaction from SenKa in how to choose a power supply   
    You're doing it wrong then.
  11. Agree
    hello_there_123 reacted to porina in Intel 660p or something else?   
    Endurance is not a problem for any regular user. To reach the 2TB's endurance in warranty you'd have to write 200GB average a day, every day. The first few days after you get it, you might exceed that as you set up games, but patches and stuff don't add up to much after. Even with typical OS usage on top you'll be hard to sustain that average.
  12. Agree
    hello_there_123 got a reaction from PSUGuru in how to choose a power supply   
    https://www.amazon.in/dp/B01MTZ96RU/?tag=pcp03-21
  13. Informative
    hello_there_123 got a reaction from Gapi in Intel 660p or something else?   
    Keep in mind that unless OP has gigabit ethernet, 100-200mb/s is much faster than his/her internet speed.
  14. Agree
    hello_there_123 got a reaction from Mattias Edeslatt in Common PSU Misconceptions   
    Yea, I could understand paying 5 euro more for a pure power 11 400W over a CX450 for the 6 pin. But what I won't understand is getting unnecessary 650W or above for "futureproofing". 
  15. Like
    hello_there_123 got a reaction from LukeSavenije in Common PSU Misconceptions   
    Some common PSU misconceptions
     
    “80+ rating is the best indicator of PSU quality”
     
    Not wholly true. You also have to look at ripple, voltage regulation, transient response, noise, fan bearing, topology, load regulation (including 12V crossloads), temperatures (includes rated temperature, 30C 40C 50C etc.). That’s why even though Raider II spinoffs like EVGA GQ and Be Quiet! Pure Power 10/11 are Gold, they are still worse than Corsair CX in quality (ACRF vs Half Bridge LLC). 
     
    “You should get a higher wattage PSU because of efficiency curve”
     
    First, let’s look at what an efficiency curve on a modern PSU looks like. 

    (Left percentage is 115VAC, right is 230VAC) Efficiency peaks at 200/650 = 31% of the rated wattage at ~91.5/93%, then decreases until it hits ~88/92%  at max load and ~87.5/91.5% above 700W. The difference between max load and peak load efficiency is 3.5/1%, which is extremely insignificant. 
    Even on older platforms, this trend of efficiency plateauing above ~20-40% load is present:

     
    Now for the real world calculations: let’s say that a person is debating on whether to get a 550W or 750W PSU for a build that consumes 400W when overclocked to its max. Also, he’s debating on whether to get the Corsair RMx 550 or 750. 
     
    These are the efficiency curves for both: 


    The 550W has ~90/92% efficiency at 400W, and the 750 also has ~90/92% efficiency at 400W. Conclusion? He should get the 550 since it’s probably much cheaper, he doesn’t need 750W anyway since his build consumes 400W max, and the efficiency is the same! Source 1 Source 2 Source 3 Source 4
     
    “Get more wattage than you need because it will last longer/running your PSU at max will significantly decrease its lifespan” 
     
    First off, most modern not trash PSUs are rated for 40C or higher, which is good for most cases. And the advertised wattage is usually CONTINUOUS, not PEAK. If a 550W continuous PSU at 40C were not able to deliver 550W constantly at 40C, then that would be false advertising and the manufacturer would get sued. Thus, the claim that running the PSU at max would significantly decrease its lifespan is generally false. Besides, most PSUs are overbuilt to ~120%+ of their rated wattage, and you can see that from where the protections are set. 
     
    More wattage = last longer isn’t necessarily true; a MOSFET or the fan can still fail under high wattage PSUs. In fact, higher wattage PSUs can be more dangerous; this is especially true for single rail PSUs, because highly set OCP (OPP’s even slower!) is much slower than lower set OCP. And in the event of a short within the PSU or within the motherboard, FET/VRM failure on the motherboard etc. the OCP/OPP on the PSU won’t be fast enough to trip before the short causes massive damage on the PSU and/or the motherboard. Source 1 Source 2
     
    “PSUs branded by [Brand name] are always good” 
     
    Every brand except Delta and Diablotek, etc have made good and bad PSUs (Delta have made only good PSUs if you don’t care about noise, and Diablotek only makes bombs). The most common brands I've seen people spout as “never having made a bad PSU” are Seasonic and EVGA. 
     
