There are two visible signs indicating an electrolytic capacitor is failing. These are bulging of the capacitor itself and leakage of the electrolyte. Since, this forum is frequented by people who work as repair technicians, it would be interesting to read about their experiences and whether there are other less obvious signs of electrolytic death.
Hi, I work repairing electronic industrial equipment. Another sign of 'lityc near the end of its live is the "unwear" of the external plastic coverage. When the heat shrink plastic with the mark and characteristics is made shorter and you can see laterally the aluminium can, the cap is usually faulty. Another case is when the plastic and/or the lettering has changed its color.
Last edited:
Pass DIY Addict
Joined 2000
Paid Member
I'm not a repair tech, just a guy who's repaired lots of consumer electronics around the house over the past few decades. The most frequent problem I find with consumer electronics is failed electrolytic caps in the power supply that causes fuses to open. Most of the caps I've had to replace due to failure EXHIBITED NO OUTWARD SIGN OF TROUBLE. They just short internally and sit there looking innocent. Do not rely on visual inspection to sort the good from the bad.
Sure, a bulging cap is an obvious sign of trouble. So is leaked electrolytic fluid (use a flashlight on the PCB to see). But absence of these physical symptoms is NOT evidence that the cap is good.
When a device is 15-20 years old and starts misbehaving, I go after the caps. No point in trying to measure/diagnose each specific cap - this is a giant pain in the neck and is often inconclusive. I know some will criticize me here, but when an old piece of gear starts to malfunction, I do a wholesale cap replacement. This has fixed ALL of my dead pieces of equipment.
Sure, a bulging cap is an obvious sign of trouble. So is leaked electrolytic fluid (use a flashlight on the PCB to see). But absence of these physical symptoms is NOT evidence that the cap is good.
When a device is 15-20 years old and starts misbehaving, I go after the caps. No point in trying to measure/diagnose each specific cap - this is a giant pain in the neck and is often inconclusive. I know some will criticize me here, but when an old piece of gear starts to malfunction, I do a wholesale cap replacement. This has fixed ALL of my dead pieces of equipment.
I once found a reference somewhere where it was said that electrolytic cap aging can be best identified by increased dielectric absorption (DA) which starts to rise way before ESR rises and capacitance decreases, in the process of the electrolyte drying up over time.
The explaination was that when the electrolyte starts to dry up in some section of the cap this part of the cap is a high impedance path and couples the remaining intact sections, the cap now becoming a parallel combination of series RC's...and that happens to be the model circuit for DA as well. Sounds reasonable to me.
The hard part is to measure the DA (at least one pin must be lifted) and to have a nominal DA value to compare too, using a fresh capacitor (same or comparable type) and the same test equipment/procedure.
The explaination was that when the electrolyte starts to dry up in some section of the cap this part of the cap is a high impedance path and couples the remaining intact sections, the cap now becoming a parallel combination of series RC's...and that happens to be the model circuit for DA as well. Sounds reasonable to me.
The hard part is to measure the DA (at least one pin must be lifted) and to have a nominal DA value to compare too, using a fresh capacitor (same or comparable type) and the same test equipment/procedure.
Last edited:
What is easy to measure, is output power of an amp. Which is correlated with output voltage, which is easily measured with an analog VOM with an AC scale of 20 v 50 v or 100 v. Measure speaker resistance, multiply by 1.2 to get impedance. Power = (v^2)/z If the name of the amp is something like "S100" and it is putting out 2 W, you're pretty sure the amp has a rail cap problem. Then change all the other caps without measuring as Eric said. You can wait til the little ones go one by one, resulting in you probably buying a new amp because the old one is broken all the time.
On televisions, squeezing in of the horizontal edges is a good sign the hori cap is going out. Unplug before it blows the transistors or something else hard to diagnose. No, the industry numbers are not on the transistor, just a house number.
On HDTV tuners, diagonal lines in the video is a bad sign. Also inability of the tuner to change channels when first turned on or other digital idiocy.
Switcher power supplies like PC's etc usually blow something expensive before they exhibit performance limitations.
I've bought an Atlas ESR meter, but all it does is confirm that what you just took out is bad. In circuit a lot of times it won't measure. Once you've lifted one leg, the calender says caps are > 20 years old, may as well lift the other leg and change it. 5 years on cheap consumer gear, 1 year on really modern consumer products.
On televisions, squeezing in of the horizontal edges is a good sign the hori cap is going out. Unplug before it blows the transistors or something else hard to diagnose. No, the industry numbers are not on the transistor, just a house number.
