Author Topic: adding capacitors to a PC power supplies output [ project ]  (Read 6628 times)

0 Members and 1 Guest are viewing this topic.

Offline forestelfTopic starter

  • Newbie
  • Posts: 5
  • Country: ca
adding capacitors to a PC power supplies output [ project ]
« on: March 14, 2019, 04:37:16 pm »
hi, im forestelf im new here and i wanted to throw a question out to knowledgeable ppls ~ !
for some background=
i own a EVGA 750W GQ semimoduar computer power supply and it has ripple level of about ~60mv on the 12v rail,
and
i was reading a review over at toms hardware
https://www.tomshardware.com/reviews/evga-gq-series-750w-psu,4396-3.html

and over on the front of the powersupplys' pcb (where the cables connect into, there is a some places designed to have additional capacitors soldered to them, but they were left empty,(presumably a cost related decision)
and in the tomshardware review, they mention "At the front side of the modular PCB is free space for installing additional filtering caps, though EVGA apparently decided it didn't need them. That's a shame, since this platform doesn't have good ripple suppression." (according to their tests)

so i had an idea to reach out to more knowledgeable ppl than myself about this,
about whether this is something beneficial to do, or if its even possible~
 
(i have soldering skills but not a massive knowledge of pcb electronic circutry)
theres alot of photos up on the tomshardware review, lots of info too,
&if theres anything youd need to know that i can help with,  please ask :)

i guess the main questions are; is it possible and is it beneficial(even if only a little bit);
but more importantly where would i find out which capacitor would be best for the job and would the existing circutry allow for these capacitors to simply be added on their own?

any help is much appreciated, thanks for droppin by my post :D <3
« Last Edit: March 14, 2019, 04:39:53 pm by forestelf »
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #1 on: March 14, 2019, 05:28:04 pm »
There's extremely small benefit to adding capacitors on the output, and 60mV is well within the acceptable range (up to 120mV for 12v I think).
Adding a bit of capacitance by the connectors may help a few more Mhz when you're overclocking video cards but that's about it.

Keep in mind that you have capacitors in the power supply and there's capacitors on the video card or motherboard as well, so you basically have C-L-C , because the wires have some inductance.
Whatever you add on the power supply would not make a huge amount of difference.

I don't know, if you want to add them... aim for polymer capacitors, anything in the 270-470uF 16v would be suitable.
 
The following users thanked this post: forestelf

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 10173
  • Country: gb
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #2 on: March 14, 2019, 05:29:20 pm »
Welcome to the forum.

Personally, I would leave well alone. With so many heavy wires coming out of the PCB, there is far too much chance of bending the existing caps over and damaging their seals, shortening their life.

Adding extra caps might reduce ripple slightly, but they would need to be the same low impedance ones as the existing ones to share current properly. The manufacturer has put in as many as are needed to meet spec. Adding more (apart from the possible mechanical damage I mentioned above), could well reduce their life more by blocking airflow and increasing temperature, by more than they save.

Is it just me, or do PC PSUs get messier the more advanced they get? I've never seen so many wires coming out of a PSU PCB.  :o
« Last Edit: March 14, 2019, 05:35:29 pm by Gyro »
Best Regards, Chris
 
The following users thanked this post: forestelf

Offline WildMOSFET

  • Contributor
  • Posts: 24
  • Country: it
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #3 on: March 15, 2019, 08:53:27 pm »
Welcome to the forum.

Personally, I would leave well alone. With so many heavy wires coming out of the PCB, there is far too much chance of bending the existing caps over and damaging their seals, shortening their life.

Adding extra caps might reduce ripple slightly, but they would need to be the same low impedance ones as the existing ones to share current properly. The manufacturer has put in as many as are needed to meet spec. Adding more (apart from the possible mechanical damage I mentioned above), could well reduce their life more by blocking airflow and increasing temperature, by more than they save.

Is it just me, or do PC PSUs get messier the more advanced they get? I've never seen so many wires coming out of a PSU PCB.  :o

I agree in everything.
If the manufacturer has decided not to put them, they probably aren't necessary.
And we're talking about EVGA not some scrapped producer.
 

