John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Jump to Latest
W

Waly

Banned
Joined 2011
PMA said:
Waly, AFAIK only simulators show Y axis in V/rt(Hz). Instruments use dB, dBu, dBV and you need to take BW and bin width into account, when calculating noise from noise spectral density.
Click to expand...

Incorrect. I still have to find a non-toyish spectrum analyzer that doesn't have the option to display V/rtHz or dBV/rtHz on the Y scale. Even free PC tools like Arta have this option, and it's also available for log sweeps (not very common in standalone spectrum analyzers). Of course, you have to know about and look it up, it is not the default setting.
 
M

Markw4

Member
Joined 2016
Dennis,
That DacMagic Plus looks like it may actually be the real deal, with performance similar to a Benchmark DAC-1, and it looks like it has a better headphone amp. There are various opinions in reviews, some people like it as well or more than DAC-1 and some less so. For $349, it is extremely tempting. So, I just ordered one to find out how I like it. Should be here in less than a week. We'll see. I can always send it back, depending.
 
Last edited:
PMA

PMA

Member
Joined 2002
pmacura.cz/audiopage.html
Paid Member

Attachments

  • dacmagic_toslink_jitter.png
    dacmagic_toslink_jitter.png
    28.6 KB · Views: 195
PMA

PMA

Member
Joined 2002
pmacura.cz/audiopage.html
Paid Member
Waly said:
No, you are just not measuring "noise", but a signal spectrum. A small but essential difference.
Click to expand...

Yes, but I understand this.

Still you can get the noise level, related to reference level and measurement bandwidth, from spectrum, when measuring output with no input signal. In Arta and similar SW, you get the "RMS" value, which is the RMS voltage noise at the actual setting, say for 48kS/s it is for 24kHz BW unweighted, unless the weighting was chosen. It can be also calculated from "noise floor", e.g. for -140dB "bottom floor" with 0dBFS = 1V, 48kS/s and 65536 samples in the FFT record (0.73Hz bin width), after some trivial calculations we get noise voltage of 19.84uV over 20kHz BW unweighted. This is S/N = 94dB/1V/20kHz or, in other words, 140nV/rtHz. Hope I did not make many mistakes in trivial maths.
 
gerhard

gerhard

Member
Joined 2005
hoffmann-hochfrequenz.de
PMA said:
Waly, AFAIK only simulators show Y axis in V/rt(Hz). Instruments use dB, dBu, dBV and you need to take BW and bin width into account, when calculating noise from noise spectral density.
Click to expand...

This is not a simulation. W&G SNA-33 spectrum analyzer, 1Hz resolution BW from
20Hz to 26 GHz. It has some DSP, but is essentially a purely analog device.

BTW this is the noise of a R&S NGT20 lab power supply, 5V from the rightmost channel.
-100 dB/rtHz is 1nV/rtHz
 

Attachments

  • ngt20_5V_1nvis-100.png
    ngt20_5V_1nvis-100.png
    508 KB · Views: 181
Status
Not open for further replies.
Top Bottom