So this is a price optimization after all
Going up to 50V, let alone 100V, doesn't always make sense. Most of the time, these caps are used to bypass 5V or 3.3V chips, in which case 16V or 25V has ample margin. If you go to 50V, you won't get higher capacitances in the handy 0603 case anymore at all; in corner cases where it is available, the price is often too much (and the real capacity under those conditions suffers more, so it's partially a marketing bluff).
Get something like this:
0603:
4.7 pF NP0, voltage irrelevant (for smaller crystals)
18 pF NP0, voltage irrelevant (for larger crystals)
100 pF NP0, voltage irrelevant (signal filtration)
1nF NP0, voltage irrelevant (signal filtration)
10nF X7R 50V (some heavier RC filtration stuff)
100nF X7R 25V (general bypassing; or get 50V if you use higher than 12V chips, e.g. opamps at high rails, but it'll be more expensive without much benefit at 3.3V, and you'll be using A LOT of these! Buy 500 at least, consider a full reel.)
1uF X7R 25V (high-power bypassing, fet drivers, fet driver bootstrap caps etc. Going over 25V shoots the price up or ups the case to 0805)
0805:
2.2uF X7R 25V (DC/DC converter / LDO output caps)
1206 or 1210:
10uF X7R 25V (DC/DC converter / LDO output caps, less than, say, 15V)
10uF X7R 50V (DC/DC converter / LDO output caps, higher than, say, 15V)
Take this as a grain of salt. I don't completely agree with myself while trying to write it down like this, but the general idea is valid.
Going up to 50V, let alone 100V, doesn't always make sense. Most of the time, these caps are used to bypass 5V or 3.3V chips, in which case 16V or 25V has ample margin. If you go to 50V, you won't get higher capacitances in the handy 0603 case anymore at all; in corner cases where it is available, the price is often too much (and the real capacity under those conditions suffers more, so it's partially a marketing bluff).
Fair points! Of course with SMDs the constrained physical sizes imposes some extra limits which are less obvious with through-holes. I'm still getting used to thinking in these terms.Get something like this:
0603:
4.7 pF NP0, voltage irrelevant (for smaller crystals)
18 pF NP0, voltage irrelevant (for larger crystals)
100 pF NP0, voltage irrelevant (signal filtration)
1nF NP0, voltage irrelevant (signal filtration)
10nF X7R 50V (some heavier RC filtration stuff)
100nF X7R 25V (general bypassing; or get 50V if you use higher than 12V chips, e.g. opamps at high rails, but it'll be more expensive without much benefit at 3.3V, and you'll be using A LOT of these! Buy 500 at least, consider a full reel.)
1uF X7R 25V (high-power bypassing, fet drivers, fet driver bootstrap caps etc. Going over 25V shoots the price up or ups the case to 0805)
0805:
2.2uF X7R 25V (DC/DC converter / LDO output caps)
1206 or 1210:
10uF X7R 25V (DC/DC converter / LDO output caps, less than, say, 15V)
10uF X7R 50V (DC/DC converter / LDO output caps, higher than, say, 15V)
Take this as a grain of salt. I don't completely agree with myself while trying to write it down like this, but the general idea is valid.
Thanks, that looks like a good starting point. I'm amazed it's possible to squeeze enough dielectric into a 0603 case to get 1uF @ 25V - or at least enough to be able to call it that; I'm totally with you on the misleading specs.
@dadler: That looks like a nice and simple way to store the components, thanks for sharing! Since I don't really have any SMD parts yet (hence this thread) apart from some LEDs and a few bits that have been bought for specific projects I haven't really got a firm plan for how best to store them. I like the look of those little flip-top boxes that can be joined up into trays, but I'm not sure. I store pretty much everything in a stack of Raaco Handy-Boxes, but while I'm a big fan of these, especially with the smooth lids and dual layer inserts, I doubt they'd work well for SMD components - I've already found some SMD LEDs that have escaped their trays...
Re. SMD resistors, I've now added 100 each E12 series of those 0805 Yaegos you suggested - total price: £35 for 85 values Based on what's been said I think I will be getting the same in 0603 OR 0402 as well. That's general purpose resistors taken care of - next up MLCCs...
First, it doesn't make sense to have every possible capacitance. Sure, have many sizes of resistors, but you don't need that many cap values, especially something with tolerances and specs (DC/temperature bias, etc.) so poor to begin with. Just have a few small C0G sizes for signal filtration and crystal loads, and then a few bigger sizes for power decoupling; get decent X7R from reputable companies (no Ebay stuff, no matter what they claim).
It might not make sense for you to have every possible capacitor value, but other people do have such requirements. Filter design is a classic case where you need a specific value to work with the limited range of available inductors. In RF filter design it is often the case that you have to tweak component values after design, to take account of inevitable implementation "suboptimalities".
Just a quick note re. storage; while I'm happy for now to just keep the lengths of tape in individual bags, I have been looking at different storage options. The "tape strips in binders" style I have already discounted: I already have too many loose SMD parts in various quantities for tape binders to be a good option - and they're no good for wider tapes, or items that come in other packaging (or loose). The flip-top lid stackable boxes in @NANDBlog's post look like they can be a nightmare to get parts out of - you'd need to pick individual parts out with tweezers. I've also looked at the SMD trays from Licefa (model "A 1"), which based on price seems to be the Rolls-Royce option, but I really don't like them sharing a single lid - way too much risk of disasters; imagine dropping one of those trays The most interesting option I've found so far is the "vial" type storage, such as the "V 11" from Licefa. These are more reasonably priced (~£25 for 130 vials) and look like they will be easy to get individual items out of, by simply tapping/pouring them out on a small dish, but would love to hear what others think about this option!
You need to grab it with tweezers anyway, to place it on the PCB??? I dont know about the quality, these are like a year old, and I havent used them "in action" they just look neat, and I thought I share it.
The "phials" are available individually (V4-1) and can be packed tighter inside other boxes
This will work much better than a bunch of loose containers.
If you have an "engineering pillbox" with dozens of linked little boxes... uhh, how do you take out more than one or two parts at a time? Do you unlink all the boxes to get the one you want? Or do you clear your entire bench, put down a sheet of paper, then tip the entire thing over?
Thanks for the tip, but I'm afraid I don't agree; I find individual containers easier to work with, allowing me to pick out just the parts I need without cluttering up my workbench with component trays. It's usually pretty crammed anyway. If you want to stock whole series of resistors and capacitors, you would typically need to have at least three of those boxes on your bench (resistors+capacitors+others) to populate a board, and enough space left to unfold at least one - with individual phials/vials you can fetch only those parts you need from the storage rack instead. My through-hole storage already works like this, and I've found it to be very convenient.
What do you do when you add or remove a value?
The only drawback is you can't label little teeny test tubes from the top of the cap
For parts that need to stay in the tape, like SOT 23
Arguing on the internet is a waste of life.
Arguing on the internet is a waste of life.