There are, as you all know, many types of capacitors and most gear seems to come with an assortment of them, film, poly, electrolytic, ceramic etc. The question I have is are the types used based on specific value ranges, price or suitability for the task? All I know about them is Voltage and capacitance value. Ok, I know a *little* more than that but not by much 😛
Are caps with more or less equal values interchangeable within reason? I ask because looking at my various circuit boards I see all kinds of caps squirted about and figured it was about time I asked why that is and what types I should think about replacing/avoiding in amplifiers (ceramics?)
Are caps with more or less equal values interchangeable within reason? I ask because looking at my various circuit boards I see all kinds of caps squirted about and figured it was about time I asked why that is and what types I should think about replacing/avoiding in amplifiers (ceramics?)
Hi isnotdynamic,
Be confident about the types and values engineers choose. So take the same dielectric type : ceramic if ceramic, films if films, lytics if lytics.
That said, some things to know. About capacitance values : precision matters when caps are here for a filter circuit or when you have two rails for stereo and you want each channel having the same behavior. I talk about lytic caps which capacitance value can be more than 20% precision. Notice they have a precision marking: M for instance for 20% iirc,. Always respect the markings when swapping a cap : LL for Long Life for instance, temperature coef. form factor: pitch between legs has a particular inductance (one of the important behavior of a cap) and should fit perfectly in the vias of the board.
Lytics caps are aging and should be swapped after 15 to 30 years (depends) in our hifi devices when at power supply areas. I would not be worry as far the top of the cap has no bump. Rest of the caps are very reliable but the tantalum wet type.
Cost reduction matters for mass market hifi devices. Often lytics are choosed cheap. When refurbishing an old amp/device you may choose better characteristics : for instance 105°C instead of 85°c of the original cap, Long Life if after the rectifier bridge and a reliable manufacterer (tons of threads here and everywhere) and distributor:Mouser, Digikey, Farnell to be sure to have genuine parts.
Changing caps may change the sounding equilibrium: for the worst or the best, often it's subjective, costly and just "reliable" in your particular hifi environment. It can only give you a good enough result if you have: time, money and experience and it's always a bad advice if you read somewhere : put here that brand & model to improve your sound in that particular device... but of course if it was chosen from the beginning of the design process by the designers. You can do better but few are able of that Most engineers will choose only measurable values because anyway the final you hear is also about the other devices in your hifi system and psychoacoustic factors. Although good results can be expected if someone is knowing well the device he is tweaking with a correct understanding of it with many time, money to test. And more and more you have smd devices where you have no room for changing anything. At the end caps is just a small part of a whole design, so no miracles can be ewpected when swapping caps but avoiding a break of an too old device (again look at bumps on the top of the lytic caps or leakings is enough most of the time, halas some caps may also dry... age of your device is a good clue.
A bit long, sorry. Just draw the contour hopping what I write is good enough not being tech myself while being an old hifi enthusiast... just being there with the same questions as yours. The best is to read the litterature about it: many sources.
Hope this helps.
Thanks for the write up, good advice.
I haven't done a lot of electronics diy of late and never really learned a whole lot about electronics because, for one it's over my head for the most part, the other is that I already have a shed-load of other hobbies that take too much learning and memory. Frankly I can't keep it all in at once so I tend to 'crash course' it when needed and forget it when I don't. I don't pretend to hear cap differences when swapping them out in the past, I generally do it for the benefit of the circuit if it's considered a good move, electrically speaking.
Other than that, I'm happy to leave as-is.
I haven't done a lot of electronics diy of late and never really learned a whole lot about electronics because, for one it's over my head for the most part, the other is that I already have a shed-load of other hobbies that take too much learning and memory. Frankly I can't keep it all in at once so I tend to 'crash course' it when needed and forget it when I don't. I don't pretend to hear cap differences when swapping them out in the past, I generally do it for the benefit of the circuit if it's considered a good move, electrically speaking.
Other than that, I'm happy to leave as-is.
Hi all,
I found a very interesting site that has links to a series of 9 articles written by a capacitor engineer (Cyril Bateman) for Electronics World back in 2002-2003. Very extensive, and some surprising results come from his extensive testing of thousands of capacitors of all sizes and types. Even he was surprised by some of the results his tests revealed. It seems like a good place to do a little research before selecting caps for that next project. Cyril Bateman's Capacitor Sound
I found a very interesting site that has links to a series of 9 articles written by a capacitor engineer (Cyril Bateman) for Electronics World back in 2002-2003. Very extensive, and some surprising results come from his extensive testing of thousands of capacitors of all sizes and types. Even he was surprised by some of the results his tests revealed. It seems like a good place to do a little research before selecting caps for that next project. Cyril Bateman's Capacitor Sound
Yup, a classic. Mr Jan D. is a Diyaudio member and had gently putt together all the Cyrill B. papers on this link.
