measuring ultra low noise powersupply on budget

[fonograph]

I want to make audio amplifier and dac,maybe even microphone pre-amplifier with as low noise as possible,atleast -120db.Lots of expensive audio gear is audiophool c.rap,but there are also very good stuff out there,especialy the high end solid state studio equipment that have impressive  low noise powersupllies.

My problems is that I have very limited budget,only instrument I currently own is Voltcraft VC890,its good modern 60000 count DMM with nice oled display and 0.03%dc acuracy.I plan to buy one low noise oscilloscope,either Siglent SDS1102x or 12bit Owon XDS3072a for 450 euro.

My question is,what kind of electronic gear is used to measure these low level ripples and noises? What should I get? I dont want to spend more than 300euro on top of the new 450e scope,so my absolute max is 800e,keep in mind only thing I have is one handheld DMM.If possible I would like to spend even less than that. I would like to know best new option,and best used option,I fear buying things out of warranty,but if its only way to satisfy my requirements I have no other choice within my limited budget range.

I am new at electronic engineering,I have been learning only 4 months so my knowledge is limited,I have been thinking agilent 34401a used could be good.but again,I am not sure what I really need,and here in eu its rare to see it being sold,shipping from US is expensive and I fear it will break since its old and that would be catastrophy for me since my income is very limited.Also as I was reading about 8.5 digits dmms ( out of curiosity,I cant afford those lol ),I saw something about microvolt DMMs from keithley,also I saw some instruments dedicated to ac rms measurements,is any of what I mentioned good for my purpose?

[Rerouter]

for measuring that low an amount of noise, it would pretty much only be doable with a spectrum analyser,

If you where talking 0.1mV per 1V of signal (80db), then that is just within the range of some oscilloscopes, but for 120db dynamic range, you are talking spectrum analyser.

To make things clear, 120db is 1uV of noise per 1V of signal, which for a microphone signal is not much

[DimitriP]

Before you go shopping, take a look at http://audio.rightmark.org/products/rmaa.shtml.
You'll also need a low(lower than some) noise soundcard

[Someone]

for measuring that low an amount of noise, it would pretty much only be doable with a spectrum analyser,

If you where talking 0.1mV per 1V of signal (80db), then that is just within the range of some oscilloscopes, but for 120db dynamic range, you are talking spectrum analyser.

Its easily measured with active probes or handmade amplifiers, this is audio range stuff not RF.

More reading:
https://www.eevblog.com/forum/beginners/question-about-preamp-for-measuring-noise/
http://www.mh-audio.nl/articles/lownoisepreamp.pdf

A cheap opamp in differential measurement can measure 14uV of noise at 500kHz bandwidth which for professional audio levels is -95dB.
https://www.eevblog.com/forum/projects/diy-100mhz-differential-probe/
Better opamps, lower bandwidth, and lower impedances could get right into the territory the OP is asking for.

[fonograph]

just to clarify,there is big difference between measuring power supply noise and measuring amp/dac output noise.I made mistake by writing about audio stuff like amps and dacs but that isnt what I had in mind!

The big difference is between measuring amplifier output noise and powersupply noise,for amplifier you need 120db dynamic range,becose its very small noise riding on very large sinewave signal and everything is in audible range so relatively low freqency,20khz maximum.

For powersupply its different becose its just noise riding on top of dc,there are also ripple,but ripple is low level,its not anywhere near as big as the sinewave signal comming from amplifier so having many bits of resolution isnt needed,what is needed is ability to go down very low like microscope to look at the little noise and also bandwidth requirements are much higher,the standart noise measurement is done to 20mega-hertz,thats 1000 times higher than 20khz needed for audio.

I want to measure powersupply dc output,not dac or amp output,I apologize for confusion.

[tautech]

FYI, a screenshot, a 4mV 100 Hz sine wave with 3V offset (My AWG wouldn't do more at these levels) shown on a SDS1102X, all other info can be read from the screenshot.

[fonograph]

Can you do that screenshot again without averaging but with eres/hi-res on maximum setting? The siglent have eres instead of boxcar and it can select 8.5,9,9.5,10,10.5,11,11.5 and 12bit mode right?

I am planning to do some 10 volt 6mhz PWM,can your waveform generator do 6mhz square wave? I would love to have a look at it.

another thing I noticed,its running at half the maximum samplerate,instead of 1gs/s its 500ks/s... do you have second channel turned on or why its like that?

[tautech]

The siglent have eres instead of boxcar and it can select 8.5,9,9.5,10,10.5,11,11.5 and 12bit mode right?

The max bit enhancement is 3.0

I am planning to do some 10 volt 6mhz PWM,can your waveform generator do 6mhz square wave? I would love to have a look at it.

5 MHz is all the AWG I have out can do in Pulse mode.

another thing I noticed, its running at half the maximum samplerate,instead of 1gs/s its 500ks/s... do you have second channel turned on or why its like that?

Don't know why it did that, only one channel on.

Can you do that screenshot again without averaging but with eres/hi-res on maximum setting?

Here's some more:
All 5 MHz 4mV 40ns pulse.

Normal acquisition


Averaging


Eres


Note this was only enhanced 2 bits as further shortens waveform peak level to a shorter duration than other acquisition types.