Offline forestelfTopic starter

  • Newbie
  • Posts: 5
  • Country: ca
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #4 on: March 20, 2019, 03:31:14 pm »
There's extremely small benefit to adding capacitors on the output, and 60mV is well within the acceptable range (up to 120mV for 12v I think).
Adding a bit of capacitance by the connectors may help a few more Mhz when you're overclocking video cards but that's about it.

Keep in mind that you have capacitors in the power supply and there's capacitors on the video card or motherboard as well, so you basically have C-L-C , because the wires have some inductance.
Whatever you add on the power supply would not make a huge amount of difference.

I don't know, if you want to add them... aim for polymer capacitors, anything in the 270-470uF 16v would be suitable.
hi thank you for the info! i compared some other psu capacitors and they were in that range,(not sure the difference in the colors and brands and whatnot, but im attaching a few photos as example of whats typically found on psu,
on this psu-teardown page about half way down theres photos and description of all capacitors that are in this psu,(there japanese aparently)
( https://www.tomshardware.com/reviews/evga-gq-series-750w-psu,4396-3.html )(also added some of those pics onto this post) with this in mind, can you recomend one on ebay for me ? the price isnt important i can do the bargain hunting, itd just be for a closer reference! im seeking those small benefits so everybit counts,(even small overclocking benefits and prolonging motherboard components lifespan) i understand 60mv is acceptable, but if theres a performance gain to be had why not go for it? im wanting to smooth the ripple closest to the source and i think this is the simplest place (happens to even be designed for it)
Welcome to the forum.

Personally, I would leave well alone. With so many heavy wires coming out of the PCB, there is far too much chance of bending the existing caps over and damaging their seals, shortening their life.

Adding extra caps might reduce ripple slightly, but they would need to be the same low impedance ones as the existing ones to share current properly. The manufacturer has put in as many as are needed to meet spec. Adding more (apart from the possible mechanical damage I mentioned above), could well reduce their life more by blocking airflow and increasing temperature, by more than they save.

Is it just me, or do PC PSUs get messier the more advanced they get? I've never seen so many wires coming out of a PSU PCB.  :o
the place these capacitors are located have perfect clearance, no extra wires are in the way, all the higher end models have them in this place and similar pcbs use them as well, the slight benefit is the goal here, im not expecting the psu to transform into the best power supply of all time or anything lolol but the case allows ample room to secure a fit without modification.

Welcome to the forum.

Personally, I would leave well alone. With so many heavy wires coming out of the PCB, there is far too much chance of bending the existing caps over and damaging their seals, shortening their life.

Adding extra caps might reduce ripple slightly, but they would need to be the same low impedance ones as the existing ones to share current properly. The manufacturer has put in as many as are needed to meet spec. Adding more (apart from the possible mechanical damage I mentioned above), could well reduce their life more by blocking airflow and increasing temperature, by more than they save.

Is it just me, or do PC PSUs get messier the more advanced they get? I've never seen so many wires coming out of a PSU PCB.  :o

I agree in everything.
If the manufacturer has decided not to put them, they probably aren't necessary.
And we're talking about EVGA not some scrapped producer.
jus feel like if theres room for improvement i should do it.  its not a high price unit or anything.(so it was perhaps an afterthought on evgas' side for reducing its cost/performance ratio as all similar models have these caps in those places its just this is not the top of the line version for the evga-GQ models) for me a couple japanese caps would presumably be a few dollars or so on ebay and is reasonable, i get that for a manufacturer catering to the budget psu consumer, that leaving out these caps on millions of units saves them alot of monies on their production and still results ina competitive unit for similar psus, but being the pcb wasent changed it leaves designated room for the caps with no physical modification to the unit being necessary so it seems like a great project right


i get it could be a hassle for many people to go through this for such a small gain, but if im willing to, and for the small benefits/ if i could find a reccomendation for a particular capacitor on ebay id really like to go for it c:!

