Does the migration to SMD mean that cap-type debate will become moot?

[Mark Hammer]

Maybe I'm underinformed about what's out there, but it strikes me that there aren't all that many type of caps available in surface-mount form.  Or at least, the types of cap materials that lend themselves to SMD aren't identical to what has been used in caps for most of the pedals we have come to know and love from the last 50 years.

This raises a couple of questions in my mind:

1) Do SMD reissues of classic pedals sound the same as thru-hole originals?
2) If they don't sound the same, is it better, worse, or just "different"?  Will we be able to describe/measure the differences?
3) If the cap materials do not cover what we are accustomed to having, are there changes to design we need to start incorporating?
4) Will the advent of SMD alter the value of older thru-hole versions?
5) If they don't sound any different, does that make all the debate over cap-type moot?

[aron]

All good points. 4 - I am sure regular caps will go up in value due to this.
5 - Maybe for our pedals.
I think that I can hear the difference between all ceramic caps vs all film caps but maybe I am imagining it. Who knows at this point. I know a lot of the pedals I used to like were mostly ceramic anyway so go figure.

[armdnrdy]

3) If the cap materials do not cover what we are accustomed to having, are there changes to design we need to start incorporating?

I just checked Mouser, punched in ceramic, electrolytic, mica, tantalum, and film.

All of these types are available in SMD.

http://www.mouser.com/Passive-Components/Capacitors/_/N-5g7r/

[bloxstompboxes]

Working at Harman, we have both through hole in some of the older products and smd caps in the newer. I have seen electrolytics with flat bottoms sitting in a plastic saucer like base with pad like leads, surface mount tants with curved under leads, and the standard chip style smd caps. The only though hole component on the product I work on now though is an optocoupler.

[R.G.]

I'm convinced that the discussions about which type of [whatever] is as good as the old ones will continue. There is simply too much money to be made  by convincing people that X is better than Y,

[Rob Strand]

I hope this is an eye opener for some people, even professional engineers:

What many people don't realize is SMD caps are fairly crap compared to the traditional polyester film.  They are essentially ceramic caps.

For small capacitance parts where you can get NPO or COG dielectrics (say < 1nF) you are fairly safe.

When you get to larger values in X7R/X5R dielectrics beyond 100nF for 0402 size or beyond 1uF for 0805 (very rough figures) then watch out.

What's going on these days is the physical size of the capacitors is so small that the electric field in the dielectric is now strong enough to saturate the dielectric and cause significant non-linearity.   What this means is as the voltage across the cap increases the capacitance reduces from the rated capacitance. 

One thing to note is a higher cap voltage rating usually does not help.   You need to use a physically large package.  It is not uncommon for the capacitance to drop to 1/10th !!! the rated capacitance at the rated voltage.   Almost making a joke out of the rating.  You will find the AVX company doesn't bother making caps that do this and they use a much more honest rating system.

The non linearity can be a real problem for audio circuits, switch-mode power supply filters and even simple oscillators.

It becomes more of an issue when:
- the DC voltage (or AC peak voltage) becomes large
- the capacitance is large (requiring closer layers and hence higher E-field in the capacitor)
- the capacitor is physically small.

The issue is reduced when you choose a larger package which is in direct conflict to the size reduction.   If you don't need small parts don't use them if the above issue affects your circuit.

Instead of yabbering on further check out these articles:

http://www.edn.com/design/analog/4402049/2/Temperature-and-voltage-variation-of-ceramic-capacitors--or-why-your-4-7--F-capacitor-becomes-a-0-33--F-capacitor

http://www.electrical-integrity.com/Paper_download_files/DC11_13-TH2Paper_Novak.pdf

Also, go here and click on "Search by characteristics"

http://product.tdk.com/capacitor/mlcc/en/

Try:
size:  CC0603
value: 1uF
Dielectric: tick X5R and X7R

- Then press search

- Click on the first capacitor in the list

- Under "Characteristic graphs" find the DC-Bias Characteristic

This link might do the above:
http://product.tdk.com/capacitor/mlcc/detailed_information.php?lang=en&ref=jp&part_no=C1608X5R1A105K080AC

Now look how badly the capacitance drops off as the voltage increases.
The part is rated for 10V but the capacitance is already dropping at 1V.

People who have designed circuits with inductors have had to deal with inductors saturating with DC currents.   Now you have watch out for capacitors saturating with "high" DC voltages ; 10V is hardly high.

[anotherjim]

Hmmm...
Capacitance can fall as voltage increases. I've learned something there. I bet that's got a lot of people thinking "I can use that"!

[bool]

I hope this is an eye opener for some people, even professional engineers:

... loadsa stuff ...

Basically AVX and Kemet are supposedly OK, AVX having a niobium type "better" ... "tants".

