Hi,
Does everyone rely on an ESR meter to truely identify a faulty cap or can I use a standard multimeter/multimeter with capacitor test?
I'd like to test electrolytic capacitors in an amp I'm trying to fault find on:
http://www.diyaudio.com/forums/solid-state/200517-amp-repair-help.html (which if anyone is able to further asist with I would really appreciate).
I don't have the luxury of being able to budget for an ESR just yet, and I wondered if the industry relies on these or whether you can just as effectively work without one?
The caps all look to be ok, but they are at least 11 years old. I'm only really relying on the fact that the tops of them look fine, it's hard to tell if the odd one is bulging or not because some are more stumpy in size than others, there is certainly no obvious bulge.
Thanks.
Does everyone rely on an ESR meter to truely identify a faulty cap or can I use a standard multimeter/multimeter with capacitor test?
I'd like to test electrolytic capacitors in an amp I'm trying to fault find on:
http://www.diyaudio.com/forums/solid-state/200517-amp-repair-help.html (which if anyone is able to further asist with I would really appreciate).
I don't have the luxury of being able to budget for an ESR just yet, and I wondered if the industry relies on these or whether you can just as effectively work without one?
The caps all look to be ok, but they are at least 11 years old. I'm only really relying on the fact that the tops of them look fine, it's hard to tell if the odd one is bulging or not because some are more stumpy in size than others, there is certainly no obvious bulge.
Thanks.
11 years is not that old. The caps are probably fine. If you scope the (power supply) caps voltage while the amp is putting out 50% of rated output and the ripple voltage is less than a few volts the caps are probably fine.
In old caps its not the ESR that becomes the problem. They start to leak current, and or have a much lower capacitance than rated. Charge the cap and see how long it takes to discharge (with no load). It should take a long time (5 minuets or so ). Tghis will show you if it leaks. The cap value is a little more difficult but doable. Discharge the cap thru a known resistance and use the time constant to figure out the capacitance.
In old caps its not the ESR that becomes the problem. They start to leak current, and or have a much lower capacitance than rated. Charge the cap and see how long it takes to discharge (with no load). It should take a long time (5 minuets or so ). Tghis will show you if it leaks. The cap value is a little more difficult but doable. Discharge the cap thru a known resistance and use the time constant to figure out the capacitance.
Last edited:
Hi RichIOM
Most of us actually do without any kind of capacitance quality evaluation like ESR testing. The age of good, industry quality electrolytic caps used within their ratings in linear power supplies and general audio application has been suggested here as 20 years from the purchase date. I can agree with that, noting that many cheap caps can fail much sooner due to inferior construction and materials. High temperature operation in some designs will also shorten cap. life considerably. If your amp runs hot and the parts too, consider a 5-10 year replacement cycle for these.
Don't be discouraged here. The simple expedient of using reliable parts and doing systematic replacements by a sensible time limit like even 20 yrs will likely be better insurance and you may never need be concerned again.
Most of us actually do without any kind of capacitance quality evaluation like ESR testing. The age of good, industry quality electrolytic caps used within their ratings in linear power supplies and general audio application has been suggested here as 20 years from the purchase date. I can agree with that, noting that many cheap caps can fail much sooner due to inferior construction and materials. High temperature operation in some designs will also shorten cap. life considerably. If your amp runs hot and the parts too, consider a 5-10 year replacement cycle for these.
Don't be discouraged here. The simple expedient of using reliable parts and doing systematic replacements by a sensible time limit like even 20 yrs will likely be better insurance and you may never need be concerned again.
If the series model capacitance (Cs), which is probably what your meter measures, is good/sufficient, the power supply will work fine. Ripple is dependent almost entirely on Cs and other losses would have to be huge before they mattered at all. Now, if other losses were huge, you might get (believe it or not) a very high reading for Cs, well above the marked value. Low is bad, unexpectedly high is bad. So, bottom line is, for mains operated supplies (not switchers), if the capacitance is about right, and if the thing isn't leaking or bulging, it's probably good. For switchers you really should be measuring loss, but selecting caps for switchers is a whole topic in itself.
If there are caps parked next to hot components then they could have dried out and gone low in capacitance without any noticeable bulging. A couple of years can be long enough for this to happen even with decent quality caps. I've lost count of the number of times I've seen this design faux pas.
Thanks for the input, the reason I'm looking at the caps as the fault is because I can't see any real fault with the amp, other than it's not very loud even when turned up.
It used to be loud and now isn't on the same equipment and speakers, it's my brother in-laws amp and I have the same speakers so I've since hooked it up to my equipment with my speakers and the volume is just the same.
There is one cap that may be suspect purely on the fact it doesn't kick when I drain it using a car lightbulb, where as the same rated cap next to it, gives off a real crack and a spark when I connect the bulb to it.
Both caps are rated at 8200uF.
I've attached a photo so you can see how close they are in proximity to one another which is why I would expect them both to be carrying similar charges?
It used to be loud and now isn't on the same equipment and speakers, it's my brother in-laws amp and I have the same speakers so I've since hooked it up to my equipment with my speakers and the volume is just the same.
There is one cap that may be suspect purely on the fact it doesn't kick when I drain it using a car lightbulb, where as the same rated cap next to it, gives off a real crack and a spark when I connect the bulb to it.
Both caps are rated at 8200uF.
I've attached a photo so you can see how close they are in proximity to one another which is why I would expect them both to be carrying similar charges?
An externally hosted image should be here but it was not working when we last tested it.
I very quickly skimmed your other thread.
My honest feeling is that there is something else making you think its not as loud. You have measured the amp output and confirmed its to spec. That pretty much rules out the 8200uF caps.
I am not familiar with the amp at all but anything like equalisers or custom sound settings can alter the level greatly. I don't really think the caps are at fault. Any source equipment that may have been changed and has a different outut level ?
If you want to prove the caps to yourself then perhaps swap them over.
My honest feeling is that there is something else making you think its not as loud. You have measured the amp output and confirmed its to spec. That pretty much rules out the 8200uF caps.
I am not familiar with the amp at all but anything like equalisers or custom sound settings can alter the level greatly. I don't really think the caps are at fault. Any source equipment that may have been changed and has a different outut level ?
If you want to prove the caps to yourself then perhaps swap them over.
Thanks for your feedback. I know what you mean by something else making me think it's not as loud, as many people can have this feeling when they get used to a system etc.
However, this system isn't mine it's my brother in-laws, who said it wasn't as loud as it used to be and sounded pretty flat until it's wound up.
I would have thought as you if I didn't then take it from him and listen to it myself. I have the exact same speakers, so I could rule that out.
I've heard it on his source equipment first, when he was showing me that it wasn't as loud. The only way I can describe how it's not that loud, is that it only starts to get loud at -25db, the dial starts at -72db, so theres a big gap before it get loud. We were even able to turn it into the +db before it was the sort of level I would expect of -20db/-15db. I would have expected +2db to be deafening, over a 5 speaker surround setup with a sub from this amp, in a small room.
It only starts to sound decent in my spare room (not big enough for a double bed), at -25db, on a pair of bookshelf 8ohm Infinity Alpha 5 bookshelf speakers.
Spec for Speakers here:
Infinity - Home Audio
One other fault I did see at my brother in-laws but haven't further tested on my own setup, as I've only been testing with a pair of speakers, was that not all speakers displayed on the display when it was on a surround setting.
So although all speakers we playing of equal volume when I put my ear up to them, the display wasn't showing all of them as lit, so I don't know if this factors into the fault in some way.
The caps in the photo aren't power supply caps, they're the two largest caps in the centre of the motherboard (I call it this because each of the input banks at the back have their own circuit board, like PC cards).
As the source has been varied and tested on every input to the amp, I think it's safe to rule the source out.
Thanks for all the help, I just really want to make progress with this as it's one of those faults I really want to fix and it's a really good learning experience for me, so I really do appreciate the input.
However, this system isn't mine it's my brother in-laws, who said it wasn't as loud as it used to be and sounded pretty flat until it's wound up.
I would have thought as you if I didn't then take it from him and listen to it myself. I have the exact same speakers, so I could rule that out.
I've heard it on his source equipment first, when he was showing me that it wasn't as loud. The only way I can describe how it's not that loud, is that it only starts to get loud at -25db, the dial starts at -72db, so theres a big gap before it get loud. We were even able to turn it into the +db before it was the sort of level I would expect of -20db/-15db. I would have expected +2db to be deafening, over a 5 speaker surround setup with a sub from this amp, in a small room.
It only starts to sound decent in my spare room (not big enough for a double bed), at -25db, on a pair of bookshelf 8ohm Infinity Alpha 5 bookshelf speakers.
Spec for Speakers here:
Infinity - Home Audio
One other fault I did see at my brother in-laws but haven't further tested on my own setup, as I've only been testing with a pair of speakers, was that not all speakers displayed on the display when it was on a surround setting.
So although all speakers we playing of equal volume when I put my ear up to them, the display wasn't showing all of them as lit, so I don't know if this factors into the fault in some way.
The caps in the photo aren't power supply caps, they're the two largest caps in the centre of the motherboard (I call it this because each of the input banks at the back have their own circuit board, like PC cards).
As the source has been varied and tested on every input to the amp, I think it's safe to rule the source out.
Thanks for all the help, I just really want to make progress with this as it's one of those faults I really want to fix and it's a really good learning experience for me, so I really do appreciate the input.
It is really useful to load the amp with the proper impedance. I use 10 ohm 225 w variable resistors, then I measure the output of the amp on the AC volts scale. 35W into 8 ohm should be about 17 VAC. If the amp is not putting that out, it has a problem.
If the caps are putting AC ripple at full power, then they are probably low capacitance. I've never owned a meter that could measure 8200 uf, or even 100 uf. The capacitance scale on $30 sears meters tops out at 1 or 2, realistically. I mostly measure B+ on tube amps, if it sags at the 4th section but not at the first it is the cap, not the tube. (had to recap my ST70 3 times for low B+, and two rectifier tubes). The same principal applies to solid state amps, only they don't have 4 gang caps typcally and SS rectifiers either work or they don't. SS rectifiers don't sag.
A good Simpson 266 VOM with 100kohm/volt 2VAC and 20 VAC scales helped me debug for 25 years before I ever found a working oscilloscope I could afford. Modern DVMs in the $30 area aren't sensitive to non-power line frequency AC, and have provided me with enough stupid answers I quit trying to use them. At 125 VAC and above at 50 and 60 hz, they are fine. The $119 Fluke meter the factory loaned me had the same problem. Besides all DVM's I could afford average over 2-4 seconds, and signals come and go faster than that.
If you actually use a scope, the voltage you calculate from the P=IV and V=IR equations is 0.7* the peak to peak voltage you see on the scope, roughly.
If the caps are putting AC ripple at full power, then they are probably low capacitance. I've never owned a meter that could measure 8200 uf, or even 100 uf. The capacitance scale on $30 sears meters tops out at 1 or 2, realistically. I mostly measure B+ on tube amps, if it sags at the 4th section but not at the first it is the cap, not the tube. (had to recap my ST70 3 times for low B+, and two rectifier tubes). The same principal applies to solid state amps, only they don't have 4 gang caps typcally and SS rectifiers either work or they don't. SS rectifiers don't sag.
A good Simpson 266 VOM with 100kohm/volt 2VAC and 20 VAC scales helped me debug for 25 years before I ever found a working oscilloscope I could afford. Modern DVMs in the $30 area aren't sensitive to non-power line frequency AC, and have provided me with enough stupid answers I quit trying to use them. At 125 VAC and above at 50 and 60 hz, they are fine. The $119 Fluke meter the factory loaned me had the same problem. Besides all DVM's I could afford average over 2-4 seconds, and signals come and go faster than that.
If you actually use a scope, the voltage you calculate from the P=IV and V=IR equations is 0.7* the peak to peak voltage you see on the scope, roughly.
Last edited:
I really don't know what to suggest to you.
If all channels behave the same then I think you can rule out an "audio" fault as such.
The next step has to be to see if it meets its rated output for its quoted input sensitivity. You have a scope so that's easy but you need to be careful you don't damage anything.
You need to apply the correct input voltage (sine wave at say 400hz) remembering that the RMS figure is what you see on the scope (peak to peak) divided by two and then multiplied by 0.707
I would initially test with no load because most commercial gear has limited heatsinking and is not specified to run for long at full output into say 8 ohms. See if the output meets its spec for the given input. If you then reduce the output to a more manageable figure of say 1 or 10 watts RMS you can apply the specified load impedance and the output should be maintained.
You are quoting the readout from the display (-db figures). This is where it gets tricky. Are you certain the amp has not got a loudness function etc and that this has in the past been on but now is not ?
Is the volume control a mechanical (motorised maybe ?) conventional potentiometer or is the volume control purely electronic and the front control just a "rotary encoder" which tell the microprocesser which way to "electronically" reduce or increase the volume ?
If the later then you next need to check that the output follows an approximate logarithmic law as it is increased. I wouldn't read to much into the display reading as I am sure it is not referenced to anything. The reading will alter with volume position whether a signal is present or not and so in terms of actual level its meaningless.
Now those caps....
Two equal value large caps next to a bridge rectifier. You say they are not PSU caps. What do you think they are ?
If all channels behave the same then I think you can rule out an "audio" fault as such.
The next step has to be to see if it meets its rated output for its quoted input sensitivity. You have a scope so that's easy but you need to be careful you don't damage anything.
You need to apply the correct input voltage (sine wave at say 400hz) remembering that the RMS figure is what you see on the scope (peak to peak) divided by two and then multiplied by 0.707
I would initially test with no load because most commercial gear has limited heatsinking and is not specified to run for long at full output into say 8 ohms. See if the output meets its spec for the given input. If you then reduce the output to a more manageable figure of say 1 or 10 watts RMS you can apply the specified load impedance and the output should be maintained.
You are quoting the readout from the display (-db figures). This is where it gets tricky. Are you certain the amp has not got a loudness function etc and that this has in the past been on but now is not ?
Is the volume control a mechanical (motorised maybe ?) conventional potentiometer or is the volume control purely electronic and the front control just a "rotary encoder" which tell the microprocesser which way to "electronically" reduce or increase the volume ?
If the later then you next need to check that the output follows an approximate logarithmic law as it is increased. I wouldn't read to much into the display reading as I am sure it is not referenced to anything. The reading will alter with volume position whether a signal is present or not and so in terms of actual level its meaningless.
Now those caps....
Two equal value large caps next to a bridge rectifier. You say they are not PSU caps. What do you think they are ?
I've sent you a pm (hope you don't mind), to confirm how best to test the output of the amp.
In answer to your other questions.
There's definately no loudness function on the amp, and I actually performed a reset of all the preset settings, following the reset instructions in the manual, to ensure it was at base settings.
I believe the volume control to be purely electronic, but I'll confirm for sure. I have the full service manual which I managed to source online.
It's not very obvious to a newb what is what however.
The only reason I quote the volume control so much is that in terms of relativity it's having to be turned quite high in order to get any sort of output, so unless there's something wrong with that part of the amp, I would imagine it to be relevant to the issue?
I'm not sure where you're seeing the bridge rectifier in the photo? I'm not experience enough to know anything about them to be honest, but the reason I say they're not PSU caps is because there's a couple of large caps in a totally different section of the amp where the power comes in. These caps are in the centre of the main motherboard.
In answer to your other questions.
There's definately no loudness function on the amp, and I actually performed a reset of all the preset settings, following the reset instructions in the manual, to ensure it was at base settings.
I believe the volume control to be purely electronic, but I'll confirm for sure. I have the full service manual which I managed to source online.
It's not very obvious to a newb what is what however.
The only reason I quote the volume control so much is that in terms of relativity it's having to be turned quite high in order to get any sort of output, so unless there's something wrong with that part of the amp, I would imagine it to be relevant to the issue?
I'm not sure where you're seeing the bridge rectifier in the photo? I'm not experience enough to know anything about them to be honest, but the reason I say they're not PSU caps is because there's a couple of large caps in a totally different section of the amp where the power comes in. These caps are in the centre of the main motherboard.
Hi,
Your testing method of the amp seems OK. The DVM is an unknown quantity in that it will probably only read what equates to an RMS voltage for sine wave voltages and also only read accurately over a limited frequency range although 100hz should be fine on any DVM. I'd take a properly calibrated scope every time though. The reading on the scope and DVM should agree too, once you have converted it from peak to peak to RMS.
The volume control...
if we take the volume as being set on full, then in order to get say 100 watts RMS output the amp will need a certain minimum voltage as an input. This is because the amplifier will have a certain level of gain that is designed into it. A typical voltage gain might be around 30. That means with 1 volt RMS applied to the input, the voltage from the output would be 30 volts RMS if the amp were capable of that. That's on maximum volume (with no attenuation). The volume control can only operate to reduce the output from that maximum figure. In reality 100 watts RMS which equates to approx 28 volts RMS across 8 ohms.
So if the input we are applying were say 5 volts then the amp would need only a little rotation of the volume control for the output to reach 28 volts RMS. It would clip as the output passed the amplifiers maximum, perhaps with the volume control rotated only a third or half way. The amp would sound and appear "loud" to the user in use because it gives the impression of having plently of power for a small rotation of the volume control.
If the input were now only say 0.5 of a volt RMS, then even on maximum volume setting the amp would not be able to deliver its rated output. The gain of the amp (in this example 30) would not be sufficient to raise that 0.5volt enough to fully drive the amp. So with the volume on full the output would be only 15 volts RMS and the amp would seem lacking in power. That 15 volts RMS would be only 28 watts RMS.
This very important input voltage figure is usually quoted in the specifications along the lines of "input sensitivity of 700mv for full output" and will be quoted for a particular input as they may all vary.
So... you need to set the volume on full and increase the input voltage (leaving the volume full) until the amp output reaches its maximum specification. Then measure with the scope what the actual voltage on the input socket is while still connected to the signal source.
That should give you a good idea as to whats what. A CD player typically puts out a maximum of 2 volts RMS. An older bit of kit like a cassette deck could be as low as 0.5 volt RMS.
I would do all the initial testing with no load as I don't think for a minute there is a problem with the amp not maintaining its output into a real load. So if the amp were 50 watts rating into 8 ohms then it should deliver 20 volts RMS. Any load of 8 ohms or higher should not alter that result.
Another test you can do is to run your normal source component (CD Player ?) through the amp and still with no load attached turn up the volume while scoping the output and see if the amp can be driven up to the point of clipping on a music signal.
Finally... could the amp be faulty.
If it were a normal resistive volume control, then I would say not as an amplifier going "low gain" would be an odd fault indeed. The gain of each audio stage, power amp, preamp etc is set by a couple of components in most cases and, for both channels to alter in the same way would be virtually unheard of.
The fact that these AV receivers sometimes have a lot of complex audio processing complicates things and it is possible I guess for a fault to affect levels somewhere but it's very unlikely. It would have to be something like the scaling of the volume attenuator to a control voltage etc or in some process if/where the audio were converted into the digital domain for processing and then converted back with a digital to analogue and for there to be some issue there... all very unlikely I think.
So all you can really do is to see if the amp meets its specification first before any real conclusions can be drawn.
Your testing method of the amp seems OK. The DVM is an unknown quantity in that it will probably only read what equates to an RMS voltage for sine wave voltages and also only read accurately over a limited frequency range although 100hz should be fine on any DVM. I'd take a properly calibrated scope every time though. The reading on the scope and DVM should agree too, once you have converted it from peak to peak to RMS.
The volume control...
if we take the volume as being set on full, then in order to get say 100 watts RMS output the amp will need a certain minimum voltage as an input. This is because the amplifier will have a certain level of gain that is designed into it. A typical voltage gain might be around 30. That means with 1 volt RMS applied to the input, the voltage from the output would be 30 volts RMS if the amp were capable of that. That's on maximum volume (with no attenuation). The volume control can only operate to reduce the output from that maximum figure. In reality 100 watts RMS which equates to approx 28 volts RMS across 8 ohms.
So if the input we are applying were say 5 volts then the amp would need only a little rotation of the volume control for the output to reach 28 volts RMS. It would clip as the output passed the amplifiers maximum, perhaps with the volume control rotated only a third or half way. The amp would sound and appear "loud" to the user in use because it gives the impression of having plently of power for a small rotation of the volume control.
If the input were now only say 0.5 of a volt RMS, then even on maximum volume setting the amp would not be able to deliver its rated output. The gain of the amp (in this example 30) would not be sufficient to raise that 0.5volt enough to fully drive the amp. So with the volume on full the output would be only 15 volts RMS and the amp would seem lacking in power. That 15 volts RMS would be only 28 watts RMS.
This very important input voltage figure is usually quoted in the specifications along the lines of "input sensitivity of 700mv for full output" and will be quoted for a particular input as they may all vary.
So... you need to set the volume on full and increase the input voltage (leaving the volume full) until the amp output reaches its maximum specification. Then measure with the scope what the actual voltage on the input socket is while still connected to the signal source.
That should give you a good idea as to whats what. A CD player typically puts out a maximum of 2 volts RMS. An older bit of kit like a cassette deck could be as low as 0.5 volt RMS.
I would do all the initial testing with no load as I don't think for a minute there is a problem with the amp not maintaining its output into a real load. So if the amp were 50 watts rating into 8 ohms then it should deliver 20 volts RMS. Any load of 8 ohms or higher should not alter that result.
Another test you can do is to run your normal source component (CD Player ?) through the amp and still with no load attached turn up the volume while scoping the output and see if the amp can be driven up to the point of clipping on a music signal.
Finally... could the amp be faulty.
If it were a normal resistive volume control, then I would say not as an amplifier going "low gain" would be an odd fault indeed. The gain of each audio stage, power amp, preamp etc is set by a couple of components in most cases and, for both channels to alter in the same way would be virtually unheard of.
The fact that these AV receivers sometimes have a lot of complex audio processing complicates things and it is possible I guess for a fault to affect levels somewhere but it's very unlikely. It would have to be something like the scaling of the volume attenuator to a control voltage etc or in some process if/where the audio were converted into the digital domain for processing and then converted back with a digital to analogue and for there to be some issue there... all very unlikely I think.
So all you can really do is to see if the amp meets its specification first before any real conclusions can be drawn.
Many thanks for your input guys, particularly Mooly, your time is appreciated in helping me learn.
I've taken another photo of the unit from above so you can see the entire unit.
Top left of the photo are the large caps I refer to as being PSU caps, as this is where the power comes into the unit and is an independant circuit, bottom left is a huge transformer. On the right at the other two large caps in the middle of the motherboard as photographed earlier in the thread. The upright boards slotted in the motherboard (albeit soldered directly to the board so not easily removable), are all input boards for auxiliary connections.
It's a surround receiver yes, and yes all menu settings have been reset back to factory defaults following the instructions in the manual to ensure I have a base setting to work from.
I agree it's all very odd, even my Dad said as much saying that for all channels to be affected at once is indicative of something at the pre stage.
I've taken another photo of the unit from above so you can see the entire unit.
An externally hosted image should be here but it was not working when we last tested it.
Top left of the photo are the large caps I refer to as being PSU caps, as this is where the power comes into the unit and is an independant circuit, bottom left is a huge transformer. On the right at the other two large caps in the middle of the motherboard as photographed earlier in the thread. The upright boards slotted in the motherboard (albeit soldered directly to the board so not easily removable), are all input boards for auxiliary connections.
It's a surround receiver yes, and yes all menu settings have been reset back to factory defaults following the instructions in the manual to ensure I have a base setting to work from.
I agree it's all very odd, even my Dad said as much saying that for all channels to be affected at once is indicative of something at the pre stage.
That was good timing lol and you're welcome...
Caps... I think we are seeing multiple power supplies all fed off the large transformer. Looking at the circuit diagram and seeing what sections they feed will reveal all as far as they are concerned but I don't think they are suspect in any way.
Don't know what to suggest really, other than the tests outlined.
Caps... I think we are seeing multiple power supplies all fed off the large transformer. Looking at the circuit diagram and seeing what sections they feed will reveal all as far as they are concerned but I don't think they are suspect in any way.
Don't know what to suggest really, other than the tests outlined.
No that's great, just wanted to feed your curiosity .
I'll crack on with testing. Won't be able to do it tonight as my multimeter only goes down to 200V ac which is useful.... So I'll either borrow one from my Dads work or wait till I've ordered a new one as I need one that's better than the one I have for jobs like this.
.I'll crack on with testing. Won't be able to do it tonight as my multimeter only goes down to 200V ac which is useful.... So I'll either borrow one from my Dads work or wait till I've ordered a new one as I need one that's better than the one I have for jobs like this.
I usually use a scope to check big electrolytics, I just monitor the supply voltage and look for serious ripple or for the supply collapsing completely. I start off with the amp just on then check as the volume goes up.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.