Electrolytic Capacitor Shoot-Out
- Thread starter patfont
- Start date
Mr:Bond
New Member
Very good!!
I would not have been able to find these KW in these measurements.
Which others do you recommend to replace the 4 SMD caps there in the middle of the Pana Caps?
Caps SMD 100uf 35v 12
Last edited:
Mr:Bond
New Member
Thanks for the help, I made the purchase of Digikey caps.
Champco
Super Member
Hi Pat. old post I know but can you expound on this.. "is at a minimum -3db down at 43hz,: a little more.. Thanks
An oldie but a goodie, Pat.
Champco
Super Member
Ok thanks. What got my attention was the -3db, which I see now meant "down to" 43hz
neg 3db is volume level
was wondering down from what??? zero? but is up from -12
getting to confusing Thanks.
Just chiming in on this very old thread as it relates to the Marantz 2230, and to add an interesting commentary on recent observations regarding some replacement caps being used in the classic and vintage equipment field, and the real world requirements for such parts.
It is a long commentary (so long it's in 2 postings). But I am not intending to hijack this thread from the OP. Only merely to add an experiential aside to it.
We have had as a 'garage' system for over 20 years now a 2230. When we first acquired it, I did a minimum restoration to protect the outputs and assure safe operation. And it ran that way for about 15 years or so, until one afternoon I noticed the OE Elna main filter was dripping electrolyte all over the counter under it. Then it finally got the full treatment that I normally lavished on my Fisher tube instruments. And it had earned it.
My rule of thumb with vintage equipment is to respect the original designer's voicing and circuit choices to preserve the device for future generations to enjoy. It's the philosophy of a trust. Heavily hacked up examples are simply not as desirable or valuable to future generations. I've seen this with many Fisher tube instruments that have been Dr. Stereo'ed and modified into oblivion, and no longer work, look, or sound like a golden-era Fisher anymore. These old classics are now marching gracefully into their sixth and seventh decades, and every generation rediscovers them anew. And there is then a fresh run on them.
And every 10-15 years or so, some enthusiasts and techs armed with their Wellers, scopes and fresh parts somehow think that an instrument that continues to admirably function 60 years out can be made 'better' with some wholesale circuit and parts changes. And so many are lost to the modification beast each time.
Today, very few intact specimens faithful to the original design now remain, and those that do are now commanding increasingly high values. And many of those made 'better' by being converted to different output tubes and heavily reworked circuits are now sitting in landfills or being used as parts donor units.
But I digress. Back to the 2230.
So interestingly, with that philosophy, on this 2230, Silmics were used in many locations not just for their excellent performance, but also because the instrument was originally built with Elnas throughout in the early 70s. And they DO sound good. Where a Silmic would not fit, a comparable quality Nichicon was used. Every board was renewed, but not reinvented, or values changed. Original active components were not replaced if they still functioned well. And most all did, as the 2230 operates them all fairly conservatively.
For that big single filter, the elderly hemorrhaging Elna was replaced with a screw terminal Sprague Powerlytic (at that time a wallet-wincing $50+ capacitor). A fine replacement that should last a very, very long time, and perform well. And it has. For the next replacement, I'll be watching from above.
The 2230 is one of the later capacitive-coupled output transistor designs by the big makers of that era. A dinosaur by today’s standards, but still a rich sounding, musically captivating one. And while all the other parts choices renewed the 2230 to its original sonic lushness, the ones that had the most impact -- by far -- were the two main couplers. The originals were starting to fade. And I was fortunate enough to still have in my dusty parts boxes two NOS Black Gates in the exact values needed.
The 2230 was about to cross the Rubycon (pun intended).
And that one singular parts replacement transformed the little garage 2230 into a giant killer. When visitors walk into that garage, their heads turn and their jaws drop. Yes, Black Gates in the direct signal path are really that good.
So the little 2230 sings along faithfully, shocking visitors, to the present day. Only a couple people knew what was its under hood secret.
...
I don't do too much bench work anymore because of my failing eyes. Most of my bench gear is now gone or packed away for the descendants. I still give advice on the care of ancient Fishers when I can. But I recently had an interesting conversation with a still very active professional tech who I occasionally send work to about what he's been seeing lately. The conversation turned to capacitors.
In recent years, 105c rated, and in particular low impedance, high ripple caps have been dropped into a lot of gear as replacements by measurement-obsessed enthusiasts and hobbyists (many of whom openly sneer at audio grade parts and define an electrolytic’s worth based solely on advertised impedance, ESR, tan, and ripple values). A great many wannabe engineers among them, who know just enough to insult and offend others (particularly curious beginners, which I hate to see).
The Panasonics in particular have gotten a lot of attention by the hobbyists from this crowd over the years. The Pannys have great specs, they’re compact, and they’re cheap (for a first-tier maker part). A lot of professional techs also use them for service work, too, mainly because they're inexpensive, readily available, and saying so keeps the customers happy.