It's a very good read but it will not provide you advices to swap blindly the caps but let you understand the types of dielectrics and their behavior for audio purpose. There is no good move for electric improvement if the circuit is already working and have no break.
It's a very good read but it will not provide you advices to swap blindly the caps but let you understand the types of dielectrics and their behavior for audio purpose. There is no good move for electric improvement if the circuit is already working and have no break.
Its horses for courses.
But sometimes some manufacturers take the cheaper option to keep prices down.
Larger values of capacitance tend to be electrolytic.
Unless it is a bi polar electrolytic then you need to be careful about polarity.
I sometimes use two electrolytics back to back.
But sometimes some manufacturers take the cheaper option to keep prices down.
Larger values of capacitance tend to be electrolytic.
Unless it is a bi polar electrolytic then you need to be careful about polarity.
I sometimes use two electrolytics back to back.
Indeed, two Elna Simic II in serie (+--+ arengement) at DC bloking is a good example, but the (green) non polar Nichicon Muse costs less and have a very low distorsion as they measured in Jan D. papers as well. That said each gives a different sounding at the end and you never know before testing both... and some will prefer a mkp cap here, a styren, a teflon, etc, etc. So unless you have years of experience and tons of caps, monney and time... better to leave your device as it is (but if a cap is leaking or has its top bumped). And the op device is perhaps costing less than a bunch of caps just for testing purpose with any improvement guaranty...
A more expensive lytic cap is NEVER a guaranty of a better sound. Even if one take audio lytics off shelves. I liked a lot the Black Gate, but they were not always good for a swap and could be worse than a cheap lytic in a device.
A more expensive lytic cap is NEVER a guaranty of a better sound. Even if one take audio lytics off shelves. I liked a lot the Black Gate, but they were not always good for a swap and could be worse than a cheap lytic in a device.
Last edited:
There are many circuits that act as unintended frequency filters, due to off value or high ESR electrolytic caps, without the caps ever leaking or popping a vent. Once a consumer electronic device sounds funny, I take that as a sign of the life of the caps purchased by the manufacturer, and change them all. Dried up rail caps or B+ caps make an amp sound polite, having no dynamics due to the limited power available. There can be timer effects too of non-leaking e-caps. Power on silence circuits never time out and allow sound, or key sustain circuits get shorter and random in time length. DC detect circuits can get too aggressive due to off value e-caps.
I haven't found any "forever" epoxy sealed electrolytic caps in any consumer device. (They were in fender guitar amps of a certain era, David of NC reports). I've put as high as 185 e-caps in one model organ for example. I've had success replacing 100 % failure 1970's tantalum caps used for interstage coupling, with COG high voltage (50) ceramics.
Paper dielectric caps from the sixties have also hit life, and I have to say, polyester is not a great substitute. I found polyester harsh in my PAS2. Leaving it alone was not an option, the channels had a ~4 db unbalence due to a paper cap. I've moved on to the upgraded RA-88a mixer as my entertainment hub, pending replacement of the polyester caps in the PAS2 with polyprophylene of the right value (non-stocked .02 uf caps). Most hammond B/C3 owners of paper equipped units (pre spring 64), have had to deal with the off value paper caps.
Last edited:
Hi Indianajo
I can't agree more about dry lytics and ESR getting worse with time. Tantalum, especially wet ones are also a nightmare cause they can fire.
For an amp, what is the safe time before lytics replacement if you can't measure ESR ? 15/20 years of daily use or more often with high voltage caps in tube stuffs ? 20/25 years with smps powersupply ? I had no problem with 35 years old cd players or preamp powersupplies though.
I have nothing against polyester caps, that's true polypropylene have far better specs but polyester can produce sometimes a better subjective result, padding things for instance or adds some 'nice" distorsion which gives the sound signature of the device in case it was choose in that purpose by the designer. Find polypro Orange drop have a close character if one want to swap a polyester in the signal path.
Did you try tin foilded styrens (not the cheap ones with plastic case and thin leads) instead NPO for coupling ? Ceramic class I is good enough but I try to use styren instead for that purpose. And a PPS smd cap may be closer to the tantalum, I mean from a subjective hearing point of view... depends, you must try to know halas... PPS are fragile with soldering and capacitance can move if too much heat. Better to avoid if you need accuracy wth capacitance.
