Author Topic: Electrolytic Capacitor Selection: Nod to Dissipation Factor or Ripple Current?  (Read 2204 times)

0 Members and 1 Guest are viewing this topic.

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Consider these two capacitors which are the same physical dimensions, same Ripple Current rating, yet with different Dissipation Factor DF and Voltage and Life ratings, both made by Nichicon:

DF=0.3, Ripple Current=210mA , 10000hr@105°C
ULD1E101MED

DF=0.19, Ripple Current=210mA, 5000hr@105°C
UPV1A101MGD

We know the calculation:

ESR=tanδ*Xc (Ω), where tanδ is the Dissipation Factor, which is the "tangent of the loss angle."  Xc = Capacitive Reactance.

We also know ESR usually affects the Ripple Current rating of a capacitor.  Yet in the case of the above two capacitors, DF is different but Ripple Current remains the same.  So it would seem that in the case of the capacitor with DF=0.3, Xc must be lower to yield the same ESR which ultimately results in the same Ripple Current.  Is this correct, or am I misunderstanding something?

Here is why I am asking.  If we are searching for high quality, aluminum electrolytic, off the shelf capacitors and we come across two capacitors such as the above 2 that would both satisfy a particular design's requirements for physical size, capacitance, voltage rating and life, should we consider selecting the capacitor with the lower DF, assuming the Ripple Current is the same?

Now consider the following 2 capacitors:

DF=0.24, Ripple Current=40mA, Life=5000hr@105°C
UTT1C100MDD1TP

DF=0.15, Ripple Current=33mA, Life=5000hr@105°C
USV1E100MFD1TP

These two are interesting because despite the significantly lower DF=0.15 of the USV1E100MFD1TP, the UTT1C100MDD1TP has a higher Ripple Current rating!  So in this case, assuming either capacitor would satisfy the design criterion, and seeing the high volume price of each is nearly the same, would you choose the capacitor with the lower DF or the capacitor with the higher Ripple Current, and why?

Thank you.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
...Yet in the case of the above two capacitors, DF is different but Ripple Current remains the same.  So it would seem that in the case of the capacitor with DF=0.3, Xc must be lower to yield the same ESR which ultimately results in the same Ripple Current.  Is this correct, or am I misunderstanding something?
...

DF goes up over time so this could very well be a case of the manufacturer setting a higher DF for the first capacitor just so it appears to meet spec at the end of its 10kH rated life - in other words, the two capacitors might very well end up at the same DF after 5kH have passed, all else being equal. Another possible difference is the (rarely specified) thermal resistance: the capacitor with the higher DF might also have a lower thermal resistance, and therefore be capable of the same ripple current.

Nichicon is good about showing the tradeoffs of each of their capacitor lines with respect to its most common substitutes and they make a very good capacitor in general. If you will be coming anywhere close to the ripple current rating - especially if ambient is 55C or higher - then I'd go with the second one (UPV series) because it is listed as a low-impedance type, but if this a general purpose application then either will be fine.

 
The following users thanked this post: JDW

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17427
  • Country: us
  • DavidH
The higher DF capacitor which is otherwise identical can have the same ripple current rating if it operates at a higher temperature which might be the case if it has an electrolyte which does not evaporate as quickly.  The lower DF capacitor might be used for lower ripple even at the expense of operating life.

Like MagicSmoker says however, the DF rating is not the nominal value when the capacitor is new.  It is the maximum DF at the end of its operating life.  If you want to know the real values when new, then you will have to measure it.
 
The following users thanked this post: JDW

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
Gentlemen, thank you for your detailed replies.  I was asking for general use purposes.  There are sometimes when i need to select a capacitor for a design I myself created, but there are other times I may be recapping a vintage computer built 30+ years ago -- recapping everything from a switching power supply to  logic board to a CRT driver board. When recapping old computers there is no schematic available, nor do I know technical details of the capacitors used beyond what is written -- usually capacitance, voltage and temperature (most often 85°C).  That's why I am often unsure when selecting replacement capacitors if I should give priority to the Ripple Current spec or the DF spec.  I do try to pick replacement caps that are rated at 5000hr@105°C or 10000hr@105°C whenever possible.  Some vintage computers might actually prefer a low ESR replacement, especially decoupling caps on a logic board, but I need to be careful when recapping their switching power supplies.  For example, using polymer electrolytics like OSCON with 40m-ohm or lower ESR could in fact cause instability in some old switchers (in theory, but I've not experimented with this), which is why I stick with regular aluminum electrolytics, even the "low ESR" kind because the ESR in the hundreds of milliohms is perhaps not too low for an old switcher.  But again, I was just trying to better understand what my capacitor choice should be in light of DF and Ripple Current specifications.  And based on your replies it would seem I perhaps should give the nod to lower DF (even if that spec pertains to end-of-life).  (Also, most all caps have a DF spec shown, but the ESR spec is often not shown in the data sheet.)

Thank you.
 

Offline Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 2077
  • Country: us
    • The Messy Basement
For line applications, go by DF. For switching PS circuits, go by 100 kHz ESR. Look at the rated life too. I've always suspected that when they optimize one parameter, something else has to suffer. Finally, measure. It seems that many cap specs are sort of standardized, and a given series might consistently do a lot better than the DF chart, whereas another series might be near the published values, yet they use the same chart.
 

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
 Thank you for your suggestions, but as I’m sure you know, most datasheets and even the specifications shown on Mouser, do not list ESR. Instead they tend to show Dissipation Factor. And even when the ESR is shown in the datasheet, it’s often rated at 120 Hz, not 100 kHz. So what is one to do?
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Thank you for your suggestions, but as I’m sure you know, most datasheets and even the specifications shown on Mouser, do not list ESR. Instead they tend to show Dissipation Factor. And even when the ESR is shown in the datasheet, it’s often rated at 120 Hz, not 100 kHz. So what is one to do?

At risk of pointing out the obvious, if a capacitor datasheet doesn't specify DF or ESR and this (or ripple current, etc.) are important to your application then don't use that capacitor... Note that while DF and ESR are directly related, DF effectively specifies the inverse Q of a capacitor - that is, the ratio of ESR/Xc - and is often only given at a low frequency, a single temperature and as a worst case value, whereas ESR might be given at a low frequency or 10kHz or 100kHz, or with a chart showing how it varies over frequency and/or a chart showing variation with temperature, and is most often given as an average or typical value. Note, also, that the manufacture might not even check DF via sampled testing, whereas ESR typically will be.
 

Offline Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 2077
  • Country: us
    • The Messy Basement
IMO, the data sheets for well known quality brands are usually good. If you don't see 100 kHz ESR, the cap wasn't intended for switching applications, so don't use it. I'd never put a cap in a switching supply if the data sheet didn't highlight that application specifically. For line/mains applications it's hard to find a cap that's actually poor unless it's a fake. I go by lifetime specs more than anything.
 
The following users thanked this post: Siwastaja

Offline JDWTopic starter

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
...if a capacitor datasheet doesn't specify DF or ESR and this (or ripple current, etc.) are important to your application then don't use that capacitor...

I typically buy from Mouser.  In almost every case where I've searched for an electrolytic, the DF is given.  It's just the ESR/impedance that is not always given.  Occasionally both are given, which is very helpful.

Note that while DF and ESR are directly related, DF effectively specifies the inverse Q of a capacitor - that is, the ratio of ESR/Xc - and is often only given at a low frequency, a single temperature and as a worst case value, whereas ESR might be given at a low frequency or 10kHz or 100kHz, or with a chart showing how it varies over frequency and/or a chart showing variation with temperature, and is most often given as an average or typical value. Note, also, that the manufacture might not even check DF via sampled testing, whereas ESR typically will be.

I don't often see data sheets with a nice graph of ESR over a broad range of frequencies.  Usually when I actually do get lucky enough to find an electrolytic datasheet with both DF and ESR, the ESR is listed in a table at a single fixed frequency.  And sometimes the ESR at that fixed frequency is shown at 20°C or 25°C as well as -10°C.

Here's an example recapping project of vintage hardware.  It's a switching power supply made by SONY in late 1980's and is used in an external Apple HD20SC hard disk enclosure.  I have no schematic and redrawing the schematic is more trouble that it's worth.  Here are photos.  And here is a list of all the electrolytic capacitors in that PSU, followed by the Mouser replacement cap which I selected:

C226: 22uF 35V, D=5.2mm, Leads=5mm -- (Mouser: UHE1H220MDD1TD)
C202: 47uF 25V, D=5.2mm, Leads=5mm -- (Mouser: EEU-FR1E470B)
C222: 47uF 25V, D=5.2mm, Leads=5mm -- (same as C202)
C109: 150uF 400V, H=32mm, D=25.8mm, Leads=10mm -- (Mouser: 860021383023)
C110: 4.7uF 350V H=32mm, D=12.8mm, Leads=5mm -- (Mouser: UPM2G4R7MHD)
C210: 330uF 16V, D=8.1mm, Leads=5mm -- (Mouser: EEU-FR1E331B)
C215: 470uF 10V, D=8.1mm, Leads=5mm -- (Mouser: EEU-FR1E471YB)
C213: 22uF 100V, D=10.2mm, Leads=5mm -- (Mouser: UBT2A220MPD1TD)
C214: 2200uF 10V, D=12.7mm, Leads=5mm -- (Mouser: UHE1C222MHD1TO)
C209: 2200uF 16V, D=12.7mm, H=30mm, Leads=5mm -- (same as C214)

CR-35 daughter card:

C181: 100uF 10V, D=5.2mm, Hmax=12.5mm, Leads=5mm -- (Mouser: EEU-FR1E101)
C182: 100uF 10V, D=5.2mm, Hmax=12.5mm, Leads=5mm -- (same as C181)

Clicking on the Mouser links of course take you to the capacitor product page, and you can see specs like DF, Life and so on.  But you can also then click on the Datasheet PDF to see more.  For example, consider the datasheet of the very first cap -- Panasonic UHE1H220MDD1TD.  On the first page you see the DF for this 50V 22uF capacitor is 0.10@20°C.  Scrolling down to the page in the same PDF marked "222" we see MAX Impedance is given at 100kHz at both 20°C (0.7Ω) and -10°C (2.8Ω).  There is no graph of ESR over a broad range of frequencies, but I consider this kind of datasheet good enough.  And although I don't have a schematic, I know the PSU is a switcher, and I believe the capacitor impedance values shown are sufficiently low, yet not too low.  In this particular case, all the data sheets are quite informative with 100kHz values for Impedance except the Wurth cap, but I picked that Wurth cap because it is one of the few long life caps available that is physically sized to fit properly.  However, there are other caps, yes even quality Nichicons, which have data sheets lacking in details such as the following:

https://www.mouser.com/datasheet/2/293/e-umv-1512793.pdf

I do get DF, but no detailed impedance info.

Here's yet another example of a Nichicon datasheet slim on impedance details:

https://www.mouser.com/datasheet/2/293/e-sv-5439.pdf

When recapping old boards, physical size of the replacement cap is a key consideration, and sometimes there is maybe only 2 or three choices, and when all the data sheets are slim on details, I need to make the best decision I can based on that limited information.  And that is why I posted my opening post, to determine what datasheet spec of an electrolytic cap should usually be given priority when I am trying to find replacements for old existing caps.
« Last Edit: August 05, 2019, 12:52:13 am by JDW »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf