Let's say I want to put 40000uf of filter caps on each voltage rail. I need them to be 100v caps. Assume I have these two options available:
1. one 100v, 40000uf cap for each rail
2. two 50v, 40000uf cap in series for each rail
Option 2 is cheaper, which is the direction I'm leaning.
Are there any disadvantages to using two 50v caps in series to make a 100v cap?
awhiteguy
1. one 100v, 40000uf cap for each rail
2. two 50v, 40000uf cap in series for each rail
Option 2 is cheaper, which is the direction I'm leaning.
Are there any disadvantages to using two 50v caps in series to make a 100v cap?
awhiteguy
Equation for series caps capacitance:
C = 1 / ( 1/C1 + 1/C2 +... )
so 2 40000uF 50V caps in series = 20000uF 100V
Parallel caps just add.
You would need to put resistors across the series caps to equalize voltage to avoid making a bomb.
(Capactors have unequal leakage currents.) I would just buy caps with proper ratings.
This has been done before, but mainly for things like 900V where electrolytics aren't available and high capacitance is needed. I have use this to put some 800V .56uF film caps in series / parallel to make a 1600V .56uF cap.
I never build normal projects.
What are you building?
Darrell Harmon
C = 1 / ( 1/C1 + 1/C2 +... )
so 2 40000uF 50V caps in series = 20000uF 100V
Parallel caps just add.
You would need to put resistors across the series caps to equalize voltage to avoid making a bomb.
(Capactors have unequal leakage currents.) I would just buy caps with proper ratings.This has been done before, but mainly for things like 900V where electrolytics aren't available and high capacitance is needed. I have use this to put some 800V .56uF film caps in series / parallel to make a 1600V .56uF cap.
I never build normal projects.What are you building?
Darrell Harmon
First you have to make sure that the caps are charged equally or you may exceed the voltage rating of the individual capacitor. This is obstacle number one.
Second, when connecting caps in series they behave exactly the opposite of resistors that is when you connect then in series the resulting capacitance will be smaller than the largest cap. The formula for series caps is as for parallelling resistors:
1/Ctot = 1/C1 + 1/C2 ..... 1/Cn
In this case you end up with half the size as they are equal and you need to invest in twice as much.
Go for the correct option. Choose the right voltage and buy a size that when parellelled gives you the total capacitance needed a nd the best price/performance.
Parallelling caps gives you a total of the sum of the individual caps so 4 times 10.000 uF is "the same" as 40.000 uF single cap "all other things being equal".
/UrSv
Second, when connecting caps in series they behave exactly the opposite of resistors that is when you connect then in series the resulting capacitance will be smaller than the largest cap. The formula for series caps is as for parallelling resistors:
1/Ctot = 1/C1 + 1/C2 ..... 1/Cn
In this case you end up with half the size as they are equal and you need to invest in twice as much.
Go for the correct option. Choose the right voltage and buy a size that when parellelled gives you the total capacitance needed a nd the best price/performance.
Parallelling caps gives you a total of the sum of the individual caps so 4 times 10.000 uF is "the same" as 40.000 uF single cap "all other things being equal".
/UrSv
Good call Darrell and UrSv, I forgot that series caps use the same equation as parallel resistors. Shame on me...
Sounds like I need to find some 100v caps. Know of any good sources?
I am planning to build one of Anthony Holton's 400w symmetrical mosfet amps for some HT subwoofers.
It will be my first project, so I'm sure I'll have lots of questions. There is so much good info on this forum, but sometimes it's hard to find. Thanks
awhiteguy
Sounds like I need to find some 100v caps. Know of any good sources?
I am planning to build one of Anthony Holton's 400w symmetrical mosfet amps for some HT subwoofers.
It will be my first project, so I'm sure I'll have lots of questions. There is so much good info on this forum, but sometimes it's hard to find. Thanks
awhiteguy
Supplier for caps:
http://www.mouser.com
http://www.digikey.com
Request a catalog from both or download the pdf. You can get about anything from one of the two.
