Electronics > Metrology

Ultra Low Noise Reference 2DW232, 2DW233, 2DW23x

<< < (126/129) > >>


I ordered 20 from the address MegaVolt supplied, and received 2 packages of 20 each, date code 15-4 (and 2 credit card charges). I may or may not have double ordered, not being used to AliExpress. Inked markings not very distinct.

With Pipelie's warning, I wanted a quick noise test so set up a comparison with a LM329. I set up a test using my LM399 standard as the source, 3K resistor for the LM329, and 750 ohms for the 2DW323, pins 1 & 2. Measured with Tek 7A22 plug-in set to 1Hz-1kHz bandwidth. Photos captured with Hantek DSO connected to 7834 Vertical Output. LM322 voltage source measures 35.6uV P-P. LM329 measures 25.2uV P-P, and 2DW232 measures 5.2uV P-P.


so you have 1:50000 amplification?

Hmm, to measure 1Hz you would need a minimum sweep time of 1 second.
Otherwise you get another pole by the limited measurement duration.

with best regards


You're correct. The high pass filter on the 7A22 plug-in ( love those) was set to 1 Hz, but the sweep time would have to be set slow to actually show really low freq noise.

I was just glad to see that the 2DW232's from that site were real Diamond components and were low noise compared to the LM329's.

I built a copy of a "low noise" 60 db amp, and was trying to add that to the test setup to get further above the scope noise when my LM399 supply started latching into a 3.5V output mode,  |O so will need to build a low noise, higher current supply to do further testing. The DW will be good for that as I will only worry about short term stability, and not 0 TC.

Chances are that even the not so good 2DW232 would be OK at the higher frequency. The real test is for the low frequency range and 1/f noise. The 1/f noise is what depends one purity and can vary between batches.

For a simple noise test one would ideally have to independent references and than measure the difference.
There is no real need to have a separate low noise supply. The normal 1 OP amplification of the ref. and getting the current from the amplified voltage should be Ok. This may even work with an LM358 on a breadboard - though poor contacts can be a hassle.
The difficulty may be more in setting the right current as 1/f noise and temperature fluctuations can look similar.


--- Quote from: Pipelie on September 22, 2020, 06:58:51 am ---be ware that not all the 2DW232 with diamond mark is the low noise one. we have tested samples from several source and we found the date code start with 12-xx and 13-xx are most likely the low noise one. the newer one is probably not, such as the 14-xx to 15-xx we tested are kind of bad.

--- End quote ---

Pipelie, what is your noise test method?


[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Go to full version