On HDTV tuners, diagonal lines in the video is a bad sign. Also inability of the tuner to change channels when first turned on or other digital idiocy.
Switcher power supplies like PC's etc usually blow something expensive before they exhibit performance limitations.
I've bought an Atlas ESR meter, but all it does is confirm that what you just took out is bad. In circuit a lot of times it won't measure. Once you've lifted one leg, the calender says caps are > 20 years old, may as well lift the other leg and change it. 5 years on cheap consumer gear, 1 year on really modern consumer products.
Last edited:
"5 years on cheap consumer gear, 1 year on really modern consumer products."
Years ago I bought a digital to analog TV converter that went bad in a little over a year. I opened it and the power supply didn't just have bulging electrolytic capacitors, some of the capacitor seals (rubber stoppers) had been forced out of the aluminum can. Those defective capacitors were really easy to find. I replaced all the capacitors with higher voltage ones and that fixed it.
Years ago I bought a digital to analog TV converter that went bad in a little over a year. I opened it and the power supply didn't just have bulging electrolytic capacitors, some of the capacitor seals (rubber stoppers) had been forced out of the aluminum can. Those defective capacitors were really easy to find. I replaced all the capacitors with higher voltage ones and that fixed it.
5 years ago I bought a HD to analog TV converter from K-mart. It lasted 4 years. I bought another for my summer camp from ****-mart. It had diagonal lines in it when first turned on the day I powered it up. It lasted a year. K-mart has closed all stores in my metroplex, ****-mart covers the earth.
2 years ago I bought a Samsung 32" TV from ***'s Club. 1 year and 50 weeks later it quit working. Always a bargain at ****-Mart.
2 years ago I bought a Samsung 32" TV from ***'s Club. 1 year and 50 weeks later it quit working. Always a bargain at ****-Mart.
As a sesoned OLD repair tech with 45+ years in the business, I can say that a lot of the previous posts sums up the issues with bad capacitors.
The internet-driven "recapping" craze is mostly overblown chatter that I find amusing at best.
Because in most cases, it's not needed - someone, perhaps some "guru" of the internet, recommended this procedure... and now it's become some biblical mandate.
"Follow the leader, do as he says" BS!
In occasional instances, a slight improvement in performance can be had, but manufacturers usually knew what they were doing - back then.
Sure, capacitors have a certain "lifespan", but the "craze" erupted in part to the early 2000's "bad capacitor syndrome", as well as Inianajo's mention (above post #10) of premature failure of modern consumer equipment.
This is because of the quality levels of components now-a-days.
A lot of parts are just not made with the quality that once was common practice.
I've had plenty of products, thousands, of 20 to 40 years of age, with hardly any bad caps, perhaps an occasional weak one, and few with visible leaks.
It depends on the manufacturer, batch, operating conditions.
And trust me, I've been in a LOT of products to know.
But I do know that the "newer" the product is, the more likely it will fail.
Blame it on the manufacturers - idiots seeking greed, instead of a quality reputation.
The internet-driven "recapping" craze is mostly overblown chatter that I find amusing at best.
Because in most cases, it's not needed - someone, perhaps some "guru" of the internet, recommended this procedure... and now it's become some biblical mandate.
"Follow the leader, do as he says" BS!
In occasional instances, a slight improvement in performance can be had, but manufacturers usually knew what they were doing - back then.
Sure, capacitors have a certain "lifespan", but the "craze" erupted in part to the early 2000's "bad capacitor syndrome", as well as Inianajo's mention (above post #10) of premature failure of modern consumer equipment.
This is because of the quality levels of components now-a-days.
A lot of parts are just not made with the quality that once was common practice.
I've had plenty of products, thousands, of 20 to 40 years of age, with hardly any bad caps, perhaps an occasional weak one, and few with visible leaks.
It depends on the manufacturer, batch, operating conditions.
And trust me, I've been in a LOT of products to know.
But I do know that the "newer" the product is, the more likely it will fail.
Blame it on the manufacturers - idiots seeking greed, instead of a quality reputation.
