Tantalum Caps

[AFF]

Okay from what I know/heard the range of good to best for caps is:

Metal Film (silver mica : is this right?)
Poly Film
Tantalum
Electrolytic
Ceramic

Please let me know if I messed this list up. Any feel free to include any if you think I left one or two or more out of the list.

I recently read that tantalum caps are a bad upgrade cap ... that's the main reason for this post. I might have thought wrong when I had to use tantalums in my pedal.

Thank you all for your insight!!!

[R.G.]

Okay from what I know/heard the range of good to best for caps is:

The problem with "what I know/heard" is that there is no underlying information there. Worse yet, changing caps out for dubious improvements has become a regular snake oil market in the hifi world. It shows every sign of doing the same here.
What matters about caps is
- the spreading resistance of the conductive electrodes compared to the charge/discharge currents they have to handle. For high power pulse caps, sheets of thick copper are the only thing with low enough resistance. For low-signal caps, a whiff of aluminum almost thin enough to see through and vacuum deposited on plastic film is fine. You have to match the spreading resistance to the job.
- the insulating material(s); some materials have quirks about them that affect signal flow in some situations but not in others. There is not a universal goodness factor which makes some sound good, some bad. After all we cannot even write down what sounds are good and what are bad.
- how the physical construction of the cap affects the workings of a cap. If the cap is made so that signals have to travel in a spiral to get from input to output that can introduce extra inductance which ...may... in some circumstances be audible.

Metal Film (silver mica : is this right?)

"Metal film" means "a layer of metal deposited by some means on an insulating film. Generally "metal film" means "metal film deposited on plastic film of one of a number of types". It is not silver mica. Silver mica is sheets of the mineral mica with metallic silver deposited on it. They're not the  same. Silver mica is generally quite high voltage and was a longtime favorite in tube radios in the last century.

"Stacked foil metal film" is a kind of metal film with stacked layers instead of wound up strips for low inductance and low spreading resistance. They're plastic film of several types.

Note that "metal film" is more accurately a term for resistors, where the conduction is done by an actual film of deposited metal. Electrolytic caps are the only caps used in electronics I know of where one of the actual electrodes is not literally a metal.

Poly Film

A variety of metal film where the insulation is one of a number of plastics. The difference in the term from "metal film" is that one describes the conductive plates (metal) and one describes the insulator (poly-) the poly being the prefix for polystyrene, polyester, polycarbonate, and poly- several others. The insulating quirks and construction quirks vary from part to part and maker to maker.

Tantalum

This is an electrolytic capacitor made from the rarer tantalum instead of the ubiquitous aluminum. The reason for this is that tantalum pentoxide has a higher dielectric constant than aluminum oxide, so for the same volume of metal, electrolyte, etc, the tantalum cap has a bigger capacitance, and since it's smaller, lower parasitics. Tantalum pentoxide is also a bit more stable than aluminum oxide in use. But there are quirks in other ways.

Electrolytic

This is the generic term for all capacitors made from metals which have an insulating oxide film grown on the outside of the metal plates by electrochemical (i.e. electrolytic) means. Of these, only aluminum and tantalum are commercially viable, and aluminum is so common that electrolytic is commonly used to mean "aluminum electrolytic". It's a way to get a very thin insulator and high capacitance per volume. You pay for that by it not being able to withstand being reverse biased, so the voltage on an electrolytic cap must always be one polarity. Still there are many applications where this can be used. Like all insulators, aluminum oxide has quirks.

Ceramic

Big surprise - the insulating material is one of a family of ceramics where the dielectric constant and other features like temperature stability and dielectric constant vary depending on what blend of ceramics you pick. Also with quirks.

Please let me know if I messed this list up.

I don't know that you messed it up - you just are expecting things to be more simple than they are.

I recently read that tantalum caps are a bad upgrade cap ... that's the main reason for this post. I might have thought wrong when I had to use tantalums in my pedal.

It is in general impossible to upgrade some pedal by subbing in highly regarded capacitors, or resistors (i.e. carbon comp versus carbon film) or transistors (putting in germanium) without also knowing what the circuit is doing and knowing that what you put in will have some noticeable effect. Put another way, you can't make a pedal sound better by rubbing germanium on it. Or tantalum, or carbon comp for that matter.