Out of the blue, my old tech friend voiced his opinion on these caps after many years in service, and how they’re doing now. And his opinion was "not well". Many units he is seeing come in with 20+ years on them are now leaking. And in many cases it is destroying the boards underneath them. He is not seeing the same problems with other first-line Japanese caps, and is now replacing the bad Pannys with Nichicons where the board can still be saved.
And as a final unsolicited comment, he added that in the last 30+ years as a tech he had yet to ever see a failed Black Gate on his bench, and in his experience they seem to be lasting, and still testing, like they will run forever. And that is consistent with some other information I understood about these remarkable, exotic, eye-wateringly expensive caps. They were originally designed to have an indefinite service life exceeding 30 years. And they apparently have. That is quite a contrast to the leaking, board-destroying, FC and FMs now crossing my tech friend's bench.
So were the BGs worth the 10x premium over an FC 20 years ago? I'll leave that for you to decide this far out.
This is not to be a knock on any particular brand of part, and in fact in today’s industry, getting 20-25 years out of an electrolytic is an excellent service life. Those Pannys delivered full value. Most consumers don't even keep their equipment that long.
But it does lead into some broader observations about what hobbyists fixate on today, verses what the designers and us now retiring (or soon to retire) hands considered relevant 30, 40 or 50 years ago.
...
The first is that most competently designed instruments and circuits, both today and 60 years ago, were designed conservatively. And most will operate perfectly well and for a very long time on general purpose electrolytics.
The hobbyist obsession over high spec, high ripple and very low ESR, impedance and leakage specification capacitors today as somehow required for good power supply operation is simply unwarranted for most equipment then, and now.
That is not to say that the capacitor manufacturer's specifications do not matter. They do. But for a well-designed supply or other circuit, the more modest performance of a general purpose capacitor works perfectly well for many years, and in fact have been used in both mainstream and in a lot of even 'high end' consumer electronics. These instruments were designed for such mainstream parts.
Crack open a fancy 25-30 year old Kilowatt-eating Classe mono block (that still works and measures great), and you'll find it stuffed full of inexpensive 85c general purpose Nichicons. Our old Cal Alpha, a now legendary 30 year old tube DAC that sold new for well over $2,000 (a black box with tubes and zero internal ventilation), is loaded with cheap general purpose Taiwanese caps.
It recently got a recap, and not with low impedance or high temp Pannys or PWs.
There are exceptions to this, and that is where a more aggressively specified part may be worth considering. Some old Counterpoint gear immediately comes to mind where the supplies were designed at the ragged edge. But absent some unconventional designer choices, most equipment will not require a Panny FM or Nichicon UPW to operate well again for many decades.
The general build quality of the replacement part and the good original design around them will determine that outcome.
Again, most competently designed equipment was designed for general purpose capacitors, not heightened specification ones. Including in the power supplies. That capacitor manufacturers make such specialized capacitors does not necessarily mean that the designers use them. Where you find an exceptional case that requires unusual servicing (i.e. a problem design), some of the higher specification replacements are often considered in the field. But otherwise, a similar general purpose part is perfectly appropriate ... and what the circuit was designed to use.
...
The second observation, and a corollary to the first, is that with few exceptions, most older equipment does not need a 105c part, and the use of a standard 85c part will operate for a very, very long time. Almost always because it was well designed to do so. Among the warmest running consumer electronics were hot-running tube gear stuffed full of 15 or 20 burning vacuum tubes (all often stuffed in a tight little cabinet). Yet I would routinely see 45 and 50+ year old Fishers, Scotts and Macs still cruising along nicely on their original 85c (or lower) rated Mallory, Sprauge or GI twist locks. The only thing that typically happened (if they weren't cold energized by a foolish collector after sitting too long) was fading capacitance after about 55 or so years. Yes, even those great old cans from the Kennedy Administration sometimes start drying out.
And I would always replace the tired small under chassis electrolytics on them with 85c Atoms, which are still soldiering on fine, some now in service for over 30 years themselves and still working great.
The same thing is true for your classic Marantz receivers, and most all other older audio electronics. And a lot of more recent production equipment, too. They were designed around 85c parts, and many go on for decades on them. A 105c replacement is a luxury, but rarely a necessity, in those applications. If you want to drop 105c replacements in your PSU build, go for it. But don't preach to others that it must be so, or that they are making some major mistake using an 85c part. That's a falsehood.
As I mentioned, there are some exceptions. But they are mainly from failures of design ... a cap practically leaning on a regulator sink or provided inadequate ventilation is not going to last as long. Or an unusual application environment: a hot running car or an oven or heater containing electronics. There, a 105c part or higher might add a couple extra years of service.
...
And that leads into the reasons why the higher temp/low impedance parts have gained favor in recent years.
What spurred on the hobbyist rage towards the high temp, low impedance/ESR caps have been two things, neither of which you will find in decent older gear: the stolen-formula capacitor plague of a couple decades ago (where the wrong stolen formula prematurely failed, destroying thousands of boards), and the trend to miniaturization involving tight and poor placement of mainly dubious quality Chinese capacitors in more recent equipment.
