Re: how to find bad caps if they measure/look good??
Well, do some work yourself.
The first Mastech review I linked to shows the meter opened up. The Peak ESR review I linked to shows the meter opened up with high res pictures in the forum. Look at them and compare.
Well, Mastech is a semi-professional meter that will be used by various people to measure inductors, capacitors and resistors at various frequencies, so it has to offer selection of frequency.
The Peak ESR is simply designed for the same purpose of other cheap meters, to tell you if a capacitor is good or bad, so 100kHz is the only one offered because that's what you usually see in datasheets so it makes it easy to check if a capacitor is going bad or not.
Actually, the ESR is NOT listed.
Datasheets tell you the Impedance value at a certain frequency, which is a different thing. Luckily, at high frequencies, if a capacitor is good, the ESR tends to have a value very close to the impedance value.
esr = dissipation factor * Xc (capacitor reactance)
impedance² = esr² + (Xl - Xc )²
The Peak ESR's purpose is - like any cheap meter - to tell you if a capacitor is going bad or not by reporting the ESR at 100 kHz.
You're supposed to use your brains and your actual past experience to determine if this esr value is suitable for the that capacitor based on where the capacitor is used in a circuit.
You're supposed to know that the circuit near the capacitor is a switching power supply using high frequencies, or it's just for filtering some low frequency signal, and so on.
Once you have some electronics experience you start to have an idea of what a circuit does based on how chips are arranged, what components are around a certain capacitor and so on.
Most circuits don't even work at 100kHz.. most cheaper chips used in power supplies run at around 52-75 kHz .. usually everything runs over around 40 kHz because lower than that, humans and animals can actually start to hear this frequency from the oscillations in the transformers.
So the point is, even if you measure a capacitor's esr at 100 kHz, in real world, when that power supply runs, the capacitor may only be subjected to 50 kHz or 75 kHz ...
Luckily, if it's a good low esr capacitor, the esr will either be somewhat flat down to about 1 kHz or slowly and linearly increase as the frequency decreases so if the ESR is low at 100kHz, it won't increase by much at 50 kHz or 10 kHz
Here's for example a graph from the link I posted a few posts above. If you have read that you would have understood this by now:
the green line is the Impedance...
the yellow line is the ESR...
That line where the A is below is 100kHz, you can see the capacitor is 11 mOhm at 100 kHz.
You can also see the impedance is actually going down a bit as frequency decreases, then stays steady and only goes back up a bit at what probably is 10 kHz point.
Meanwhile, the ESR is actually lower and flatter throughout these high frequencies, only going up as frequency goes below 1 kHz.
But you can see at 100kHz, the ESR and Impedance value are very close, just a few mOhm difference between them, that's why again it's easy enough to put an equal sign between the impedance at 100kHz value reported in datasheets and the ESR. For our not 100% precise, not 100% accurate needs, those values are close enough.
So if you know it's a low esr capacitor from a good manufacturer, you can assume the esr value won't change dramatically at frequencies just a bit lower than 100 kHz.
Peak ESR will report something close to the impedance, not the actual ESR, it's not good and accurate enough to compute the real "ESR", as seen on that graph..
But it makes no difference to you, if it's a low esr capacitor, even some value close to the impedance value is enough to help you determine if the capacitor is going bad or not.
A bad capacitor will not show 11 mOhm anyway, it will show 400 mOhm, or 3 ohm.. so you'll know it's bad.
As for why only at 100kHz... that's simple. It's the one most often found in datasheets as Impedance, which like I said, with low esr capacitors the ESR tends to be close to Impedance at 100kHz.
LCR meters like that Mastech have a pair of specialized custom chip with lots of circuitry to handle those frequencies, made only to do LCR measurements. The frequencies also matter for measuring inductors, and capacitors are measured in a different way compared to how Peak ESR measures them.
The Peak ESR has an off the market 2-3$ microcontroller chip in it that has limited memory, limited processing power, limited "disk" space for the application in it, and has limited circuitry besides the microcontroller chip.
There's often not enough room in that microcontroller to add formulas for different frequencies, and adding more chips besides the microcontroller to handle lower frequencies would make the whole thing more expensive and the profit lower. Like I said, Peak ESR is too expensive for what it does, should be a 50-70$ thing, like other meters are.
Also, at 100kHz these cheap esr meters can get away with not using quite a perfect 100kHz sine wave signal (they actually use a square wave) to measure what they measure, while at lower esr frequencies the method they use wouldn't give results close enough to be of value to you. There's often not enough room in the processor used in the meter to implement a better method to measure.
And have a look back at the graph above. As Peak ESR and other cheap meters actually measure something close to impedance and not esr, at lower frequencies this value reported by the esr meter will be higher and you may incorrectly assume the capacitor is bad due to high value reported by the meter. All the time, the capacitor's ESR is flat, the peak esr and other cheap meters just can't measure it though.
Regular caps don't have esr listed too, can't they be checked?
Capacitors that don't have impedance at 100 kHz listed are not designed to be used in circuits where low esr is important.
These capacitors will actually be unable to hold much capacitance at high frequencies... if you watch the last Mastech review you probably saw that a 10uF capacitor at 100kHz only reported 2.8 uF.
They behave like inductors at high frequencies. So you have to use your brains and check the board where the capacitor is to determine if the circuit requires low esr capacitor or not.
Well, do some work yourself.
The first Mastech review I linked to shows the meter opened up. The Peak ESR review I linked to shows the meter opened up with high res pictures in the forum. Look at them and compare.
I'm sorry perhaps my english is really bad, I wanted to say that with mastech you first need to switch the meter to 100kHz measure esr then, switch to say 100Hz or 120Hz and measure capacitance. It's a 2 step process more accurate but it would be nice if it could be selected like semi auto mode.
The Peak ESR is simply designed for the same purpose of other cheap meters, to tell you if a capacitor is good or bad, so 100kHz is the only one offered because that's what you usually see in datasheets so it makes it easy to check if a capacitor is going bad or not.
I checked the bad cap thread
https://www.badcaps.net/forum/showthread.php?t=388
Downloaded some datasheets, and what do you know. ESR is listed only for low ESR caps, some manufacturers list their ESR at 1khz or anything else not 100khz.
https://www.badcaps.net/forum/showthread.php?t=388
Downloaded some datasheets, and what do you know. ESR is listed only for low ESR caps, some manufacturers list their ESR at 1khz or anything else not 100khz.
Datasheets tell you the Impedance value at a certain frequency, which is a different thing. Luckily, at high frequencies, if a capacitor is good, the ESR tends to have a value very close to the impedance value.
esr = dissipation factor * Xc (capacitor reactance)
impedance² = esr² + (Xl - Xc )²
So it seems in those cases peak could not be used as the value can't be compared to actual datasheet.
You can't switch the frequency on peak even to a lower one like 100Hz, or 1khz, 10kHz I wonder why?
Sure if you have mastech then it is possible to test at any frequency they specify in datasheet. But some do not specify it at all. What to do then?
You can't switch the frequency on peak even to a lower one like 100Hz, or 1khz, 10kHz I wonder why?
Sure if you have mastech then it is possible to test at any frequency they specify in datasheet. But some do not specify it at all. What to do then?
You're supposed to use your brains and your actual past experience to determine if this esr value is suitable for the that capacitor based on where the capacitor is used in a circuit.
You're supposed to know that the circuit near the capacitor is a switching power supply using high frequencies, or it's just for filtering some low frequency signal, and so on.
Once you have some electronics experience you start to have an idea of what a circuit does based on how chips are arranged, what components are around a certain capacitor and so on.
Most circuits don't even work at 100kHz.. most cheaper chips used in power supplies run at around 52-75 kHz .. usually everything runs over around 40 kHz because lower than that, humans and animals can actually start to hear this frequency from the oscillations in the transformers.
So the point is, even if you measure a capacitor's esr at 100 kHz, in real world, when that power supply runs, the capacitor may only be subjected to 50 kHz or 75 kHz ...
Luckily, if it's a good low esr capacitor, the esr will either be somewhat flat down to about 1 kHz or slowly and linearly increase as the frequency decreases so if the ESR is low at 100kHz, it won't increase by much at 50 kHz or 10 kHz
Here's for example a graph from the link I posted a few posts above. If you have read that you would have understood this by now:
the green line is the Impedance...
the yellow line is the ESR...
That line where the A is below is 100kHz, you can see the capacitor is 11 mOhm at 100 kHz.
You can also see the impedance is actually going down a bit as frequency decreases, then stays steady and only goes back up a bit at what probably is 10 kHz point.
Meanwhile, the ESR is actually lower and flatter throughout these high frequencies, only going up as frequency goes below 1 kHz.
But you can see at 100kHz, the ESR and Impedance value are very close, just a few mOhm difference between them, that's why again it's easy enough to put an equal sign between the impedance at 100kHz value reported in datasheets and the ESR. For our not 100% precise, not 100% accurate needs, those values are close enough.
So if you know it's a low esr capacitor from a good manufacturer, you can assume the esr value won't change dramatically at frequencies just a bit lower than 100 kHz.
Peak ESR will report something close to the impedance, not the actual ESR, it's not good and accurate enough to compute the real "ESR", as seen on that graph..
But it makes no difference to you, if it's a low esr capacitor, even some value close to the impedance value is enough to help you determine if the capacitor is going bad or not.
A bad capacitor will not show 11 mOhm anyway, it will show 400 mOhm, or 3 ohm.. so you'll know it's bad.
As for why only at 100kHz... that's simple. It's the one most often found in datasheets as Impedance, which like I said, with low esr capacitors the ESR tends to be close to Impedance at 100kHz.
LCR meters like that Mastech have a pair of specialized custom chip with lots of circuitry to handle those frequencies, made only to do LCR measurements. The frequencies also matter for measuring inductors, and capacitors are measured in a different way compared to how Peak ESR measures them.
The Peak ESR has an off the market 2-3$ microcontroller chip in it that has limited memory, limited processing power, limited "disk" space for the application in it, and has limited circuitry besides the microcontroller chip.
There's often not enough room in that microcontroller to add formulas for different frequencies, and adding more chips besides the microcontroller to handle lower frequencies would make the whole thing more expensive and the profit lower. Like I said, Peak ESR is too expensive for what it does, should be a 50-70$ thing, like other meters are.
Also, at 100kHz these cheap esr meters can get away with not using quite a perfect 100kHz sine wave signal (they actually use a square wave) to measure what they measure, while at lower esr frequencies the method they use wouldn't give results close enough to be of value to you. There's often not enough room in the processor used in the meter to implement a better method to measure.
And have a look back at the graph above. As Peak ESR and other cheap meters actually measure something close to impedance and not esr, at lower frequencies this value reported by the esr meter will be higher and you may incorrectly assume the capacitor is bad due to high value reported by the meter. All the time, the capacitor's ESR is flat, the peak esr and other cheap meters just can't measure it though.
Regular caps don't have esr listed too, can't they be checked?
These capacitors will actually be unable to hold much capacitance at high frequencies... if you watch the last Mastech review you probably saw that a 10uF capacitor at 100kHz only reported 2.8 uF.
They behave like inductors at high frequencies. So you have to use your brains and check the board where the capacitor is to determine if the circuit requires low esr capacitor or not.
Comment