Homebrew PCBs

John,
Now we're getting there. I went to Mouser's site and checked on a
capacitor I typically buy for construction of crystal filters in 
transmitters and receivers, the Kemet Golden Max series, which is
monolythic ceramic, NP0/C0G.  Lately I've used them in VFOs as well.

I linked to the data sheets, and while they didn't specify K, Q or
ESR, they did specify dissipation factor, which is the inverse of Q
and therefore has ESR in its equation (I looked at another site for
that).  So, now when an application specifies high Q for a cap, I
simply look at its dissipation factor.  It's odd I didn't know that;
I've used Q in other  applications, obviously inductors, and in
crystals themselves, but I'd just never thought about it in terms of caps.

I guess after close to 40 years in electronics, I haven't got all the
dumbness out yet!

Thanks,
Ted

--- In Homebrew_PCBs@yahoogroups.com, "John Popelish" <jpopelish@...>
wrote:
---snip---
> At the low end of that range many film caps are fine, as long as they
> are low inductance construction.  At the top and of that frequency
> range, I think mylar film (the most common and lowest cost film caps)
> start to show the losses of mylar, and polystyrene, polypropylene, and
> teflon are better, but larger and more costly.  And at all
> frequencies, these dielectrics have better temperature stability.
> So the big concerns for film (and really, for any dielectric)
> capacitors for RF applications is the losses and stability of the
> dielectric, and the inductance and series resistance of the
> construction.  To compare a pair of cases, fairly, you realy need to
> pour over the data sheets.  If the data sheet doesn't mention
> dielectric losses, stability, inductance or resistance, assume the
> news is not good or do your own testing.
> 
---snip---
> 
> Ceramic covers a lot of variation.  The low K (NPO COG) dielectric is
> low loss and high stability, as well as having capacitance that is
> pretty stable with respect to applied voltage.  The high K types (Z5U
> Y5V) are not only variable with temperature and voltage, but loose
> capacitance over time and often have high resistivity metalization (in
> the multi layer, monolythic types).  The middle K types (X5R X7R) are
> intermediate in character (pretty good, pretty cheap and pretty
> small).  I have given up on the really high K types, and choose
> between the low k and mid K types, depending on capacitance and loss
> requirements.
> 
> Good silver micas are very good, and very expensive.  The cheapest
> micas have the annoying characteristic of producing small charge steps
> when voltage is applied.  It is a very specific noise for the type.  I
> think it has to do with the mica being made up of many distinct
> layers, and charge may get trapped in or released from any given layer.
>