I am now trying the pc interface with the soundcard with REW,and the scope that commes with it.
I put in a LM4562 wich has better noise data than the original one.
I now with the regulator connected get about 100uV,
But I dont know about the calibration...
I put in a LM4562 wich has better noise data than the original one.
I now with the regulator connected get about 100uV,
But I dont know about the calibration...
If I disconnect everything from my soundcard (https://www.inet.se/produkt/6303370/asus-xonar-essence-stx) I still get a 100uV signal, could it be that the soundcard has a noiseflor of 100uV?
If I recall correctly, the FFT on these scopes is pretty bad. It's even horrible on the LeCroy WaveSurfer 40xxHD which is a 12-bit scope based on the top Siglent platform. You would expect better for the price, doesn't come close to the old Keysight/Agilent MSOX4000 series, in FFT performance anyway.
As mentioned, the probe connection is critical. It's best if you can get a "bayonet" style ground connection right at the output. See: https://electronics.stackexchange.c...w-do-you-attach-an-oscilloscope-ground-spring.
Even better would be a coax connector on your PCB.
As mentioned, the probe connection is critical. It's best if you can get a "bayonet" style ground connection right at the output. See: https://electronics.stackexchange.c...w-do-you-attach-an-oscilloscope-ground-spring.
Even better would be a coax connector on your PCB.
They want to sell SA's too, make the 'scope too good and fewer would buy the SA's or mixed-signal scopes!!
Having said that, the dynamic range on an 8-bit scope is too low to really bother with for spectral analysis in many situations, a 12 bit ADC would be the minimum to really give nice results I think, but even that's very coarse for audio of course.
This is what I get with the input disconnected and all the software volumes set to zero.
Does this make sense?
make sense?
Does this make sense?
100uV from a bench supply is quite low.
As for calibration, replace the supply you are measuring with a 100uV source* and see if you get 100uV measured.
Jan
* for instance from a 1V signal generator with the output attenuated with 10k - 1 ohm.
As for calibration, replace the supply you are measuring with a 100uV source* and see if you get 100uV measured.
Jan
* for instance from a 1V signal generator with the output attenuated with 10k - 1 ohm.
Or, if you have a decent SG, set it for a -67 dBm output and the scope for 50 ohm input. And be very careful about noise introduced from other sources such as lighting, wireless APs or computers - anything with a switching supply. Also, as Jan pointed out, 100uV is a good noise level from a linear supply and a very good one from a SMPS.
A very good reference is: https://xdevs.com/doc/HP_Agilent_Keysight/HP AN 372-1 Application Note.pdf
The recommended connections for measuring PARD are on page 9.
A very good reference is: https://xdevs.com/doc/HP_Agilent_Keysight/HP AN 372-1 Application Note.pdf
The recommended connections for measuring PARD are on page 9.
Another good reference from the late Jim Williams AN-47 (then Linear Technologies, now Analog Devices) has an entire chapter on probing: https://www.analog.com/media/en/technical-documentation/application-notes/an47fa.pdf
I have a FY6900 signalgenerator,and if I set it to 2v out the Siglent SDS1202x reds 2,16v.If I messaures the signal with the DMM i get 0,700 vRMS =1,974v P-P,Is it safe to say thet the scope messaures to much then?
How important is it use this terminations ? https://www.ebay.com/itm/3934092959...T0UU9sUpgpfgkIg1y9AT0vxu3g==|tkp:BFBMsuigqohg
This is a messaurement I did on my PC nothing connected to the sound card ,and the input grounded.
The 50 hz is quite large here,could it be that one I see on the next meassaurement?Wich is also PC with grounded input.
How important is it use this terminations ? https://www.ebay.com/itm/3934092959...T0UU9sUpgpfgkIg1y9AT0vxu3g==|tkp:BFBMsuigqohg
This is a messaurement I did on my PC nothing connected to the sound card ,and the input grounded.
The 50 hz is quite large here,could it be that one I see on the next meassaurement?Wich is also PC with grounded input.
Another good resource which illustrates with effects of various windowing on noise measurements. https://www.ap.com/technical-library/fft-scaling-for-noise/
I've yet to see an FFT built into a scope that was worth the powder to blow it to hell. Sound cards usually do better. Noise is difficult. If the level is low, there won't be enough gain to see what's going on. When I look at noise I use an analog scope and a differential plug-in with filtering (lest what you want to look at be buried in irrelevant high frequency noise) and a gain as high as 10 uV/division, though audio never requires going quite that far. Looking at noise without filtering is a fool's errand and putting a number to noise always requires specifying the measurement bandwidth. A really good project is to build a scope preamp with some filtering. It can be a 1-opamp project with a couple batteries. That will give you something that moves the noise into the capabilities of the scope.
Tektronix 7A22N! -- or its smaller cousin, the 5A22N.
As this is DIY, the schematics are available on the boat anchor archive (BAMA) and it wouldn't be difficult to replicate some of the functions with a pair of LT1113 JFET dual opamps in differential mode. Just remember to protect the inputs with clamp diodes. You can "post filter" with less expensive opamps for the bandwidth desired -- use Analog Devices "Filter Wizard", or Tex Instruments "Filter Pro".
I'm far from being an expert but you should leave your scope ON at least 30 minutes before you do any accurate measurement. Also, the oscilloscopes are not supposed to be very accurate on the vertical scale so I would trust the DMM as long as it's a True RMS (or you measure a sinewave).
My understanding is that the 50 Ohm termination is necessary for frequencies well above the audio range (i.e in the MHz region).
Again, take all this with a grain of salt. Maybe someone more experienced will chime in and correct me 🙂