People don't realize just how hot 85c is, let alone 105c. Way too hot to even touch. Those are temps more associated to a heavily leaking or shorted part. Most semiconductors have difficulty even operating at those temps. And virtually no decent consumer equipment ever runs near that warm. Again, there are exceptions, but I am taking about the norm.
A few years ago, a poster on one of the we-know-best-DIY-audio-science boards ridiculously pronounced that the best replacement capacitors for these consumer electronics supplies were automotive grade capacitors. Now maybe if your Marantz is sitting on a running gasoline engine, or your preamp runs three inches from a red hot exhaust manifold, those caps might make a difference. Or when the cheap Chinese caps on the oven controls of your Kitchen Aid oven crap out after an ill-advised 900F self-clean cycle.
But in the usual consumer electronic operating environment of 40-60c, a 135c automotive part is gross overkill, both in specification and to the wallet.
And applying the usual temperature/life exponential calculations, an 85c part will still last many decades in most equipment. A 105c one will likely outlive the installer, at least as to ambient temperature related declines.
Again, if you want to use a 105c part, do so. But don't deride the use of 85c parts that are just as good from a practical perspective in most situations. It's what the designer used in virtually every case ... unless they were working around a problematic design.
It is a long commentary (so long it's in 2 postings). But I am not intending to hijack this thread from the OP. Only merely to add an experiential aside to it.
We have had as a 'garage' system for over 20 years now a 2230. When we first acquired it, I did a minimum restoration to protect the outputs and assure safe operation. And it ran that way for about 15 years or so, until one afternoon I noticed the OE Elna main filter was dripping electrolyte all over the counter under it. Then it finally got the full treatment that I normally lavished on my Fisher tube instruments. And it had earned it.
My rule of thumb with vintage equipment is to respect the original designer's voicing and circuit choices to preserve the device for future generations to enjoy. It's the philosophy of a trust. Heavily hacked up examples are simply not as desirable or valuable to future generations. I've seen this with many Fisher tube instruments that have been Dr. Stereo'ed and modified into oblivion, and no longer work, look, or sound like a golden-era Fisher anymore. These old classics are now marching gracefully into their sixth and seventh decades, and every generation rediscovers them anew. And there is then a fresh run on them.
And every 10-15 years or so, some enthusiasts and techs armed with their Wellers, scopes and fresh parts somehow think that an instrument that continues to admirably function 60 years out can be made 'better' with some wholesale circuit and parts changes. And so many are lost to the modification beast each time.
Today, very few intact specimens faithful to the original design now remain, and those that do are now commanding increasingly high values. And many of those made 'better' by being converted to different output tubes and heavily reworked circuits are now sitting in landfills or being used as parts donor units.
But I digress. Back to the 2230.
So interestingly, with that philosophy, on this 2230, Silmics were used in many locations not just for their excellent performance, but also because the instrument was originally built with Elnas throughout in the early 70s. And they DO sound good. Where a Silmic would not fit, a comparable quality Nichicon was used. Every board was renewed, but not reinvented, or values changed. Original active components were not replaced if they still functioned well. And most all did, as the 2230 operates them all fairly conservatively.
For that big single filter, the elderly hemorrhaging Elna was replaced with a screw terminal Sprague Powerlytic (at that time a wallet-wincing $50+ capacitor). A fine replacement that should last a very, very long time, and perform well. And it has. For the next replacement, I'll be watching from above.
The 2230 is one of the later capacitive-coupled output transistor designs by the big makers of that era. A dinosaur by today’s standards, but still a rich sounding, musically captivating one. And while all the other parts choices renewed the 2230 to its original sonic lushness, the ones that had the most impact -- by far -- were the two main couplers. The originals were starting to fade. And I was fortunate enough to still have in my dusty parts boxes two NOS Black Gates in the exact values needed.
The 2230 was about to cross the Rubycon (pun intended).
And that one singular parts replacement transformed the little garage 2230 into a giant killer. When visitors walk into that garage, their heads turn and their jaws drop. Yes, Black Gates in the direct signal path are really that good.
So the little 2230 sings along faithfully, shocking visitors, to the present day. Only a couple people knew what was its under hood secret.
...
I don't do too much bench work anymore because of my failing eyes. Most of my bench gear is now gone or packed away for the descendants. I still give advice on the care of ancient Fishers when I can. But I recently had an interesting conversation with a still very active professional tech who I occasionally send work to about what he's been seeing lately. The conversation turned to capacitors.
In recent years, 105c rated, and in particular low impedance, high ripple caps have been dropped into a lot of gear as replacements by measurement-obsessed enthusiasts and hobbyists (many of whom openly sneer at audio grade parts and define an electrolytic’s worth based solely on advertised impedance, ESR, tan, and ripple values). A great many wannabe engineers among them, who know just enough to insult and offend others (particularly curious beginners, which I hate to see).
The Panasonics in particular have gotten a lot of attention by the hobbyists from this crowd over the years. The Pannys have great specs, they’re compact, and they’re cheap (for a first-tier maker part). A lot of professional techs also use them for service work, too, mainly because they're inexpensive, readily available, and saying so keeps the customers happy.
Out of the blue, my old tech friend voiced his opinion on these caps after many years in service, and how they’re doing now. And his opinion was "not well". Many units he is seeing come in with 20+ years on them are now leaking. And in many cases it is destroying the boards underneath them. He is not seeing the same problems with other first-line Japanese caps, and is now replacing the bad Pannys with Nichicons where the board can still be saved.
And as a final unsolicited comment, he added that in the last 30+ years as a tech he had yet to ever see a failed Black Gate on his bench, and in his experience they seem to be lasting, and still testing, like they will run forever. And that is consistent with some other information I understood about these remarkable, exotic, eye-wateringly expensive caps. They were originally designed to have an indefinite service life exceeding 30 years. And they apparently have. That is quite a contrast to the leaking, board-destroying, FC and FMs now crossing my tech friend's bench.
So were the BGs worth the 10x premium over an FC 20 years ago? I'll leave that for you to decide this far out.
This is not to be a knock on any particular brand of part, and in fact in today’s industry, getting 20-25 years out of an electrolytic is an excellent service life. Those Pannys delivered full value. Most consumers don't even keep their equipment that long.
But it does lead into some broader observations about what hobbyists fixate on today, verses what the designers and us now retiring (or soon to retire) hands considered relevant 30, 40 or 50 years ago.
...
The first is that most competently designed instruments and circuits, both today and 60 years ago, were designed conservatively. And most will operate perfectly well and for a very long time on general purpose electrolytics.
The hobbyist obsession over high spec, high ripple and very low ESR, impedance and leakage specification capacitors today as somehow required for good power supply operation is simply unwarranted for most equipment then, and now.
That is not to say that the capacitor manufacturer's specifications do not matter. They do. But for a well-designed supply or other circuit, the more modest performance of a general purpose capacitor works perfectly well for many years, and in fact have been used in both mainstream and in a lot of even 'high end' consumer electronics. These instruments were designed for such mainstream parts.
Crack open a fancy 25-30 year old Kilowatt-eating Classe mono block (that still works and measures great), and you'll find it stuffed full of inexpensive 85c general purpose Nichicons. Our old Cal Alpha, a now legendary 30 year old tube DAC that sold new for well over $2,000 (a black box with tubes and zero internal ventilation), is loaded with cheap general purpose Taiwanese caps.
It recently got a recap, and not with low impedance or high temp Pannys or PWs.
There are exceptions to this, and that is where a more aggressively specified part may be worth considering. Some old Counterpoint gear immediately comes to mind where the supplies were designed at the ragged edge. But absent some unconventional designer choices, most equipment will not require a Panny FM or Nichicon UPW to operate well again for many decades.
The general build quality of the replacement part and the good original design around them will determine that outcome.
Again, most competently designed equipment was designed for general purpose capacitors, not heightened specification ones. Including in the power supplies. That capacitor manufacturers make such specialized capacitors does not necessarily mean that the designers use them. Where you find an exceptional case that requires unusual servicing (i.e. a problem design), some of the higher specification replacements are often considered in the field. But otherwise, a similar general purpose part is perfectly appropriate ... and what the circuit was designed to use.
...
The second observation, and a corollary to the first, is that with few exceptions, most older equipment does not need a 105c part, and the use of a standard 85c part will operate for a very, very long time. Almost always because it was well designed to do so. Among the warmest running consumer electronics were hot-running tube gear stuffed full of 15 or 20 burning vacuum tubes (all often stuffed in a tight little cabinet). Yet I would routinely see 45 and 50+ year old Fishers, Scotts and Macs still cruising along nicely on their original 85c (or lower) rated Mallory, Sprauge or GI twist locks. The only thing that typically happened (if they weren't cold energized by a foolish collector after sitting too long) was fading capacitance after about 55 or so years. Yes, even those great old cans from the Kennedy Administration sometimes start drying out.
And I would always replace the tired small under chassis electrolytics on them with 85c Atoms, which are still soldiering on fine, some now in service for over 30 years themselves and still working great.
The same thing is true for your classic Marantz receivers, and most all other older audio electronics. And a lot of more recent production equipment, too. They were designed around 85c parts, and many go on for decades on them. A 105c replacement is a luxury, but rarely a necessity, in those applications. If you want to drop 105c replacements in your PSU build, go for it. But don't preach to others that it must be so, or that they are making some major mistake using an 85c part. That's a falsehood.
As I mentioned, there are some exceptions. But they are mainly from failures of design ... a cap practically leaning on a regulator sink or provided inadequate ventilation is not going to last as long. Or an unusual application environment: a hot running car or an oven or heater containing electronics. There, a 105c part or higher might add a couple extra years of service.
...
And that leads into the reasons why the higher temp/low impedance parts have gained favor in recent years.