[fonograph]

Awesome!  tautech thank you,thats so good you can show me how it looks like,the siglent seems to have very low noise frontend and with that adjustable eres mode it seems like best scope in this pricerange for these kind of low level measurements.

Can I ask you in future to upload more pictures of siglent doing things?

[tautech]

Awesome!  tautech thank you,thats so good you can show me how it looks like,the siglent seems to have very low noise frontend and with that adjustable eres mode it seems like best scope in this pricerange for these kind of low level measurements.

Can I ask you in future to upload more pictures of siglent doing things?

Sure, but it just happens that I've had some time today, not all days are like that.
I do have better AWG's with a better "in house" selection of waveforms and higher frequencies.

There's a good selection of SDS1kX capailities in this thread:
https://www.eevblog.com/forum/testgear/siglent-sds1000x-series-oscilloscopes/

[David Hess]

Power supply output noise is trivial to measure if you build a quick and dirty AC coupled low noise amplifier with an operational amplifier.  The only issues are that gain-bandwidth product will limit high frequency response and input leakage or bias current will limit low frequency response.  Linear Technology has some application notes which discuss measuring regulator and reference noise that will be helpful.

If you only care about the audio range, then you can get good results using a sound card input and FFT calculation.

I use a Tektronix 7A22 or AM502 to make RMS spot noise measurements up to 1 MHz and down to 10uV/div.

[fonograph]

how much bandwidth you need for measing powersupply? I know standard rating is for noise from DC to 20mhz,I was reading some Texas Instruments article about switch mode voltage regulators and they were saying something about 500mhz ringing,thats raised big concern becose I planed to buy budget 1gs/s 100mhz scope like Siglent SDS1102x or Rigol DS1054Z and now I worry it will be not enough,for 500mhz you need minimum 5gs/s scope like Keysight 3000,that is 6x times more expensive.

So my question is this,is it normal for switch mode regulator to generate 500mhz ringing? Cant I just put capacitor into my design and asume it will filter it without looking at it with expensive scope?

here is the link to that article http://powerelectronics.com/regulators/low-noise-alternatives-linear-regulators

[mk_]

I am new at electronic engineering,I have been learning only 4 months so my knowledge is limited....
....is any of what I mentioned good for my purpose?

Hi

buy the book "Art of Electronics" and study it... because even if you have some equipment it looks like you don`t know what to do with it...

regards

Michael

[tautech]

how much bandwidth you need for measing powersupply? I know standard rating is for noise from DC to 20mhz,I was reading some Texas Instruments article about switch mode voltage regulators and they were saying something about 500mhz ringing,thats raised big concern becose I planed to buy budget 1gs/s 100mhz scope like Siglent SDS1102x or Rigol DS1054Z and now I worry it will be not enough,for 500mhz you need minimum 5gs/s scope like Keysight 3000,that is 6x times more expensive.

So my question is this,is it normal for switch mode regulator to generate 500mhz ringing? Cant I just put capacitor into my design and asume it will filter it without looking at it with expensive scope?

here is the link to that article http://powerelectronics.com/regulators/low-noise-alternatives-linear-regulators

All you need be worried about is output ripple, but when using SM regulators that operate at these sort of HF's both layout and use of the manufacturers recommended inductors (brands and values) will get it operating close to datasheet spec. You less need to see the waveform per se as the amount of ripple on the DC and tweak component values (caps) or add additional to get the ripple to acceptable levels.
I had to go through this process not so long ago and the MLCC's on their own wouldn't get the result desired so additional bulk capacitance was added to the rail in the form of a SMD Tantalum, just as fast as MLCC's but they're available in much higher values.

[David Hess]

how much bandwidth you need for measuring power supply? I know standard rating is for noise from DC to 20mhz,I was reading some Texas Instruments article about switch mode voltage regulators and they were saying something about 500mhz ringing, that's raised big concern because I planed to buy budget 1gs/s 100mhz scope like Siglent SDS1102x or Rigol DS1054Z and now I worry it will be not enough,for 500mhz you need minimum 5gs/s scope like Keysight 3000,that is 6x times more expensive.

So my question is this,is it normal for switch mode regulator to generate 500mhz ringing? Cant I just put capacitor into my design and assume it will filter it without looking at it with expensive scope?

here is the link to that article http://powerelectronics.com/regulators/low-noise-alternatives-linear-regulators

A bandwidth of 20 MHz is suitable for measuring switching ripple which is proportional to the output capacitor ESR but the switching artifacts require a much higher bandwidth and proper probing techniques including watching for ground loops.  Verifying their presence before and after filtering requires as much oscilloscope bandwidth as possible and on a budget, this is a good place for an old high bandwidth analog oscilloscope.  Spectrum analyzers can also be used for this.

In practice the higher frequency switching artifacts are usually ignored because their higher frequency content allows them to be effectively filtered with small passive components; even a short length of printed circuit board trace has enough inductance to be useful for this.  Some switching regulators have provisions to control voltage and current slew rates which helps a lot but sacrifices some efficiency.

Linear Technology has a bunch of application notes discussing these issues and one really good one which I cannot find at the moment.  When you get to this level of performance, all kind of other factors like magnetic and electrostatic leakage became a problem.