Hi,
I would like to measure the noise of a power supply, like LT3045 based, Salas UltrBiB etc. My oscilloscope (RIGOL 1054z) is not suitable for that kind of measurement, but I've found something.
Noise Measurement Amplifier
Is it ok for that kind of measurements?
I would like to measure the noise of a power supply, like LT3045 based, Salas UltrBiB etc. My oscilloscope (RIGOL 1054z) is not suitable for that kind of measurement, but I've found something.
Noise Measurement Amplifier
Is it ok for that kind of measurements?
The maximum input DC voltage is 6.3 V, according to the datasheet. Does that suffice?
Otherwise it seems quite suitable to me. You will have to correct for the noise of the amplifier itself.
Otherwise it seems quite suitable to me. You will have to correct for the noise of the amplifier itself.
Hi,
Is it ok, if I put a good quality filmcapacitor and keep the AC component if I would measure more than 5V?
Is it ok, if I put a good quality filmcapacitor and keep the AC component if I would measure more than 5V?
0.1 Hz to 100 kHz according to the datasheet. You can switch the upper limit. The scope has an FFT function, so one can also measure noise densities.
I would not spend so much money on this single instrument. Better buy a good USB soundcard connected through a 100V film cap. The PC with REW does the measurement of noise with FFT down to the uV region.
The noise of power supply it is measured in 20MHz bandwidth. That amplifier is not suitable for standard measurement of power supply.
I think soundcard is not suitable for that. The noise of the soundcard itself in the uV region, andthe bandwith is very limited.
I disagree. The limited BW is what you need for audio related noise measurements on power supplies. Higher frequencies are filtered easyly and do not contribute to audible noise. I did these measurements several times and never had any voltage source below the noise of the soundcard. There is definitely no sub-uV-noise to be explored.
@gaszto: Hi ... maybe this would work?
https://www.diyaudio.com/community/...inear-audio-vol-3-spare-boards.287604/page-18
Without knowing I would guess that rsavas might have some spare PCBs since he bought from JLCPCB and the lowest amount of PCBs that can be ordered is 5 pieces.
Cheers, Jesper
https://www.diyaudio.com/community/...inear-audio-vol-3-spare-boards.287604/page-18
Without knowing I would guess that rsavas might have some spare PCBs since he bought from JLCPCB and the lowest amount of PCBs that can be ordered is 5 pieces.
Cheers, Jesper
Possibly it is not the best instrument, but since it is present almost for free in any PC it can be used. And it can give quite a lot of info about a 20 Hz - 20 kHz frequency range.
I don't know of any scopes with noise density readout on the FFT. For that the scope needs to adjust for the FFT bandwidth etc. for calibration. Not real hard but I have only seen that on specialized instruments.
For a power supply circuit I shared here some years ago I had to build a really low noise preamp. I wond up using a transformer to get the noise into the pV/rtHz range I needed. Well below anything a soundcard can do.
The gadget looks interesting and you get a lot for the money. The internal noise floor is not specified which is dissapointing. Maybe write to them and ask what the noise floor is, input shorted and with a 10K resistor across the input. Voltage and current noise can be deduced from those numbers.
For a power supply circuit I shared here some years ago I had to build a really low noise preamp. I wond up using a transformer to get the noise into the pV/rtHz range I needed. Well below anything a soundcard can do.
The gadget looks interesting and you get a lot for the money. The internal noise floor is not specified which is dissapointing. Maybe write to them and ask what the noise floor is, input shorted and with a 10K resistor across the input. Voltage and current noise can be deduced from those numbers.
Read the datasheet:
https://www.analogtechnologies.com/document/ATNMA2.pdf
Table 1 on page 3:
150 nV, 0.1 Hz to 10 Hz
600 nV, 0.1 Hz to 1 kHz
800 nV, 0.1 Hz to 100 kHz
Hi,
heaving read the datasheet I miss on crucial specs/parameters, especially regarding noise figures.
Their Table 1. lists ´Amplifier Noise´ figures without specifying the conditions like gain, etc.
Isn´t that bit fruitles? Besides, 150nV (p-p or rms, or whatever?) would be quite ambitious.
Could it honestly be as good as the Datasheet suggests?
The 200µF DC-blocking cap into 10kR should raise the 0.1-10Hz noise to higher Values than indicated.
Also gain values above 60-80dB seem a bit exaggerated with a device that seemingly doesn´t feature a decent screening casing?
The best I see so far, is the nice mechanism of the test probes.
jauu
Calvin
heaving read the datasheet I miss on crucial specs/parameters, especially regarding noise figures.
Their Table 1. lists ´Amplifier Noise´ figures without specifying the conditions like gain, etc.
Isn´t that bit fruitles? Besides, 150nV (p-p or rms, or whatever?) would be quite ambitious.
Could it honestly be as good as the Datasheet suggests?
The 200µF DC-blocking cap into 10kR should raise the 0.1-10Hz noise to higher Values than indicated.
Also gain values above 60-80dB seem a bit exaggerated with a device that seemingly doesn´t feature a decent screening casing?
The best I see so far, is the nice mechanism of the test probes.
jauu
Calvin
I missed that. It translates (1 KHz band) into an equivalent 20K Ohms or 18 nV/rtHz. Thats 10X to 20X the equivalent noise of a low noise opamp today. Meaning that you need at least 10X gain for a low noise opamp to be able to see its noise with this on its output. However Its not clear that the spec is the shorted input noise. It happens to be about the noise of the 20K input resistors.
It's obviously not white noise, otherwise the value would increase by a factor of ten for each 100 times increase of the bandwidth.
Assuming that all values are RMS (if they were quasi-peak-peak it would be poor marketing not to mention that) and input-referred with shorted input, and that white noise dominates above 1 kHz, the white part of the equivalent input noise voltage would be ((800 nV)2 - (600 nV)2)/(100 kHz - 1 kHz) ~= 2.8283 10-18 V2/Hz or 1.6817 nV/√Hz.
Assuming that all values are RMS (if they were quasi-peak-peak it would be poor marketing not to mention that) and input-referred with shorted input, and that white noise dominates above 1 kHz, the white part of the equivalent input noise voltage would be ((800 nV)2 - (600 nV)2)/(100 kHz - 1 kHz) ~= 2.8283 10-18 V2/Hz or 1.6817 nV/√Hz.
Last edited:
Should it? Integrated from 0 to infinity, the noise of the input RC coupling should be √(kT/C) ~= 6.4159 nV for C = 100 µF (as there are two effectively in series) at 25 degrees Celsius. That's less than 150 nV.
this is true but do you think the op. is capable of soldering one up? It is all smt and a fairly tight layout. I have provided a few to another person but it’s easy enough to order more if needed