I had posted here a newbie query of measurement of old amplifier performance.
https://www.diyaudio.com/forums/solid-state/343349-40-stock-optonica-receiver-worry.html#post5930088
member zjjwwaa has answered some of the question. On the net I found article about old amplifiers measured and it showed increased distortion. link here
Two vintage Technics amps measured: SU-V900 and SU-V90D | Audio Science Review (ASR) Forum
Has anyone measured basic specs of an old working amplifier 'as is' and possibly after recapping etc.
thanks and regards.
https://www.diyaudio.com/forums/solid-state/343349-40-stock-optonica-receiver-worry.html#post5930088
member zjjwwaa has answered some of the question. On the net I found article about old amplifiers measured and it showed increased distortion. link here
Two vintage Technics amps measured: SU-V900 and SU-V90D | Audio Science Review (ASR) Forum
Has anyone measured basic specs of an old working amplifier 'as is' and possibly after recapping etc.
thanks and regards.
In many cases a failing electrolytic will give no external sign of this. It will simply reduce in value and gain some ESR, and may or may not also increase leakage current. For a good electrolytic in a benign environment this can take a number of decades. For a bad electrolytic or a good one being misused it can take a few years. Routine replacement of capacitors as a 'repair' or 'upgrade' technique is silly, unless you know you have a badly designed unit or one with bad caps.
Yeah, the MacIntosh rep measured my dynaco ST70 in 1971 at 7 W/ch and 10% HD. He said I probably needed new e-caps and a rectifier. The next year after re-e-capping he measured it at 30 w/ch and 2% HD. He said I probably needed new output tubes. Sounded better after I installed them. I didn't do the negative cathode supply e-cap at first, 2 years later the cathode current went out of control. Should have done it the first time.
The Allen organ S100's went silent in a church service 2017. The next Monday I measured it at 2 W/ch (2.8 vac on a 4 ohm speaker). New rail caps pushed them up to 25 W/ch. All new electrolytics pushed them to 100 w/ch. Measured the rail caps later with the Atlas ESR meter, the 14000 were .8 to .9 ohms ESR. Pro techs on organforum tell everybody Allens never need new e-caps. Wong! After 30 years they do.
I have to argue with pros like wiseoldtech. I view this attitude as minimizing the service charge while maximizing the number of service calls one gets off one customer. I bought a Hammond H110 organ, the dates on the 4 cathode bypass caps in the power amp indicated they had done 4 service calls between 2 and 5 years after the organ was built. In another H110 the motor run cap had shorted off in a messy explosion, the tech didn't do the mains caps which left the 30 w bass amp putting out about 5 W and the 12 W treble channels putting out 2. The string bass and percussion (attack) functions weren't working at all. The harp sustain had all sorts of random sustain times except one key that cyphered. That organ will rattle the chandelier now, and everything works.
In 2011 I re-ecapped a fuzzy insensitive 1979 FM radio that sounded great when my dad bought it. Without realignment it again sounded great and could tune the 100 W high school station from the next county, which no other radio would pick up. But I couldn't replace one e-cap under the dial cord. Last winter that one started making the tuning wander as the room temperature changed. I've changed it but can't get the dial cord to tune more than 1/6 the dial.
I took apart Peavey PV-1.3k and PV-4c. These were built in the 90's The mains caps, while not exploded, were about 4x correct ESR. The little ones are randomly correct or 2x high. My CS800s blew a fuse in the switcher supply because of a little shorted e-cap. The rail caps were 3x correct ESR, haven't changed them yet since they are specially "low" ESR. Since I'm not performing with an amp, just a wood piano.
An Allen organ switcher supply blew a NTCR they were using as a fuse. New e-caps, new NTCR, it was putting out rated dc current on a resistor. Allen wanted $1100 for a new power supply.
So the origin of this craze for me goes back to the MacIntosh salesman. Do them all, use long service life, leave it alone for 30 more years. The ST70 needed 4 sets of e-caps using **** caps from the TV repair shop available before debit cards. Life is too short. Now I can get the 10000 hour service life ones from Newark! Be done with it.
Last edited:
I have worked as a technologist for a large clinical/environmental lab for 36 years. I have replaced hundreds of E caps in all matter of lab equipment including a lot of tube based equipment in the early days. The problem caps were visibly defective maybe 10% of the time.
The last few years I have checked out and repaired a lot of 70s Kenwood, Pioneer and Sansui integrated amps etc.
I only do a total recap when it is asked for. A lot of time is required to replace 30+ small E caps in a typical integrated amp. Many folks don’t want to spend that much. I rebuilt a KA3500 amp to give to a friend and did a total recap. When I measured the amp afterwards it did meet all the factory specs.
The last few years I have checked out and repaired a lot of 70s Kenwood, Pioneer and Sansui integrated amps etc.
