Importance of replacing the electrolytic capacitors in any vintage gears

[nanook]

I use the "Rosin" solder wick from Chemtronics.

Soder-Wick Rosin Flux Desoldering Wick

Soder-Wick brand solder wick is designed for today’s heat sensitive electronic components using lighter mass, pure copper braid construction that allows for better thermal conductivity, even at low temperatures.

www.chemtronics.com

 

[nanook]

I have only messed around with through hole and find the solder hard to melt to remove a component without damaging the board. I think the "low temp solder" and some flux will go a long way. I don't use no lead solder but most devices made during the last 20 years or so seem to.

With lead-free solder you need ca. 360 °C and a relatively short and broad tip. I use a short flat tip either 2.4mm or 3.2mm broad. If the solder joint is in a plane (with or without thermal relief) then you might want to use an even broader tip.
It's important that the tip is properly covered with solder - adding flux will help. In case you use the Rosin solder wick from Chemtronics, the amount of flux in the solder wick may suffice.

What soldering iron do you use?
Temperature regulated?
Wattage?

 

[JeffS7444]

I don't recall anyone mentioning the "capacitor plague" era, which began in 1980s or 90s, and seemingly was being felt well into the 21st century. Some of the electrolytics used by Apple and others in the 1990s were terrible, and I've needed replace all of the SMD electrolytics from Mac Classic II and Portable motherboards, and wash away any electrolyte residue, lest it cause irreparable damage. I chose solid tantalum chip capacitors as replacements, and the last time I checked, my Classic II was still working just fine.

Even well into the aughties, I read a number of reports of blown power supply capacitors in the Behringer DEQ2496. So while mine was still new, I recapped the SMPS, and for good measure, the CPU board too. The only electrolytics that I didn't touch were smaller value ones used for power supply decoupling, figuring those were minimally-stressed parts.

Hafler DH101: As received, mine looked terrible, and the thrift store must have thought so too, because they had it sitting on the floor, rather than on the shelves. Not going to attempt to power it up until I've recapped it, and maybe I'll upgrade the resistors to metal film while I'm at it.

Sony STC-7000: Quite the find for me, and in lovely condition too. Very likely it will work as-is, but because it's kind of special to me, and it's modular design makes servicing really easy, I may as well do so. It also incorporates some Sony proprietary ICs, and I don't wish to risk those.

Where space allows, I like to replace smaller value electrolytics, including bipolar types as well as tantalum, with film capacitors. I used to be a stickler for audiophile-approved dielectrics like polypropylene, but those can be bulky, so I've relaxed my standards to include pretty much any type of film capacitor which is sufficiently compact and reasonably priced, on the assumption that even a mediocre film capacitor will outperform an outstanding electrolytic.

Although I do sometimes fixate on electrolytic capacitor specs including MTBF, I was surprised to find in build notes for one of his designs, Douglas Self was using house-brand parts from a popular UK supplier like Farnell: I figure there's lesson to be learned there.

 

[JeffS7444]

A few years ago, I finally upgraded my 1980s Weller soldering station for a Hakko FX951, which is a newer sort of design which integrates temperature sensor, heating element, and tip into a single unit. Because while I could fit a smaller-sized tip onto my old iron, it couldn't necessarily deliver sufficient heat to a tiny area, whereas the newer iron is designed to address that. Some newer, less costly designs like Pinecil have a similar arrangement as the FX951.

Generally speaking, if I'm working on RoHS-era circuits or a new build, I'll use RoHS solder, and leaded solder for older stuff. Either way, I like to use silver-bearing solder, because there's almost always some device or switch with silver-plated leads.

 

[levimax]

Where space allows, I like to replace smaller value electrolytics, including bipolar types as well as tantalum, with film capacitors.

This will often work but sometimes the lower ESR of the film caps can cause problems. Some circuits are very tolerant of different components and some are not and it is hard to know in advance and even harder to troubleshoot if something goes wrong. I like to use parts that are as similar as possible to the original parts whenever possible.

