Erik, Just a thought. The 7805 in your schematic appears to be a big part of the signal chain. I've run into issues with 78xx series regulators being noisy. Spent a few weeks chasing down noise issues in some equipment and the 78xx regulator was a part of the problem. Replaced it with LT3042 as part of the solution. Also note, not all 7805's are created equal. Different vendors may be better/worse than others. Good luck! Dan On 7/12/2022 9:45 PM, time-nuts-request@lists.febo.com wrote: > > I'm struggling with the noise floor. > First tests where done with a 5nV/sqrt(Hz) opamp. Noise floor with > shorted mixer output at 10kHz was -140dBc/Hz. Then I tried with > 1nV/sqrt(Hz) opamp, but that made no difference, noise floor at 10kHz > was still -140dBc/Hz > The setup was simplified to this schematic: > http://athome.kaashoek.com/time-nuts/PNA/SSPNA.JPG > The REF_buffer creates a virtual ground, the Audio_LNA amplifies into > the differential audio output . > Why did the lower noise opamp not make a difference? > Also the setup is acting like a nice microphone. Tapping the housing is > clearly audible. Which component may be causing the microphony?

On 7/13/22 5:24 AM, Dan Kemppainen via time-nuts wrote: > Erik, > > Just a thought. The 7805 in your schematic appears to be a big part of > the signal chain. I've run into issues with 78xx series regulators > being noisy. Spent a few weeks chasing down noise issues in some > equipment and the 78xx regulator was a part of the problem. Replaced > it with LT3042 as part of the solution. > > Also note, not all 7805's are created equal. Different vendors may be > better/worse than others. > > Good luck! > > Dan +1 for the 200 mA LT3042 (and the higher current 0.5A LT3045). There are now negative voltage versions, too, LT3094, also 0.5A. Noise from 10Hz to 100kHz is around 0.8 microvolts.  At 10 kHz 2 nV/sqrt(Hz).  But the real thing these things are good at is PSRR at higher frequencies.  >70dB at 1 MHz They have enable inputs, can be paralleled willy-nilly, etc > > > > On 7/12/2022 9:45 PM, time-nuts-request@lists.febo.com wrote: >> >> I'm struggling with the noise floor. >> First tests where done with a 5nV/sqrt(Hz) opamp. Noise floor with >> shorted mixer output at 10kHz was -140dBc/Hz. Then I tried with >> 1nV/sqrt(Hz) opamp, but that made no difference, noise floor at 10kHz >> was still -140dBc/Hz >> The setup was simplified to this schematic: >> http://athome.kaashoek.com/time-nuts/PNA/SSPNA.JPG >> The REF_buffer creates a virtual ground, the Audio_LNA amplifies into >> the differential audio output . >> Why did the lower noise opamp not make a difference? >> Also the setup is acting like a nice microphone. Tapping the housing >> is clearly audible. Which component may be causing the microphony? > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe send an email to time-nuts-leave@lists.febo.com >

Am 2022-07-13 15:50, schrieb Lux, Jim via time-nuts: > On 7/13/22 5:24 AM, Dan Kemppainen via time-nuts wrote: >> Just a thought. The 7805 in your schematic appears to be a big part of >> the signal chain. I've run into issues with 78xx series regulators >> being noisy. Spent a few weeks chasing down noise issues in some >> equipment and the 78xx regulator was a part of the problem. Replaced >> it with LT3042 as part of the solution. >> >> Also note, not all 7805's are created equal. Different vendors may be >> better/worse than others. > +1 for the 200 mA LT3042 (and the higher current 0.5A LT3045). There > are now negative voltage versions, too, LT3094, also 0.5A. > > Noise from 10Hz to 100kHz is around 0.8 microvolts.  At 10 kHz 2 > nV/sqrt(Hz).  But the real thing these things are good at is PSRR at > higher frequencies.  >70dB at 1 MHz > > They have enable inputs, can be paralleled willy-nilly, etc I have measured the noise density of an assortment of regulators under the same conditions and in the proposed circuits of their data sheets. I don't believe in different brands of 7805; they are all ugly. < https://www.flickr.com/photos/137684711@N07/24070698809/in/album-72157662535945536/ > You can increase Cset on the 3042 to 100u tantalum. That helps at the low frequency corner. More does not help. For the output capacitor, 4u7 is optimum. More produces a noise peak at 100 KHz or so. The trace of the 3042 is sandbagged at the low end because of the input capacitor issue with my then-preamp I explained a few days ago. Paralleling lt304* is not cheap, althoug it drives the noise further down. The circuit from the data sheet with the external NPN is about as good as a single 3042. I tested it with 1.5A. It takes a little bit more input voltage, abt 6V4 for 5V out, but is cheaper than many LT304* in parallel. Gerhard