Can't find low-ESR cap - alternative solution?

[sb1370]

Hi
I'm repairing my monitor power board and I found two blown capacitors which are from low ESR series. I could find a replacement for 680uf 25v cap which is 50v and have to fit it someway. But my problem is I couldn't find a replacement for 1000uf 25v with low ESR.
I thought about using two 470uf in parallel to reduce ESR but according to the ESR formula it won't work cause by halving capacity the ESR doubles (in theory).
I know I can order from the internet shops but the shipping costs outweighs the capacitor price and won't worth it.
What's your suggestions? Also is it a right thing to not touching the other caps cause as I said I can't find the proper replacement.

[ogden]

There is no such manufacturer part number "capacitor which is from low ESR series". Please be specific - what exactly capacitors you are talking about. I doubt that low ESR is key property in your case. Put new generic capacitors in and most likely you will be fine.

[sb1370]

What? I have seen many of them. BTW it's GF series of Samxon.

[wraper]

I thought about using two 470uf in parallel to reduce ESR but according to the ESR formula it won't work cause by halving capacity the ESR doubles (in theory).

470uF capacitor will have higher ESR than 1000uf from the same series, however when you connect 2 capacitors in parallel, ESR basically halves. So two 470uF caps instead of one 1000uF is a viable solution. Or if those two 1000uF caps are in parallel, better use three 680uF capacitors instead.

[precaud]

All of the 1000uf 25v radial polar 'lytics of current manufacture I've measured recently have pretty low ESR, in the 15 to 25 mOhm range. To get one that isn't, you'd have to buy the cheapest junk available, or a series that is made to a higher-ESR spec, such as Nichicon PM series.

[wraper]

All of the 1000uf 25v radial polar 'lytics of current manufacture I've measured recently have pretty low ESR, in the 15 to 25 mOhm range. To get one that isn't, you'd have to buy the cheapest junk available, or a series that is made to a higher-ESR spec, such as Nichicon PM series.

It's not so much about ESR as such but that general purpose capacitors will not survive there for more than a few months due to high ripple current.

[precaud]

It's not so much about ESR as such but that general purpose capacitors will not survive there for more than a few months due to high ripple current.

Perhaps, but that's not his question.

[wraper]

It's not so much about ESR as such but that general purpose capacitors will not survive there for more than a few months due to high ripple current.

Perhaps, but that's not his question.

His question is about suitable replacement. General purpose capacitors are certainly not suitable to be there.

[precaud]

His question is about suitable replacement. General purpose capacitors are certainly not suitable to be there.

Where is "there"? He hasn't said where in the circuit they are. If they're post-regulator, ripple current will be negligible.

[wraper]

Where is "there"? He hasn't said where in the circuit they are. If they're post-regulator, ripple current will be negligible.

Here there:

I'm repairing my monitor power board

Capacitors on secondary side of SMPS/power input of CCFL inverter.

[sb1370]

I thought about using two 470uf in parallel to reduce ESR but according to the ESR formula it won't work cause by halving capacity the ESR doubles (in theory).

470uF capacitor will have higher ESR than 1000uf from the same series, however when you connect 2 capacitors in parallel, ESR basically halves. So two 470uF caps instead of one 1000uF is a viable solution. Or if those two 1000uF caps are in parallel, better use three 680uF capacitors instead.

Have you tried it in practice? because in theory ESR=tan δ / (C*f*2π) says otherwise. Unfortunately I don't have a ESR meter to measure it.

All of the 1000uf 25v radial polar 'lytics of current manufacture I've measured recently have pretty low ESR, in the 15 to 25 mOhm range. To get one that isn't, you'd have to buy the cheapest junk available, or a series that is made to a higher-ESR spec, such as Nichicon PM series.

Well you are right the DF between general purpose and Low-impedance/ESR series isn't much different so in theory their ESR shouldn't be much different I can even get higher voltage ones for lower ESR, but I don't know how they are in practice. I think the more important factor here is "Ripple Current". Ripple current of a low-impedance series is twice the general one with the same ratings.

