how to find bad caps if they measure/look good??

[smile]

Hello, I recently have replaces some caps in my PSU , the were easy to spot because they were bulged and when replaced the PSU does not make whining noise anymore and works ok.

However I have Agilent multimeter that can measure caps, so I measured it and it reads just fine. So the question is how can i identify a capacitor if it looks OK because I read that they also fail looking fine.

Until now I was sure that desoldering and measuring them was reliable but it seems not???

[tom66]

Re: how to find bad caps if they measure/look good??

You need an ESR meter, not just a capacitance meter.

[raresvintea]

Re: how to find bad caps if they measure/look good??

You need a ESR meter. It's the best way to measure the internal resistance. You measure the capacitance but with a clasical multimeter you can't measure ESR.

Succes!

[Khron]

Re: how to find bad caps if they measure/look good??

Also, the manufacturer / brand can also be an indicator The "good" ones are very few, while the crap ones are apparently more and more numerous.

[budm]

Re: how to find bad caps if they measure/look good??

Two of many cap's parameters are ESR (Equivalent Series Resistance), ESL (Equivalent Series Inductance).

[tom66]

Re: how to find bad caps if they measure/look good??

ESL's pretty much fixed and doesn't change though.

[smile]

Re: how to find bad caps if they measure/look good??

Isn't ESR and impedance the same thing?
The Agilent U1252B can measure impedance

[mariushm]

Re: how to find bad caps if they measure/look good??

No, it's not quite the same thing.

However, when selecting capacitors to replace broken ones, you can use the impendance values in the datasheets as if it's ESR.

At low frequencies, the impedance can be quite different than the ESR but at high frequencies like 100kHz (where usually low esr also matters), the differences between impedance and ESR are usually minimal (less than 1%), especially if the capacitors are high quality.

Thing is the Agilent won't measure the impedance at 100 kHz, so the value it measures won't be equivalent with the ESR value of the capacitor. You need a cheap ESR meter to measure a close enough value or an LCR meter with ESR measurement for an accurate value, depending how much money you're willing to spend.

Z = Impedance

Z² = ESR² + (XL -Xc)²

At high frequencies, the second part is close to 0.

ESR = DF * Xc = DF / ( 2 * pi * frequency * C)

DF = tan δ in the datasheets

and so on and so forth..

Also see

http://en.wikipedia.org/wiki/Equival...ies_resistance

https://www.badcaps.net/forum/showth...460#post119460

http://en.wikipedia.org/wiki/Dissipation_factor

[smile]

Re: how to find bad caps if they measure/look good??

Thank you for reply.

Anyone knows what meter is that :
http://www.youtube.com/watch?v=WjkYq4Auvk8

[mariushm]

Re: how to find bad caps if they measure/look good??

Looks like a meter made from a design from pro-radio, probably the one here: http://pro-radio.ru/measure/3288-89/

It probably uses the same math and "thinking" to determine the ESR as ESR Micro or Blue ESR.

[smile]

Re: how to find bad caps if they measure/look good??

Looks like a meter made from a design from pro-radio, probably the one here: http://pro-radio.ru/measure/3288-89/

It probably uses the same math and "thinking" to determine the ESR as ESR Micro or Blue ESR.

My russian is not that good, is it any good? Can't find any links to buy one (check the price).
I have also found another made in chine (like everything else these days) meter
but it's lcr meter

M4070 AutoRanging LCR (it seems to be able to measure electrolytic caps too)
http://www.ebay.com/itm/M4070-AutoRa...item4abf4b9411

Any information on this one?

Another one is nice but more expensive and can't measure electrolytic caps? Why make one line this at all, german made or not not a single 1% resistor, caps are doubtful too, not yellow type.
http://www.ebay.com/itm/AE20204-High...item3ccba90f67

[mariushm]

Re: how to find bad caps if they measure/look good??

The ones on pro-radio are not commercial, they're sort of designs participating in contests. If you want you can make one yourself if they attached the design, or you can register on their forum and talk to them.

But it's easier to just contact RusMike here on the forum and talk to him about ESR Micro, he's the one selling those.