    First off, both have their fair share of bad PSUs. Look no further than Seasonic’s S12ii/M12ii/EVO series or their S12III series, and EVGA’s G3 (protections are set too high)/BQ/BT etc. 
     
    Secondly, EVGA doesn’t even make PSUs (that job goes to FSP, HEC, Super Flower, RSY, Andyson etc.), and Seasonic along with Super Flower often outsource to RSY because their manufacturing capacity is small. If Seasonic/Super Flower choose to manufacture a certain PSU themselves, the Quality Control will be worse on those because every part is outsourced, compared to FSP, CWT etc who do not outsource the PCB-A. 
     
    This misconception also falls under statements such as “just get a brand name PSU, they’re all fine”. 
     
    “Modularity is extremely important for the consumer” 
     
    Truth: unless the person in question is doing Cablemod, Mod vs Semi-Mod will make absolutely no difference, as the cables that are preattached on semimod PSUs will be necessary cables, such as the 24pin for motherboard. Also, in my and many other people’s experiences, on lower wattages (<=550W) nonmod is actually EASIER to work with than semi/full mod because there’s only a few cables left over from nonmod, which are easily ziptied and tucked under the PSU shroud (or left out in the open). Of course the issue of nonmod vs semi-mod is subjective, but at least I have found that the ~2 cables to ziptie after all the cables are attached on nonmod PSUs are easily dealt with. 
     
    “Japanese Caps are the most important factor of a PSU's quality” 
     
    A PSU is not just a box of caps. There is also the bridge rectifier, transformer, MOSFETs (particularly the 12V ones), fan quality, and topology to worry about when judging a PSU’s quality. The capacitor plague is over; for most modern PSUs, tier 2 Teapo/whatever else caps are as good as Japanese caps, and MOSFETs/the fan are usually the first to fail anyway (Stefan Payne). 
     
    “PSU calculators are good for estimating your system’s wattage” 
     
    Ever notice how PSU calculators have affiliate links for their recommended PSUs? Yeah, they’re not there for YOUR benefit, they’re there to make money. And in this case, it’s by overestimating the wattage a lot so that they make money off higher wattage PSUs, which cost more thus bringing better profit margins. Outervision is by far the worst offender here; when I put in an RX 5700 XT, 3600, 2x8gb RAM, 1 SSD, 1 HDD, I got ~550W for my power consumption during gaming. However, when I look at Tom’s Hardware power consumption measurements, I see that the 3600 draws about 70W when PBO is turned on during gaming, and RX 5700 XT full OC draws 250W during gaming. Add 30W for everything else (because everything else draws minimal wattage) and we get 350W, which is FAR from the 550W that Outervision proclaimed. 

    Be Quiet!’s PSU calculator doesn’t brashly overestimate as much as Outervision’s; when I put in the same specs, I got 387W, which is still overestimating by ~37W but not as much as 550W. However, Be Quiet! Calculator is still overestimating for their own benefit, not yours. Moral of the story? Use THW’s power consumption measurements for the CPU and GPU, then tack on 30W for your true power consumption when gaming and overclocking. Source 1 Source 2 Source 3 Source 4
     
    “Get a PSU with higher wattage than you need for headroom” 
     
    Nothing about the current trend of CPU and GPU power consumption says that their power consumption will go up in the future. In fact, with die shrinks, the opposite is likely to happen, as same core CPU from Zen+ to Zen 2 power consumption has dropped, similarly with GCN to RDNA. 
     
