Betreff: RE: [sdiy] Re: 4049 and 4069

[Czech Martin]

Depends on what we are considering.
A CD4000 series should have improved over the years.
Assuming same technology size but better material and
machines and quality control. I think that this is 3um
CMOS or so.

Up to 0.5u geometry there was still some improvement.
Below that the transistors look very wierd.
Not only that the geometry shrinks, very short channels
are in effect, but also substrate/well doping rises,
and leakage rises. And vt goes down and down.
The measured curves look very odd, not at all like the
textbook characteristics.
These transistors are designed as leaky switches.
Not for analog performance. But still analog circuits
have to been made on the same chip. The price is:
large area devices. So at some point analog circuits
grow in area instead of shrinking when you downscale further.
At least the downscaling stops for analog.
The only advantage is higher speed, making marvels like
80 MHz sampling rate with decend resolution possible.

The noise vs. hum was: it is possible that modern fabricated
CD4000 have improved in noise so much, that other noise
sources (like trafo hum, stray fields, etc.) which have nothing
to do with your circuit make it difficult to get good hissing
noise without any other artefacts.

Or in other words: your noise circuit might have a bad S/N . ;->

m.c.



> -----Original Message-----
> From: jhaible at debitel.net [mailto:jhaible at debitel.net]
> Sent: Dienstag, 20. Januar 2004 13:45
> To: Czech Martin
> Cc: synth-diy at dropmix.xs4all.nl
> Subject: RE: Betreff: RE: [sdiy] Re: 4049 and 4069
> 
> 
> Does this this mean very old ones are too noisy
> because of bad material, and modern ones are
> too noisy becaus eof shrinked transistor
> size, and there is an optimal period from 
> where to get "vintage CMOS" ?
> 
> Didn't understand the noise vs. hum part, btw.
> 
> JH.
> 
> 
> 
> > And the year of production is interesting, too.
> > 
> > Surface problems and gate oxide integrity
> > have improved considerably since 1970 or so.
> > 
> > Perhaps, modern stock CMOS is not useable for noise
> > because of bad S/N (noise voltage too low compared to hum
> > etc...).
> > 
> > m.c.
> > 
> > > -----Original Message-----
> > > From: jhaible at debitel.net [mailto:jhaible at debitel.net]
> > > Sent: Dienstag, 20. Januar 2004 13:20
> > > To: Czech Martin
> > > Cc: synth-diy at dropmix.xs4all.nl
> > > Subject: RE: Betreff: RE: [sdiy] Re: 4049 and 4069
> > > 
> > > 
> > > > > Are _all_ MOSFETs so bad, noise-wise?
> > > > 
> > > > No, the noise corner can be controlled by gate oxide quality and
> > > > surface quality within some limits. 
> > > 
> > > 
> > > Then it would be interesting to see which brand of 4069
> > > has the "lowest" noise, compared to others, at least.
> > > 
> > > Alot of manufacturers make (or used to make) 4069s.
> > > Has anybody compared them, noise-wise?
> > > 
> > > JH.
> > > 
> > > -------------------------------------------------
> > > debitel.net Webmail
> > > 
> > 
> > 
> 
> 
> 
> 
> -------------------------------------------------
> debitel.net Webmail
>