Capacitance versus voltage for X7S caps?
in [Design]
|
Prev: Automotive datalogger power(down) supply strategy
Next: What do you call it/where do you get it?
From: legg on 31 Mar 2010 22:34 On Wed, 31 Mar 2010 11:12:18 -0700, Joerg <invalid(a)invalid.invalid> wrote: >legg wrote: >> On Tue, 30 Mar 2010 14:07:58 -0700, Joerg <invalid(a)invalid.invalid> >> wrote: >> >>> legg wrote: >> <snip> >>>>> Do you know how they take hardcore freezes, like north of the Klondike >>>>> in January? >>>> Oscon aluminum organic polymer are characterized to -55degC., with >>>> less than 10% cap reduction. At 4.7uF, ESR is in the 80mR range. >>>> >>>> Later developers tend to follow this. >>>> >>> I have used Oscons on another project. They are very expensive and in >>> this case I need 100V and the height cannot be more than 2.5mm or around >>> 0.100". >> >> I'm not sure if you've actually stated your requirement at some time >> or another. You seem to be merely floating generalized but limited >> questions regarding part characteristics. >> > >No, I have in a follow-up post in this thread, that the graph ends at >50V but I need to know higher. Oscons don't get into that range much. > > >> What is the actual requirement? >> > >I need to replace an electrolytic that was used so far because it's too >large and also prone to failure. Total 100uF/100V, although in a pinch >we could scrape by with a 63V rating if we zener any larger excursions >than 55V or so. High pulsed load, easily two amps. Hi-rel application >meaning stress fracture is an issue with very large SMT footprints. I >don't want to get up to 2220, better stay with 1812. Height restriction >2.5mm or just under 0.100". > >Nominal voltage is 50V but we have to (ideally) allow excursions up to >70V because this is the only capacitance at the end of a cable that can >easily be 250ft long in some situations. > >If X7S really drops by around 50% as I have a hunch it will then we'll >have to see if the remaing 50uF is going to be enough. X7R in 2220 would >be better but we can't really fit that many, plus the stress fracture >issue. So, the classic situation of being between a rock and a hard spot. You need to know: Permissible real estate and profile limits. Permissible capacitive droop under pulsed loading conditions. Source transient capability under the same pulsed loading. If unregulated - the source impedance after 250 feet. Repetition rate of pulsed load. RMS source ripple current at the conversion frequency. Permissible body surface demperature rise For voltage dependancy of the TDK parts above 50V, click the 'display data' control button when viewing the 'zoom' window. The data is presented over the ratings in a plain text list. I guess the GUI guy had a hard time fitting a variable graph axis. You can make a POL converter in the volume occupied by some 100uF 100V capacitors....... RL
From: Joerg on 31 Mar 2010 21:41 legg wrote: > On Wed, 31 Mar 2010 11:12:18 -0700, Joerg <invalid(a)invalid.invalid> > wrote: > >> legg wrote: >>> On Tue, 30 Mar 2010 14:07:58 -0700, Joerg <invalid(a)invalid.invalid> >>> wrote: >>> >>>> legg wrote: >>> <snip> >>>>>> Do you know how they take hardcore freezes, like north of the Klondike >>>>>> in January? >>>>> Oscon aluminum organic polymer are characterized to -55degC., with >>>>> less than 10% cap reduction. At 4.7uF, ESR is in the 80mR range. >>>>> >>>>> Later developers tend to follow this. >>>>> >>>> I have used Oscons on another project. They are very expensive and in >>>> this case I need 100V and the height cannot be more than 2.5mm or around >>>> 0.100". >>> I'm not sure if you've actually stated your requirement at some time >>> or another. You seem to be merely floating generalized but limited >>> questions regarding part characteristics. >>> >> No, I have in a follow-up post in this thread, that the graph ends at >> 50V but I need to know higher. Oscons don't get into that range much. >> >> >>> What is the actual requirement? >>> >> I need to replace an electrolytic that was used so far because it's too >> large and also prone to failure. Total 100uF/100V, although in a pinch >> we could scrape by with a 63V rating if we zener any larger excursions >> than 55V or so. High pulsed load, easily two amps. Hi-rel application >> meaning stress fracture is an issue with very large SMT footprints. I >> don't want to get up to 2220, better stay with 1812. Height restriction >> 2.5mm or just under 0.100". >> >> Nominal voltage is 50V but we have to (ideally) allow excursions up to >> 70V because this is the only capacitance at the end of a cable that can >> easily be 250ft long in some situations. >> >> If X7S really drops by around 50% as I have a hunch it will then we'll >> have to see if the remaing 50uF is going to be enough. X7R in 2220 would >> be better but we can't really fit that many, plus the stress fracture >> issue. So, the classic situation of being between a rock and a hard spot. > > You need to know: > > Permissible real estate and profile limits. > Permissible capacitive droop under pulsed loading conditions. > Source transient capability under the same pulsed loading. > If unregulated - the source impedance after 250 feet. > Repetition rate of pulsed load. > RMS source ripple current at the conversion frequency. > Permissible body surface demperature rise > All those I know, except some of the other pulsed loads to some extent. Simulated it, woiks :-) > For voltage dependancy of the TDK parts above 50V, click the 'display > data' control button when viewing the 'zoom' window. The data is > presented over the ratings in a plain text list. I guess the GUI guy > had a hard time fitting a variable graph axis. > Well, we had to press on and the layout is going to be done by tonight. Not much choice considering real estate and the possibility of stress fractures. > You can make a POL converter in the volume occupied by some 100uF 100V > capacitors....... > I've made a converter but not in the shape of a capacitor can, because it has to be as flat as a PCMCIA card. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