 

[rdenney]

I have recapped a vintage Holman preamp, and all the removed caps measured well and there was no improvement in the (excellent) performance. But I had the caps and it gave me a chance to really clean it up and study it.

I have also recapped a late-80's Lin Axis turntable, and it needed it. Half the caps were out of spec, one was open, and most had high ESR. The speed control was not functioning properly and replacing the caps solved the issue. (It was broke and I fixed it.) The power supply in that turntable is a class A amplifier creating an AC signal of variable voltage and frequency for controlling the synchronous motor. It lacked a proper power switch and also proper thermal control for a device that could not be powered off.

I have a buzz in a the line-stage amp in the preamp that the Holman replaced in my main system. I'll bet that is a cap, but I have not yet explored it.

Aaaand, I replaced some caps in some test equipment, but I don't think those devices really needed it. But the class X and class Y RIFA filter caps in my HP 5334b frequency counter were an explosion waiting to happen--the epoxy matrix had cracked, which usually lets moisture in eventually resulting in a short.

Replacing the caps in my old Advent speakers had no effect, as expected.

The 400V-rated paper-in-oil snubber caps in my old Teac tape deck measured fine until... They sound like cherry bombs when they blow apart. Ask me how I know. I replaced all of those with 630V-rated film caps.

Rick "unwilling to generalize one way or the other" Denney

 

[nanook]

Edward Muntz would disagree, more like 90% serve purposes not directly observable to most.

Just read it up. Interesting approach...

I agree, many parts are "angst-parts", i.e. you add them to know you're on the safe side, especially if you have limited impact on the PCB artwork or don't want to spend time on it.

On the other hand removing parts may make a circuit marginally stable and with component variation during production or temperature the circuit may indeed oscillate. Missing bypassing at medium/high speed Opamps is a typical example. Making sure the circuit is safely on the stable side is quite some effort, so you rather spend a couple of cents and you know it's stable.
What did I tell my colleagues again and again (we were developing high performance measurement-amplifiers): "You can only see whether a circuit will oscillate when it oscillates."
Of course there are several techniques to assess if the circuit is close to oscillating, but the simple observation that a circuit does not oscillate should not lull you into a false sense of security that it is stable.

 

[nanook]

But the class X and class Y RIFA filter caps in my HP 5334b frequency counter were an explosion waiting to happen--the epoxy matrix had cracked, which usually lets moisture in eventually resulting in a short.

Good point!
Since these are "self-healing" they will not necessarily blow the fuse. I had to ventilate my hobby lab for 2 weeks - the capacitor of an old Philips oscilloscope had peacefully burned down almost completely....
Since then I change these "sweets" first, at least if the device has been stored in damp conditions

 

[rdenney]

Good point!
Since these are "self-healing" they will not necessarily blow the fuse. I had to ventilate my hobby lab for 2 weeks - the capacitor of an old Philips oscilloscope had peacefully burned down almost completely....
Since then I change these "sweets" first, at least if the device has been stored in damp conditions

I just worked on an old educational lab AC/DC power supply. I'm afraid my $25 was wasted, if getting a filtered AC power supply was the objective rather than just the fun of playing with the thing. It uses a variac transformer to variably adjust the AC voltage, and then rectifies and filters that output to create the DC output. The AC is unfiltered and reflects the same harmonic distortion present from the power company (which is significant). Cleaning that up would be difficult.

But the rectifier is made up of single-plate selenium diodes, an archaic technology that is actually pretty useful in this application. But the stink of an exploded capacitor is not on the same planet of nasty-smelling as a blown selenium diode, not to mention that the selenium smoke is poisonous at some level.

