Choosing capacitors voltage rating

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi folks, I need your's advise. I am going to make power supply for a IRS2092s amplifier (-/+71 VDC) with a 1200va toroid transformer (50-0-50 dual ac). I have ordered a pcb for capacitors, and the pcb can hold 48 pieces of 18mm diamiter's of capacitors. I want to use 80v 1500uf capacitors in the board, because they are 18mm of diamiter and perfectly fit in to the board and and also these are lot cheaper than 100v capacitors. My question is should I use 80v capacitors with +/-71 VDC ? I'm really nervous to do so if it going to be exploded or so. I would be more comfortable with 100v or higher voltage rating capacitors but they are expensive and hard match up the diameter of 18mm. Please guys help me out. 80+80=160 which is kinda close to 142(+/-71vdc).
 
Because I had previous experience with 16v capacitors in series. with a 20vac (20-0ac) single power transformer with fullwave rectifiers bridge with 2 16v 6800uf capacitors in series , I was getting around 28.6 DC voltage but after 20-25seconds later both of those 16v capacitors got exploded.
 
80% of rated wdc working volts..

so if you are using +-71 volts, i will use 100vdc caps,
what is the primary line voltage when your rails are +-71
how high can your primary line voltage go?
knowing the answers to these questions will help you decide better,
while 80vdc caps seems fine, then ask yourself, what if the line voltage increases?
 
I do it all the time. How else can you get a cap that works at 900VDC?
If you NEED to do that, you ...
Use sharing resistors across each cap to equalize the charging. I use 300k.
But the OP didn't say he/she/they did that, and by the results they got I suspect they did not. Another thing to do if you're going to do that is match pairs so their capacitance are the same or close to each other, so there's less imbalance of voltage to start with, as well as using voltage-balancing resistors.

https://www.illinoiscapacitor.com/pdf/Papers/voltage_balancing_resistors.pdf
How to size balancing resistors – Nanoramic Laboratories
Balancing Resistor Values for Series Capacitors - Electrical Engineering Stack Exchange
 
Because I had previous experience with 16v capacitors in series. with a 20vac (20-0ac) single power transformer with fullwave rectifiers bridge with 2 16v 6800uf capacitors in series , I was getting around 28.6 DC voltage but after 20-25seconds later both of those 16v capacitors got exploded.
80VDC cap for 50VAC nominal voltage is ok. 100V - could be ok too :).
I usually prefer to choose the higher one, because it's more reliable (but it's my own problem).
Let's do some math.
50VAC nom *1.1 *1.41 = 77 VDC max (it's maximum VDC, but it won't be so high in average).
You should test your mains voltage if it is usually at a higher limit (118-120V or higher) then you really can consider 100 VDC capacitors. Else, if it isn't so high - 80V will work ok.

And don't use caps in series except for very high voltages (when really there aren't such nominals - it's the only reason. Of cause don't forget to put parallel resistors with caps to align voltage).
 
Last edited:
Account Closed
Joined 2018
There is a school of thought which says that series capacitors will adjust their share of the total voltage so that they each draw the same leakage current, hence there is no need for balancing resistors. I am not entirely convinced by this.


Indeed, that "school of thought" is the self-proclaimed internet guru's without any education to back up their claims.
These days the phrase used is "fake news". :eek:
 
There is a school of thought which says that series capacitors will adjust their share of the total voltage so that they each draw the same leakage current, hence there is no need for balancing resistors. I am not entirely convinced by this.


that school of thought is valid if the two caps in series are exactly alike in capacity and esr, otherwise, they will not balance out..for safety's sake i will never do that without the resistors...

i saw some of these work out when repairing atx psu with half bridge topology, the caps were supposed to share 300 vdc to 150 each, but did not, so what i did was to measure them in an esr meter and true enough, they were not the same as they were labeled, so replacing them with identical caps, restored balance...

theory is fine until some factors come in to upset them...
 
Yep, you gotta have quality data on leakage variations, both part to part and lot to lot. Then choose your balancing resistors to swamp that by 3:1 or more. Big caps often have big leakages; one shipment could be better or worse by 10 or 15:1.

Still better (already mentioned / credit due others), only use them in series if there's no other option.

Regards
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.