( first 3 photos show this 750w psus capacitor types on the main board last 3 photos show various modular psu pcb with typical cap location,)
« Last Edit: March 20, 2019, 03:43:26 pm by forestelf »
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 10173
  • Country: gb
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #5 on: March 20, 2019, 05:22:54 pm »
Feel free to spend your money.

One caution, please don't expect to find suitable genuine capacitors on ebay - they will most likely be fake. To need to identify the manufacturer's parts used and purchase from a proper component distributor.

Best Regards, Chris
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #6 on: March 20, 2019, 06:55:18 pm »
I was going to post that those spots only look large enough for polymer aluminum electrolytic capacitors and the second set of photographs shows exactly that in most cases.

I do not think there will be any problem adding those capacitors and while they will reduce high frequency noise, I doubt this will make any practical difference to the system.  If you are having a problem with reliably, those capacitors will not make up for the larger bulk output capacitors being worn out or too small.  If there is a real problem, I would replace the bulk output capacitors first or in addition to adding the smaller capacitors.
 
The following users thanked this post: forestelf

Offline forestelfTopic starter

  • Newbie
  • Posts: 5
  • Country: ca
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #7 on: March 24, 2019, 04:27:42 pm »
Feel free to spend your money.

One caution, please don't expect to find suitable genuine capacitors on ebay - they will most likely be fake. To need to identify the manufacturer's parts used and purchase from a proper component distributor.


do you know some reliable online places peoples here would consider in general?
 

Offline forestelfTopic starter

  • Newbie
  • Posts: 5
  • Country: ca
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #8 on: March 24, 2019, 04:35:48 pm »
I was going to post that those spots only look large enough for polymer aluminum electrolytic capacitors and the second set of photographs shows exactly that in most cases.

I do not think there will be any problem adding those capacitors and while they will reduce high frequency noise, I doubt this will make any practical difference to the system.  If you are having a problem with reliably, those capacitors will not make up for the larger bulk output capacitors being worn out or too small.  If there is a real problem, I would replace the bulk output capacitors first or in addition to adding the smaller capacitors.
mhm :3 (this might be simple question but is there difference between the blue and red polymer capacitors?voltage or something maybe?)  im not having issues with noise, just felt like a project~
would this likely be the main (bulk output?) filtering capacitor? =

ps im prob not going to touch it without discharging it or being grounded first~
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #9 on: March 25, 2019, 12:43:14 am »
Yes, that's big capacitor is the main one which will have up to 400-420v DC on it while the power supply runs.
Right near it, there's a "bleeder resistor"which slowly discharges the capacitor as soon as the psu is turned off.  However, sometimes these resistors aren't installed or they're broken, so it's possible to have high voltage on that capacitor after psu is shut down. So NEVER assume that capacitor is discharged.
Safest would be to get a multimeter and MEASURE the voltage on its terminals before you get close to it. If there's more than around 10-20v on it, discharge it. You COULD just use a screwdriver and make a short circuit between the two terminals, but if there's a lot of energy in the capacitor you can damage the screwdriver and in rarer cases even the circuit board if you just short the capacitor with a screwdriver.
Safer method would be to get a resistor (let's say 100 ohm 1w or higher) and carefully connect the resistor between the terminals for a few seconds and then measure voltage again with multimeter.

regarding color on polymer capacitors - it's mostly related to who manufacturers them, some company may prefer RED (for example nichicon), other may prefer light blue (I think teapo for example), others use the plain black or gray.  You can often tell which manufacturer is by the series - for example FP is an indicative of "nichicon fp series", fp being short for "functional polymer" or something like that.
In one of your posts, there's one with FP and 49cr written on it - FP is series, 49cr is factory code, like manufacturing date or extra series information, and then there's 2 numbers ... the one ending with 1 is the capacitance in short form ... like for example 471 means  47 x 101 = 470uF  , and the last line 16 means voltage, 16v.
In another picture there's some light blue capacitors, those are made by Nippon Chemi Con / United Chemi Con  - i can tell that because of the tiny "hat", the small rectangle like thing over the text
Here's an example (click on picture to see what i mean) : https://www.digikey.com/product-detail/en/united-chemi-con/APSC6R3ELL152MJB5S/565-3229-ND/1826693
In your picture, you easily see the capacitance (330uF - in long form because the capacitor is wide enough to fit more than 3 digits), and 16 which means voltage rating.  The series may be determined (or not) from the letter and numbers above capacitance.

