• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Upper limit of frequency response of $150 or so sound interfaces?

Anchorman

Member
Joined
May 2, 2024
Messages
5
Likes
0
I’m interested in measuring frequency response, distortion, etc of signals from dc to 100kHz or so with a spectrum analyzer, and digital oscilloscope, and wondering where to find upper and lower frequency limits of various USB audio interfaces. I’d like a little more range dynamic range than my current setup, which is a Syscomp circuit gear mini. It has a maximum of 60db range as a network analyzer, and is limited also in the noise floor for distortion tests.

Many of the sound card interfaces can do much better for s/n measurements, but are limited to about standard audio range of about 20-22khz for both tone generation and audio input. I’ve seen a few examples here where there are interfaces useful up to about 55khz such as here: https://daqarta.com/dw_gguu.htm

Looking for info on something like the creative x6 to see what upper and lower frequency ranges are, and whether there are other, better options out there for not much money.

Primarily I’m interested in looking at parasitic oscillations in amplifiers and quantifying distortion products/harmonics in lower frequency oscillators, looking at bode plots of filters, transducers, etc. some for audio, some for “science”. The new nano-VNA devices satisfy these needs from 100khz to the gigahertz range, but anything useful as a stand alone device better than the Syscomp circuit gear for the sub 100khz range is not really justifiable for hobby use.

I think I would be plenty happy with a device that is able to achieve between 90-100 db sinad though higher would be nicer), and usable frequency to 192khz, being able to calibrate the response in software, so perfectly flat probably not necessary.

Thanks in advance for any tips on what to look at!
 
Interfaces usually have rising noise past the usual 20-20Khz.

Here's an example with my ancient E-MU but it's a trend I see to much newer ones.

That's the exact same measurement of a 1Khz sine at 192Khz with 20Khz and 96Khz BW:

20Khz.PNG

20Khz BW


192Khz.PNG

96Khz BW
 
That amount of noise probably won’t bother me, and useful frequency range to 70khz with noise below -100db is still better than useful range to only 22khz. I realize there are tradeoffs between absolute accuracy, noise floor, upper and lower frequency limits and cost.
 
The MOTU M2 or M4 is probably one of the best at lower prices. According to Julian Krause it is flat to 90 khz or so.
Look around the 3:30 mark in the video. The Cosmos ADC is pretty close to SOTA in every parameter.
 
Last edited:
My motherboard's cheap onboard audio is flat up to 80kHz (DAC->ADC Loopback):
MSI Z170A Krait 3X (ALC1150) Loopback measurment at 192kHz.png

So I'm sure you can get >100kHz BW with a dedicated interface, even an inexpensive one.

Just keep in mind that the maximum sample rate dictates the achievable bandwidth.

For >100kHz BW, an interface with a sample rate of 384kHz or higher is required.

FWIW, the Cosmos ADC, with the calibration files provided by the manufacturer, has a flat frequency response all the way up to 384kHz.
 
My motherboard's cheap onboard audio is flat up to 80kHz (DAC->ADC Loopback):
View attachment 372577

So I'm sure you can get >100kHz BW with a dedicated interface, even an inexpensive one.

Just keep in mind that the maximum sample rate dictates the achievable bandwidth.

For >100kHz BW, an interface with a sample rate of 384kHz or higher is required.

FWIW, the Cosmos ADC, with the calibration files provided by the manufacturer, has a flat frequency response all the way up to 384kHz.

How do we find out if it has analog filtering in place that will negatively affect bandwidth?
 
How do we find out if it has analog filtering in place that will negatively affect bandwidth?
By measuring it. You can do a loop back with an on board card. Or by feeding the input to the on board card with a wideband signal from another source.
 
The MOTU M2 or M4 is probably one of the best at lower prices. According to Julian Krause it is flat to 90 khz or so.
Look around the 3:30 mark in the video. The Cosmos ADC is pretty close to SOTA in every parameter.
I think he's more interest in SINAD.
FR is usually flat to all of them up to a point:

FR.PNG

E-MU is -1dB at about 86Khz and it's 20yo.
 
I think he's more interest in SINAD.
FR is usually flat to all of them up to a point:

View attachment 372578

E-MU is -1dB at about 86Khz and it's 20yo.
Lots of the cheaper interfaces top out around 30 khz. Not all do. As you get into more expensive devices you can pretty well count on them being limited by the sample rate. Amir I think measured the M2 as good to 84 khz at 192 khz sample rate on the ADC. It had a SINAD of 96 db, but that was from distortion. The noise levels were lower and so the dynamic range would be as well. As you point out some have rising noise above 20 khz. The M2 seems okay in that regard.

A Focusrite 3rd gen 2i2 has flat response to 48 khz, but steeply rising noise above 30 khz. Even so it is still -70 db or more down at the 80 khz range. OTOH, Focusrite 4th gen devices don't have response beyond about 30 khz.

 
Back
Top Bottom