SMD you can do practically - ceramics, tants (AVX having an audio-friendly type), small poly-film types (panasonic looks OK), and "classic" 'lytics, and "polymer" 'lytics (how is leakage current with these? last I checked it wasn't suitable for low-power circuits imho).

So there is some choice.

The obvious question is, as has always been - what's the SMD "mojo factor" compared to TH? Or P-T-P?

I know my own gut-reaction.

SMD has the "cheaply made in China" vibe, and the classic TH hass less of it.


(For RF immunity though, a well done SMD circuit has obvious advantages. For nerds).

[Mark Hammer]

I'm hopelessly naive in such matters.  All I know is that when I look in my parts drawers, some kinds of caps, of identical age, are dramatically larger than other types.  Granted, caps in virtually all types of materials have shrunk in size over the last 50 years, and higher voltage-rated caps will be larger than lower-rated ones.  Still, I ask myself, if they could have made these smaller years ago, wouldn't they have done so?  And it's certainly not because of the space requirement of thru-hole.  I have small 100nf ceramic caps with leads that are orders of magnitude smaller than metal film and polyester film types of the same value and voltage-rating.

That was, in fact, the basis of my question.  As all those remarkably feature-packed 1590A pedals amply demonstrate, SMD clearly offers advantages with respect to packing a lot of stuff in a small space.  But does it mean that our choices in caps become limited by the drive towards miniaturization?

If everything is still hunky-dory with little SMD ceramics, fine.  After all, we forget just how many classic pedals from the 60's and 70's relied on those big ugly flat discs.  But if there IS anything whose sonic qualities or behaviour either depend on, or are enhanced by, particular types of cap material, then we, and manufacturers have some choices to make.

I suppose we could debate the things that Rob mentioned.  The challenge to be addressed is not whether those things are true or not, but rather whether they actually matter, in audible ways, in our context.  I have no basis for thinking Yes or No.

Carry on.

Oh, and thanks for checking out Mouser, Larry.  I looked up Digikey, and saw something similar.

[Rob Strand]

The envelope of physics is definitely being pushed these days.  In the old days it there was no reason to go this far.  That being said larger valued ceramics were never great from a linearity perspective.

As bool mentioned "small poly-film types (panasonic looks OK)".  These seem OK to me too.  The thing is though they are actually quite large.  When open up equipment and see SMD caps and they are usually those little brown MCC types!

[Puguglybonehead]

I've got a couple of "Learn To Solder SMD" kits here at home, but I haven't built up the courage to go at it yet. I also have a LPB-1 SMD kit (courtesy of a member here) and a (partly) SMD tube preamp kit (that uses a Russian direct-heated submini tube that is supposedly highly resistant to microphonics) waiting to be built.

I'm pretty excited about getting into SMD components. The main attration for me is that they make it possible to adapt a lot of the popular effects circuits to onboard guitar mounting. (or whatever else you want to mount them in) Artec already has a large lineup of onboard effects available. Their onboard distortion sounds pretty good. Planning on mounting mine in "metal" guitar one of these days when I get around to building it. GFS also has a line of onboard stuff, including their own onboard chorus and an onboard delay. (which may or may not be based on designs from this DIY community)

SMT isn't all that new. It's been standard in the computer motherboard world for many years. Cellphones and tablets wouldn't exist in their current form without it. I get the impression that reliability isn't that much of an issue. It probably means having to learn more about choosing the appropriate components. At least PT2399s are easy to find in SMD format. I just wish MN3007s and MN3207s came as SMD. I'm sure there will be a lot of hybrid circuits being attempted too. I could see an onboard germanium fuzz being a fun thing to have, especially if you put it in something like one of those Mosrite "Strawberry Alarm Clock" guitars.

[PRR]

> When you get to X7R/X5R dielectrics ..... then watch out.
> Under "Characteristic graphs" find the DC-Bias Characteristic

This:


Yes, X7R has ALWAYS been bad for audio.

I can't hear different caps; except I CAN often hear an X7R coupling cap. Even a 3/4" disk-cap from 50 years ago.

I dunno about "electric field in the dielectric". You may be right. My caveman explanation is that NP0 ceramics are "nearly perfect, inert", like glass; and that the super-high-value ceramics are "salted" with some stuff which is NOT stable.

X7R is a fine bypass cap when you need "lots" of capacitance but do not really care how MUCH you have. If 0.01 or 0.005 is all the same to you, X7R may be fine. But if you need capactiance consistent to a semitone (so an audio filter doesn't wander or warble), X7R is not your dog.
_________________________________________

For "16V SMD cap", Mouser is showing:
* Aluminum Electro (1,177)
* Film (423)
* Ceramic (8,987)

i.e., in SMD ceramics out-number film and e-caps more than 5:1.