There's usually enough room on these capacitors to have capacitance and voltage written on them
Yes, there's differences between them, like ripple current values, esr etc, but in power supplies you want polymer with very low esr and high current ripple so there's a bunch of series with similar values, therefore you don't have to use the exact series from the exact maker.

Don't mess with the main capacitor, won't make any difference and it's from a quality manufacturer.
 

Offline forestelfTopic starter

  • Newbie
  • Posts: 5
  • Country: ca
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #10 on: March 27, 2019, 04:06:17 pm »
Yes, that's big capacitor is the main one which will have up to 400-420v DC on it while the power supply runs.
Right near it, there's a "bleeder resistor"which slowly discharges the capacitor as soon as the psu is turned off.  However, sometimes these resistors aren't installed or they're broken, so it's possible to have high voltage on that capacitor after psu is shut down. So NEVER assume that capacitor is discharged.
Safest would be to get a multimeter and MEASURE the voltage on its terminals before you get close to it. If there's more than around 10-20v on it, discharge it. You COULD just use a screwdriver and make a short circuit between the two terminals, but if there's a lot of energy in the capacitor you can damage the screwdriver and in rarer cases even the circuit board if you just short the capacitor with a screwdriver.
Safer method would be to get a resistor (let's say 100 ohm 1w or higher) and carefully connect the resistor between the terminals for a few seconds and then measure voltage again with multimeter.

regarding color on polymer capacitors - it's mostly related to who manufacturers them, some company may prefer RED (for example nichicon), other may prefer light blue (I think teapo for example), others use the plain black or gray.  You can often tell which manufacturer is by the series - for example FP is an indicative of "nichicon fp series", fp being short for "functional polymer" or something like that.
In one of your posts, there's one with FP and 49cr written on it - FP is series, 49cr is factory code, like manufacturing date or extra series information, and then there's 2 numbers ... the one ending with 1 is the capacitance in short form ... like for example 471 means  47 x 101 = 470uF  , and the last line 16 means voltage, 16v.
In another picture there's some light blue capacitors, those are made by Nippon Chemi Con / United Chemi Con  - i can tell that because of the tiny "hat", the small rectangle like thing over the text
Here's an example (click on picture to see what i mean) : https://www.digikey.com/product-detail/en/united-chemi-con/APSC6R3ELL152MJB5S/565-3229-ND/1826693
In your picture, you easily see the capacitance (330uF - in long form because the capacitor is wide enough to fit more than 3 digits), and 16 which means voltage rating.  The series may be determined (or not) from the letter and numbers above capacitance.

There's usually enough room on these capacitors to have capacitance and voltage written on them
Yes, there's differences between them, like ripple current values, esr etc, but in power supplies you want polymer with very low esr and high current ripple so there's a bunch of series with similar values, therefore you don't have to use the exact series from the exact maker.

Don't mess with the main capacitor, won't make any difference and it's from a quality manufacturer.

i appreciate you sharing your knowledge;' if this project entailed removing caps i wouldent of considered it+the main capacitor isnt something i was really planning on anyways to be honest, ( also lol@damaging the screwdriver XD thts last priorities there :P )

so i have a question, on capacitors you linked, the c 929 marking at the top of it, what does it mean, dosent match the short form you explained as the number is allready written below but if i compare it to the one in my second post or so the one there is c 529 , is this number a specification or something not as relevant such as factory or date made kinda thing?
i figure there suitable for the task of the front panel location and the price was fine so i was trying to compare it down to the last details -

also what does ripple current at low frequency represent? (looking at caps trying to find something) if its over 600ma at 120khz will it not have effect on the 60mv ripple
 

