CAD CAM EDM DRO - Yahoo Group Archive

re:Re: tantalum caps

Posted by Wally K
on 2000-11-21 02:43:11 UTC
--- In CAD_CAM_EDM_DRO@egroups.com, ballendo@y... wrote:
> Wally K, others, and list,
>
> Okay, someone mentioned this has been beat to death. But not much
in
> the way of useful CNC "real" circuit info presented, IMO. :-(
>
> Results:
> 1)When to use tantalum?
> long cycle timing circuits.
> timing circuits (with one dissenter, who was told by an expert that
> drift made them unsuitable).
>



As i stated earlyer tantalums make poor timming circuits, however
never say never. In special circumstances like long timming circuits
you may have to be design with them. But you must go into the
design with your eyes open. Shooting from the hip i would say you
have to plan for a 30% drift over time.



> 2) When NOT to use tantalums?
> When power supply "inrush" is possible, unless a series resistor is
> added, which negates the low ESR of the tant. cap.
> When circuit design shows (by failure) that they won't work.



Tantalums are routinly specked to be placed at the output of a 3
terminal regulator by every regulator manufacture i can think of. Do
you think tantalums are specked in these circuits because the
regulator manufacture wants the circuit to fail.

This is the type of thing the tantalum cap was designed for. This is
were there are BILLIONS of them in use around the world. If you read
the other posts carfully no one else has admitted to seeing one of
these fail.



>
> 3) Additional considerations.
> They will fail soon, if they are to fail. (this is true of almost
all
> electronic components!)
> "Good" design and "proper" usage will lead to success with tant.
caps
> (again, this is true of nearly all electronic components)
> Multi-layer ceramics have improved to the point of being a suitable
> replacement for tant. caps (but tolerance/tempco is poor)
>
> So, getting back to Jon E's post "I never use 'em" and MY post "HOW
> and WHEN to use 'em; What is the answer?
>
> Since the concensus seems to be "it's OK to use 'em", WHERE and HOW
> should they be spec'd? WHAT is the "good design" and "proper
usage"
> alluded to, but never stated?!?



Well since it appears i am the only one who knows how to use
tantalums i will answer your question.

With modern electronics power supplys became smaller physically and
had less current demands. with the big push to make things smaller
and more realiable, tantalums were devoloped. Almost everyone wanted
a replacment for the electrolitic that was large and prone to
failure. Like most new components they were tried in many designs
with or without the manufactures blessing. Over time it has been
acknowledged that tantalums should not be used in hi current or hi
current pulse applications. This is why tantalums have found a home
in the 3 terminal regulator power supply design. With a sprinkling
of tantalums around a PCB to help transit responce. In the usual 3
terminal regulator design you are not drawing large slugs of
current. When you do draw large slugs of current somthing else
must be used.

To try to be more specific use a tantalum in 3 terminal regulator
design when the current will be a more or less steady value.

Use a electrolitic in applications were you draw large currents or
slugs of current.

Take a look at the DigiKey cataloge if you look at electrolitic caps
you will see a spec for ripple current. As you see in the cataloge
the values range from under 100 mA to a couple of amps. your cap
must have a higher current rating than your application draws. This
is why many designs fail because this spec was ignored.


If you also look in the Digikey catalog for this ripple current spec
for tantalums you will see none is given. Why, because they are not
meant to give large slugs of current. They are meant to be used in a
more or less steady state mode, like the use of 99% of all 3 terminal
regulator power supplies.

>
> Ballendo ( who understands that sometimes simple questions don't
have
> simple answers)
>
>
> Wally K writes:
> >Do not get paranoid about tantalums.
> >This thread needs a little common sense.


Wally K.

Discussion Thread

ballendo@y... 2000-11-20 20:28:18 UTC re:Re: tantalum caps Jon Elson 2000-11-20 21:53:56 UTC Re: [CAD_CAM_EDM_DRO] tantalum caps Wally K 2000-11-21 02:43:11 UTC re:Re: tantalum caps Mariss Freimanis 2000-11-21 08:42:07 UTC Re: tantalum caps Jon Elson 2000-11-21 12:07:50 UTC Re: [CAD_CAM_EDM_DRO] Re: tantalum caps