What spurred on the hobbyist rage towards the high temp, low impedance/ESR caps have been two things, neither of which you will find in decent older gear: the stolen-formula capacitor plague of a couple decades ago (where the wrong stolen formula prematurely failed, destroying thousands of boards), and the trend to miniaturization involving tight and poor placement of mainly dubious quality Chinese capacitors in more recent equipment.
People don't realize just how hot 85c is, let alone 105c. Way too hot to even touch. Those are temps more associated to a heavily leaking or shorted part. Most semiconductors have difficulty even operating at those temps. And virtually no decent consumer equipment ever runs near that warm. Again, there are exceptions, but I am taking about the norm.
A few years ago, a poster on one of the we-know-best-DIY-audio-science boards ridiculously pronounced that the best replacement capacitors for these consumer electronics supplies were automotive grade capacitors. Now maybe if your Marantz is sitting on a running gasoline engine, or your preamp runs three inches from a red hot exhaust manifold, those caps might make a difference. Or when the cheap Chinese caps on the oven controls of your Kitchen Aid oven crap out after an ill-advised 900F self-clean cycle.
But in the usual consumer electronic operating environment of 40-60c, a 135c automotive part is gross overkill, both in specification and to the wallet.
And applying the usual temperature/life exponential calculations, an 85c part will still last many decades in most equipment. A 105c one will likely outlive the installer, at least as to ambient temperature related declines.
Again, if you want to use a 105c part, do so. But don't deride the use of 85c parts that are just as good from a practical perspective in most situations. It's what the designer used in virtually every case ... unless they were working around a problematic design.
...
So after all of that commentary, how does it pertain to the OPs use of Silmics and similar audio grade electrolytic caps, even in the power supplies of vintage equipment? A controversial and derided practice elsewhere.
It means this: Most audio grade electrolytics are simply general purpose specification capacitors of premium build and materials manufacture.
And there is not a damn thing wrong with that. Or with them.
For emphasis, I'll state it again:
Audio grade electrolytics are principally general purpose capacitors in specification, but made with premium build quality and materials. They are not necessarily high ripple, low impedance or high temperature designs in a meaningful way.
And the practical fact of the matter is, they almost always don't need to be. Because few audio instruments, including their power supplies, demand more than what a general purpose capacitor delivers in such applications. Despite what the measurement obsessed hobbyists and self-anointed capacitor and circuit experts tell you. The reality in the field is very different.
Whether it is a waste of money spending an extra buck on a fancy Silmic or KA is a separate -- and electronically irrelevant -- conversation.
So more power to the OP and anyone else who chooses to use them. Just make sure the requirements of the original circuit do not require something more, and they are perfectly fine to use. Just like I did 30 and 40 years ago in using Atoms, and in not tearing out every 85c twist lock I saw.
And they just might sound better. Your ears will decide that. Most oscilloscopes are just not equipped with ears yet and don't measure hearing. Because something is not measurable today does not mean it is not there. It may mean the operative parameter has not yet been identified and quantified into measurability yet. Throwing psychoacoustics smears at those with a different experience doesn't solve the dispute. Double-blind testing is an interesting conversation changer in those situations.
Measurements, meet statistics.
I remember when the causes of digital timing jitter and its consequences to the early receiver chips were ignored or still misunderstood. And also ignored were the sound complaints about the earliest digital audio equipment in the 70s, because the measuring naysayers weren't measuring the relevant parameters yet. They weren't even aware of them. So maintain the perspective of history, and maintain an open mind when something is not readily measurable at the moment, and never deride the poor lay messenger. Sometimes, they get the last laugh.
Were we to listen to our gear only with our oscilloscopes, we would all be perfectly happy with our Panny FMs everywhere and forever ... until they finally leaked.
First-line Japanese capacitor manufacturers are not selling audio electrolytics to steal money selling "snake oil" to a handful of "audiophools". They are doing well enough selling billions of standard capacitors every month to large industrial customers. Trust me, they don't need the extra small money from selling a minuscule extra number of audio rated caps to enthusiasts.
It is because there are audio enthusiasts working at some of these companies who were in a position to pursue some more premium products to support the application, and pursue their passions. It's the same reason some engineers at JRC came out with the Muse series op amps, and TI actually does have an engineering "listening room" for their OPA models. Some electrical engineers are audio enthusiasts, and the two are not mutually exclusive.
These are also halo products to their line cards. Just like Chevrolet still makes the Corvette, but still depends on truck sales to stay relevant. Complaining that the Vette is too expensive and won't tow the boat like the trunk can is missing the point.
The most legendary of the audio electrolytics, the Rubycon Black Gate, is now long gone.
And to those following the industry, the lovely Elna Silmic line is now being discontinued, too.
So use them now if you want them. For a buck or so more than a general purpose cap, and if a general purpose cap was originally equipped, there's no harm whatsoever in doing so. And a lot of potential upside.
...