    For example, 450W is enough to run a Vega 56/64 and 8700k overclocked to 4.9ghz (both @Stefan Payne and @LienusLateTips did this). That shows that most builds really don’t need 550W+. 
  16. Like
    hello_there_123 got a reaction from Vejnemojnen in Common PSU Misconceptions   
    Some common PSU misconceptions
     
    “80+ rating is the best indicator of PSU quality”
     
    Not wholly true. You also have to look at ripple, voltage regulation, transient response, noise, fan bearing, topology, load regulation (including 12V crossloads), temperatures (includes rated temperature, 30C 40C 50C etc.). That’s why even though Raider II spinoffs like EVGA GQ and Be Quiet! Pure Power 10/11 are Gold, they are still worse than Corsair CX in quality (ACRF vs Half Bridge LLC). 
     
    “You should get a higher wattage PSU because of efficiency curve”
     
    First, let’s look at what an efficiency curve on a modern PSU looks like. 

    (Left percentage is 115VAC, right is 230VAC) Efficiency peaks at 200/650 = 31% of the rated wattage at ~91.5/93%, then decreases until it hits ~88/92%  at max load and ~87.5/91.5% above 700W. The difference between max load and peak load efficiency is 3.5/1%, which is extremely insignificant. 
    Even on older platforms, this trend of efficiency plateauing above ~20-40% load is present:

     
    Now for the real world calculations: let’s say that a person is debating on whether to get a 550W or 750W PSU for a build that consumes 400W when overclocked to its max. Also, he’s debating on whether to get the Corsair RMx 550 or 750. 
     
    These are the efficiency curves for both: 


    The 550W has ~90/92% efficiency at 400W, and the 750 also has ~90/92% efficiency at 400W. Conclusion? He should get the 550 since it’s probably much cheaper, he doesn’t need 750W anyway since his build consumes 400W max, and the efficiency is the same! Source 1 Source 2 Source 3 Source 4
     
    “Get more wattage than you need because it will last longer/running your PSU at max will significantly decrease its lifespan” 
     
    First off, most modern not trash PSUs are rated for 40C or higher, which is good for most cases. And the advertised wattage is usually CONTINUOUS, not PEAK. If a 550W continuous PSU at 40C were not able to deliver 550W constantly at 40C, then that would be false advertising and the manufacturer would get sued. Thus, the claim that running the PSU at max would significantly decrease its lifespan is generally false. Besides, most PSUs are overbuilt to ~120%+ of their rated wattage, and you can see that from where the protections are set. 
     
    More wattage = last longer isn’t necessarily true; a MOSFET or the fan can still fail under high wattage PSUs. In fact, higher wattage PSUs can be more dangerous; this is especially true for single rail PSUs, because highly set OCP (OPP’s even slower!) is much slower than lower set OCP. And in the event of a short within the PSU or within the motherboard, FET/VRM failure on the motherboard etc. the OCP/OPP on the PSU won’t be fast enough to trip before the short causes massive damage on the PSU and/or the motherboard. Source 1 Source 2
     
    “PSUs branded by [Brand name] are always good” 
     
    Every brand except Delta and Diablotek, etc have made good and bad PSUs (Delta have made only good PSUs if you don’t care about noise, and Diablotek only makes bombs). The most common brands I've seen people spout as “never having made a bad PSU” are Seasonic and EVGA. 
     
    First off, both have their fair share of bad PSUs. Look no further than Seasonic’s S12ii/M12ii/EVO series or their S12III series, and EVGA’s G3 (protections are set too high)/BQ/BT etc. 
     
    Secondly, EVGA doesn’t even make PSUs (that job goes to FSP, HEC, Super Flower, RSY, Andyson etc.), and Seasonic along with Super Flower often outsource to RSY because their manufacturing capacity is small. If Seasonic/Super Flower choose to manufacture a certain PSU themselves, the Quality Control will be worse on those because every part is outsourced, compared to FSP, CWT etc who do not outsource the PCB-A. 
     
    This misconception also falls under statements such as “just get a brand name PSU, they’re all fine”. 
     