All of the 1000uf 25v radial polar 'lytics of current manufacture I've measured recently have pretty low ESR, in the 15 to 25 mOhm range. To get one that isn't, you'd have to buy the cheapest junk available, or a series that is made to a higher-ESR spec, such as Nichicon PM series.

It's not so much about ESR as such but that general purpose capacitors will not survive there for more than a few months due to high ripple current.

I agree with you, however I looked up the power board of another monitor newer but close to my model from the same brand and there they likely had used regular capacitors.

[ogden]

It's not so much about ESR as such but that general purpose capacitors will not survive there for more than a few months due to high ripple current.

It is very much about low ESR and high ripple current - combined. Low ESR capacitor will heat-up less, thus live longer. My consideration of suggesting generic (105oC) cap - most monitors do not consume much, thus ripple current may be tolerable for generic to live some 5 years or longer. There's always option to replace capacitor again

I would not even consider putting multiple capacitors where is place for only one. What's the point of adding low ESR capacitors through long, inductive leads which have higher ESR/impedance than generic caps? 

[mzzj]

Where is "there"? He hasn't said where in the circuit they are. If they're post-regulator, ripple current will be negligible.

Answer to "where?" is "Iran" and it might explain if OP has problems sourcing any components. 

[precaud]

Answer to "where?" is "Iran" and it might explain if OP has problems sourcing any components.

Good point...

Some cap makers list ripple current rating as a function of case size. So, if there's an absence of detailed info, the OP could put the largest cap he can find that will fit in there. Voltage rating can be higher than 25V, of course.

[wraper]

I agree with you, however I looked up the power board of another monitor newer but close to my model from the same brand and there they likely had used regular capacitors.

I've never seen LCD monitor with general purpose capacitors in PSU on 12V+ rail. And overall they are almost nonexistent in SMPS, sometimes placed on some power rail with low current, while the rest use LOW ESR.

[wraper]

My consideration of suggesting generic (105oC) cap - most monitors do not consume much, thus ripple current may be tolerable for generic to live some 5 years or longer. There's always option to replace capacitor again

Even some LOW ESR caps do not survive there for long. General purpose caps will pop very fast, hopefully without electrolyte spraying over the board. If OP can get 470uF LOW ESR, using general purpose capacitors is dumb.

I would not even consider putting multiple capacitors where is place for only one. What's the point of adding low ESR capacitors through long, inductive leads which have higher ESR/impedance than generic caps?

The point is they will not fail. Longer leads will make no noticeable difference.

[ogden]

If OP can get 470uF LOW ESR, using general purpose capacitors is dumb.

Yes. - It will work when one of two capacitors is right where 1000uF needs to be and other is doing low freq bulk storage work.

Longer leads will make no noticeable difference.

Hold your horses. Extra length of capacitor leads will increase impedance at high frequencies, thus increase ripple amplitude on the PSU output. It's SMPS design basics - do not use extension cords for your filtering capacitors

[wraper]

Hold your horses. Extra length of capacitor leads will increase impedance at high frequencies, thus increase ripple amplitude on the PSU output. It's SMPS design basics - do not use extension cords for your filtering capacitors

Then look at length of traces on those one layer PCBs. Also it's not like there are super high frequencies.

[ogden]

Hold your horses. Extra length of capacitor leads will increase impedance at high frequencies, thus increase ripple amplitude on the PSU output. It's SMPS design basics - do not use extension cords for your filtering capacitors

Then look at length of traces on those one layer PCBs.

Look where exactly? Filtering capacitor connected to power rail through "length of traces"? Could you please show what you are talking about? 

Also it's not like there are super high frequencies.

Broadband noise of SMPS easily reach 20MHz. Inductance of 1cm long 0.5mm dia conductor is 7nH, impedance of such @10MHz equals 0.4Ohms. Take a moment to consume this.