As a final aside, I'll leave you all with this. We have two classic CAL tube DACs. Both are pushing 30, and I don't want to deal with them a few years from now. So the ancient Taiwanese caps in them are getting replaced now. The other is a Sigma II. And the Sigma is far better laid out for parts placement than the Alpha (go figure). Plenty of room everywhere for just about anything. So it is getting some spare Silmics I still have on hand for the bulk supply caps. Big, huge caps, serving as main bulk filters right off the supply rectifiers. And the price is right.
Since CAL went out of business many years ago, there is very little technical data available on their designs, even among people still servicing them. Not even schematics or parts specs. It all has to be reverse-engineered off the board layouts. So in an abundance of caution, I checked into the original Elite branded caps originally installed in these DACs three decades ago. Elite is a Chinsan brand, an established Taiwanese maker of decent quality parts for over 50 years, and still sold today. Not as renowned or as respected as the first-tier Japanese or US makers. But way better than the mainland junk.
And after a lot of digging, it turns out that the Silmics have better ripple and ESR behavior than the original Elites did. The Elites were just general purpose miniature electrolytics (no surprise if you've followed this far). All these two classic DACs need are good general purpose caps to carry on for another 30 years. And the Silmics meet that task that just fine.
So would any of the Nichicon audio grade caps as well.
And any Nichicon or Rubycon general purpose one would, too.
Whether any particular one sounds any better is my business. Maybe, maybe not. But I know electronically, they are all just fine replacements.
Would a low impedance UPW or FC really do that any better, or longer, than the old Elites did? Probably not. For all the reasons covered above.
And the Elites didn't leak.
And if anyone wants to repost this very long commentary from a retired hand in the field over at the DIYaudio and Audio Science fora, you have my blessing to do so (I am not a member at those places, and I'm too old to start now). I'm sure all the angry measurement curmudgeons there would be suitably agitated by having this reality shoved up their test probe noses.
So after all of that commentary, how does it pertain to the OPs use of Silmics and similar audio grade electrolytic caps, even in the power supplies of vintage equipment? A controversial and derided practice elsewhere.
It means this: Most audio grade electrolytics are simply general purpose specification capacitors of premium build and materials manufacture.
And there is not a damn thing wrong with that. Or with them.
For emphasis, I'll state it again:
Audio grade electrolytics are principally general purpose capacitors in specification, but made with premium build quality and materials. They are not necessarily high ripple, low impedance or high temperature designs in a meaningful way.
And the practical fact of the matter is, they almost always don't need to be. Because few audio instruments, including their power supplies, demand more than what a general purpose capacitor delivers in such applications. Despite what the measurement obsessed hobbyists and self-anointed capacitor and circuit experts tell you. The reality in the field is very different.
Whether it is a waste of money spending an extra buck on a fancy Silmic or KA is a separate -- and electronically irrelevant -- conversation.
So more power to the OP and anyone else who chooses to use them. Just make sure the requirements of the original circuit do not require something more, and they are perfectly fine to use. Just like I did 30 and 40 years ago in using Atoms, and in not tearing out every 85c twist lock I saw.
And they just might sound better. Your ears will decide that. Most oscilloscopes are just not equipped with ears yet and don't measure hearing. Because something is not measurable today does not mean it is not there. It may mean the operative parameter has not yet been identified and quantified into measurability yet. Throwing psychoacoustics smears at those with a different experience doesn't solve the dispute. Double-blind testing is an interesting conversation changer in those situations.
Measurements, meet statistics.
I remember when the causes of digital timing jitter and its consequences to the early receiver chips were ignored or still misunderstood. And also ignored were the sound complaints about the earliest digital audio equipment in the 70s, because the measuring naysayers weren't measuring the relevant parameters yet. They weren't even aware of them. So maintain the perspective of history, and maintain an open mind when something is not readily measurable at the moment, and never deride the poor lay messenger. Sometimes, they get the last laugh.
Were we to listen to our gear only with our oscilloscopes, we would all be perfectly happy with our Panny FMs everywhere and forever ... until they finally leaked.
First-line Japanese capacitor manufacturers are not selling audio electrolytics to steal money selling "snake oil" to a handful of "audiophools". They are doing well enough selling billions of standard capacitors every month to large industrial customers. Trust me, they don't need the extra small money from selling a minuscule extra number of audio rated caps to enthusiasts.
It is because there are audio enthusiasts working at some of these companies who were in a position to pursue some more premium products to support the application, and pursue their passions. It's the same reason some engineers at JRC came out with the Muse series op amps, and TI actually does have an engineering "listening room" for their OPA models. Some electrical engineers are audio enthusiasts, and the two are not mutually exclusive.
These are also halo products to their line cards. Just like Chevrolet still makes the Corvette, but still depends on truck sales to stay relevant. Complaining that the Vette is too expensive and won't tow the boat like the trunk can is missing the point.
The most legendary of the audio electrolytics, the Rubycon Black Gate, is now long gone.
And to those following the industry, the lovely Elna Silmic line is now being discontinued, too.
So use them now if you want them. For a buck or so more than a general purpose cap, and if a general purpose cap was originally equipped, there's no harm whatsoever in doing so. And a lot of potential upside.