However 1,000+ e-caps is hardly a narrow selection. 400+ film caps is plenty for most projects. I remember when DigiKey's *entire* cap-list fit on one page, say 200 caps, and we thought that was good. I've built a lot of serious audio using only RadioShack stock of a few dozen (at most) caps.

I'm not sure we have fewer "good" caps in SMD. We sure have a lot more "maybe bad (for audio)" caps to weed through.

And maybe half of those "SMD ceramic" caps may be fine, except mostly under 1,000pFd and we only have one or two small caps in a project.

[PRR]

> if they could have made these smaller years ago, wouldn't they have done so?

Generally, yes. Cap size is not vital in your grandmother's kitchen radio; the speaker and tuning-dial dominate the space.

But smaller stuff is *usually* cheaper, due to less material.

OTOH, too-small is too fiddley, takes skilled operators, is more prone to rejects.

So for any general construction, there is a "minimum" size and a "lowest cost" size.

Of course in your accumulated stock-box you have both types.

> some kinds of caps, of identical age, are dramatically larger than other types.

There may be differences you can't see.

In "film caps" there are various films and foils.

It seems to be easy to roll polystyrene very thin yet uniform. PS caps can be smaller than most. (But soldering can ruin them.) Paper can't be thin and uniform, so paper caps are larger. Paper largely vanished when Mylar and friends got affordable.

The obvious conductor is metal foil. But you can only roll foil so thin; or rather, hyper-thin foil tears and blows-away in the winder. They developed a process to "spray" molten metal onto plastic film, "metalized film". Not only is the conductor thinner, they felt they could make the film thinner. If that makes weak-spots in the film, which arc-over, the arc quickly burns-out a bit of the very thin metalization, and the breakdown "heals". This does happen in power circuits, though not in audio circuits.

AC electric motors have Start and Run caps. Start-cap may be 150uFd, Run cap may be 15uFd, yet they may be similar size. The Start cap is typically an electrolytic, compact but with high losses, and would blow-up if worked steady, but Start is only 1 or 2 seconds. The Run cap has to work forever, so was historically a film cap, large but with low internal losses.

There is another extreme. Caps for welding and especially for A-bombs must pass huge peak currents. Common caps either won't deliver high peak current or will die trying. Takes very hefty construction to deliver 100,000 Amps for a micro-Second for precision detonator timing.

[Rob Strand]

I dunno about "electric field in the dielectric".

See Page 15.
http://pages.venkel.com/rs/venkelltd/images/Testing%20and%20Measurement%20Practices%20of%20High%20Capacitance%20Capacitors.pdf

I forgot to mention a "thicker" part often helps because the capacitor has more volume.

In defense of MCC caps:  They have very low loss and are great for switch-mode power supplies.  I've built switch-mode power supplies with 2 to 4A output using only a couple of output caps.  A small 4A power-supply with very little power dissipation definitely wouldn't happen in the old days.

[StephenGiles]

When all these SMD fanatics get tired eyes later in life, they will all wish it had never been invented. Luckily I have enought normal stock - thanks TI!!!

[Transmogrifox]

What the responses in this thread indicate is that the cap-type debate will not be come moot.  It just got a lot more interesting.

With more quirks in components being used there is bound to be more potential for mojo.

When all these SMD fanatics get tired eyes later in life, they will all wish it had never been invented.

A magnifier lamp does wonders for relaxing my eyes working with SMT.  If my eyes get worse, and the strain is still too much then perhaps I will invest in a microscope.

I think I am going to suffer from arthritis doing fine work with tweezers and always having my hands bunched up like that before my eyes bother me.

In either case I agree there's a price to be paid -- but it's pretty insignificant compared to other hobbies.

[Metaldestroyer]

WIMA caps come in SMD but they're not tiny and they're not cheap.

[wavley]

What I can tell you, maybe not so relevant to the discussion, is that the Kemet MLCC .1uF caps hold their capacitance better than Panasonic .1uf film caps when cooled to 15 Kelvin, so they make a better cryogenic decoupling cap.  Also, somehow the ringing of a certain new cryogenic refrigerator found the resonant frequency of the panasonic caps and excited them to the point of breaking gold bond wires attached to them, which is why we had to try the Kemet caps and found they were better upon cooling.

I know that nobody here is cooling a pedal to near absolute zero, but those little Kemet 0805 MLCC caps are tough little boogers. 

[Puguglybonehead]

I have seen standard through-hole electrolytics being used on some SMD boards. No idea why. Maybe there is an issue there.

[karbomusic]

I know that nobody here is cooling a pedal to near absolute zero, but those little Kemet 0805 MLCC caps are tough little boogers.

The mental image of a near-absolute zero pedal on a stage made me giggle. I could just see the band rider for that show.