From: JosephKK on 31 Mar 2010 23:23 On Wed, 31 Mar 2010 11:12:18 -0700, Joerg <invalid(a)invalid.invalid> wrote: >legg wrote: >> On Tue, 30 Mar 2010 14:07:58 -0700, Joerg <invalid(a)invalid.invalid> >> wrote: >> >>> legg wrote: >> <snip> >>>>> Do you know how they take hardcore freezes, like north of the Klondike >>>>> in January? >>>> Oscon aluminum organic polymer are characterized to -55degC., with >>>> less than 10% cap reduction. At 4.7uF, ESR is in the 80mR range. >>>> >>>> Later developers tend to follow this. >>>> >>> I have used Oscons on another project. They are very expensive and in >>> this case I need 100V and the height cannot be more than 2.5mm or around >>> 0.100". >> >> I'm not sure if you've actually stated your requirement at some time >> or another. You seem to be merely floating generalized but limited >> questions regarding part characteristics. >> > >No, I have in a follow-up post in this thread, that the graph ends at >50V but I need to know higher. Oscons don't get into that range much. > > >> What is the actual requirement? >> > >I need to replace an electrolytic that was used so far because it's too >large and also prone to failure. Total 100uF/100V, although in a pinch >we could scrape by with a 63V rating if we zener any larger excursions >than 55V or so. High pulsed load, easily two amps. Hi-rel application >meaning stress fracture is an issue with very large SMT footprints. I >don't want to get up to 2220, better stay with 1812. Height restriction >2.5mm or just under 0.100". > >Nominal voltage is 50V but we have to (ideally) allow excursions up to >70V because this is the only capacitance at the end of a cable that can >easily be 250ft long in some situations. > >If X7S really drops by around 50% as I have a hunch it will then we'll >have to see if the remaing 50uF is going to be enough. X7R in 2220 would >be better but we can't really fit that many, plus the stress fracture >issue. So, the classic situation of being between a rock and a hard spot. From the link i already provided it drops at least 60%. Please continue.
From: JosephKK on 2 Apr 2010 00:07 On Thu, 01 Apr 2010 10:07:24 -0700, Joerg <invalid(a)invalid.invalid> wrote: >JosephKK wrote: >> On Tue, 30 Mar 2010 09:03:17 -0700, Joerg <invalid(a)invalid.invalid> wrote: >> >>> JosephKK wrote: >>>> On Sun, 28 Mar 2010 13:47:48 -0700, Joerg <invalid(a)invalid.invalid> wrote: >>>> <snip> >>>> About 1/2 cubic inch any way you pack it. >>> >>> They aren't that big anymore. You can get X7R in 2220 but that size is >>> already causing concern WRT stress fracture. >>> >> 1/2 cubic inch refers to film caps. Ceramic has not needed that much >> volume for 4.7 uF for 20 years. >>>> Alternatively i might try the aluminum polymer electrolytics. >>> >>> If I could get them in 2.5mm height or less. Plus there's the issue of >>> freezing. >> >> How would freezing affect a polymer (solid) electrolyte? ... > > >That's what I don't know (yet). Unequal expansion in materials can cause >seals to let go a bit, moisture gets in, and so on. Can't take a risk >here and with the "responsiveness" of today's support lines I can't >really wait. Layout is done now ;-) I suppose you could get some samples and play with temperature and voltage. > >BTW, it's been almost a week now and AVX hasn't bothered to answer at >all. It used to be such a great company. Anyone knows what happened? > IIRC they got bought up not too long ago. > >> ... Height may be >> the issue. Perhaps you could end using 50 2.2 uF caps instead. > > >We might have to. Although 4.7uF X7S dropping to half capacitance would >be essentially the same then, except they are a bit less expensive.
From: Joerg on 2 Apr 2010 16:00
JosephKK wrote: > On Thu, 01 Apr 2010 10:07:24 -0700, Joerg <invalid(a)invalid.invalid> wrote: > >> JosephKK wrote: >>> On Tue, 30 Mar 2010 09:03:17 -0700, Joerg <invalid(a)invalid.invalid> wrote: >>> >>>> JosephKK wrote: >>>>> On Sun, 28 Mar 2010 13:47:48 -0700, Joerg <invalid(a)invalid.invalid> wrote: >>>>> > <snip> >>>>> About 1/2 cubic inch any way you pack it. >>>> They aren't that big anymore. You can get X7R in 2220 but that size is >>>> already causing concern WRT stress fracture. >>>> >>> 1/2 cubic inch refers to film caps. Ceramic has not needed that much >>> volume for 4.7 uF for 20 years. >>>>> Alternatively i might try the aluminum polymer electrolytics. >>>> If I could get them in 2.5mm height or less. Plus there's the issue of >>>> freezing. >>> How would freezing affect a polymer (solid) electrolyte? ... >> >> That's what I don't know (yet). Unequal expansion in materials can cause >> seals to let go a bit, moisture gets in, and so on. Can't take a risk >> here and with the "responsiveness" of today's support lines I can't >> really wait. Layout is done now ;-) > > I suppose you could get some samples and play with temperature and voltage. Well, it's done and done. We are well beyond the point of no return by now :-) >> BTW, it's been almost a week now and AVX hasn't bothered to answer at >> all. It used to be such a great company. Anyone knows what happened? >> > IIRC they got bought up not too long ago. Sad. That's when things often go downhill, and fast. [...] -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM. |