I spent some time imagining how I would replace that rectifier with silicon without undermining the good features (including the ability to rectify and move current at very low voltages and currents, even if inefficient), and even more time trying to figure out how I would build a steep 70Hz low-pass filter that could handle 75 watts of power at very low impedance. Schottky diodes would handle the low outputs but at rated outputs the junction temps are too high for any examples I could find at Digikey. Silicon diodes stop rectifying and moving current below about half a volt. A low-pass filter to clean up the AC would likely ring and oscillate without a lot more design work I'm not up for. Not worth the trouble in both cases. But for now the rectifier works, though the ripple is much greater (0.5 V RMS at 15V/5A output into my calibrated load) than theory would suggest. Interestingly, the ripple voltage was a much cleaner 120-Hz sine wave than the 60-Hz AC unrectified power from the variac. Back to topic: the filter is a T network with two inductors in series and a shunt capacitor (2500 mfd) in between. I replaced that cap with no effect on the ripple. The old cap was a Mallory aluminum electrolytic made probably 65 years ago, and it still measures reasonably well. I might replace it with a much larger cap to see if I can get the thing to release its magic poisonous smoke, but for now it's on the shelf.

Rick "$25 of fun, but a working low-voltage variac" Denney

 

[nanook]

Schottky diodes would handle the low outputs but at rated outputs the junction temps are too high for any examples I could find at Digikey.

@rdenney: There should be Schottky diodes around that can easily handle 5A. All of the SMPS use Schottky diodes on the secondary side. Well, they usually have a small heat-sink attached, but that's probably just more economic. 5A and ca. 0.5V drop (it's actually more complicated, but take it as a zero - order estimate) mean roughly 2.5W. Sounds feasible.

Afaik selenium diodes have a fairly high series resistance. See e.g. https://www.radiomuseum.org/forum/selen_gleichrichter_1.html

You might need to add a series resistor between rectifier and caps (or between transformer and rectifier) to compensate for the low differential resistance of the schottky diodes.

Did you already blow off the rectifier or can you still do measurements on a single diode and thus choose a resistance value that takes the schottky diode close to her outdated and bulky successor?

 

[rdenney]

@rdenney: There should be Schottky diodes around that can easily handle 5A. All of the SMPS use Schottky diodes on the secondary side. Well, they usually have a small heat-sink attached, but that's probably just more economic. 5A and ca. 0.5V drop (it's actually more complicated, but take it as a zero - order estimate) mean roughly 2.5W. Sounds feasible.

Afaik selenium diodes have a fairly high series resistance. See e.g. https://www.radiomuseum.org/forum/selen_gleichrichter_1.html

You might need to add a series resistor between rectifier and caps (or between transformer and rectifier) to compensate for the low differential resistance of the schottky diodes.

Did you already blow off the rectifier or can you still do measurements on a single diode and thus choose a resistance value that takes the schottky diode close to her outdated and bulky successor?

I haven’t unsoldered the bridge to measure an individual diode, but an adding a series resistor would be easy. I had expected that. Rated power output (after filters) is 75W, but I wouldn’t want any less than twice that in a Schottky bridge. I’ll get interested an again in a couple of weeks—mind is on travel right now.

Rick “also looked into a synchronous rectifier with MOSFETs” Denney

 

[DanielT]

A practical example of doing something or not, that's the question. What would you have done?

I wonder what else I can do with my vintage Luxor 7082A amplifier. An amplifier that is now 46 years old. Half a century. Sounds crazy if you put it that way.

The thing is that it works flawlessly as far as I can hear. No strange popping sounds, no noise in the speakers. I've checked the DC offset and it's ok. I've cleaned the potentiometers with electronics spray because there was a little scratch in the volume control. I've had it for three years, playing with it sometimes. Switching between a few different amplifiers/receivers. In other words, since I don't use it daily, I don't wear it down that much.

It looks visually good, I say with my layman's eyes, no capacitors look like they're starting to leak or anything. The big capacitors in the power supply part: 50 V, 10000uF, 85 C, brand Nippon.