[wraper]

Broadband noise of SMPS easily reach 20MHz. Inductance of 1cm long 0.5mm dia conductor is 7nH, impedance of such @10MHz equals 0.4Ohms. Take a moment to consume this.

Take a moment to consume there is easily 4-5 cm or more trace length from transformer to capacitor (each lead). And there is no solid ground plane as well. And PCB is located in a steel cage, so don't worry about electromagnetic emissions.

[ogden]

Broadband noise of SMPS easily reach 20MHz. Inductance of 1cm long 0.5mm dia conductor is 7nH, impedance of such @10MHz equals 0.4Ohms. Take a moment to consume this.

Take a moment to consume there is easily 4-5 cm or more trace length from transformer to capacitor (each lead). And there is no solid ground plane as well.

Right. So what. I talk about prolonged leads/tracks between filter capacitor and DC bus, capacitor connected to DC bus "through extension cord". If this does not explain what I mean then just leave it there and we better end this fruitless chat.

And PCB is located in a steel cage, so don't worry about electromagnetic emissions.

So what. There is condcted emissions as well. If you do not filter hi-frequnecy ripple in DC bus, then wires going out of steel cage become antennas. Also downstream circuit my not like unfiltered ripple.

[wraper]

So what. There is condcted emissions as well. If you do not filter hi-frequnecy ripple in DC bus, then wires going out of steel cage become antennas. Also downstream circuit my not like unfiltered ripple.

Don't be ridiculous, additional centimeter or two from electrolytic capacitor leads will not make any significant difference in SMPS operating at about 30-60 kHz. It's not like extending leads of MLCC in a high frequency device. If you are so concerned about 20MHz, place 0.1uF MLCC with short leads in parallel  and it will have much lower impedance than electrolytic cap at such frequency.

[ogden]

additional centimeter or two from electrolytic capacitor leads will not make any significant difference in SMPS operating at about 30-60 kHz.

You need to learn, kid.

[wraper]

additional centimeter or two from electrolytic capacitor leads will not make any significant difference in SMPS operating at about 30-60 kHz.

You need to learn, kid.

You need to learn to be realistic grandpa. Suggesting general purpose caps in a place where they will certainly fail instead of LOW ESR on somewhat extended leads is lame. Not that he will bring that monitor to EMC certification to begin with. And apparently you have completely no idea about monitor SMPS and how stressed those caps are. They run damn hot and at high ripple current.

My consideration of suggesting generic (105oC) cap - most monitors do not consume much, thus ripple current may be tolerable for generic to live some 5 years or longer.

[bdunham7]

Hi
I'm repairing my monitor power board and I found two blown capacitors which are from low ESR series. I could find a replacement for 680uf 25v cap which is 50v and have to fit it someway. But my problem is I couldn't find a replacement for 1000uf 25v with low ESR.
I thought about using two 470uf in parallel to reduce ESR but according to the ESR formula it won't work cause by halving capacity the ESR doubles (in theory).
I know I can order from the internet shops but the shipping costs outweighs the capacitor price and won't worth it.
What's your suggestions? Also is it a right thing to not touching the other caps cause as I said I can't find the proper replacement.

I recently repaired my monitor power (brick) and had similar issues.  The original caps were made to fit tightly and just barely meet specs, so they cooked.  I was able to do a little grinding and make way for 35 volt caps, which were bit taller and had better ripple ratings.  I don't know if you have the room for that.

As for ESR and ripple...

Don't obsess about the ESR in particular,  instead try to find caps that have a ripple rating at least equal to the output current of the power supply circuit.  And parallel caps do add up, so if you sub in two 470uF for 1000uF, just multiply the ripple rating X2.  If you find caps that meet that requirement, the ESR is going to be fine--high ripple and low--or low enough--ESR ratings usually go together.  I would make specific suggestions, but I have no idea what your parts availability is there.  I'm sure it isn't easy.