Capacitor Life expectancy

[999999999]

Re: Capacitor Life expectancy

Therefore as cap ESR decreases more ripple flows through the cap and less through the CPU.

This is possible, but will vary per circuit, is not necessarily true for at least a couple of reasons.

1) The cap will still follow the peak voltage even if it does so slower, it only has to rise to peak value within the period which it may still do. If it cannot, then we more often have #2

2) If it can't do this, a higher ripple current into the CPU means more current consumed from the higher voltage, and when the switching goes through zero crossing there is a greater voltage depression from lesser charging of the cap and higher impedance back into the circuit, so the voltage is then lower and the regulator operates at a higher duty cycle, really I mean pulse width in most cases, which A) Gives the higher ESR cap even more time to charge to that peak voltage and B) higher power consumption, which may make the higher ESR cap as susceptible through higher voltage peaks vs it's higher resistance = ripple current through it may be lower, same, or even higher depending on circuit specifics.

The key here is we can't assume the circuit a constant, since the cap is a key filtration element in that circuit which has feedback to control it. Poor cap - effects voltage sensed - feedback loop compensates for it - poor cap even worse off now but at least the circuit had kept regulation as well as the design allowed.

[davmax]

Re: Capacitor Life expectancy

Sorry 999999999 you are making up your own theory. You clearly do not understand a VRM switching regulator. And certainly have not understood the method of fixing a value when calculating. This method in no way means that the operation of the circuit is constant. Learn some good sound theory.

[davmax]

Re: Capacitor Life expectancy

In my last post comparing 10mm and 12.5mm caps it was seen that there would be only a small life gain if only considering core loss heat. No one questioned where the case does create gain in reducing temperature. It certainly did not come to me straight away. So below is the answer that I believe calculation reveals. I have posted the complete comparison section the first part you have seen already.


10mm to 12.5mm comparison
Here it has been shown that for the same ripple current (2.2A) the temperature rise is reduced by 2.23 oC (6.02-3.79) when using the larger 12.5mm capacitor. The gain is found in the reduced core loss and the 22.8% improvement in case to air thermal resistance. This will lower both the case and core temperatures, life gains are 43% for 5 oC and 100% for 10 oC therefore a 2.23oC improvement will not gain much in terms of life. Again it is clear that external temperatures are the dominant determinant of capacitor life time.

The above does not consider the fact that when using a capacitor of lower ESR the ripple current will increase. Assuming the AC ripple source voltage remains the same (constant). For the 12.5mm example the ripple rise could cause a heat loss rise from 43.56mW to 77.44mW. If this does in fact happen there will be an improvement in ripple removal but the core temperature rise will be 6.7oC and there will no life gain over the 10mm at all.

So where are the big gains in case size? Clearly it makes little difference when only considering the core loss. The case size makes the greatest gain when external heat is injected into the capacitor. Injection temperatures can easily reach 60oC and at this level of temperature the effect of case to air thermal resistance is more significant e.g the 12.5mm cap having a 22.8% low thermal resistance the gain can be more than 10oC (at least doubling life). So here again it can be seen that external temperature plays a major role in capacitor life.

[999999999]

Re: Capacitor Life expectancy

Sorry 999999999 you are making up your own theory. You clearly do not understand a VRM switching regulator. And certainly have not understood the method of fixing a value when calculating. This method in no way means that the operation of the circuit is constant. Learn some good sound theory.

You are the one who only weeks ago began to understand these cap issues I was familiar with 20 years ago. You want to arbitrarily decide some things are constasts and others aren't - to suit some argument, instead of recognizing that when it is the cap change being considered, the effect on the circuit cannot be a constant as it is the very thing being considered.

What you have done is leap from hypothesis to calculations, and too quickly at that. We know, factually and most certainly, that if we have two different caps in the circuit, the ripple will NOT be the same, yet you want to claim same as some way to contrast different caps. You used a similar theory about cap temp, ignoring that the whole purpose was to understand temp and that it's different.

I have to say that your interpretation of the information is loosely based on random evidence, but improperly applied onto the point of being counterproductive continually.

Did you do any tests? We can in fact see that if the ripple voltage is higher the (CPU, for example) will consume more current, and with less charging of the cap, the regulator controller will have a higher duty cycle to keep the voltage up. This much is univerally true in the example circuit, upstream of the CPU. A short "I don't understand" doesn't quite cut it, nor do (even pages) calculations when based upon false premise.

[davmax]

Re: Capacitor Life expectancy

999999999 I can understand your reaction to my stating that you need to study good theory. This was not done lightly because much of the circuit operations you describe are not correct and as before with the Elna document you reject simple, practical and proven circuit theory that describes practical operation. Indeed in your last post you are in fact classing proven theory and the Elna equations and tables as hypotheses of my own. I have consistently had to refer you back to hard data that you tend to ignore for the purpose of your argument, in fact you have been doing what you claim I am doing. That is incredible. You gave a good positive input with your experience using larger diameter caps. Yes and I made some mistakes exploring the ELNA data whilst investigating a number of cap issues raised and I thanked you for pointing out a legitimate error. This I welcome. But when I see incorrect statements particularly those not supported by fact I point them out and point back to FACT. Part of the two way exchange.



Yes I do make measurements and I do design systems that work first time. I will not tackling a design without first fully understanding the theory of operation.



Lets look at two of our recent respective posts.



My post.

My last post acknowledged that confusion could occur over the words "constant voltage" However some clarification as to what this means is appropriate.
When calculating circuit conditions it is essential to use a specific a fixed value such as in the example cap calculations a ripple current of 2.2 amps was used. This is fixed and therefore "constant". To obtain this current value the ripple voltage must also be "constant". I think it will be understood that calculation would be very difficult if values were not held "constant" ie fluctuating. So the term Constant Voltage source is appropriate, although it can be misunderstood, and it can apply to either a DC or AC voltage source. Clearly in the cap examples ripple is an AC component.



First I acknowledged that the wording I used might be confusing. I then described how circuit theory is applied in calculation, not how the circuit operates under all conditions. It is simply explaining the calculation method and is correct with no fancy hypothesis. Nowhere does it state or imply that multiple capacitors must have the same ripple current unless they have identical ESR. Nor does it prohibit variability of circuit operation.

Previously I have stated that the filter cap/s and processor have ripple current flowing through each in proportion to their resistive value (ESR in the case of caps).This holds true for ANY value of ripple voltage during circuit operation. This is simple and practical resistive theory.





Your post

This is possible, but will vary per circuit, is not necessarily true for at least a couple of reasons.

1) The cap will still follow the peak voltage even if it does so slower, it only has to rise to peak value within the period which it may still do. If it cannot, then we more often have #2 Yescap charges at any voltage value greater than the current charge i.e at and near peak voltage.

2) If it can't do this, a higher ripple current into the CPU means more current consumed from the higher voltage, and when the switching goes through zero crossing there is a greater voltage depression from lesser charging of the cap and higher impedance back into the circuit, so the voltage is then lower and the regulator operates at a higher duty cycle, really I mean pulse width in most cases, which A) This explanation is erroneous. The CPU operates on the mean DC voltage. There is no zero crossing effect on the DC side of a switching or any other regulator. Yes capacitance voltage does drop when charge reduces (see choke operation) and rises again when charge occurs again. This is ripple. In the case of a VRM regulator this is minimised by the choke in the DC output and a diode that conducts during the MOSFET OFF period, the two combine to maintain load current during the MOSFET OFF period thus minimising cap discharge and ripple. The duty cycle of the regulator is not determined by the high ripple frequency that is outside the regulating response time. The regulator responds to the average DC output voltage and the speed of the response in largely governed by the bulk capacitance of the output filter. Gives the higher ESR cap even more time to charge to that peak voltage ESR mainly determinesthe ripple current through the cap and also the cap voltagedropwhen there is a CPU transient load, yes really bad caps may be limited in charging. The output choke resists rapid changes in charge current and is therefore a major current determinant. The choke is designed to smooth out current changes. Any sudden demand for current from the CPU is supplied by the bulk capacitance. and B) higher power consumption, which may make the higher ESR cap as susceptible through higher voltage peaks vs it's higher resistance = ripple current through it may be lower, same, or even higher depending on circuit specifics. Higher power consumptionwill increase ripple in the whole output filter and CPU as cap discharge increases, the average DC voltage will be maintain by the regulator. A high ESR cap will draw less ripple current at ANY given ripple voltage (if the ESR becomes infinite then no ripple current will flow). This is simple resistive theory and well proven. There should be no “higher voltage peaks” instead higher cap voltage drop. Voltage peaks can be seen on VRM outputs and these are generated by a resonant ringing that is determined circuit reactive components not by any value of ESR.

The key here is we can't assume the circuit a constant (it has never been stated that the circuit operation is constant), since the cap is a key filtration element in that circuit which has feedback to control it. Yes the caps are key Poor cap - effects voltage sensed - feedback loop compensates for it - poor cap even worse off now but at least the circuit had kept regulation as well as the design allowed. Yes the regulator will still maintain average DC with poor caps. This will mean high ripple and high voltage drop during CPU transient load and will lead to CPU malfunction.

[999999999]

Re: Capacitor Life expectancy

999999999 I can understand your reaction to my stating that you need to study good theory. This was not done lightly because much of the circuit operations you describe are not correct

Yes they are, to the same extent that the ripple current changes, it is true. You are quite rude to continually try to claim that if someone doesn't agree with your misleading examples that they must be needed to study good theory or otherwise wrong it very vague ways. I have time and time again supplied information that you have no counter for unless you contrive inappropriate scenarios, like caps that couldn't be same temp or circuits where ripple couldn't be same (which is the whole point of a capacitor). You have it backwards, you are just now learning some new things about capacitors and have not had enough time to use the information and look back in retrospect to formulate more than theories. A good theory is only a beginning, it then necessarily depends on some accumulation of real evidence, testing in the appropriate scenario, not just a rewording of a spec sheet. If all it took was the spec sheet, it would have been in that spec sheet.

and as before with the Elna document you reject simple, practical and proven circuit theory

No, recall that I was the one who linked that document to be taken as it was. You are the one who immediately begain trying to reword it to align better with your original suspicions in a new PDF you were producing. I again pointed out nonlinearity in aging with another manufacturer document link and still you try to use impossibly contrived scenarios.

I have not read the rest of what you wrote, because it is unnecessary and because it is rude for you to not concisely state things, you should not be an author of epic novels and if you can't state a fact concisely then you are trying to again twist the truth. Truth needs no misleading context painted.

There is no need for pages of this dialog, all we needed was the original manufacturers documents but then you went and tried to reword them to suit some ego-driven agenda. When you did that, I pointed out the errors in what you were doing. End of story, you are not a cap manufacturer and should leave their documents alone and run your own tests instead of taking a guess and trying to only think it through in your mind. The problem with theory is you don't then do the experiments. Theory is only that, until proven, and completely inappropriate to use to produce derivative documents without solid testing and reproducible test results. By deliberately ignoring this you are not being practical nor proving, yet claiming another person is rejecting practical and proven theory. It is a house of cards you are building.

Anyone can twist information to suit their argument, which is why we have to reject your claims and consider the actual implementations as the testing scenario, and reject your theories until proven because thus far they are full of holes. I urge you to start over, to follow the manufacturers documents instead of trying to reword them to support your guesses. Original documents are in this case, better than a 3rd person interpretation.

[davmax]

Re: Capacitor Life expectancy

Yes 9999999 I have expressed gratitude a number of times for your good inputs and the Elna document. What I have found difficult is that I use the Elna formulae and plugin data from manufacturers data sheets and you seen to claim this is my theory , dream, guesses etc. When you express concern about rounding of life to 500 hours, something I have yet to see, I checked the life curves presented for a number of caps on the Elna paper and advised that these indicate no problem and giving you the reference. You ignore Elna data. I explain some circuit principles as simply as I can and you do not act as though you understand.

I made a slip in my last post concerning the increase in cap ESR in that I essentially said that there was no increase in ripple voltage. Of course this is not correct, the load ripple is affected upward, only the ripple source voltage is unaffected. Ripple source voltage increases with load and decreases with increased in filter capacity.

I am willing to acknowledge any error because we all make mistakes. I felt you made a major mistake in describing a VRM regulator as controlling ripple within the control loop.

I make no claim to be perfect or to know everything about all the physical and temperature related cap properties. However I do know how to apply them and thanks to you, I am exploring the Elna and other information to achieve objectives that I believe most cannot confidently define.

They are:
1. The maximum operating temperature. This is controversial with different inputs producing varying conclusions
2. To determine the relationship of case to core temp. It seems pretty clear in the Elna paper.
3. The Lifetime impact of maximum rated core loss. Looks pretty clear with Elna
4. To verify that Elna based theory does indeed line up with your measured gains of larger diameter caps. A verification of practical and theory. This is the case.
5.The quantifying the gain difference between increasing diameter of standard caps versus long life caps. Clarifying the relative benefits and applications.

Now a lot of the above can be simply expressed but I would like to understand what is right and certainly not who is right.

I do really think it amusing that you urge me to folow manufacturers documents when this is what I have asked of you time and time again. I note you only snipe at what you claim I am doing. There is no postive feedback of correction that is manufacturer or data based. You have even quoted a dimensionless life curve against the real Elna ones I referenced for you (looking at factual a data).

I apologise for telling you to study theory but your theory was wrong and I was losing patience with such inacurracy and lack of recent good feedback. I do not like saying such things.

I will carry on to completion the objectives I have set. It will prove most informative.

[999999999]

Re: Capacitor Life expectancy

To be frank I think our discussion has reached as fruitful an end as it may. I am content that we may agree on some things and disagree on others.

[davmax]

Re: Capacitor Life expectancy

Thank you 999999999 we agree on this matter. I am sure that we will try to keep on topic and constructive.

[davmax]

Re: Capacitor Life expectancy

There has been disagreement over the 105C max temperature for caps.

One party stated it is core temp others thought ambient or case. I have searched manufacturers information for an absolute statement on this. The Elna document that I evaluated appeared to indicate 105C was ambient, however this can only be deduced.
Now there is a Nippon paper that supports all the ELNA information, but additionally makes it very clear that the 105C specification is ambient.

Looking at the Nippon paper it proves that my break down of the Elna formulae was 100% correct and not some fanciful twisting of the facts as has been suggested.

Lets now make it clear that the 105C rating is ambient temperature.

The Nippon paper can be found here: http://www.chemi-con.co.jp/pdf/catal...n-e-060905.pdf

And the defining extract is attached.