Check the pictures, see the one where I put a small lighter. Damn how close the small capacitors are. I hadn't tried to remove them and put new capacitors in with my rudimentary soldering skills (first remove the old ones, de soldering, ugh..). It would be a bit tricky (for me). God knows what could go wrong:


Could this be something to check? According to the service manual, which can be found on Hifi engine :
Zero current check
Remove fuse F301A or F301b (2.5 AT), depending on which channel is to be measured. Connect a multimeter between the contacts on the fuse holder. Turn on the multimeter and check (after a minute) that the meter shows 20-30 mA.

And yes ,I am aware of the danger of getting an electric shock if I mess it up.

Luxor 7082 Stereo Integrated Amplifier Manual | HiFi Engine

www.hifiengine.com

Edit:
Just to clarify, I personally will not initiate any recap/exchange process, EVEN if they had to. My more knowledgeable friend would do that in that case.
(then I would help him with something else to repay the favor)

 

[levimax]

A practical example of doing something or not, that's the question. What would you have done?

I wonder what else I can do with my vintage Luxor 7082A amplifier. An amplifier that is now 46 years old. Half a century. Sounds crazy if you put it that way.

The thing is that it works flawlessly as far as I can hear. No strange popping sounds, no noise in the speakers. I've checked the DC offset and it's ok. I've cleaned the potentiometers with electronics spray because there was a little scratch in the volume control. I've had it for three years, playing with it sometimes. Switching between a few different amplifiers/receivers. In other words, since I don't use it daily, I don't wear it down that much.

It looks visually good, I say with my layman's eyes, no capacitors look like they're starting to leak or anything. The big capacitors in the power supply part: 50 V, 10000uF, 85 C, brand Nippon.

Check the pictures, see the one where I put a small lighter. Damn how close the small capacitors are. I hadn't tried to remove them and put new capacitors in with my rudimentary soldering skills (first remove the old ones, de soldering, ugh..). It would be a bit tricky (for me). God knows what could go wrong:
View attachment 471730View attachment 471731View attachment 471732View attachment 471733View attachment 471734View attachment 471735

Could this be something to check? According to the service manual, which can be found on Hifi engine :
Zero current check
Remove fuse F301A or F301b (2.5 AT), depending on which channel is to be measured. Connect a multimeter between the contacts on the fuse holder. Turn on the multimeter and check (after a minute) that the meter shows 20-30 mA.

And yes ,I am aware of the danger of getting an electric shock if I mess it up.

View attachment 471686View attachment 471687View attachment 471688

Luxor 7082 Stereo Integrated Amplifier Manual | HiFi Engine

www.hifiengine.com

Edit:
Just to clarify, I personally will not initiate any recap/exchange process, EVEN if they had to. My more knowledgeable friend would do that in that case.
(then I would help him with something else to repay the favor)

In addition to the "Zero Current Check" mentioned most service manuals will have "voltage check points" shown on the schematic with the expected voltage range. I would check those (carefully as shorting out the test leads can damage things in addition to making a big "flash bang"). If the voltages are within spec I would just leave well enough alone and enjoy it. If there are voltages that are out of range then you probably want to investigate further before performance degrades and damage happens especially since this is quite a nice and valuable unit.

 

[MAB]

A practical example of doing something or not, that's the question. What would you have done?

I wonder what else I can do with my vintage Luxor 7082A amplifier. An amplifier that is now 46 years old. Half a century. Sounds crazy if you put it that way.

The thing is that it works flawlessly as far as I can hear. No strange popping sounds, no noise in the speakers. I've checked the DC offset and it's ok. I've cleaned the potentiometers with electronics spray because there was a little scratch in the volume control. I've had it for three years, playing with it sometimes. Switching between a few different amplifiers/receivers. In other words, since I don't use it daily, I don't wear it down that much.

It looks visually good, I say with my layman's eyes, no capacitors look like they're starting to leak or anything. The big capacitors in the power supply part: 50 V, 10000uF, 85 C, brand Nippon.

Check the pictures, see the one where I put a small lighter. Damn how close the small capacitors are. I hadn't tried to remove them and put new capacitors in with my rudimentary soldering skills (first remove the old ones, de soldering, ugh..). It would be a bit tricky (for me). God knows what could go wrong:
View attachment 471730View attachment 471731View attachment 471732View attachment 471733View attachment 471734View attachment 471735

Could this be something to check? According to the service manual, which can be found on Hifi engine :
Zero current check
Remove fuse F301A or F301b (2.5 AT), depending on which channel is to be measured. Connect a multimeter between the contacts on the fuse holder. Turn on the multimeter and check (after a minute) that the meter shows 20-30 mA.

And yes ,I am aware of the danger of getting an electric shock if I mess it up.

View attachment 471686View attachment 471687View attachment 471688

Luxor 7082 Stereo Integrated Amplifier Manual | HiFi Engine

www.hifiengine.com

Edit:
Just to clarify, I personally will not initiate any recap/exchange process, EVEN if they had to. My more knowledgeable friend would do that in that case.
(then I would help him with something else to repay the favor)

Something that is often overlooked; measure the left/right performance of the volume control. These old volume controls weren't particularly good when new, likely have degraded resistive elements that can't be fixed. You can even directly probe the resistance of the control, compare left to right to see the mistracing. It's highly likely you have >1dB left-right, especially in the part of the volume control used most often, which causes wear. Even new, these controls are often not well matched left to right. If I were going to 'upgrade', it would likely be a new volume control, which is particularly easy on your machine. Can be done without touching the PCB or modifying a thing.

The capacitors is a tricky subject. Some units have lots of overstressed caps, caps in high heat areas. I tend to not own units like that any more. The vintage units I still own have either had a only few caps replaced, or none at all. More commonly have to replace relays, since they are responsible for all sorts of measured and audible noise and distortion, and can be very intermittent. They can also be a pain if direct replacement isn't available.

You should check every breakable ground point, refresh it with contact cleaner and preservative. Make sure grounds are gas-tight. If the manufacturer had used therad-sealer, consider replacing that.

Replacing capacitors is not first on my list of things to do on vintage equipment, since most of the significant audible and measurable issues seem to lie elsewhere. That's where the hobby part comes in though, since I think you can also get a good sense of accomplishment when a piece is fully refreshed. Or at least the capacitor part of the piece...

 

[nanook]

I agree, as for aluminum electrolytic caps it's hard to give a generalized answer. My impression is that lifetime differs a lot for caps from that era, so it's really hard to tell.
Measuring the ESR is the solution here. A good 10uF cap will have a couple of ohms, a good 100uF cap maybe 1 ohm and a good 470uf or more definitely below 0.5 Ohms. A better approach than the pure ESR is to compare the ESR of several nominally equal caps (same values, same rating, same size, same manufacturer, same series = "same capacitor"). If they stray only little (ratio <1.5), then I would consider them good.
Most likely it's the physically small caps that wear out (usually dry out) first. 105°C caps were not around yet asfaik but it's worth a look if there's a temperature rating on the sleeve. There's a good chance that the 105°C rated ones are still good.

Apart from the caps I would follow @MAB's recommendations. Replace or clean relay contacts, treat pots and switches with "Oszillin Teslanol T6" contact spray (it's been renamed a couple of years ago and is called "Contact spray" on the can but is advertised as Teslanol T6 in many cases to avoid confusion). This spray has been a hint from a Tektronix service technician about 30 years ago and I have only good experience using it since. It's able to dissolve stiff grease in switches (push-button as well as rotary) it protects surfaces and for whatever reason it is able to remove the oxide. It's necessary to operate the switch or the potentiometer a couple of 10 times after applying the Teslanol T6.
The only adverse effect I know of is with slider- pots. Since it dissolves the grease, the slider will likely loose this smooth mechanical resistance when you operate it. This might apply to rotary pots as well but never happened to me.

Important: Keep away from "Kontakt 60" and similar aggressive contact cleaners although they are advertised for this purpose. They will ruin the pots and contacts long term (I wouldn't even use it on contacts in my car). "Kontakt 60" has to be thoroughly washed with "Kontakt WL" from the same company (Kontakt Chemie), but you will never really accomplish getting it washed out completely. And then you would still need another spray to protect and lubricate the surfaces.

If you decide to clean relay contacts, use a stripe of paper (it's abrasive enough in many cases) with some Teslanol T6.

Adjusting the bias current of the output stage (the "zero current check" is definitely a good idea. Make sure there's no speaker connected (would give a faulty measurement) and give the unit a decent time to fully warm up (30 minutes).
Remark: In some bias current adjustment circuits the trim-pot is connected in the wrong way (I even encountered this in a Studer Pro-Audio Amp), such that the bias current jumps to huge values when the old trim-pot intermittently does not make contact. This will often blow up the output stage transistors.
The usual countermeasure is to connect an incandescent lamp ca. 25 W in series to the mains supply (between wall outlet and amplifier). When the output stage draws a lot of current, the lamp lights up and the voltage in the amplifier breaks down to safe values.
In case you found a circuit diagram / schematic diagram of this amplifier, I can look up for you if this risk applies to your amplifier.

 

[nanook]

I haven’t unsoldered the bridge to measure an individual diode, but an adding a series resistor would be easy. I had expected that.

In case you pick up that topic again. The series resistance seems to be huge compared to Silicon or Silicon-Schottky diodes.
For now: have fun with traveling!

https://www.radiomuseum.org/forum/selen_gleichrichter_1.html

 

[DanielT]

I would like to thank everyone for their advice and tips regarding checking currents (zero current check), cleaning, lubricating with Oszillin Teslanol T6" contact spray and possibly replace components like MABs suggestion regarding the volume pot.
The currently installed is a 200 kOhm Bx2 volume potentiometer. If there is a good 200 kOhm Alps pot that fits, I can definitely imagine it.
My friend and I can fix what you all mention, including possibly replacing the volume pot, this fall when we visit each other.

The now existing volume potentiometer:

Amp with the lid on:


Saw another thing. In my 7082A amp there is an NJM 4458D op amp.From the 7082A service manual:






This talk about replacing op amps that keeps popping up here on ASR. Not something I planned to do, just thought I'd mention it in passing. As a curiosity.

A little about the Luxor 7082A from a Kjell who worked at Luxor, #44 from the thread linked below:

'Classic Integrated Amplifiers!'

Sep 19, 2023

I'm a fan of classic integrated amplifiers, so I thought I would start a thread to celebrate these great devices!
The 1970's and '80s seemed to me as the heyday for integrated amplifiers. Several manufacturers had models with flagship performance and features. These features made it down to models lower in the line and were fantastic alternatives to receivers, and often had better performance and more flexible features. The minimum set for an integrated was typically two or three line-level inputs, a phono input, and a tape loop, speaker selection, plus a headphone jack. High end...

• MAB
• Replies: 215
• Forum: Amplifiers, Phono preamp, and Analog Audio Review

• 

 

[KSTR]

I agree, as for aluminum electrolytic caps it's hard to give a generalized answer. My impression is that lifetime differs a lot for caps from that era, so it's really hard to tell.
Measuring the ESR is the solution here. A good 10uF cap will have a couple of ohms, a good 100uF cap maybe 1 ohm and a good 470uf or more definitely below 0.5 Ohms. A better approach than the pure ESR is to compare the ESR of several nominally equal caps (same values, same rating, same size, same manufacturer, same series = "same capacitor"). If they stray only little (ratio <1.5), then I would consider them good.

I recall having read a paper that said dielectric absorption goes way up long before ESR does. The reason is that when the electrolyte has been drying up in like just a few percent of volume, the ESR is still almost unchanged but the dried up sections are very high impedance and thus the usual absorption test of discharging a charged and settled cap followed by measuring open voltage will give a high reading after some time because the 10 second short was too short to drain the dried-up regions with their high parasitic resistance.

 

[nanook]

I recall having read a paper that said dielectric absorption goes way up long before ESR does. The reason is that when the electrolyte has been drying up in like just a few percent of volume, the ESR is still almost unchanged but the dried up sections are very high impedance and thus the usual absorption test of discharging a charged and settled cap followed by measuring open voltage will give a high reading after some time because the 10 second short was too short to drain the dried-up regions with their high parasitic resistance.

Interesting - I was not aware of this effect. But shouldn't a capacitance measurement 100 Hz or 1 kHz (and even more the ESR measurement at an equivalent of 100 kHz) reveal this too?

And of course this means to unsolder the cap and I for one would replace it when it's unsoldered already.

 

[nanook]

My question:
Which parameters of an aluminium electrolyitic capacitor change when it wears out / dries out?

This is the KI answer:

When an aluminium electrolytic capacitor wears out or dries out, several of its electrical parameters degrade. The most critical changes are:

---

1. Capacitance (C)​

• Decreases as the electrolyte dries out.
• The dielectric layer (aluminium oxide) remains, but the reduction in effective surface area and increased spacing from electrolyte loss reduces capacitance.

---

2. Equivalent Series Resistance (ESR)​

• Increases significantly with aging.
• Dried-out electrolyte increases internal resistance, leading to higher ESR.
• This is one of the most noticeable and damaging effects.

---

3. Leakage Current​

• Can increase or decrease depending on how the dielectric is affected.
• Often, in dried-out caps, leakage current decreases, but if the oxide layer is damaged, leakage may increase.

---

4. Dissipation Factor (tan δ)​

• Increases with ESR and dielectric degradation.
• Higher dissipation factor indicates more energy loss as heat.

---

5. Impedance​

• Increases, especially at higher frequencies.
• Due to increased ESR and decreased capacitance.

---

6. Ripple Current Handling​

• Decreases, making the capacitor more prone to overheating and failure under normal conditions.

---

Summary Table:​

Parameter	Change when drying out
Capacitance	↓ Decreases
ESR	↑ Increases significantly
Leakage Current	↑ or ↓ (varies)
Dissipation Factor	↑ Increases
Impedance	↑ Increases
Ripple Current Rating	↓ Decreases

---

Diagnostic Tip:​

An ESR meter is often the most effective tool to detect worn-out aluminium electrolytics in-circuit, as ESR is the parameter most sensitive to electrolyte degradation.
Let me know if you want to know how this affects circuits like power supplies or audio paths.

My question: How about dielectric absorption?

KI answer:

Great question!

Dielectric Absorption (a.k.a. “soakage”)​

Dielectric absorption is the phenomenon where a capacitor, after being discharged, slowly regains some voltage due to charge retention in the dielectric material.

---

What happens to dielectric absorption in a worn-out aluminium electrolytic capacitor?​

Short answer:
Dielectric absorption may slightly change, but it is not a dominant failure mode in dried-out aluminium electrolytics.

---

Details:​

1. Electrolytic capacitors use an oxide layer (Al₂O₃) as the dielectric, which is formed electrochemically.
2. The electrolyte's role is to maintain this oxide and serve as part of the anode system.
3. As the electrolyte dries out, the oxide may degrade, and the internal structure changes.
4. This might cause minor changesin dielectric absorption, but:
   • ESR, capacitance, and leakage current degrade much more significantly.
   • Dielectric absorption is not usually measured or used as a diagnostic indicator of aging in electrolytic caps.

---

In practice:​

• For precision applications (like sample-and-hold circuits), dielectric absorption is a concern — but aluminium electrolytics are already poor choices for such applications, even when new.
• Other capacitor types (e.g., film, ceramic, tantalum) are evaluated more closely for dielectric absorption.

---

Summary:​

Parameter	Changes with aging	Diagnostic relevance
Dielectric Absorption	Slight, unclear	Low
Capacitance	Decreases	High
ESR	Increases	Very high
Leakage Current	Can increase	Medium

---

If you're troubleshooting a circuit, dielectric absorption is not worth testing in electrolytics. Focus on ESR and capacitance instead.
Want a quick guide on testing electrolytics in-circuit?