Offline TimNJ

  • Super Contributor
  • ***
  • Posts: 1720
  • Country: us
Re: adding capacitors to a PC power supplies output [ project ]
« Reply #11 on: March 29, 2019, 02:26:40 am »
Yes, that's big capacitor is the main one which will have up to 400-420v DC on it while the power supply runs.
Right near it, there's a "bleeder resistor"which slowly discharges the capacitor as soon as the psu is turned off.  However, sometimes these resistors aren't installed or they're broken, so it's possible to have high voltage on that capacitor after psu is shut down. So NEVER assume that capacitor is discharged.
Safest would be to get a multimeter and MEASURE the voltage on its terminals before you get close to it. If there's more than around 10-20v on it, discharge it. You COULD just use a screwdriver and make a short circuit between the two terminals, but if there's a lot of energy in the capacitor you can damage the screwdriver and in rarer cases even the circuit board if you just short the capacitor with a screwdriver.
Safer method would be to get a resistor (let's say 100 ohm 1w or higher) and carefully connect the resistor between the terminals for a few seconds and then measure voltage again with multimeter.

regarding color on polymer capacitors - it's mostly related to who manufacturers them, some company may prefer RED (for example nichicon), other may prefer light blue (I think teapo for example), others use the plain black or gray.  You can often tell which manufacturer is by the series - for example FP is an indicative of "nichicon fp series", fp being short for "functional polymer" or something like that.
In one of your posts, there's one with FP and 49cr written on it - FP is series, 49cr is factory code, like manufacturing date or extra series information, and then there's 2 numbers ... the one ending with 1 is the capacitance in short form ... like for example 471 means  47 x 101 = 470uF  , and the last line 16 means voltage, 16v.
In another picture there's some light blue capacitors, those are made by Nippon Chemi Con / United Chemi Con  - i can tell that because of the tiny "hat", the small rectangle like thing over the text
Here's an example (click on picture to see what i mean) : https://www.digikey.com/product-detail/en/united-chemi-con/APSC6R3ELL152MJB5S/565-3229-ND/1826693
In your picture, you easily see the capacitance (330uF - in long form because the capacitor is wide enough to fit more than 3 digits), and 16 which means voltage rating.  The series may be determined (or not) from the letter and numbers above capacitance.

There's usually enough room on these capacitors to have capacitance and voltage written on them
Yes, there's differences between them, like ripple current values, esr etc, but in power supplies you want polymer with very low esr and high current ripple so there's a bunch of series with similar values, therefore you don't have to use the exact series from the exact maker.

Don't mess with the main capacitor, won't make any difference and it's from a quality manufacturer.

i appreciate you sharing your knowledge;' if this project entailed removing caps i wouldent of considered it+the main capacitor isnt something i was really planning on anyways to be honest, ( also lol@damaging the screwdriver XD thts last priorities there :P )

so i have a question, on capacitors you linked, the c 929 marking at the top of it, what does it mean, dosent match the short form you explained as the number is allready written below but if i compare it to the one in my second post or so the one there is c 529 , is this number a specification or something not as relevant such as factory or date made kinda thing?
i figure there suitable for the task of the front panel location and the price was fine so i was trying to compare it down to the last details -

also what does ripple current at low frequency represent? (looking at caps trying to find something) if its over 600ma at 120khz will it not have effect on the 60mv ripple

Don't worry about the CXXX marking on the cap. The 'C' probably just indicates PSC, the series from United Chemi-Con. The other digits are likely a manufacturing date code or factory identifier as you mentioned.

The ripple current rating of an electrolytic/polymer capacitor basically says: "You can run this capacitor at both the stated ripple current spec and at the rated operating temperature (105C typically), and expect to get the stated life expectancy (1,000 - 15,000 hours typically).

The ripple current spec is fundamentally a function of the capacitors equivalent series resistance (ESR). The lower the ESR, the higher the ripple current spec. This is pretty intuitive because a higher ESR means the capacitor will dissipate more heat internally, leading to a higher core temperature, and shortened life.



 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf