EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: doodle43 on June 07, 2020, 11:50:42 pm
-
My Samsung TV broke, presumably because of this leaking capacitor. It is a Sam Young NXE 1200uf 25V. I am totally new to all of this btw, and I don’t even know how to solder yet. But I’d like to try to fix it.
I don’t know the exact specs, but I found this data sheet - https://www.samyoung.co.kr/download/new/NXE.pdf (https://www.samyoung.co.kr/download/new/NXE.pdf)
I haven’t measured the physical dimensions but I am thinking it is the 12.5 x 20 one. So it would be 3000 mArms ripple current and the listed ESR:
Ωmax./20°C,100kHz - 0.014
Ωmax./-10°C,100kHz - 0.042
I could be wrong but it’s the only data sheet I could find.
I was thinking of replacing it with this Panasonic FR series - https://www.mouser.com/ProductDetail/667-EEU-FR1E122L (https://www.mouser.com/ProductDetail/667-EEU-FR1E122L)
The only problem is it seems to be not as wide, lower ripple current, and possible higher impedance and ESR. I’m not sure because I’m a noob to all this. But I am having trouble finding a better replacement. Please help
-
Use 1200uF 35V rated FR series cap to be safe if 5mm larger height fits. Or find different series, or use 1500uF 25V caps from different series if cannot find appropriate 1200uF, say Panasonic FM.
-
Don't worry too much about it.
FR series is very good, much better than those capacitors.
If you're paranoid about it, you can use versions with higher rated voltage like 35v for example... and if they're too tall, you can place the capacitor flat on the circuit board. Technically the leads should be as short as possible but in practice it's not gonna be an issue.
1200uF 35v FM series is in stock : https://eu.mouser.com/ProductDetail/Panasonic/EEU-FM1V122L?qs=S7VFkBl86Ap4U438Gy8KPQ%3D%3D
It's 30mm tall, but like i said, you could lay it flat if too tall... then it's only 12mm in diameter
-
I don’t even know how to solder yet. But I’d like to try to fix it.
Almost 10 years ago, I started trying to fix electronics. The first was an Optiquest Q9 19 inch monitor. It had bad caps and I replaced them with Rubycon. Details at
https://www.badcaps.net/forum/showpost.php?p=103476&postcount=16 (https://www.badcaps.net/forum/showpost.php?p=103476&postcount=16)
The best soldering video I found when I started was
https://www.youtube.com/watch?v=IpkkfK937mU (https://www.youtube.com/watch?v=IpkkfK937mU)
For my first soldering iron, I just ordered whatever was cheapest on ebay and cheapest 60/40 lead solder. I think it was a total of $5 USD. While basic, it did the job since my monitor is still working almost 10 years later.
As others have mentioned, either Panasonic FR or FM are good high quality low ESR caps.
-
I don’t even know how to solder yet.
Sometimes desoldering is more difficult than soldering. That board doesn't look too bad.
Decent, professional, succinct soldering videos that show how to do it and what happens when you get it wrong: https://www.youtube.com/playlist?list=PL926EC0F1F93C1837 (https://www.youtube.com/playlist?list=PL926EC0F1F93C1837)
You will need to practice before you attack attempt something valuable to you.
-
I’m not sure if a 30mm height would fit. I think the original one is 20mm. Maybe I’ll take the back off tomorrow and have a look and get the actual dimensions. I don’t think it would fit on its side. See the photo of the whole board.
I see that the 35V FR seems to have a higher ripple current and lower impedence/ESR compared to the 25V, maybe closer to the original for those specs? I was hoping to get as close a match as possible for all specs, so would the higher voltage one be better? A friend of mine told me not to go higher voltage, but I see it recommended a lot.
https://www.mouser.com/ProductDetail/Panasonic/EEU-FR1V122?qs=ATUT3BIXiFhuEXBCnw4wqg%3D%3D (https://www.mouser.com/ProductDetail/Panasonic/EEU-FR1V122?qs=ATUT3BIXiFhuEXBCnw4wqg%3D%3D)
Thank you for your replies and the videos. I really appreciate it!
-
A friend of mine told me not to go higher voltage, but I see it recommended a lot.
Tell your friend he had no clue what he was saying. Higher voltage rating does not hurt anything (except size), only makes it more reliable. And with it you get larger can, lower ESR and higher ripple current rating. FR 1200uF 35 V is 25mm high. You don't need exact match, You want ripple current rating not lower than original and about the same or lower ESR.
-
Plenty of room to place the capacitor flat on the board if needed.
See image below.
Perfectly safe and fine to use capacitors rated for higher voltage... no idea why people advise against it in this context.
[attachimg=1]
-
I think my friend’s reasoning was that a higher voltage would create more heat, and something about the zener diode. But I am definitely leaning towards that FR 1200uf 35V. It looks like it is 0.5mm wider diameter, 5mm taller which is probably fine, higher ripple current, and slightly larger ESR (15 mOhms). So would that be ok?
Is it recommended to replace any of the other capacitors? They don’t look swollen to me.
-
I think my friend’s reasoning was that a higher voltage would create more heat, and something about the zener diode.
:palm:
-
Is it recommended to replace any of the other capacitors? They don’t look swollen to me.
Any cracked RIFAs should be replaced on sight, before they make you aware of their non-existence. Plenty of horror-story pictures around.
-
There's no RIFA capacitors in that monitor, the monitor is too new.
Even though it's not swollen, I'd also replace the 1200uF 10v capacitor to the left of the swollen capacitor. Can be replaced with 16 or 25v rated capacitor as well, and you can probably also go up in capacitance to 1500uF without any risks.
The reason for this suggestion is that capacitor degradation is accelerated by heat, and just like that other swollen capacitor, that 1200uF 10v capacitor is close to the heatsink and subjected to heat.
Even though it's not swollen, it could be somewhat degraded ... maybe pressure released through the bottom instead, or maybe there's just not enough electrolyte inside to cause enough pressure to raise the capacitor top
What else ... probably not needed, but as they're so cheap, maybe consider the capacitors that are part of the startup circuit for the controller chips ... if you have issues like monitor randomly not wanting to start when you press the button (but starts and works fine at consecutive attempts) those capacitors can go back.
If I analyze correctly those capacitors should be around 10-47uF and rated for 50v or something like that... I would guess the capacitor near the J825 link (by the blue rectangle x2 capacitors) and probably the yellow cap by the CM804, or the one by CM808... can't be sure.
These don't have to be low esr, low impedance... something just low impedance would be fine.... and again you can go up in voltage rating ... typically the operating voltage of the controller chips is 10-20v but they use 50v rated capacitors because as such low capacitance values there's almost no benefits to using lower voltage ratings, it's not a board where diameter and height of capacitors is critical.
-
I'd also replace the 1200uF 10v capacitor to the left of the swollen capacitor.
Can be replaced but not needed IMHO. AFAIK and IME only discontinued Samwha WB series had serious issues. Other series are OK.
-
I think my friend’s reasoning was that a higher voltage would create more heat, and something about the zener diode.
Maybe he was saying to you that running the capacitor at higher voltage than its rating causes heating and it kind of works like a zener diode. After all, that's what you are writing in your quote, but note this is the complete opposite of what we are discussing in this thread, we are discussing about running the capacitor at lower voltage relative to its rating - by increasing the rating.
So maybe you understood it the wrong way around. The problem with all these "rules of thumb" and hearsay (instead of understanding) is that even a small miscommunication or a typo completely messes everything up. What you should have done is to discuss this with your friend until you reach an understanding what he's saying, even if it takes hours to do that. Fun times, such discussions! At this point you can also see if he has a clue; teaching something requires quite a lot of understanding and asking to be taught demonstrates such abilities or lack thereof.
Getting a capacitor with excess voltage rating usually helps increase the lifetime (it has smaller ESR so less loss - and heat dissipation happens over larger area because it will be physically larger), as does getting a capacitor with excess capacitance rating (for exactly same reasons!).
Obviously, the choice of getting the exact same C and V rating makes sense as well; if you pick a higher quality product such as a well-known Panasonic FR or FM series, it's likely to solve the lifetime problem even without increasing either C or V rating.
Improving cooling is another option.
ESR is the most important parameter but typically not defined in the datasheets as-is. Sam Young datasheet is a nice exception to the rule.
So by all means, replace the 1200uf 25V with a Panasonic FR or FM 1200uF 25V component. If there is enough space on the board, getting either 1800uF 25V, or 1200uF 35V replacement is likely equally beneficial for lifetime. Or why not 1500uF 35V.
-
ESR is the most important parameter but typically not defined in the datasheets as-is. Sam Young datasheet is a nice exception to the rule.
Nope, ripple current spec is the most important in most cases as it tells how hard you can stress the capacitor. And in no way it's exception to the rule. Basically ALL low ESR capacitors have ESR and ripple current ratings in the datasheet.
So by all means, replace the 1200uf 25V with a Panasonic FR or FM 1200uF 25V component.
Panasonic makes high quality caps and likely would be OK but still it would be significant downgrade it terms or ripple current and ESR specs, so I would not advise it.
-
I think my friend’s reasoning was that a higher voltage would create more heat, and something about the zener diode.
:palm:
To tanslate the above response into plain, British English: that's a load of bollocks!
Explaination: a capacitor will only ever charge up to the voltage, provided by the circuit, never more. The voltage marked on the can is the maximum voltage rating, which means it's the highest voltage it can safely charge up to, without being damaged. Using a capacitor rated to a higher voltage, than the original is no problem and can result in a longer life. On the other hand, you shouldn't use a capacitor, with a lower voltage rating, because it could reduce the life or explode. It's a similar principle to a hoist or jack's maximum safe working load.
-
Nope, ripple current spec is the most important
Yes, you are of course right, ripple current spec is there because the capacitor has ESR, which causes heating. With the ripple current spec, the math of calculating loss (Iac_rms^2*ESR) then calculating for internal temperature rise through the thermal resistances is already done - or they have just simply tested the part to see if it lasts. With ripple current rating available, you don't "need" to know ESR or Rth_core-to-ambient to select the capacitor for the job; but sometimes you would want to know them anyway, to do your own loss analysis, for efficiency calculation for example. Or just for pedantic and logical design procedure.
My point was, electrolytic capacitors for switch mode power supplies are primarily selected for ESR as the physical property - aka. ripple current rating; these two are directly related!, rather than voltage or capacitance. V rating obviously needs to exceed the highest possible voltage on the bus, and C must be enough for the filtering, but with alu elcaps, the chosen C often ends up being even an order of magnitude more because the ripple current rating (ESR) drives the choice. For example, the ripple voltage calculation, control loop stability analysis and step response calculation would reveal that you need a 123.4uF capacitor, but you need 2A ripple current rating so must pick a 1000uF capacitor instead because the closest 150uF cap was only rated at say 300mA.
Basically ALL low ESR capacitors have ESR .. ratings in the datasheet
So far, I have only seen those in polymer caps. Panasonic FR, FM datasheets for example do not list ESR values. They use dissipation factors instead, Panasonic datasheets list DF only at 120Hz, which you can use to calculate ESR at 120Hz (ESR = tan delta * 1/(2pi*120Hz*C)), but then you need to correct the ESR for 100kHz by utilizing the ratios in the ripple current frequency correction table, and realize the existence of second exponent in that ratio.
Panasonic makes high quality caps and likely would be OK but still it would be significant downgrade it terms or ripple current and ESR specs, so I would not advise it.
Looking again at the Sam Young datasheet, you are right; that cap has some very respectable specs, total world-class for a classic aluminium electrolytic (not polymer). I made the mistake of assuming Sam Young is a low-class brand, I have never used their parts. Assuming these specs are real, you need to choose something with similar or preferably even higher ripple current rating (if too low of a ripple current rating was the cause of original failure). That may mean having to choose a part with higher C than the original 1200uF, or a higher voltage rating, or both.
-
Based on how often I saw Samyoung and Samwha capacitors fail in power supply, they're either sensitive to heat (as the power supply is passively cooled and more or less sealed inside the monitor, minimal or non existent airflow), or they use a very unstable electrolyte formula making them prone to failure (like rumors say it's the case for KZG series from united chemi con, another ultra low esr series)
My personal thinking is that those capacitors started to degrade a long time ago, yet the power supply and monitor kept working reasonably well for a long time, so the circuit and everything is most likely not that sensitive about those technical parameters.
Even if the capacitor is not quite as good, it would probably still be within some safety range.
Also, keep in mind you're replacing a capacitor rated for 3000-4000h at 105c with a capacitor that's rated for 10000h@105c (for 1000uF 25v, endurance varies with cap. volume) ... so even if you have an accelerated "degradation" due to not so good specs (which I don't think it's the case here) ... the capacitor will probably still last quite a longer time than your current capacitors.
Basically, it would probably still take years to fail again.
-
So far, I have only seen those in polymer caps. Panasonic FR, FM datasheets for example do not list ESR values. They use dissipation factors instead, Panasonic datasheets list DF only at 120Hz, which you can use to calculate ESR at 120Hz (ESR = tan delta * 1/(2pi*120Hz*C)), but then you need to correct the ESR for 100kHz by utilizing the ratios in the ripple current frequency correction table, and realize the existence of second exponent in that ratio.
They clearly have it, they call it impedance @ 100kHz instead of ESR though.
-
I am really enjoying this discussion, thank you all. If it helps, what happened with my TV is that it never recovered from a power outage. It was plugged into a surge protector. The standby light comes on, and it will respond to the remote (standby light will go off when I turn it on), but it won’t respond at all to the physical power button. When it is “on”, there is no picture, backlight, or audio. It is about 10 years old I think and had no trouble until now.
Regarding the ESR- it is listed on the Mouser page. It says 15 mOhms for the Panasonic FR 1200uf 35V. Which seems to be slightly higher than the Sam Young. But the ripple current rating is higher too.
-
So far, I have only seen those in polymer caps. Panasonic FR, FM datasheets for example do not list ESR values. They use dissipation factors instead, Panasonic datasheets list DF only at 120Hz, which you can use to calculate ESR at 120Hz (ESR = tan delta * 1/(2pi*120Hz*C)), but then you need to correct the ESR for 100kHz by utilizing the ratios in the ripple current frequency correction table, and realize the existence of second exponent in that ratio.
They clearly have it, they call it impedance @ 100kHz instead of ESR though.
Are you sure, source or reasoning?
-
So far, I have only seen those in polymer caps. Panasonic FR, FM datasheets for example do not list ESR values. They use dissipation factors instead, Panasonic datasheets list DF only at 120Hz, which you can use to calculate ESR at 120Hz (ESR = tan delta * 1/(2pi*120Hz*C)), but then you need to correct the ESR for 100kHz by utilizing the ratios in the ripple current frequency correction table, and realize the existence of second exponent in that ratio.
They clearly have it, they call it impedance @ 100kHz instead of ESR though.
Are you sure, source or reasoning?
Trying actually looking into datasheet at least once before making your claim (it was first) about the lack of such?
[attachimg=1]
-
I've been through that datasheet a million of times, don't worry.
I'm asking for a source or reasoning for the claim that by impedance, they mean ESR. I mean, I hope you are right!
Physically, these two terms have different meanings. ESR would be even shorter to write if they wanted to do that. You have circler "impedance", which is not ESR, unless either they use the wrong term for some reason I don't understand, or the ESR happens to equal the impedance. But I would like to hear your stance on whether you think this is the case.
I have been thinking that these two could be close to each other if 100kHz happens to be the self-resonant frequency. But SRF can't be close to 100kHz with all parts listed, hence I expect impedance is close to ESR at 100kHz only on some parts.
The reason I'm asking is that as a part of a Uni course I was teaching at, I had to teach people about this exact thing as well, using the Panasonic FM capacitor datasheet, and I still can't say for sure what they mean with "impedance". Looked at a lot of appnotes as well, and one Panasonic appnote confirmed my way of back-calculating ESR from the DF and ripple current frequency correction table, but doing that gives a different result.
If you have insight on this, please elaborate.
-
I've seen through that datasheet a million of times, don't worry.
I'm asking for a source or reasoning for the claim that by impedance, they mean ESR.
Physically, these two terms have different meanings. ESR would be even shorter to write if they wanted to do that. You have circler "impedance", which is not ESR, unless either they use the wrong term for some reason I don't understand, or the ESR happens to equal the impedance. But I would like to hear your stance on whether you think this is the case.
I have been thinking that these two could be close to each other if 100kHz happens to be the self-resonant frequency. But SRF can't be close to 100kHz with all parts listed, hence I expect impedance is close to ESR at 100kHz only on some parts.
The reason I'm asking is that as a part of a Uni course I was teaching at, I had to teach people about this exact thing as well, using the Panasonic FM capacitor datasheet, and I still can't say for sure what they mean with "impedance". Looked at a lot of appnotes as well, and one Panasonic appnote confirmed my way of back-calculating ESR from the DF and ripple current frequency correction table, but doing that gives a different result.
If you have insight on this instead of guesses, please share.
Most ESR meters actually measure impedance, not ESR. And even if considering that listed ESR is true ESR and not impedance, the difference is minuscule, usually <10% which is nothing considering ESR is not precise parameter. Also you cannot calculate ESR @ 100 kHz from Tan(δ) @ 120Hz. Neither you can calculate ripple current rating from either ESR or Tan(δ).
-
From tan delta at 120Hz, you can calculate ESR at 120Hz. Sadly, ESR is a function of frequency (and temperature, and age, and so on)...
The necessary information is on the datasheet, though: the ripple current correction factors. We need to make an assumption, though: that because the ripple current ratings must exist due to heating caused by ESR, we assume that the manufacturer has made that particular table so that the listed ripple currents cause same amount of internal heating, at different frequencies. Then you can work backwards from Ploss = I2*ESR.
For example, the ripple current correction table says 0.80 for 120Hz and 1.0 at 100kHz. Assuming Ploss is same for both,
I1202 * ESR120 = I100k2 * ESR100k
ESR100k = I1202 * ESR120 / I100k2 = ESR120 * (I120/I100k)2 = ESR120 * 0.64
When I went through the math and needed confirmation on the feasibility of the calculation, I was finally able to find a strange Engrish Panasonic appnote which I could download by registering, which confirmed this calculation! I can see if I can still find that file.
The issue is, this calculation gives higher ESR values than the "impedance" number listed on the datasheet. To me, it seems, based on temperature measurements, that the ESR must be more than the value listed in the "impedance" column, and that the ESR calculated by this formula from DF, seems to be in the right ballpark. Simply put, I have used 16V 1000uF Pana FR caps on a buck converter at the ripple current rating, i.e., about 1.5A. They do get very warm! From the size of the components and general experience on component dissipation with similar surface areas, I can say this is something roughly around half a watt of dissipation. Using the "Impedance" number as ESR, though, would result in only 1.5^2 A^2 * 30mOhm = 67 mW being dissipated. This possibly can't be the case. Thus I think the "impedance" must mean something else, but can't confirm what it is!
Neither you can calculate ripple current rating from either ESR or Tan(δ).
Indeed you can't, you would need Rth-capacitor-core-to-ambient and maximum Tcapacitor-core. They have done all this for you.
The reason I'm mentioning this is to draw an analogue how you work with calculating MOSFET ratings. There, the current rating number is usually totally bogus (useless), but accurate-enough Rds(on) allows you to calculate P, and values for Rth are given, allowing you to calculate Tj; finally, max Tj is given. With elcaps, you can't do this because you lack information, but OTOH you have a realistic current rating given to you which satisfies most purposes - but doesn't allow efficiency calculation!
-
The issue is, this calculation gives higher ESR values than the "impedance" number listed on the datasheet. To me, it seems, based on temperature measurements, that the ESR must be more than the value listed in the "impedance" column, and that the ESR calculated by this formula from DF, seems to be in the right ballpark.
ESR cannot be higher than total impedance, only lower.
Indeed you can't, you would need Rth-capacitor-core-to-ambient and maximum Tcapacitor-core. They have done all this for you.
It's not as simple as just heating. There may be funny stuff happening chemically when high ripple current flows through capacitor. Usually you get higher ripple current with lower ESR but it's not a direct relation. And you may have higher ESR and higher ripple current at the same time when comparing different capacitors.
EDIT: Also at high frequencies capacitor basically stays at the same charge all of the time and ESR is dominant. At low frequencies capacitor goes through large charge/discharge cycles and capacitance is dominant.
-
ESR cannot be higher than total impedance, only lower.
Exactly - and conversely, total impedance, at every point, must be higher or equal to the ESR!
Please reread, I made an edit (a bad habit of mine) where I question the possibility of ESR being as low as what the "impedance" column says, based on actual experience of using these capacitors.
Run the ESR heating calculations for these "impedance" numbers and you'll see the dissipated power will be very low, at the maximum rating, similar to the ratings of tiny SMD resistors. Yet we all know these parts do heat up, and are part of the losses significant enough to matter in efficiency calculations.
Actually, these "impedance" numbers are really close to the "ESR" numbers given for similar polymer type capacitors, while the latter are supposed to be much better.
There must be something I'm missing, but I do think that the actual ESR must be higher than the "impedance" number, and if this is the case, that "impedance" number cannot be impedance either, because as we agree, impedance must be at least the ESR.
The point of other damaging chemical reactions than just those caused by heating is a good one. Additionally, the thermal conductivity from the capacitor's hottest spot to the outer surface varies between products somewhat.
The problem I have, I think I have good understanding on all this stuff, but this one number on the Panasonic datasheets is not fitting in my reality of things.
-
I’ve measured the bad capacitor just to validate, it is indeed the 12 x 20 mm. It’s currently sticking up an extra 2-3mm from the board due to the swelling. Unfortunately I measured 27, maybe 28mm from the board to the back plastic of the TV. Is that enough clearance for a 25mm tall capacitor? How much is recommended? I don’t really want to put it down flat as the only position would be diagonal very close to the heat sink. Or can I put it at a tilted angle like some of the smaller ones are?
-
I’ve measured the bad capacitor just to validate, it is indeed the 12 x 20 mm. It’s currently sticking up an extra 2-3mm from the board due to the swelling. Unfortunately I measured 27, maybe 28mm from the board to the back plastic of the TV. Is that enough clearance for a 25mm tall capacitor? How much is recommended? I don’t really want to put it down flat as the only position would be diagonal very close to the heat sink. Or can I put it at a tilted angle like some of the smaller ones are?
If it fits, that's fine. You don't need clearance from plastic.