    “Modularity is extremely important for the consumer” 
     
    Truth: unless the person in question is doing Cablemod, Mod vs Semi-Mod will make absolutely no difference, as the cables that are preattached on semimod PSUs will be necessary cables, such as the 24pin for motherboard. Also, in my and many other people’s experiences, on lower wattages (<=550W) nonmod is actually EASIER to work with than semi/full mod because there’s only a few cables left over from nonmod, which are easily ziptied and tucked under the PSU shroud (or left out in the open). Of course the issue of nonmod vs semi-mod is subjective, but at least I have found that the ~2 cables to ziptie after all the cables are attached on nonmod PSUs are easily dealt with. 
     
    “Japanese Caps are the most important factor of a PSU's quality” 
     
    A PSU is not just a box of caps. There is also the bridge rectifier, transformer, MOSFETs (particularly the 12V ones), fan quality, and topology to worry about when judging a PSU’s quality. The capacitor plague is over; for most modern PSUs, tier 2 Teapo/whatever else caps are as good as Japanese caps, and MOSFETs/the fan are usually the first to fail anyway (Stefan Payne). 
     
    “PSU calculators are good for estimating your system’s wattage” 
     
    Ever notice how PSU calculators have affiliate links for their recommended PSUs? Yeah, they’re not there for YOUR benefit, they’re there to make money. And in this case, it’s by overestimating the wattage a lot so that they make money off higher wattage PSUs, which cost more thus bringing better profit margins. Outervision is by far the worst offender here; when I put in an RX 5700 XT, 3600, 2x8gb RAM, 1 SSD, 1 HDD, I got ~550W for my power consumption during gaming. However, when I look at Tom’s Hardware power consumption measurements, I see that the 3600 draws about 70W when PBO is turned on during gaming, and RX 5700 XT full OC draws 250W during gaming. Add 30W for everything else (because everything else draws minimal wattage) and we get 350W, which is FAR from the 550W that Outervision proclaimed. 

    Be Quiet!’s PSU calculator doesn’t brashly overestimate as much as Outervision’s; when I put in the same specs, I got 387W, which is still overestimating by ~37W but not as much as 550W. However, Be Quiet! Calculator is still overestimating for their own benefit, not yours. Moral of the story? Use THW’s power consumption measurements for the CPU and GPU, then tack on 30W for your true power consumption when gaming and overclocking. Source 1 Source 2 Source 3 Source 4
     
    “Get a PSU with higher wattage than you need for headroom” 
     
    Nothing about the current trend of CPU and GPU power consumption says that their power consumption will go up in the future. In fact, with die shrinks, the opposite is likely to happen, as same core CPU from Zen+ to Zen 2 power consumption has dropped, similarly with GCN to RDNA. 
     
    For example, 450W is enough to run a Vega 56/64 and 8700k overclocked to 4.9ghz (both @Stefan Payne and @LienusLateTips did this). That shows that most builds really don’t need 550W+. 
  17. Like
    hello_there_123 got a reaction from Slayer3032 in Common PSU Misconceptions   
    Some common PSU misconceptions
     
    “80+ rating is the best indicator of PSU quality”
     
    Not wholly true. You also have to look at ripple, voltage regulation, transient response, noise, fan bearing, topology, load regulation (including 12V crossloads), temperatures (includes rated temperature, 30C 40C 50C etc.). That’s why even though Raider II spinoffs like EVGA GQ and Be Quiet! Pure Power 10/11 are Gold, they are still worse than Corsair CX in quality (ACRF vs Half Bridge LLC). 
     
    “You should get a higher wattage PSU because of efficiency curve”
     
    First, let’s look at what an efficiency curve on a modern PSU looks like. 

    (Left percentage is 115VAC, right is 230VAC) Efficiency peaks at 200/650 = 31% of the rated wattage at ~91.5/93%, then decreases until it hits ~88/92%  at max load and ~87.5/91.5% above 700W. The difference between max load and peak load efficiency is 3.5/1%, which is extremely insignificant. 
    Even on older platforms, this trend of efficiency plateauing above ~20-40% load is present:

     
    Now for the real world calculations: let’s say that a person is debating on whether to get a 550W or 750W PSU for a build that consumes 400W when overclocked to its max. Also, he’s debating on whether to get the Corsair RMx 550 or 750. 
     