I think this amp may be a bit much for a first project. I think it's best to learn with a smaller cheaper amp before you smoke alot of expensive parts. (I have) I started with the pass labs zen.
Darrell Harmon
http://www.mouser.com
http://www.digikey.com
Request a catalog from both or download the pdf. You can get about anything from one of the two.
I think this amp may be a bit much for a first project. I think it's best to learn with a smaller cheaper amp before you smoke alot of expensive parts. (I have) I started with the pass labs zen.
Darrell Harmon
awhiteguy said:
To make it work properly, you must have charge equalizing resistors across the caps. This distributes the voltage.
http://www.elfa.se/elfa/produkter/se/20/2021348.htm
The resistor above fits "computer grade" caps from RIFA. They can be bought but it works perfectly with any kind.
I just looked at the mouser and digikey catalogs. It looks like $38 for 20000uF 100V computer grade is about the best. It would be cheaper to do something like use a bunch of 4700uF caps in parallel, but it is messy. I have 16 10000uF caps in my ps and it is messy. You might try newark or allied for better prices since they have more selection of caps.
Darrell Harmon
Darrell Harmon
That's why I wanted to use the lower voltage caps, because the 100v caps are so expensive. Man, I'm going to end up with $200 just in the power supply. Oh, well.
What are the thoughts on used caps? Sounds risky to me since brand new ones are only good for like 2000 or 3000 hours. I have wondered why the lifespan rating is so low. The life is rated at the max temperature and ripple current. So how does this change when the temperature and ripple current are lower?
Has anyone seen any derating curves for capacitor life verses operating temperature/ ripple current?
awhiteguy
What are the thoughts on used caps? Sounds risky to me since brand new ones are only good for like 2000 or 3000 hours. I have wondered why the lifespan rating is so low. The life is rated at the max temperature and ripple current. So how does this change when the temperature and ripple current are lower?
Has anyone seen any derating curves for capacitor life verses operating temperature/ ripple current?
awhiteguy
I got some surplus caps. I would't go for used caps unless they were very new, and If I got surplus caps I would check the date code. Don't even consider anything older than 5yrs. They only last 10 - 20 years. The seals can be imperfect and they dry out. Those lifetimes are at 105C and max rated ripple current. They should last much longer in your amp since you are rating them very conservatively. My scope is about 20 years old an I am thinking it might be about time to replace the electrolytics before one fails.
Darrell Harmon
Darrell Harmon
awhiteguy said:
You might find som info of interest in these tech. notes:
http://www.evoxrifa.se/europe/technotes_electrolytics.htm
Found them yesterday and have only had a quick look at them,
so I can't give you any summary yet, but they seem to have
some of the info you are asking for.
Many cities have electronic supply stores that deal with used and surplus parts. Often they will have the type of capacitors you need for a fraction of the cost of ones from a regular store. The downside is that you may not know how old they are or exactly what their condition is, but paying $5 vs. $38, the risk may be worth it. You can also try attending a Ham Fest (a gathering of Amateur Radio operators) where electronics things are being sold, or you can ask some of the guys there where they go to get various odds and ends of electrical parts (other than Radio Shack, etc.)
The 3000 hour lifetime figures are quoted at the capacitors maximum operating temperature - usually 85'C or 105'C. When operated at more realistic temperatures, say 50'C, the lifetime increases dramatically.
The ripple current also has a significant affect on lifetime, but not as great as temperature. Paralleling capacitors can help here as the ripple current will split between the number of caps you've used. This lower ripple current also reduces self-heating caused by the cap's ESR.
BHC Capacitors have an interactive lifetime calculator on their website; try entering various figures and see what you get:
http://www.bhc.co.uk/
Nice one,
David.
The ripple current also has a significant affect on lifetime, but not as great as temperature. Paralleling capacitors can help here as the ripple current will split between the number of caps you've used. This lower ripple current also reduces self-heating caused by the cap's ESR.
BHC Capacitors have an interactive lifetime calculator on their website; try entering various figures and see what you get:
http://www.bhc.co.uk/
Nice one,
David.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.