• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Selecting Replacement Capacitors

Plan "c"

Found this. Another example of careful what you ask for.

Application Guide, Aluminum Electrolytic Capacitors (CDE)

Seems to be THE definitive white paper on the subject. A couple interesting quotes ...

RATED DC VOLTAGE

Rated DC voltage is the voltage marked on the capacitor,
and it is the maximum peak voltage including ripple voltage
that may be applied continuously between the terminals
and over the rated temperature range. Higher rated
voltage capacitors may be substituted for lower rated voltage
capacitors as long as case size, DF, and ESR ratings
are also compatible.


FORMING

The anode foil carries the capacitor’s dielectric. The
dielectric is a thin layer of aluminum oxide, Al2O3, that is
chemically grown on the anode foil during a process called
“formation.” Formation is accomplished by pulling the
anode foil on rollers through an electrolyte bath and continuously
applying a DC voltage between the bath and the
foil. The voltage is 135% to 200% of the final capacitor’s
rated voltage. The thickness of the aluminum oxide is
about 1.4 to 1.5 nm for each volt of the formation voltage,
e.g., the anode foil in a 450 V capacitor may get a formation
voltage in excess of 600 V and have an oxide thickness
of about 900 nm. That’s less than a hundredth the thickness
of a human hair.

Formation reduces the effective foil surface area because
the microscopic tunnels are partially occluded by the oxide.
The tunnel etch pattern is adjusted by choice of foil and
etching process so that low-voltage anodes have dense tunnel
patterns compatible with thin oxide and high-voltage
anodes have coarse tunnel patterns compatible with thick
oxide. The cathode foil is not formed and it retains its high
surface area and dense etch pattern.


Granted - a quick read by me, but here's my take on it. Voltage isn't all that important - maybe once, but the forming process that was once a big bugaboo is no longer an issue. The caps are formed in the factory at much higher voltages than the real world will ever throw at them and should stay stable over a broad range.

With that - common sense would dictate one uses as close to the original rated voltage as is available. With that in mind, sticking with "audio grade" capacitors designed for signal path applications <should> be better than substitution of general purpose caps of the rated voltage value.

So there. I think. <G>

That doesn't cover the point that the manufacturers can play voltages close to the vest to keep costs down. It does address the fact that out of spec voltages can cause oscillations and other not so good nastiness depending on the application.

PLAN C

- Elna Silmic II where available (any voltage)
- WIMA poly where they'll fit OR Nichicon KW when all else fails

Bypass caps strictly Polypropylene, either .047 or .1 depending on availability. The Panasonic ECQ come highly recommended for bypass caps, but come to find out those are polyester? Still a bit confuzed there.

So ... any closer to right?

Baby steps ... baby steps.
 
Last edited:
Panasonic Bypass Caps

Found these on Mouser

Panasonic ECQUA (Class X2) .1µf @275v

Now we're talkin' ...

Polypropylene, and I like the price too - less than half of the polyester cap. Package is a bit bigger than the plain old red gumdrop, but probably not enough to cause any major issues.

Couldn't find them in .047, which might be why I see the .1 recommendation here and there?
 
Found these on Mouser

Panasonic ECQUA (Class X2) .1µf @275v

Now we're talkin' ...

Polypropylene, and I like the price too - less than half of the polyester cap. Package is a bit bigger than the plain old red gumdrop, but probably not enough to cause any major issues.

Couldn't find them in .047, which might be why I see the .1 recommendation here and there?

X2 caps are generally used for AC (positive to ground not chassis) to reduce RFI/EFI. You see them in use on your AC mains in before and after the power switch...

Cheers,
Bob
 
OK then ... Plan "B" ... v2

Had these suggested to me.

Panasonic ECQ-V - .047µf - Polyester Stacked Metalized Film

And right behind that ... Why bypass?

Once again, I <hear> it tends to smooth transients and add presence. Which may (or may not) any longer be true what with upgrading to audio grade electrolytics and improvements in the quality of what you can get now as compared to what was available back in the day.

<punt!>
 
Hi, I've ben a voyeur on this thread 'till now. I have this to add.

sKiZo, it's quite likely that you're over-thinking the issues. Just buy some caps and put 'em in. Just make sure that:

1) the new caps are rated at a higher voltage than the old ones AND

2) that they will fit in the allotted space.
 
sKiZo, it's quite likely that you're over-thinking the issues. Just buy some caps and put 'em in. Just make sure that:

1) the new caps are rated at a higher voltage than the old ones AND

2) that they will fit in the allotted space.
:yes: I agree :)
Just go for Elna Silmic, Nichicon when elna are not available, and Panasonic FM in large power supply caps, or just the Elnas. It will be OK anyway.

Elnas are quite big, bigger than older caps in many cases, so check the size before.

Regarding the by-passes, I-ve read that 1/1000 of the main value is a good starting point to by-pass large electrolytics.
I'm not an expert, but I guess a 0.1uF by-pass in a 2500 uF cap won't do any difference... I've added 2.2 to my 1000 uF
 
:yes: I agree :)
Regarding the by-passes, I-ve read that 1/1000 of the main value is a good starting point to by-pass large electrolytics.
I'm not an expert, but I guess a 0.1uF by-pass in a 2500 uF cap won't do any difference... I've added 2.2 to my 1000 uF

Yes - I am overthinking things here, but I'm like that ... somebody shoots me, please!! <G>

I did place an order for Elna Silmics last night. Between Parts Connection and Handmade, I was able to get good matches on most of what I need. Couple exceptions, but apparently the stereo gods aren't into making it TOO easy. I may have to get creative with a couple of the off the wall values.

BYPASS CAPS - There, the internal debate rages on. Do I even need them with new quality caps installed? Once again, with all the fun this thing's gonna be, I want it to smoke ... er ... strike that ... scream! It's looking like the Panasonic ECQ-v stacked metalized polyester gumdrop may be the best candidates. Which leaves size.

I'm familiar with the 1/1000 rule, but ... I also see yet another "rule of thumb" that says it isn't all that critical in the signal path and that you can usually get good results with either .047µf OR .1µf ... so ... which one? There again, I <hear> that depends on what you're going for. If you're looking to tighten the bottom end, go with the .1µf ... if you're looking to smooth the high midrange and add presence (opening up the soundstage), use the .047µf. In this case, I'd then go with the .047µf as bass isn't an issue, and the 9001 seemed a bit shrill on the top compared to what I like to hear. I'd also planned to concentrate on the .47µf thru 4.7µf as I'm told that's the critical range in the signal path. Please don't quote me on any of this, but DO chime in if there's any truth to it.

Good time to add a disclaimer to this thread:

HALF THE LIES I TELL YOU ARE NOT TRUE


Here's some more interesting reading (as if things aren't muddy enough already). <G>

http://www.intersil.com/data/an/an1325.pdf
 
Last edited:
Hi, I've ben a voyeur on this thread 'till now. I have this to add.

sKiZo, it's quite likely that you're over-thinking the issues. Just buy some caps and put 'em in. Just make sure that:

1) the new caps are rated at a higher voltage than the old ones AND

2) that they will fit in the allotted space.

+2. What ever way you go, what you put in will be better then what is in there now and 90% (at least) probability they are better then the parts were when they were new.

It looks like you are letting the best become the enemy of the good. You could end up stressing over this so long that it never gets done.

It will sound great as long as the capacitance is right, the parts are installed correctly and the replacement part is rated at or over the original. Keep in mind - you work is making things better.

Shelly_D
 
the best way to decide which camp you may be in would be to have a test situation where you can listen without bypass caps and then with them on some key capacitor(s) in your test amp, changing only that one thing.

Problem there is what works with one circuit design might not work with another. Arghhhhh!!!

<ohmmmmmmmmmmmmmm> There ... I'm better now.

It looks like you are letting the best become the enemy of the good.

The good of the one outweighs the good of the many? <G>

OK ... uncle! I'm thinking it would be best at this point to let the debate rage on, forget the bypass caps, and just go with the original design. When all else fails, go with what a dedicated group of professionals tasked with creating a TOTL receiver came up with.

Well ... maybe just the tone boards ... :cool:

** Speaking of tests ... I do happen to have an 8080DB sitting on the shelf. Basically the same design as the 9001, minus a couple channels. Same tone boards and amp. It's even got the problematic pass through connections on one of the boards. I really should experiment with that one first. Except for the fact that it's easily as complex as the 9001 and just one of these may have me bouncing off rubber walls ... <G>:nutz:
 
A pause in the QRX action. I pulled a QR4500 off the shelf and will be working on that while I overthink whether I'm overthinking things here.

Got another thread going on that one.

http://www.audiokarma.org/forums/showthread.php?t=401471

I already had the caps for the QR sitting in a box for a while, but this thread had me making a couple minor tweaks to the list. Had to make an order to Mouser for what I couldn't cover in the Silmic line for the QRX, so what da hey, eh. The QR4500 will be a good candidate for running a comparison on bypass caps vs no bypass, as I just so happen to have two of them to play with ... well ... three really, if you count a donor box.
 
sKiZo:

For the past week you have been going back and forth about these caps and it seems (at least to me) like you have made little to no progress. Just how audible do you expect these changes to be. Consider this: Most speaker systems cause distortion. Typically they come in at somewhere between 4% to 10% depending upon the design. Even if yours are an incredibly low 2% distortion, they will mask the changes likely based upon the selection of cap 1 over cap 2. Many excellent amps, well respected on this forum are manufactured with the parts selection you see on Mouser, Allied, Newark, et al. Whatever you decide WILL be an improvement over what you have now. In the end, you will enjoy the music long after you forget exactly what you have been agonizing over this past week.

Stop worrying over whether the Sprague or the Panasonic is better at this point. Pick the devices, install them and lean back and listen. You will be better off in the long run without the stress this is causing you.

Shelly_D
 
STRESS?
STRESS??
WHAT STRESS?????


You'll be happy to know I've found ways of dealing with it. Why, just today, I spent a good couple hours playing with my nuts.

(Had a good crop off the hickory tree this year. Got them all shucked and laid out in the greenhouse. Should be dry enough for roasting by spring. Mmmmm ...)

And yes, I tend to agree on the "when is enough enough" argument. Then again, if I pay attention to the details in the receiver, that may be just enough to compensate for all the other horrible inconsistencies I introduce with my post-processing. The patch cables alone have got to hurt - last count, I'm talking 46 of em. <G>

Oh well. Got my care package in from Handmade yesterday, so I'm just waiting on the rest from Canada. Maybe while I wait, I'll go over and play with my neighbor's nuts.

(She's got an english walnut tree and says I can have all I want.)
 

Some interesting reading ... a few of the highlights:

BYPASSING

Bypassing is used to reduce the noise current on power supply lines.

Bypassing reduces the rate of change of the current through the inductor.

Bypassing provides a secondary high frequency path that shares as little inductance with the primary path as possible.

More is better - better to use many small caps than one large one. The result is much lower inductance.

Just tends to reinforce a lot of what I've heard about how bypassing can help smooth a lot of the transients that can bleed into the listening environment.

DAMPING

It is not necessary to provide damping for every bypass capacitor; rather a single large capacitor should be added to each board and damping provided for that capacitor. This is shown in Figure 9. The single large capacitor must be at least 16 times larger than the sum of all of the high speed bypass capacitance connected to the supply to completely
dominate in the damping resistor calculation. If the capacitor is large, and the inductor is small, then often times the equivalent series resistance (ESR) of the capacitor will be large enough to damp the network. Damping is typically provided on the power supply side by the large electrolytic capacitors at the output of the regulators.

XFORMER was talking earlier about having ONE bypass bridging the rails on the main filter/rectifier assembly. I imagine that's what this is all about then?

* As always, I probably saw pretty much what I wanted to see and ignored the rest. Also probably true that I've managed to paraphrase most of this completely out of proportion. Hey ... read it yourself. <G>
 
For the 3.3uf caps in the signal path someone mentioned Wima, but if you want you can go a step further and add a couple of 3.3uF 250Vdc Claritycap ESA Range polypropylene in some key positions, only 23mm x 28mm.
 
A whole bunch of steps forward - not sure which is more impressive ... the specs or the price! $9.33 x29 ... that's, er ... a LOT! Almost what I've got into the whole project already.

I'm thinking if I went that way, the old box just got a whole bunch less "key" positions. <G>
 
If it has input blocking caps, that's a nice place to start :)

Just a couple in each channel on the main amp board could make a nice difference, don't you think?

A lovely receiver by the way :thmbsp:
 
Last edited:
Most assuredly - Also, speaking of speakers ... a good thing to keep in mind when I get around to recapping the crossovers on my McIntosh XR16s. Those are getting a bit long in the tooth too. There we're only talking about $300.

Which is, come to think of it, pretty amazing for SIX capacitors. That's more than I recently paid to get both the big woofs rebuilt. wOOF!

If anything, I'm now getting a fairly good idea of where I'm willing to draw the line ... and it's got as much to do with the pocketbook as it does audio quality. <G>
 
What caps are ya' going to use in the McIntosh's?

I used pretty cheap caps (Obligatto) in my little speakers, but still set me back over $200.
I agree, budget plays a big part.
 
75µF100V
20µF100V
50µF100V

Some fairly substantial values. Times two of course.

That also raises another issue in that I do run a lot of post processing. I could put the <cue the trumpets> BEST possible caps in the 9001 and still have to deal with 30 year electrolytics in all the rest of the gear.

Just took a peek inside my equalizer. The cap count on that bad boy is right up there too. Heck, I could spend the rest of my days just cleaning wipers.

yamaha-q2031.jpg


And there is that little thing with the 46 patch cords.

At least my OPPO SACD is fairly new ... although ... could it be ... better with better caps?

:yikes:

(somebody shoots me please)

Which leaves me leaning towards relatively good audio grade caps for ALL! The Elna Silmic IIs fall into that category for the values available. Best bang for the buck all across the board(s).
 
Back
Top Bottom