I only do a total recap when it is asked for. A lot of time is required to replace 30+ small E caps in a typical integrated amp. Many folks don’t want to spend that much. I rebuilt a KA3500 amp to give to a friend and did a total recap. When I measured the amp afterwards it did meet all the factory specs.
This is my experience too. I am still using electrolytics that have been installed in 1986. Reading online about failing electrolytic capacitors, I suspected an electrolytic may have failed, but hadn't.DF96 said:
So, even in stressful conditions, you expect a good electrolytic to last a few years. Since, electrolytics are essentially an electrolysis cell, I make it a point to respect their voltage and temperature ratings. Very powerful current surges can also be damaging. An electrolytic achieves its high capacitance value by modifying the metal surface in contact with the dielectricric. The effect increases the effective area by a large factor. Moreover, a 'wet' electrolytic capacitor is an electrolysis cell in series with a capacitor. So, in the case of very powerful current surges, a large amount of ions migrate to both electrodes causing temperature rise, which can lead to capacitor failure.
This is true. A properly used electrolytic should not repeatedly fail. The same can be said about a power transistor: if it fails, it means, the operating conditions are not according to specifications.
Last edited:
Please do not assume, suggest, or speculate how I performed the service for my customers.
I'm well aware that there are some repair shops out there that do shoddy work, and driven by greed - I've seen some of their work when it's eventually brought to me.
Because in reality, over the decades, I've had maybe a handful of "recalls" on equipment that I serviced for them.
And some of them were even due to customer-related mishaps which I certainly don't have any control over.
Such as incorrect speaker hookups, abuse related to using equipment for purposes not intended, and under warranty "tinkering" of adjustments by people who somehow felt that my work needed more "help" - like messing around with bias settings (I suspect they listened to some "internet guru" that led to failure).
I've always "went by the book" and beyond in my work, and gained quite a following.
Letters of appreciation, word-of-mouth referrals, and sometimes Christmas cards with notes all came back to me over the years.
Past customers have called me a godsend, a stickler, a dying breed, etc.
I still get the Christmas cards with "the stereo/etc is still playing beautifully" from people 10+ years ago.
In fact, I unknowingly gained a "5 Star Rating" on Angie's List and was only aware of it because the website sent me a card to the shop, stating that pleased customers have sent in their positive opinions of my work.
The bottom line is - I performed my servicing like it was something that I owned, and I'm fussy about my own stuff.
Aluminum Electrolytic Capacitors are subject to aging based in Arrhenius Law (roughly twice as fast for each 10C of operating temperature increase), thus failing periodically. When high quality and properly sized, periodically = 20 years of normal use or so. In aged equipment, depending on design choices, aged electrolytics might be stealing performance, SNR, reliability or LF.
Catastrophic failure is often detected by visual inspection. However aging failure requires a capacitance measurement, and sometimes a ESR measurement.
When capacitance starts to be 20% down it's time for replacement. When some caps of a certain model and value are 20% down it's a good idea to replace all of that model and value. When further servicing is not required it's a good idea to replace all caps of any model and value whose capacitance is 10% or more down.
To determine accurately how much % capacitance has been reduced one has to be aware that electrolytics have production tolerances, so comparing caps of the circuit of same value subject to more ripple and other subject to less ripple is the most accurate reference. However, modern electrolytics often measure +/-5% when new, except rare batches that are off from nominal value when new.
Regarding design reliability with electrolytics, there are 2 types of circuits, the ones that shut down, give a warning, fail a test, or perform poorly as the capacitors start to be under 50% capacitance or so, and the ones that just explode and blow a dozen chips or transistors! (China style.)
Catastrophic failure is often detected by visual inspection. However aging failure requires a capacitance measurement, and sometimes a ESR measurement.
When capacitance starts to be 20% down it's time for replacement. When some caps of a certain model and value are 20% down it's a good idea to replace all of that model and value. When further servicing is not required it's a good idea to replace all caps of any model and value whose capacitance is 10% or more down.
To determine accurately how much % capacitance has been reduced one has to be aware that electrolytics have production tolerances, so comparing caps of the circuit of same value subject to more ripple and other subject to less ripple is the most accurate reference. However, modern electrolytics often measure +/-5% when new, except rare batches that are off from nominal value when new.
Regarding design reliability with electrolytics, there are 2 types of circuits, the ones that shut down, give a warning, fail a test, or perform poorly as the capacitors start to be under 50% capacitance or so, and the ones that just explode and blow a dozen chips or transistors! (China style.)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.