There are usefull polyester caps as the russian K73-16, of course not "neutral" as all the polyesters but may give (or not) a good subjective result when swapping an older polyester cap.
That's true old papered caps were no consistant with specs when aging: moisture, humidity, less isolation of the paper with time (dry wax?).
Btw, asking myself if todays Arizona caps (polyester+paper) could be a good swap in Hammond organs in lieu of the paper caps?? Jentzen are too expensive... I liked a lot the Amphom with paper, were not too much expensive (but difficult to source), halas not made anymore.
Sometimes a circuit is proof to cap replacment type (your ceramic class I swap as coupling capacitor instead of the former tantalum? just asking myself if the tantalum was here for some damping behavior due to their highish ESR? Cause orange ceramic class II disc were cheap if needed instead.)
I can't agree more about dry lytics and ESR getting worse with time. Tantalum, especially wet ones are also a nightmare cause they can fire.
For an amp, what is the safe time before lytics replacement if you can't measure ESR ? 15/20 years of daily use or more often with high voltage caps in tube stuffs ? 20/25 years with smps powersupply ? I had no problem with 35 years old cd players or preamp powersupplies though.
I have nothing against polyester caps, that's true polypropylene have far better specs but polyester can produce sometimes a better subjective result, padding things for instance or adds some 'nice" distorsion which gives the sound signature of the device in case it was choose in that purpose by the designer. Find polypro Orange drop have a close character if one want to swap a polyester in the signal path.
Did you try tin foilded styrens (not the cheap ones with plastic case and thin leads) instead NPO for coupling ? Ceramic class I is good enough but I try to use styren instead for that purpose. And a PPS smd cap may be closer to the tantalum, I mean from a subjective hearing point of view... depends, you must try to know halas... PPS are fragile with soldering and capacitance can move if too much heat. Better to avoid if you need accuracy wth capacitance.
There are usefull polyester caps as the russian K73-16, of course not "neutral" as all the polyesters but may give (or not) a good subjective result when swapping an older polyester cap.
That's true old papered caps were no consistant with specs when aging: moisture, humidity, less isolation of the paper with time (dry wax?).
Btw, asking myself if todays Arizona caps (polyester+paper) could be a good swap in Hammond organs in lieu of the paper caps?? Jentzen are too expensive... I liked a lot the Amphom with paper, were not too much expensive (but difficult to source), halas not made anymore.
Sometimes a circuit is proof to cap replacment type (your ceramic class I swap as coupling capacitor instead of the former tantalum? just asking myself if the tantalum was here for some damping behavior due to their highish ESR? Cause orange ceramic class II disc were cheap if needed instead.)
Last edited:
There are some simple rules that can help you use the perfect cap:
If it's signal coupling
a: if there's a dc potential there: use polarized electrolytic from decent manufacturers if you can't afford Nichicon Muse or Elna Silmic or other expensive brands.Even Chinese caps can be a lot better than you might think!
If you have an old circuit measure the capacitors.Many capacitors of the past from great manufacturers are actually better than the new ones...Sad but True!
You might actually like the roll off or the muffled sound of some old ones more than the bright sound due to the new ones!
b.If there's barely any or no potential use bipolar.Any bipolar will do, from any company.
c.Use the lowest capacitance value that can give you the needed lower bandwidth pass in high impedance circuits and the highest capacitance in the smallest size you can fit when low impedance circuit is met. If you have 2vrms passing through a circuit, don't use 50v capacitors if you have a 10-16v capacitor at hand unless you have an expensive one like Nichicon Muse .Even so, what i really mean is that the physical measurements of a coupling capacitor is very important for noise specs in high impedance circuits.
Sometimes you can use 1/10th of its value as parallel non polarized capacitor.Any will do.
Don't use very high voltage non-polarized capacitors for signal coupling if a smaller electrolytic can do the job.Actually, if there's no potential or a very small one, there's a real chance that an electrolytic can sound 1000 times better than the most outrageously expensive non polarized capacitor on the market. That is because of a few factors:
high voltage nonpolarized capacitors have thick dielectric and high loses or
they can show microphonic behaviour ,
they are very big and can pick-up way more noises than an electrolytic,
small electrolytics actually have no microphonics due to the thin liquid layer between electrodes, they have very low loses when new and
Changing electrolytics can be fun too 😉
If it's signal coupling
a: if there's a dc potential there: use polarized electrolytic from decent manufacturers if you can't afford Nichicon Muse or Elna Silmic or other expensive brands.Even Chinese caps can be a lot better than you might think!
