[time-nuts] Re: 10 MHz TCXO periodically jumping 20 mHz up and down

[ASSI]

On Freitag, 18. Februar 2022 22:52:27 CET Attila Kinali wrote:
> Quick side note: On modern, cheap TCXO systems, expecially those
> for RTCs, frequency control is being done by a bank of switched
> capacitors. This is mostly because capacitors can be implemented
> in a standard digital CMOS process (or analog CMOS process).

I can think of better reasons.  :-)

This type of tuning requires ridiculously low power in comparison to anything 
else you'd be able to come up with, which is of utmost importance if you are 
talking about RTC.  For TCXO (or just XO) the excellent linearity of MOM or 
MIM capacitors at high signal amplitude would probably rank high on the 
requirement sheet.

> A varactor, on the other hand, would need a special doping
> profile (high-doping concentration with an abrupt transition).

No, any CMOS process provides you with a MOS varactor (at least two different 
ones depending on how deep into the woods you will go).  It has a limited 
tuning range and some biasing requirements that can be a headache (also 
significant temperature dependence).  These are also the ones with the highest 
capacitance per area in a modern CMOS process, so the cheapest option overall.

> Or in other words an additional half a dozen processing steps
> which cost a lot of time and money... especially for a single
> diode on a chip.

Yes, ideally you'd want a hyperabrupt junction for highest tuning range.  In a 
CMOS integrated process you just take whatever you already have that gets you 
closest to that ideal, maybe add one more implant if the varactor diode is 
something someone actually pays for.  It's usually not quite good enough for 
GHz RF applications if you care about Q, but it'd totally work at 10MHz.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf rackAttack V1.04R1:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada