standard size for smt tantalums?

You can now get a 100 uF 1206 ceramic, but they have rotten tc's and voltage/capacitance coefficients.

There are five or six standard surfmount tantalum sizes. DON'T use them for power supply bypasses; they tend to detonate. The new polymer tantalums are OK.

John

"Tam/WB2TT" wrote in news: snipped-for-privacy@comcast.com:

The part I mentioned is a 22uf 6.3V tantalum in a 0805 package.

You're right - I forgot that ceramics are normally multiple orders of magnitude smaller in capacity than tantalums.

I don't have the specs handy, but I suspect the largest size tantalum that would fit in a 1206 package would be around 1 ufd at 10V.

Check the price on a 1 ufd ceramic; and where would you buy one that is 100 ufd?

Tam

How about aluminum electrolytics? Since they're tall, you get at least as much capacitance per pcb footprint as you do with tantalums.

John

John Larkin wrote in news: snipped-for-privacy@4ax.com:

So for power supply bypassing I should use polymer tantalums? Or is there something better than that? I should mention that size is a bit of a constraing here... The smaller the parts I can use the better! I was looking on the Digi-Key catalog and found a couple different series' of surface mount polymer tantalums:

Kemet: (Digi-Key page 970:

T520 Series - KO-CAP Polymer Tantalum T525 Series - KO-CAP Polymer Tantalum T530 Series - KO-MAT Polymer Tantalum

Epcos: (Digi-Key page 976:

B45294R & B45296R Polymer Ultra-Low ESR Series B45496R Polymer Multiple Anode Ultra-Low ESR Series

Do any of these look like good choices for power supply bypass? (or any others that I missed - though I looked fairly carefully)

Thanks!

-Michael J. Noone

In general, tantalums have problems with surge currents, which are usually (not always) on the *inputs* to power supplies. Power supply outputs usually don't exhibit surge *into* the output bypass caps (at least for switchers, which tend to have soft start control), so tants tend to be ok here. That said, I derate rms ripple by a factor of 3 even here to be safe.

For input bypassing, I usually use a mix of electrolytics and ceramics (depending on input voltage range) and for output a mix of tants and ceramics. As with all general rules, it has to be evaluated for each design.

As to 'standard' sizes, most surface mount tantalums come in the EIA sizes 'A' through 'E', with the more recent 'R' (smaller package) available.

The A size equates closely to a 1206 package.

In mm, the sizes are, ignoring height and pad style:

A. 3.2 x 1.6 (1206) B. 3.5 x 2.8 C. 6 x 3.2 D & E. 7.3 x 4.3 ( Difference is height)

R. 2.05 x 1.3 ( 0805)

Cheers

PeteS

You got me looking at the Mouser catalog, and on the MuRata page I see they are pushing ceramics as tantalum replacements. They look expensive, but they claim that in replacing a tantalum, if the frequency is above a few hundred KHz, you can use a much smaller value of ceramic. I would go to the MuRata or AVX, or whatever web site and see what they say.

As somebody else mentioned, tantalums can have their own problems. We used to derate them X2 voltage wise to be sure they did not blow up. Running them at 90% of rated voltage made for fireworks.

BTW, there are standard sizes for SM tantalums. 3216 is a common size, once you get above the 0805 and 1210 sizes.

Tam

Yep. Even though we do burn in, a fair amount of these standard tantalums have decided to explode at our customers' sites.

The polymer tantalums seem to be working okay as supply bypassers, for us.

Bob

John Larkin wrote in news: snipped-for-privacy@4ax.com:

Will electrolytics work as well for power supply bypassing? Tantalums don't have any special characteristics that other types of capacitors don't have?

By the way - how does a Digi-Key 399-3178-1-ND look? It's a 10V 33uf Kemet T520 Series KO-CAP Polymer Tantalum, datasheet here:

$file/F3102T520.pdf

It's a B package - just slightly bigger than a 1206, so I shouldn't have much trouble fitting it on my board. Not to spendy at a little over $1 in single quantities.

-Michael

Hi Michael,

They have standard-sizes. Only the smallest fits a 1206 pad. (in millimeters:

1.2 x 0.1 inch = 3.05mm, 0.6 x 0.1 inch = 1.53mm so an 1206 equals a (non-existing) 3116 sized tantalum...) The sizes are coded A or in Numbers 3216 (which means 3.2mm x 1.6mm), B (3228), C (6032), D (7343) and E (same as D but greater hight). Sometimes additional sizes are available (depending on the Manufacturer. Check out AVX, Murata, TDK e.g.

As others alredy mentioned... these might be ceramic caps.

Ceramics work much better (lower ESR, long-term stability) but are much more expensive (you mentioned $2 each). Many (not cost-sensitive) designers prefer to use ceramics, since they are much more reliable than tantalums.

cu Bernhard Spitzer

--
bash.org - Top 100...
 hm. I\'ve lost a machine.. literally _lost_. it responds to ping, it 
works completely, I just can\'t figure out where in my apartment it is.

There is a very coherent discussion of different capacitor chemistries in