    These are the efficiency curves for both: 


    The 550W has ~90/92% efficiency at 400W, and the 750 also has ~90/92% efficiency at 400W. Conclusion? He should get the 550 since it’s probably much cheaper, he doesn’t need 750W anyway since his build consumes 400W max, and the efficiency is the same! Source 1 Source 2 Source 3 Source 4
     
    “Get more wattage than you need because it will last longer/running your PSU at max will significantly decrease its lifespan” 
     
    First off, most modern not trash PSUs are rated for 40C or higher, which is good for most cases. And the advertised wattage is usually CONTINUOUS, not PEAK. If a 550W continuous PSU at 40C were not able to deliver 550W constantly at 40C, then that would be false advertising and the manufacturer would get sued. Thus, the claim that running the PSU at max would significantly decrease its lifespan is generally false. Besides, most PSUs are overbuilt to ~120%+ of their rated wattage, and you can see that from where the protections are set. 
     
    More wattage = last longer isn’t necessarily true; a MOSFET or the fan can still fail under high wattage PSUs. In fact, higher wattage PSUs can be more dangerous; this is especially true for single rail PSUs, because highly set OCP (OPP’s even slower!) is much slower than lower set OCP. And in the event of a short within the PSU or within the motherboard, FET/VRM failure on the motherboard etc. the OCP/OPP on the PSU won’t be fast enough to trip before the short causes massive damage on the PSU and/or the motherboard. Source 1 Source 2
     
    “PSUs branded by [Brand name] are always good” 
     
    Every brand except Delta and Diablotek, etc have made good and bad PSUs (Delta have made only good PSUs if you don’t care about noise, and Diablotek only makes bombs). The most common brands I've seen people spout as “never having made a bad PSU” are Seasonic and EVGA. 
     
    First off, both have their fair share of bad PSUs. Look no further than Seasonic’s S12ii/M12ii/EVO series or their S12III series, and EVGA’s G3 (protections are set too high)/BQ/BT etc. 
     
    Secondly, EVGA doesn’t even make PSUs (that job goes to FSP, HEC, Super Flower, RSY, Andyson etc.), and Seasonic along with Super Flower often outsource to RSY because their manufacturing capacity is small. If Seasonic/Super Flower choose to manufacture a certain PSU themselves, the Quality Control will be worse on those because every part is outsourced, compared to FSP, CWT etc who do not outsource the PCB-A. 
     
    This misconception also falls under statements such as “just get a brand name PSU, they’re all fine”. 
     
    “Modularity is extremely important for the consumer” 
     
    Truth: unless the person in question is doing Cablemod, Mod vs Semi-Mod will make absolutely no difference, as the cables that are preattached on semimod PSUs will be necessary cables, such as the 24pin for motherboard. Also, in my and many other people’s experiences, on lower wattages (<=550W) nonmod is actually EASIER to work with than semi/full mod because there’s only a few cables left over from nonmod, which are easily ziptied and tucked under the PSU shroud (or left out in the open). Of course the issue of nonmod vs semi-mod is subjective, but at least I have found that the ~2 cables to ziptie after all the cables are attached on nonmod PSUs are easily dealt with. 
     
    “Japanese Caps are the most important factor of a PSU's quality” 
     
    A PSU is not just a box of caps. There is also the bridge rectifier, transformer, MOSFETs (particularly the 12V ones), fan quality, and topology to worry about when judging a PSU’s quality. The capacitor plague is over; for most modern PSUs, tier 2 Teapo/whatever else caps are as good as Japanese caps, and MOSFETs/the fan are usually the first to fail anyway (Stefan Payne). 
     
    “PSU calculators are good for estimating your system’s wattage” 
     
    Ever notice how PSU calculators have affiliate links for their recommended PSUs? Yeah, they’re not there for YOUR benefit, they’re there to make money. And in this case, it’s by overestimating the wattage a lot so that they make money off higher wattage PSUs, which cost more thus bringing better profit margins. Outervision is by far the worst offender here; when I put in an RX 5700 XT, 3600, 2x8gb RAM, 1 SSD, 1 HDD, I got ~550W for my power consumption during gaming. However, when I look at Tom’s Hardware power consumption measurements, I see that the 3600 draws about 70W when PBO is turned on during gaming, and RX 5700 XT full OC draws 250W during gaming. Add 30W for everything else (because everything else draws minimal wattage) and we get 350W, which is FAR from the 550W that Outervision proclaimed. 

    Be Quiet!’s PSU calculator doesn’t brashly overestimate as much as Outervision’s; when I put in the same specs, I got 387W, which is still overestimating by ~37W but not as much as 550W. However, Be Quiet! Calculator is still overestimating for their own benefit, not yours. Moral of the story? Use THW’s power consumption measurements for the CPU and GPU, then tack on 30W for your true power consumption when gaming and overclocking. Source 1 Source 2 Source 3 Source 4
     
    “Get a PSU with higher wattage than you need for headroom” 
     
    Nothing about the current trend of CPU and GPU power consumption says that their power consumption will go up in the future. In fact, with die shrinks, the opposite is likely to happen, as same core CPU from Zen+ to Zen 2 power consumption has dropped, similarly with GCN to RDNA. 
     
    For example, 450W is enough to run a Vega 56/64 and 8700k overclocked to 4.9ghz (both @Stefan Payne and @LienusLateTips did this). That shows that most builds really don’t need 550W+. 
  18. Like
    hello_there_123 got a reaction from Fasauceome in Common PSU Misconceptions   
    Some common PSU misconceptions
     
    “80+ rating is the best indicator of PSU quality”
     
    Not wholly true. You also have to look at ripple, voltage regulation, transient response, noise, fan bearing, topology, load regulation (including 12V crossloads), temperatures (includes rated temperature, 30C 40C 50C etc.). That’s why even though Raider II spinoffs like EVGA GQ and Be Quiet! Pure Power 10/11 are Gold, they are still worse than Corsair CX in quality (ACRF vs Half Bridge LLC). 
     
    “You should get a higher wattage PSU because of efficiency curve”
     
    First, let’s look at what an efficiency curve on a modern PSU looks like. 

    (Left percentage is 115VAC, right is 230VAC) Efficiency peaks at 200/650 = 31% of the rated wattage at ~91.5/93%, then decreases until it hits ~88/92%  at max load and ~87.5/91.5% above 700W. The difference between max load and peak load efficiency is 3.5/1%, which is extremely insignificant. 
    Even on older platforms, this trend of efficiency plateauing above ~20-40% load is present:

     
    Now for the real world calculations: let’s say that a person is debating on whether to get a 550W or 750W PSU for a build that consumes 400W when overclocked to its max. Also, he’s debating on whether to get the Corsair RMx 550 or 750. 
     
    These are the efficiency curves for both: 


    The 550W has ~90/92% efficiency at 400W, and the 750 also has ~90/92% efficiency at 400W. Conclusion? He should get the 550 since it’s probably much cheaper, he doesn’t need 750W anyway since his build consumes 400W max, and the efficiency is the same! Source 1 Source 2 Source 3 Source 4
     
    “Get more wattage than you need because it will last longer/running your PSU at max will significantly decrease its lifespan” 
     
    First off, most modern not trash PSUs are rated for 40C or higher, which is good for most cases. And the advertised wattage is usually CONTINUOUS, not PEAK. If a 550W continuous PSU at 40C were not able to deliver 550W constantly at 40C, then that would be false advertising and the manufacturer would get sued. Thus, the claim that running the PSU at max would significantly decrease its lifespan is generally false. Besides, most PSUs are overbuilt to ~120%+ of their rated wattage, and you can see that from where the protections are set. 
     
