[time-nuts] Building a DMTD/phase noise set in the 21st century

[Gerhard Hoffmann]

Am 24.08.19 um 23:49 schrieb Jan-Derk Bakker:
> there may be some flicker noise present, I've not personally seen the poor
> DC performance in many S/D audio converters either (apart as artefacts from
> the front-end electrolytic coupling capacitors).


Yes, these electrolytic capacitors can produce ugly effects.


purple = DigiKey 493-4582-1-ND organic polymer Alus 4700u/2.5V

blue    = Digikey 1189-2962-ND  2200u/25V

green  = DigiKey 1189-1056-ND 3300u/10V Rubycon, cheap


The winner is clearly the cheap

< https://www.digikey.de/products/de?keywords=1189-1056-ND      >


The test circuit is an Interfet IF3602 based low noise baseband amplifier.

I definitely wanted no bias loop via the gate and also for some reasons no

feedback into the source, so I enforced the drain current with a current 
mirror.

That is ugly, because the source must be AC-grounded with huge 
electrolytics.

In the test setup, there were 4 in parallel. There is only 500 mV across 
the capacitors.

Even the 2V5 organic polymer is not challenged voltage-wise.


The traces are the own input voltage noise of the amplifier. The 
horizontal part

of the traces is about 250 pV/rtHz noise density. The bad capacitors 
both show a 1/f**3 rise,

in textbook quality. At 1Hz is the lower passband corner of the 
following stages.

I wonder what process  produces that. Generation/recombination and most 
others do only 1/f**2.


A 4700u/25V wet slug tantal showed no weakness. But at $100 a pop I like 
to avoid them.

Vishay costs even 3 dB more than AVX.  :-(


regards, Gerhard

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 3Caps.png
Type: image/png
Size: 24105 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20190825/8b376009/attachment.png>