Seems you still don't understand about capacitors... THere's several things to measure about them, just two of them being capacitance (the value written on them) and the ESR.

Even plain multimeters can measure capacitance with some error margin, just like ESR Micro does. Capacitance is not really important. valuable to know, because from the start electrolytic capacitors can have +/- 20% of the capacitance value written on label. Also, when they go bad, capacitance can actually increase in value.

So you can't reliably know if a capacitor is going bad, just by measuring capacitance, and that's what both of them you linked to do.. They don't mention anything about ESR.

It's more important to know the ESR value, which is short for equivalent series resistance.

It's like having a resistor inside the capacitor all the time. Usually, with low esr capacitor, this resistor is like it's not there, or there's very small resistance, 0.02 ohm for example, which is very small.

When capacitors go bad, this resistor inside the capacitor increases in value and the resistor can interact with other resistors near the capacitor and mess up the circuit.
Also, because electricity goes in and out the capacitor, the resistor.. well.. resists and heats up, so it affects the capacitor even faster, makes it go bad faster.

So it won't matter that you know the capacitance, because you don't know the ESR and those meters don't tell you that.
Even a 12$ meter can measure capacitance but it's pointless: http://www.ebay.com/itm/Digital-Mult...item3f1cc35209

Now, there are LCR meters that HAVE esr measurement, but even there you have to be careful.
The value of this resistor in the capacitor is heavily dependant on the frequency of the circuit the capacitor is used in.

For example, your house's mains frequency is 50 Hz, because you're in Europe. If you put a transformer and a bridge rectifier to get DC voltage, your DC voltage is now at 120 HZ (the diodes in the rectifier double the frequency). So if you use a capacitor to smooth out the voltage, this capacitor is subjected to 120Hz.

But capacitors in switching power supplies are subjected to 60-100 kHz, which is 50-100 times the frequency above.
The ESR value will be one thing for 120 Hz, and a totally another value for 50 kHz , 60 kHz, 100 kHz etc

If you do want a LCR meter, you need to be careful to have ESR measurement function, but even then you have to pay attention to see what kind of frequencies it supports.
Most cheap ones only support 100 and 120 Hz or at best 1 kHz, which is kinda pointless when you work with power supplies that work at least at 60 kHz.

The cheapest, tested and reviewed LCR meter than can do 100kHz and measure ESR that I know of is Mastech MS5308: http://www.ebay.com/sch/i.html?_odkw...s5308&_sacat=0

A good manufacturer, but untested/reviewed meter so far so I can't guarantee it's good inside is this one, a bit cheaper :

http://www.ebay.com/sch/i.html?_trks...at=0&_from=R40

If you want brand names, even a 1kHz model costs more than 200$ and for a 100kHz model, you look at $500 or more.
For example, probably the cheapest Agilent meter with ESR and 100kHz would be this one:

http://uk.farnell.com/agilent-techno...1mm/dp/1903368


These esr meters will give you a more accurate value for ESR compared to ESR Micro or Blue ESR or other home made meters, but these later meters can approximate the ESR good enough for repairing stuff and only cost 50-60$ or more, not $150-200 or thousands.

[Wester547]

Re: how to find bad caps if they measure/look good??

For example, your house's mains frequency is 50 Hz, because you're in Europe. If you put a transformer and a bridge rectifier to get DC voltage, your DC voltage is now at 120 HZ (the diodes in the rectifier double the frequency). So if you use a capacitor to smooth out the voltage, this capacitor is subjected to 120Hz.

But capacitors in switching power supplies are subjected to 60-100 kHz, which is 50-100 times the frequency above.
The ESR value will be one thing for 120 Hz, and a totally another value for 50 kHz , 60 kHz, 100 kHz etc

Aren't input capacitors in half wave power supplies only subject to half that frequency, though? (50/60Hz as opposed to double the mains, 100/120Hz, or full wave rectification) Also, when you say they're subject to 60-100KHz, do you mean that happens during a power surge, power outage, power spike, or during some sort of interference, or that they do regularly experience that? I thought it was always 120/100Hz ripple going across the secondary capacitors in full wave and half that in half wave. Also, does that mean that it's still 100/120Hz ripple going out to the computer or is it more than that (unless you're talking bout square wave and higher, and would linear regulation change this too?)?