    More wattage = last longer isn’t necessarily true; a MOSFET or the fan can still fail under high wattage PSUs. In fact, higher wattage PSUs can be more dangerous; this is especially true for single rail PSUs, because highly set OCP (OPP’s even slower!) is much slower than lower set OCP. And in the event of a short within the PSU or within the motherboard, FET/VRM failure on the motherboard etc. the OCP/OPP on the PSU won’t be fast enough to trip before the short causes massive damage on the PSU and/or the motherboard. Source 1 Source 2
     
    “PSUs branded by [Brand name] are always good” 
     
    Every brand except Delta and Diablotek, etc have made good and bad PSUs (Delta have made only good PSUs if you don’t care about noise, and Diablotek only makes bombs). The most common brands I've seen people spout as “never having made a bad PSU” are Seasonic and EVGA. 
     
    First off, both have their fair share of bad PSUs. Look no further than Seasonic’s S12ii/M12ii/EVO series or their S12III series, and EVGA’s G3 (protections are set too high)/BQ/BT etc. 
     
    Secondly, EVGA doesn’t even make PSUs (that job goes to FSP, HEC, Super Flower, RSY, Andyson etc.), and Seasonic along with Super Flower often outsource to RSY because their manufacturing capacity is small. If Seasonic/Super Flower choose to manufacture a certain PSU themselves, the Quality Control will be worse on those because every part is outsourced, compared to FSP, CWT etc who do not outsource the PCB-A. 
     
    This misconception also falls under statements such as “just get a brand name PSU, they’re all fine”. 
     
    “Modularity is extremely important for the consumer” 
     
    Truth: unless the person in question is doing Cablemod, Mod vs Semi-Mod will make absolutely no difference, as the cables that are preattached on semimod PSUs will be necessary cables, such as the 24pin for motherboard. Also, in my and many other people’s experiences, on lower wattages (<=550W) nonmod is actually EASIER to work with than semi/full mod because there’s only a few cables left over from nonmod, which are easily ziptied and tucked under the PSU shroud (or left out in the open). Of course the issue of nonmod vs semi-mod is subjective, but at least I have found that the ~2 cables to ziptie after all the cables are attached on nonmod PSUs are easily dealt with. 
     
    “Japanese Caps are the most important factor of a PSU's quality” 
     
    A PSU is not just a box of caps. There is also the bridge rectifier, transformer, MOSFETs (particularly the 12V ones), fan quality, and topology to worry about when judging a PSU’s quality. The capacitor plague is over; for most modern PSUs, tier 2 Teapo/whatever else caps are as good as Japanese caps, and MOSFETs/the fan are usually the first to fail anyway (Stefan Payne). 
     
    “PSU calculators are good for estimating your system’s wattage” 
     
    Ever notice how PSU calculators have affiliate links for their recommended PSUs? Yeah, they’re not there for YOUR benefit, they’re there to make money. And in this case, it’s by overestimating the wattage a lot so that they make money off higher wattage PSUs, which cost more thus bringing better profit margins. Outervision is by far the worst offender here; when I put in an RX 5700 XT, 3600, 2x8gb RAM, 1 SSD, 1 HDD, I got ~550W for my power consumption during gaming. However, when I look at Tom’s Hardware power consumption measurements, I see that the 3600 draws about 70W when PBO is turned on during gaming, and RX 5700 XT full OC draws 250W during gaming. Add 30W for everything else (because everything else draws minimal wattage) and we get 350W, which is FAR from the 550W that Outervision proclaimed. 