If you KNOW that noise is a problem in a given pedal and you KNOW what parts to replace, you might get lower noise by changing to metal films. If you THINK that carbon comp resistors will make a pedal sound "brown" or vintage or tubey, you are very likely to be disappointed if you're honest with yourself and highly likely to get more noise from the swap. If you KNOW you're having problems with the quirks of aluminum caps you MIGHT help with subbing in plastic film (AKA metal film or poly). If you THINK that polypropylene caps improve the sound of everything that has them, you're likely to be disappointed.

What some people want you to do is to upgrade your pedals by rubbing banknotes on them.  Sometimes that works, sometimes not.

[El Heisenberg]

In other words, dont sweat the small stuff.

[MicFarlow77]

What some people want you to do is to upgrade your pedals by rubbing banknotes on them.  Sometimes that works, sometimes not.

Best way I have heard it put yet!!!!!

[AFF]

F-ing AWESOME STUFF HERE!!!
Answered every single question I could have thought of asking.

LOL I still like the way the tantalums look. Anyways I understand almost completely now and I can save even more on my mods!!!

Correct me if I am wrong ... (if the expense is minimal) say I swap a .047uF cap for a .001uF and instead of using a ceramic I bought a tantalum for 1 cent more than a new ceramic. This would be okay to upgrade. The reason I am asking is because you mentioned a downgrade to get a more vintage sound (like a MIJ mod getting closer to earlier specs) I just feel like I might slightly alter the tone if I "downgrade" in quality.

All this I kind of take from running sound for awhile ... if your tone has some issues for example some weird buzz ... when I mic you into the PA it's gonna pick it up and make it even worse. I guess I was worried the same thing would happen in a pedal.

I think that has satisfied all my questions (except the one in the beginning of this reply)


Thank You so much!!!
-AFF

They shoudl move this post somewhere so more people can read it!!! Good info!!!

[Taylor]

RG, I once saw you post something like your above post, except towards the end you said something like "but don't just buy ceramics because they are cheap". Can you elaborate on that a little? Why should I not use ceramics if the others aren't really a noticeable upgrade?

[R.G.]

LOL I still like the way the tantalums look. Anyways I understand almost completely now and I can save even more on my mods!!!

Liking how they look is a more logical reason to do some mods that some pseudo-explanations I've heard.  

Correct me if I am wrong ... (if the expense is minimal) say I swap a .047uF cap for a .001uF

It depends. That's where that stuff about needing to know what the circuit needs comes in. If the 0.001 was being used for frequency sensitive filtering, no, don't do the swap because that will throw the filter frequency way off. If it's being used for power supply decoupling, in general more is better, except that all capacitors have a maximum frequency they can bypass, and the bigger the cap value, the lower that frequency is. And for decoupling very high frequencies, ceramic disks happen to have a very low internal resistance, which makes them better for decoupling. But ceramic disks drift in capacitance a lot, so they're not so good for specific-frequency filters; poly films and especially polystyrene are good for that. Polystyrene is especially stable.

and instead of using a ceramic I bought a tantalum for 1 cent more than a new ceramic. This would be okay to upgrade.

Probably. Especially for power filters. Tantalum is good for power decoupling. It's just expensive and polarized.

The reason I am asking is because you mentioned a downgrade to get a more vintage sound (like a MIJ mod getting closer to earlier specs) I just feel like I might slightly alter the tone if I "downgrade" in quality.

You might alter the tone, but it's usually not in the somewhat-vague "quality" that it happens. It's usually in the tolerance of the capacitance. Capacitors are much lower-precision devices than the resistors we use. Anything worse than 5% is a poor tolerance resistor. But 5% tolerance is about the best you can commonly get capacitors. Electrolytic aluminum is better than it used to be, but it was at one time +80%/-20%. Plastic films are generally where you get the best performance per dollar in terms of tolerance, but NPO or COG ceramic caps are high precision (5% to 2%) and very stable over time, drifting very little with temperature. On the other hand, Z7R composition ceramics pack in a lot of capacitance, but are 10% tolerance and drift about 5% with temperature. Even tiny changes in the frequency-determining caps of EQ sections (like tone stacks, EQ setting caps, etc.) can make noticeable differences in tone.

I guess what I'm getting at is that it's quite difficult to look at, for instance, a 0.1uF cap in a circuit and say with any assurance whether going to the expense of 0.1uF polystyrene or teflon would help, whether stacked foil would improve performance, or whether tantalum would be an upgrade or downgrade. Upgrade and downgrade to me don't mean "costs more/less and has more precision/reserve capacity". To me they mean "meets the circuit's functioning needs and precision of operation better."  Kinda like car wheels. Let's say you could get a car wheel made of pressed steel, cast and machined aluminum, or solid gold. Which is an upgrade? The gold certainly isn't unless you can pull it off and sell it for the weight of the gold. Gold is too heavy and too soft for that use, and the increased unsprung weight would eat up tires and suspension elements, as well as taking a ding in the rim from every pothole you hit and probably caving in on the forces from hard cornering. Aluminum might be an upgrade in performance, as aluminum is 1/3 the weight, 1/3 the strength, and three times the price of the equivalent function in steel on a structural basis. You might get a lower unsprung weight if you're careful and the same strength. But mostly aluminum is a visual-only upgrade. Steel wheels are generally the best overall compromise in performance, strength and expense.

Compromise. That's a worthwhile concept. "Upgrade" implies that you can simply make a change and things will get somehow better. Unfortunately, very often there are competing requirements for each part or its value beyond just what is most or least costly.

All this I kind of take from running sound for awhile ... if your tone has some issues for example some weird buzz ... when I mic you into the PA it's gonna pick it up and make it even worse. I guess I was worried the same thing would happen in a pedal.

You're correct there - a tonal oddity which might be forgiven in a small combo will sound horrible when it's close miked and given some real muscle.

With experience on a board you probably also know about compromises and knowing your materials (i.e. performers and their equipment) too. Some singers have a voice that almost doesn't need amplification. Some sing very quietly and it becomes quite a challenge. I heard that it was very difficult to set up Donovan Leitch in concert because his singing was very, very soft and quiet. A good mike for someone with a really loud voice with operatic-level projection would not be an upgrade for miking Donovan, no matter what the tonal qualities. A mike that's got a little bit of a dull top end might work OK with someone miking an overly bright, ice-pick-y small amp.

[R.G.]

RG, I once saw you post something like your above post, except towards the end you said something like "but don't just buy ceramics because they are cheap". Can you elaborate on that a little? Why should I not use ceramics if the others aren't really a noticeable upgrade?

A couple of things. One is that tolerances thing. Some - but not all - ceramics drift a lot with temperature and age, and have a bad tolerance to start with. It makes little sense to carefully design, say, a speaker simulator EQ and then have your caps muck up the filter alignment. It might even sound good - but the second and succeeding copies will NOT sound the same.

Then there's the whole "what is an upgrade?" issue. An upgrade is easy to idealize if you are dealing with single-variable issues where there is a clear target to be achieved. In a passive LC filter, for instance, 1% capacitors are an upgrade over 5% caps, no question. But is 2% ceramic better than 4% poly? Maybe not. How much drift, how much aging, do you allow tweaking inductors by hand, will the caps be potted (and thereby change their value slightly)?

For things where there are competing requirements, it is darned hard to say what "upgrade" means. It may have to mean "mediocre at everything, but really bad at nothing". A good example of that is a tube amp output transformer. Many of the requirements flatly contradict: thin laminations make for lower eddy current losses (Good!) but decrease primary inductance (Bad!), have lower iron distortion (good!) but make you use more copper (bad) to make up for the loss of the inductance. More turns of copper wire give a higher primary inductance (good) but increase leakage inductance (bad) at the same time. Splitting a coil horizontally reduces leakage inductance (good) but increases parasitic capacitance (bad). I use the OT example just because it's so easy to get so many competing requirements.

