Hi all,
A fellow member of DIYaudio.com requested the schematic of a very low noise preamp I made. I made it a year ago, maybe more, to measure harmonics, after a -100dB notch filter. Feeding this to my sound card enabled me to measure THD down to -140dB.
Here goes, any comments?
A fellow member of DIYaudio.com requested the schematic of a very low noise preamp I made. I made it a year ago, maybe more, to measure harmonics, after a -100dB notch filter. Feeding this to my sound card enabled me to measure THD down to -140dB.
Here goes, any comments?
Attachments
Hi,
I would find it extremly instructive if you would post a photo of the stuffed board, since the physical layout is for sure one key for the impressive performance of your circuit!
Rüdiger
I would find it extremly instructive if you would post a photo of the stuffed board, since the physical layout is for sure one key for the impressive performance of your circuit!
Rüdiger
jackinnj said:
Yeh... 4 x 30pF x gain in 1st stage... hmmm about 1.2nF?.. but if you want to measure low noise, your output impedance has to be very low to begin with..
DIY offcourse!
I never really reached -140db, but close, somewhere around -130dB... again using a passive parrallel-T notch filter to filter out the 2nd and 3rd harmonic. I routinely reached -120dB but the last 10dB take some tweaking and patience...
The amplier does enable me to reach down to -140dB but I don't have a pure enough oscilator.. that explains my interest in the Uber-pure-oscilator (see Jackini) and ultra high feedback opamp in general...
jackini are you reading this? How about a super-opamp/LM3875 combo? Can we put togher a 180dB feedback chipamp?
I never really reached -140db, but close, somewhere around -130dB... again using a passive parrallel-T notch filter to filter out the 2nd and 3rd harmonic. I routinely reached -120dB but the last 10dB take some tweaking and patience...
The amplier does enable me to reach down to -140dB but I don't have a pure enough oscilator.. that explains my interest in the Uber-pure-oscilator (see Jackini) and ultra high feedback opamp in general...
jackini are you reading this? How about a super-opamp/LM3875 combo? Can we put togher a 180dB feedback chipamp?
Thx John!
I actually spotted a mistake. I thought that I had two 47K in parrallel as feedback resistor, depicted as 22K in the schematic, but now I see it is actualy a 47k adn a 4k7 resistor that cab be sattached parrallel to it.
I measured gain yesterday, at 1KHz it's 7680x (47K fb), and about 700x (47K||4k7).
grtz,
Thijs
I actually spotted a mistake. I thought that I had two 47K in parrallel as feedback resistor, depicted as 22K in the schematic, but now I see it is actualy a 47k adn a 4k7 resistor that cab be sattached parrallel to it.
I measured gain yesterday, at 1KHz it's 7680x (47K fb), and about 700x (47K||4k7).
grtz,
Thijs
Bridge null techniques (with care) should easily give you that extra 20dB rejection. You need to watch the quality of your external components, but then you need to in your twin-t anyway.
scott wurcer said:
Yep, that's right! I made quiet a performance jump when I started using only good quality film caps.. then I put the amplitude stabilizing lamp bulb in a styrofoam box, that help quiet a bit too..
Could you point me to some more information about the 'Bridge Null techniques' you mention.. ? I am not sure I understand what you mean.
grtz,
Thijs
At least for the non-inverting case. Let's say gain of ten. Think of subtracting passively the input from the output/10. With some phase correction you can get a pretty good null certainly better than -20dB over the audio BW.
Or say for a 1kHz fundamental you could deep null each harmonic frequency and go back and forth probably get more resolution than you could ever trust.
Or say for a 1kHz fundamental you could deep null each harmonic frequency and go back and forth probably get more resolution than you could ever trust.
OK , I get it. It read that method somewhere before, but never tried it. If I understand correctly, you can try to substract the output from your oscilator, from the output of your Device Under Test.. That might very wel help, thx for the tip!
One of the lovely things about an AP analyzer is the sound of the relays clacking away. I dare not open the thing up...
The question of an ULN measurement amplifier begs the question of just how precisely are you specifying the bandwidth. If you can get your hands on the Dolby paper "CCIR/ARM: A Practical Noise Measurement Method" -- you'll see why the authors prefer CCIR to A-Weighting because some parts of the noise spectrum just don't matter that much to our ears. The paper probably circulates on the web, is available on the AES website. The filter circuit for the HP 8903B which implements CCIR can be found on the Agilent website. The CCIR filter for the Boonton 1120 is somewhere on diytube.com
Someone more knowledgeable please hop in, but I think that the noise we ought care more about is radiated noise from our own power supplies, the switcher stuff which comes from your wireless telephone etc., motor transients which migrate into our systems.
...and oh yeah, there is noise we generate unwittingly when we try to gild the power supply lily and make something simple complex.
The question of an ULN measurement amplifier begs the question of just how precisely are you specifying the bandwidth. If you can get your hands on the Dolby paper "CCIR/ARM: A Practical Noise Measurement Method" -- you'll see why the authors prefer CCIR to A-Weighting because some parts of the noise spectrum just don't matter that much to our ears. The paper probably circulates on the web, is available on the AES website. The filter circuit for the HP 8903B which implements CCIR can be found on the Agilent website. The CCIR filter for the Boonton 1120 is somewhere on diytube.com
Someone more knowledgeable please hop in, but I think that the noise we ought care more about is radiated noise from our own power supplies, the switcher stuff which comes from your wireless telephone etc., motor transients which migrate into our systems.
...and oh yeah, there is noise we generate unwittingly when we try to gild the power supply lily and make something simple complex.
jackinnj said:
My main noise nightmare comes from those 🤐 fluorescent light bulbs.
Canada introduced national standards for lighting efficiency to come into force by 2012. After that, no regular bulbs will be sold except for very few applications such as in some medical lighting situations, oven lights, etc... Right now, when I'm doing measurements I have to shut off every bulb in the house, and this will only get worse 🙁
The second source of noise in my lab are TV's (plasma, LCD and CRT) radiating line frequency harmonics.
We were just looking at low level signals at work over the weekend, and saw this unexplained 28kHz waveform. Sure enough it was the overhead fluorescents with their solid state ballasts. At home I use a CFL in my workbench lamp, but I always keep a regular bulb handy for low noise work.
- Status
- Not open for further replies.