If you have an old circuit measure the capacitors.Many capacitors of the past from great manufacturers are actually better than the new ones...Sad but True!
You might actually like the roll off or the muffled sound of some old ones more than the bright sound due to the new ones!
b.If there's barely any or no potential use bipolar.Any bipolar will do, from any company.
c.Use the lowest capacitance value that can give you the needed lower bandwidth pass in high impedance circuits and the highest capacitance in the smallest size you can fit when low impedance circuit is met. If you have 2vrms passing through a circuit, don't use 50v capacitors if you have a 10-16v capacitor at hand unless you have an expensive one like Nichicon Muse .Even so, what i really mean is that the physical measurements of a coupling capacitor is very important for noise specs in high impedance circuits.
Sometimes you can use 1/10th of its value as parallel non polarized capacitor.Any will do.
Don't use very high voltage non-polarized capacitors for signal coupling if a smaller electrolytic can do the job.Actually, if there's no potential or a very small one, there's a real chance that an electrolytic can sound 1000 times better than the most outrageously expensive non polarized capacitor on the market. That is because of a few factors:
high voltage nonpolarized capacitors have thick dielectric and high loses or
they can show microphonic behaviour ,
they are very big and can pick-up way more noises than an electrolytic,
small electrolytics actually have no microphonics due to the thin liquid layer between electrodes, they have very low loses when new and
Changing electrolytics can be fun too 😉
Last edited:
On amps, maximum power output is easier to measure than ESR. My ST70 was down to 7 watts/channel when i took it to the McIntosh clinic 9 years after construction. Caps got it to about 25 W, a new rectifier tube got the other 10. That amp has needed caps 3 other times, but now I can buy 3000 hour caps or even 10000 hour in the 100 under uf class.
P=V^2/Z where Z is speaker impedance - about 20% higher than measured resistance usually.
Other than that, I have listened for dull, stupid sounds. On reproduced music I use pianos & cymbals as a standard, since they have a sound you can hear in the concert hall. Also, as I said, "polite" amps with no power can't reproduce the hammer hit of a piano.
Different brands have different cap lives depending on what service life the manufacturer put in. GE radios have gone 30-35 years in my experience, store brands and cheapos can go all mushy and useless at 5 years. The cheapo computer displays all were useless after 4 years, I have a Philips I picked up somewhere that is fine at 15+ years?
My 89$ RCA CD player I bought about 1987 is still playing fine. The 22$ ones I bought at K-mart have all gone by the wayside. Not saying that all RCA products were like that one, but they were charging a premium for a new product in 1987.
ESR tells the tale however, and can save some time in devices like PC's where there are a hundred or so e-caps. This Dell loses the cursor after 12 hours, I may dive in and find that one cap some day.
Hammond B/C3 organs were so respected in part because they had no e-caps - they were in the Leslie, and they were screw in replacement. The ones post spring 94 have polyester caps, and are fine at this late date. Some purists don't like how bright those are, they prefer the duller paper cap sound.
Atlas Peak ESR meters are only about $120, I got away with leaving some caps in a Peavey amp this winter because they weren't bad at 20 years. But the rail caps were shot, the $4.50 each ones.
My Hammond H100 organs cost 1/20th the price of a C3, because at 30 years they sounded like wet kazoos. H100 have cheap e-caps, C3 had paper caps. 77 e-caps later one sounds like the instrument that cost more than a Mustang Convertible in 1968. Another 110 and I'll have the key sustain working right. The Wurlitzers are so hammered by e-caps they only moan. Allens had a better grade cap than wurlitzer, the 300 made noise at 30 years age but the silence circuit cap silenced it once, and the rail caps in the power amp took it to 2 watts and then silent 2 months later. Why it is good to do them all when one goes bad. So you don't have to change one after the other with a dead instrument over & over.
Train your ear in a concert hall, real instruments are a powerful test for accurate sound. And not the electronic ones, those can sound like anything.
Where I put the 10 uf ceramics, replacing tantalum, I didn't have room to put 10 uf 63 v polyester. With 2 v signals on 50 v caps, they don't sound that weird, IMHO. I haven't noticed symptoms of microphonics anywhere, although I don't play my guitar amp at 11 all the time either with caps 6" from the speaker. The speakers are 3' away from the ceramics in the H110 organ, and 8' away from the ceramics in my ST-120 hifi amp.
Last edited:
- Status
- Not open for further replies.