At the bottom of this is the question - what is quality? Zen and the Art of Motorcycle Maintenance aside, there is a good answer. That is - quality is conformance to requirements. Spending more on parts with exotic materials or just different materials doesn't help if they move your circuit's performance away from what it's supposed to be doing. Likewise, if your circuit is already doing what it's supposed to, swapping in something for more money doesn't help and may hurt. However, you have control of the requirements. If one of your requirements is "looks cool", then by all means, swap in Screamin' Yaller Zonker caps. If one of the requirements is "make the other circuit hackers think you're an incredibly wize, clever and adroit circuit wizard because you can get wired made of pure drawn unobtainium", by all means get the wires, and feature them. But be honest with yourself about what you're improving. You can save yourself a lot of time and trouble by investigating what you're changing in the circuit when you install those snakeskin covered Molex-and Mustard capacitors which have been cryogenically soaked in a 12% solution of artichoke jelly and iridium oxide. There may be simpler and more inexpensive ways to do the same thing.

Finally, if you know the circuit and know that $0.02 ceramics will do as well as $5.00 polypropylenes, go for the ceramics!

[Taylor]

Screamin' Yaller

Ahem. http://en.wikipedia.org/wiki/High_yellow




...Sorry, now I'm just messing with you.  Thanks for the response.

[thereverend]

i dont see the allure of using fancy-schmancy parts that look unique or whatever since once they're in the enclosure you arent going to ever see them again, and if you do see them again you'd be too busy debugging/fixing something to care.

[Taylor]

I'm with you, but a lot of people really do care. Check out the Pictures! thread for lots of stuff like that - eyelet boards, terminal strips, carbon comp resistors, cloth-covered wire, etc.

I guess it's the same thing that makes car guys want to chrome out their engines, get custom radiators in the shape of George Jefferson, that kind of thing. Even though it's under the hood, that's where they spend most of their lives, so I guess they want it to look nice.

Plus, there are only so many permutations of Tube Screamer imaginable. Once you're done trying out endless opamp subs, eventually you have to start fooling with stuff like cool-looking caps or transistors that look like mushrooms.

[Toddy]

RG, you should seriously think of taking what you just posted and put it on your web site. That was phenomenal to say the least! Sometimes my head explodes when folks try to explain something they make things more complicated. I'm a teacher and have seen it done and have done it myself. And, AFF...great question!

Now if someone can only decode  or cross reference the transistors in my original germanium Orpheum Fuzz!! (Q1 = GE6638 033, Q2 = GE A4056 713) (2N408s and 2N2613s do not work).

Thanks!
Toddy

[Gus]

A good writeup.
http://uk.geocities.com/cyrilb2@btinternet.com/

It depends, cost, design life, size constraints, what you can find, etc.

I like tants in the gain control - leg of the Rat.  It might be the ESR and measured uf value vs low value resistance like 47ohms.  Look up Al and Tant electros specs of the same value.

[R.G.]

Screamin' Yaller

Ahem. http://en.wikipedia.org/wiki/High_yellow
...Sorry, now I'm just messing with you.  Thanks for the response.

Consider me messed with. I never heard that one!

[noelgrassy]

R.G., Do you have any experience with "wet Tantalum" caps? What on earth could make a relatively low voltage, low capacitance cap cost $48+ ? These can be seen in the Mouser catalogue but my guess is it's only the Defense industry or aerospace folks that require them. I honestly don't know.
I just realized this may be covered on Wiki so I'll search there and report back before you put your usual thoughtful and gracious answer to my query. Your work in the F/X field has been such help to me and I'm sure multitudes of others who would still be thumbing through impossibly large tomes trying to glean a morsel of these concepts.

Noel Grassy.

re: Wiki-Nope, nada. There's a thread about "wet slug" caps made many years ago but I didn't think that was similar.

[AFF]

This has been the most informative topic of caps I have read.

This is great stuff here. Thank you R.G. ... I believe the R and G in your screen name must mean Real Genius!!!

The analogies helped me to visualize what you were saying also.

Perfect! You have saved me time and money. Thank You so much!!!

I can definately say that I will look closer into the circuit to know what will work and provide what I need ... before I would have probably just bought the "best" and hope for at least half "best" results.

Good Job R.G. Thank You!!!