...
As a final aside, I'll leave you all with this. We have two classic CAL tube DACs. Both are pushing 30, and I don't want to deal with them a few years from now. So the ancient Taiwanese caps in them are getting replaced now. The other is a Sigma II. And the Sigma is far better laid out for parts placement than the Alpha (go figure). Plenty of room everywhere for just about anything. So it is getting some spare Silmics I still have on hand for the bulk supply caps. Big, huge caps, serving as main bulk filters right off the supply rectifiers. And the price is right.
Since CAL went out of business many years ago, there is very little technical data available on their designs, even among people still servicing them. Not even schematics or parts specs. It all has to be reverse-engineered off the board layouts. So in an abundance of caution, I checked into the original Elite branded caps originally installed in these DACs three decades ago. Elite is a Chinsan brand, an established Taiwanese maker of decent quality parts for over 50 years, and still sold today. Not as renowned or as respected as the first-tier Japanese or US makers. But way better than the mainland junk.
And after a lot of digging, it turns out that the Silmics have better ripple and ESR behavior than the original Elites did. The Elites were just general purpose miniature electrolytics (no surprise if you've followed this far). All these two classic DACs need are good general purpose caps to carry on for another 30 years. And the Silmics meet that task that just fine.
So would any of the Nichicon audio grade caps as well.
And any Nichicon or Rubycon general purpose one would, too.
Whether any particular one sounds any better is my business. Maybe, maybe not. But I know electronically, they are all just fine replacements.
Would a low impedance UPW or FC really do that any better, or longer, than the old Elites did? Probably not. For all the reasons covered above.
And the Elites didn't leak.
And if anyone wants to repost this very long commentary from a retired hand in the field over at the DIYaudio and Audio Science fora, you have my blessing to do so (I am not a member at those places, and I'm too old to start now). I'm sure all the angry measurement curmudgeons there would be suitably agitated by having this reality shoved up their test probe noses.
I have read the above two posts and will respond back when I have time.
The OP is from 2011 so obviously have way more experience with fit function etc.
Real quick I will say:
There are certain locations on the 2230/2245/2270 where cap choice impacts the sonics.
Also to be said is what I have enjoyed learning from the Pioneer/Sansui forums.
Pioneer in its "Tuning Fork" series has a section on cap choice.
It shows what cap variety was used in each location, and of note where in the circuit a "low leakage " cap is to be used.
See post #11
https://audiokarma.org/forums/index.php?threads/pioneer-tuning-fork-service-guide.249745/
Over on the Pioneer Forum @markthefixer rebuild rules noting:
Quote from Pioneer Forum
"UPW Series for Power supply and NON SIGNAL PATH.
UKL series for Signal path. UHE's for one or the other (fuzzy and tired now.)
Panasonic ECQ for bipolar stacked film to replace small value electrolytics in the signal path (dreaded blue Sanyo's) in 2.2uf and below."
I have done the above with very good results.
I also have rebuilt many many Advent 300 receivers and have come to learn much about the rebuild process.
Will check back in with the above location difference makers.
The OP is from 2011 so obviously have way more experience with fit function etc.
Real quick I will say:
There are certain locations on the 2230/2245/2270 where cap choice impacts the sonics.
Also to be said is what I have enjoyed learning from the Pioneer/Sansui forums.
Pioneer in its "Tuning Fork" series has a section on cap choice.
It shows what cap variety was used in each location, and of note where in the circuit a "low leakage " cap is to be used.
See post #11
https://audiokarma.org/forums/index.php?threads/pioneer-tuning-fork-service-guide.249745/
Over on the Pioneer Forum @markthefixer rebuild rules noting:
Quote from Pioneer Forum
"UPW Series for Power supply and NON SIGNAL PATH.
UKL series for Signal path. UHE's for one or the other (fuzzy and tired now.)
Panasonic ECQ for bipolar stacked film to replace small value electrolytics in the signal path (dreaded blue Sanyo's) in 2.2uf and below."
I have done the above with very good results.
I also have rebuilt many many Advent 300 receivers and have come to learn much about the rebuild process.
Will check back in with the above location difference makers.
Last edited:
I am following this thread by @patfont 
and for sure this is the holy grail for recap audio, especially for 70's SS cap coupled receivers and amps.
I have to do a recap on both Pioneer LX-880 and Sansui 350A, they are both cap coupled designs with 1000uF 35V filter caps and the main "big" 2200uF 63V (which I want to bump up to 3300uF) in PSU. The Pioneer is all Elnas stock inside and the Sansui is all Nippon Chemi-con. So what are you suggesting to put.
I live in Europe so I can order Elna Silmic's II that are going to 1000uF 50V and they have only Elna Lao 3300uF 80V. There are also Mundorf M-Lytics, and Nichicons FG,KG, KS available for ordering so please suggest what would you put in my place.
The Pioneer LX880 with the old Elnas sound sweet, lush, musical, open clear mids and highs and enough transparent
, but I feel that the lower bass is missing to enjoy the music at its fullest
.
I also have at home some unused old stock Nippon Chemi-con 3300uF 100V SME and 1200uF 63V LXZ, but after reading the disappointment that @patfont found with Nippon Chemi-cons I am not willing to solder them in my units.
and for sure this is the holy grail for recap audio, especially for 70's SS cap coupled receivers and amps.I have to do a recap on both Pioneer LX-880 and Sansui 350A, they are both cap coupled designs with 1000uF 35V filter caps and the main "big" 2200uF 63V (which I want to bump up to 3300uF) in PSU. The Pioneer is all Elnas stock inside and the Sansui is all Nippon Chemi-con. So what are you suggesting to put.
I live in Europe so I can order Elna Silmic's II that are going to 1000uF 50V and they have only Elna Lao 3300uF 80V. There are also Mundorf M-Lytics, and Nichicons FG,KG, KS available for ordering so please suggest what would you put in my place.
The Pioneer LX880 with the old Elnas sound sweet, lush, musical, open clear mids and highs and enough transparent
, but I feel that the lower bass is missing to enjoy the music at its fullest.I also have at home some unused old stock Nippon Chemi-con 3300uF 100V SME and 1200uF 63V LXZ, but after reading the disappointment that @patfont found with Nippon Chemi-cons I am not willing to solder them in my units.
Last edited:
gwiazd0r
New Member
Great reading, educational and on time for me as Im doing the recap of Luxman L-200. I have replaced all the caps exept the PSU once and still trying to source the 10,000uF 50V replacement (original is Elna with this specs). Im almost convinced I will try Nchicon KGs but having problem sourcing them. I live in Australia and there is no chance getting them here. Scanned rest of the world with same result. Well. Aliexpress got them but should i trust they original!
Can anyone suggest where to sourece them. Im also thinking up them to 63V to start with. Also some experts suggested to even use 12,000uF instead too as Nichicon are always on lower end os specs.
Looking forward for the help
Can anyone suggest where to sourece them. Im also thinking up them to 63V to start with. Also some experts suggested to even use 12,000uF instead too as Nichicon are always on lower end os specs.
Looking forward for the help
Great! You might want to take a look at this website also for suggestions, lots of good info on the L-200. (I have an L-10 & T-12)
https://www.hilberink.nl/codehans/luxman9.htm
What is the physical size of the original filter caps (D x H). Also, PCB mount, solder tabs, or screw terminals?
A lot of the Nichicon and Elna caps have been discontinued in the past year or so. Be careful not to increase the capacitance too much as it will draw more inrush current on turn-on and may stress the bridge diodes.
Looks like you can order from Digikey in Australia.
https://www.digikey.com.au/
Also Mouser.
https://au.mouser.com/
https://www.hilberink.nl/codehans/luxman9.htm
What is the physical size of the original filter caps (D x H). Also, PCB mount, solder tabs, or screw terminals?
A lot of the Nichicon and Elna caps have been discontinued in the past year or so. Be careful not to increase the capacitance too much as it will draw more inrush current on turn-on and may stress the bridge diodes.
Looks like you can order from Digikey in Australia.
https://www.digikey.com.au/
Also Mouser.
https://au.mouser.com/
Last edited:
You can also try Michael Percy in the US. I got some Nichicon KG Super Throughs from him last year. http://percyaudio.com/
In general, modern aluminum electrolytics are physically smaller than similarly rated caps from decades ago. Finding a cap with the same can diameter will usually require increasing ratings, whether it's working voltage, capacitance, operating temperature range, lifespan, or all of the above. Since the big power supply filter caps are often mounted using clamps, this can be challenging to find something that physically isn't too small.
First: It is always safe to increase the working voltage rating on power supply filter caps. In your case (originally 50V working voltage) you should not hesitate to go up to 63V, 80V or even 100V if it helps you find a cap with the correct diameter to fit the original clamps without modifications.
Second: Increasing the capacitance of power supply filter caps by 25% to 50% is a no-brainer. And you can safely increase by even more, but you are chasing diminishing returns. Increasing the capacitance will improve the smoothing of the unregulated DC voltage (decreases the ripple voltage), and will also help you with finding a cap with the correct diameter to fit the original clamps. Yes, do note that there is an increase in the inrush current at power up, but this will not overtax the rectifiers or the power transformer since it is only a transient event. If you want to read more about this, refer to the estimable Rod Elliot (I have no affiliation). He has an excellent "no-BS" article on his website regarding power supply design and addresses the myths around increasing power supply filter capacitance. If you are interested, look at section 5.3 here: Linear Power Supply Design - 5.3 Major Myth Regarding Capacitance
Good luck. Cheers! -Steve
I agree there is quite the mindless zeitgeist for the modern 105 c caps. I am guilty of this myself when it is obvious that recapping with these caps changes the sound, though it is interesting to note, that many times where the old caps are in good shape, they have esr, tan, etc... as good or better than the new modern caps.