    Be Quiet!’s PSU calculator doesn’t brashly overestimate as much as Outervision’s; when I put in the same specs, I got 387W, which is still overestimating by ~37W but not as much as 550W. However, Be Quiet! Calculator is still overestimating for their own benefit, not yours. Moral of the story? Use THW’s power consumption measurements for the CPU and GPU, then tack on 30W for your true power consumption when gaming and overclocking. Source 1 Source 2 Source 3 Source 4
     
    “Get a PSU with higher wattage than you need for headroom” 
     
    Nothing about the current trend of CPU and GPU power consumption says that their power consumption will go up in the future. In fact, with die shrinks, the opposite is likely to happen, as same core CPU from Zen+ to Zen 2 power consumption has dropped, similarly with GCN to RDNA. 
     
    For example, 450W is enough to run a Vega 56/64 and 8700k overclocked to 4.9ghz (both @Stefan Payne and @LienusLateTips did this). That shows that most builds really don’t need 550W+. 
  19. Informative
    hello_there_123 reacted to OrionFOTL in is a 530w psu enough for i7 8700 w 1070ti mini?   
    Thermaltake Smart SE Gold 530W is a rebranded CWT GPN 550W (it's actually 550W, not 530W). It's a low end platform with DC-DC converters. It's not garbage but its transient response isn't quite known, and if it's bad, your computer may shut down due to high transient spikes from top end graphics card. I'd keep your PSU and see if you encounter any problems.
     
    Note to others: please notice it's not Thermaltake Smart SE Bronze (based on group regulated CWT GPM), but Smart SE Gold.
  20. Like
    hello_there_123 reacted to Fasauceome in Looking for recommendations on parts for new PC   
    what a coincidence, just made a list in this price range.
    PCPartPicker Part List Type Item Price CPU AMD Ryzen 5 3600 3.6 GHz 6-Core Processor $264.50 @ Vuugo Motherboard MSI B450M GAMING PLUS Micro ATX AM4 Motherboard $109.99 @ Memory Express Memory G.Skill Ripjaws V Series 16 GB (2 x 8 GB) DDR4-3200 Memory $94.99 @ Newegg Canada Storage Intel 660p Series 1.02 TB M.2-2280 NVME Solid State Drive $119.99 @ Canada Computers Video Card Gigabyte Radeon RX 5700 XT 8 GB GAMING OC Video Card $578.99 @ PC-Canada Case Thermaltake Versa J23 RGB ATX Mid Tower Case $126.89 @ Amazon Canada Power Supply Corsair TXM Gold 550 W 80+ Gold Certified Semi-modular ATX Power Supply $109.50 @ Vuugo   Prices include shipping, taxes, rebates, and discounts     Total $1404.85   Generated by PCPartPicker 2019-09-16 22:29 EDT-0400    
    for your suggested price range you'll have to make compromises to accommodate a CPU that expensive.
  21. Agree
    hello_there_123 got a reaction from LogicWeasel in Gaming PC First Build   
    I also personally find the Kingston A400 to be much more reliable than the 970 pro. See what subjectivity does without any facts such as failure rates? 
  22. Agree
    hello_there_123 reacted to 5x5 in Smart BX1 RGB 550W   
    It's a good ballistic missile warhead. Not a great power supply, though
  23. Agree
    hello_there_123 reacted to VEXICUS in Are these SSDs from unknown (to me) brands any good?   
    The tcsunbow x3 drive is pretty good for the most part.
    It has pretty decent internals with a sm2258 controller and also houses in a 1gb ddr3 DRAM cache from micron (d9mnt).
  24. Like
    hello_there_123 got a reaction from ZBVirus24 in BitFenix Formula 550W or be quiet! Pure Power 11 500W?   
    Classic turtle rig, lol. Both have a 5 year warranty
  25. Agree
    hello_there_123 got a reaction from dgsddfgdfhgs in Are these SSDs from unknown (to me) brands any good?   
    Do you just automatically assume that everything you're not familiar with is bad? 
    And what's inherently bad with TcSunBow being a Chinese brand? 
    If you would've looked at Newmaxx's SSD Spreadsheet that I linked earlier, you would see that the X3 is an SM2258 drive, same controller as the MX500, just with worse binned and 32L nand vs 64L nand, probably a smaller SLC cache and worse firmware, none of which are likely to affect most people. 
×