[mariushm]

Re: how to find bad caps if they measure/look good??

Aren't input capacitors in half wave power supplies only subject to half that frequency, though? (50/60Hz as opposed to double the mains, 100/120Hz, or full wave rectification)

have you seen any have wave rectified computer power supply lately? that's the equivalent of using a single diode to rectify ac voltage to dc? They're only still around as small cellphone chargers/ usb chargers...

I think you're confusing concepts...

Also, when you say they're subject to 60-100KHz, do you mean that happens during a power surge, power outage, power spike, or during some sort of interference, or that they do regularly experience that?

I thought it was always 120/100Hz ripple going across the secondary capacitors in full wave and half that in half wave. Also, does that mean that it's still 100/120Hz ripple going out to the computer or is it more than that (unless you're talking bout square wave and higher)?

Linear power supplies get the AC voltage into the primary side of a transformer and out comes ac voltage on the secondary side but lower voltage, which then gets rectified by diodes and reservoir capacitors are used to store energy when the diodes aren't conducting or when you get near the 0 level of the AC wave.
These capacitors would see 100-120 Hz, double your mains AC frequency. These power supplies are very heavy due to the transformer which also needs a lot of copper, which is expensive.

Computers and monitors and so on use switching power supplies because it's unrealistic to use linear power supplies for this, the transformers would be HUGE to do 300-500w, at so many voltages.

So switching power supplies work by first rectifying ac voltage from mains AC to DC voltage, still at high voltage 240v x 1.41 = 350v or so ... then send them to an IC which chops it in pulses that get send to a transformer. These pulses can be from something like 20k to 100k pulses each second - the IC adjusts the width of the pulses depending on how much current the computer needs based on feedback from the seconday side of the power supply.
The transformer can therefore be much smaller and lighter, because it's designed to work at higher frequency, much higher than 120 Hz..

It's usually above 30-40 kHz because otherwise you or your animals would hear the transformer noise, due to the magnetic field and plates and wires in the transformer shrinking and expanding from it.

So on the seconday side of the transformer then will be diodes which rectify the DC and then you have capacitors smoothing out this dc voltage which comes in pulses, up to 100k or even more pulses a second.

[Wester547]

Re: how to find bad caps if they measure/look good??

Well, I was speaking for older power supplies when I say half wave rectification (a ripple frequency equal to that of the mains) - by that I mean a forward converter without any PFC circuit at all. What you're saying, though, is that every component in a switch mode power supply has a ripple frequency much higher than 100/120Hz going across it? And I am curious though... is that sort of ripple (30-40 KHz+) going to the load and to the computer (from the PSU) or is it still 100/120Hz (and are linear regulated parts of the PSU - +5VSB and linear regulated DC-DC conversion - operating at 100/120Hz?)?

[smile]

Re: how to find bad caps if they measure/look good??

Thank you for explanations, the meter must be an ESR meter.
I have looked and the best seems to be Atlas ESR PLUS Model ESR70.

- It can safely discharge caps before measuring them
- measures capacity and ESR, also mesures rezistors
- small and compact

However, I went to local radioshack and they have shown me this meter,
https://smarttweezers.3dcartstores.c...51AD_p_59.html

Obviously at that price level you have to be living in india (where its a cultural thing to fix everything) or must be replacing capacitors 24/hrs a day for clients and not even the bad ones to justify the price.

It escapes me how the Atlas ESR70 is not a tweezers type meter why bother to make it so small its almost hard to use? If you need 2 hands to hold the probes it slows measurement to a crawl, naturally you hold the PCB with one hand and the meter with another, this is impossible to do if it's a test leads only version. Oh an there are attachment like tweezers I know, but since the Atlas ESR70 is so small it will be hard to read the damn LCD screen.

The 30sec timeout is nonsense either. Look at how many times they had to switch it on on youtube just for demo purposes.

If it would be tweezers type meter and you could attach leads to it if you need them then the price could be justified, but for single function device it's crazy to ask 1/4 of Agilent U1252B their top of the line DMM.

Another thing is that since it is so small there is no capacitor table on it, while I understand its a compromise for the size, usually the capacitors are laid out in patterns, same voltage for given area.

So then why the meter designers Atlas ESR70 included can't make a design using two function keypad (like every calculator on the market) to chose the most common cap voltage with a single key press and to allow to enter capacitor voltage for uncommon types by hand? Is it only me that thinks this is more easy, especially if they make sounds that show the ESR is low or bad etc.

Saying something like the meter can't read voltage required for ESR measurement so you must look at the tabe is nonsense IMHO. It's just shows that making 2 buttons on whole device is more easy and cheap, but counterproductive.

The meter of this price level could at least be made like an Ipod or something, with touch screen, unlimited cap presets that click with you finger to select what you are going to measure to avoid placing real buttons. This way you could ask the meter if the cap is good or not. But to use damn charts to make simple measurements heck it's cheaper to replace all capacitors. An exclusion might be only a UPS unit where there hundreds of them.

And meters from china like this one http://www.fvstore.com/AutoRanging-I...R/p204720.html
Is not encouraging at best, it's like measuring resistors with typical cheap multimeters when the error of the DMM is larger than 5% resistor and nobody seems to notice that, on the other hand I had to order the 1% resistors since everyone thinks it's a luxury item or something. Only power resistors can be 5%, 10% etc IMHO.

[smile]

Re: how to find bad caps if they measure/look good??

Oh and could someone compare the Atlas ESR60 and ESR70 models, the older model seems to be chaper but is it missing only 2x lower resistance mode?

[mariushm]

Re: how to find bad caps if they measure/look good??

I may sound like advertising but..

ESR Micro measures resistors as well.
ESR Micro is also compact
ESR Micro also discharges capacitors but at somewhat low voltage, I may be wrong but I think it can handle 16-25v capacitors with a bit of charge in them - it's recommended to discharge capacitors first which is simple, just get a 1k resistor and put it across the leads of the capacitor for a few seconds.

Be warned that Atlas advertises discharging capacitors, but it's not fool proof - you really shouldn't test - let's say - 100v+ rated capacitors without discharging them first.

As for the 30s timeout.. that's easy to explain. They went with a 12v battery (probably A23 type, it says which type on the bottom) which are usually almost as if they have several wrist watch battery cells stacked in them.
So the capacity of these batteries is very small, usually under 100mAh - wikipedia says 55mAh : http://en.wikipedia.org/wiki/List_of_battery_sizes

This is very small and the meter also uses about 3-5mA for measurements and it also uses a linear regulator to get the 12v down to 5v or whatever it uses internally, which decreases efficiency.

ESR Micro, for comparison, uses 2 x CR2032 (or something very similar, I'm too lazy to open mine to see). It works internally at 5v or 3.3v, two batteries give it 6v and each battery is actually around 150-250mAh .. 3-4 times the capacity of Atlas' battery, maybe even more.
later edit: It also turns off after about 1 minute if I remember correctly, after the last measurement... and it turns off quite fast, within 1-2 seconds.

As for your capacitor tables... those are generally pointless. They're just estimations, something to help you until you learn what to expect.
The microcontroller used in the Atlas probably also don't actually have the memory (flash space or eeprom, it probably doesn't even have eeprom at all) to store all those values.

Besides, like i said, it's irrelevant.

You can have... for example.. a Panasonic FM 470uF 16v with esr 0.04 ohm and a Rubycon RX30 470uF 16v which has 0.15 ohm

If your table says the esr should be under 0.1, does that mean the Rubycon is faulty? No, it's just a low esr capacitor optimized for handling heat (it's a 130c capacitor).

later edit: oh, and the meter can't differentiate between tantalum capacitors, ceramic capacitors, film/mica capacitors, polymer capacitors.. tantalum capacitors will have low esr even at low uF and low v rating, which may confuse the meter tables, ceramics can have high esr/resistance and the meter may report the capacitor as faulty if it were to use some internal values without knowing the type.

later edit: in reply to your question about esr60 and esr70.. honestly, i don't know, i never bothered to check. I think it's something about discharging and something about speaker (sounds) but I'm not sure.

You can see the ESR70 reviewed here: http://www.youtube.com/watch?v=LZ49GCnf0V4

And the ESR60 is still available on Farnell and you have the datasheet there to compare it with ESR70: http://uk.farnell.com/peak/esr60/met...0mm/dp/7830254

[smile]

Re: how to find bad caps if they measure/look good??

I may sound like advertising but..

I might just get ESR-micro V4.0 too, are the probes OK? They do look like the ones made in china that are very bad?

Since you have the meter, what good, bad can you say about it. The price is very OK, since it is made in russia or since cheaper components are used? The specs are better then ESR 70

Although this does not matter for ESR meter, I once purchased made in china probes that looked like Agilent original ones, and the wire used in them is almost as thin as in headphones, I wonder what will happen if somebody measures 10A with them.

BTW how did you pay for the meter, the international money by bank transfer costs as much as the meter itself.

Comment from another board:

This is interesting comment: Discussion of ESR measurement and accuracy

Technicians new to ESR measurement frequently get confused about what is a good reading and what is bad. There is no dividing line. The ESR of a good capacitor depends upon it's construction, value and voltage rating. All of the ESR meters compared here have some kind of indication of what is good, fair and bad. But those are only guidelines, not hard and fast rules.

Common electrolytic capacitors in the mid range of capacitance and voltage will have low ESRs. Bad ones are easy to detect and all of the meters evaluated here will do the job.

But a good 1 ufd, 200 volt aluminum electrolytic can have an ESR of 100 ohms or more. That is beyond the measurement range of all of the meters listed above!

Film capacitors make matters worse. A good 1 ufd, 200 volt metallized polypropylene capacitor can have an ESR of less than 50 milliohms. Yet it's capacitive reactance at 100 Khz is approximately 1.6 ohms, at 50 Khz it is 3.2 ohms. The result is that all of the meters evaluated here except the Atlas ESR60 will indicate 1 ohm or more for such a capacitor even though there is nothing wrong with it. The ESR60 analyzes the capacitors real and imaginary impedances eliminating the reactive component so it can measure the ESR of any capacitor.

A tech can do three things:

Limit ESR measurements to mid-range electrolytes only. Then it is easy, but you have given up on parts of this valuable diagnostic tool.
Learn the characteristics of the different types of capacitors. The capacitor manufacturers make all this information available. Once you to learn it you can get the most out of your equipment.
Spring for the Atlas ESR60 but add the third hand if you are going to working on boards with short capacitor leads.

[mariushm]

Re: how to find bad caps if they measure/look good??

I contacted RusMike directly on this forum and he sent me the meter with probes and even some spare parts (protection diodes that may break down if I test charged capacitors and other small npn transistors or mosfets that are on the board) but without the plastic case for about $50.
At the time, he was waiting on pre-ordered cases to arrive and I simply didn't want to wait until he received the plastic cases for the meter, so I toolk it without one and I made a simple case on my own here.

The probes RusMike imho are above average, they're not the typical cheap Chinese probes. Tips are sharp, they even have removable caps, cable is flexible and feels good, they work good when using them on another meter.

Though, I can't guarantee you'll get the same type of probes, I bought mine a while ago.

I actually decided they were too good to cut and solder them on the ESR Micro board - I had an older 2-3$ 830b multimeter i bought from some 99 cents type store, cut the ends of those probes and soldered the cheap probes on the meter. I then re-calibrated the meter to take into account the leads' resistance - by default it's calibrated for the connector already on the meter (you can desolder capacitors and just insert the leads of the desoldered capacitor on the socket like you see in the manual here on page 2: ).

It works just fine, not much to say, the components on the board are good, not no-name Chinese parts as far as I can tell (though you may have hard time identifying them simply because they're so small smd parts), he didn't scratch out the microcontroller identification on the chip like some Chinese do... overall it's quite simple tool.

I'm sure if something breaks, he would be willing to send you some parts or tell you what part is that is you contact him through the forum or that website, if you don't want to send it to Russia to get it fixed.