EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: lescro on August 04, 2021, 10:36:22 am
-
When I use the headphones and listen to the input signal of my audio interface (Steinberg MR816CSX) it sounds slightly distorted.
The levels are all set correctly so there is no clipping. No EQ or COMPRESSION is being used.
I want to determine if it is actually distorting or not using my test equipment which is :
o Brymen BM869 multimeter
o Rigol DS1104 oscilloscope
o UNI-T UTG962E function generator https://www.uni-trend.com/html/product/NewProducts/UTG900E%20Series/UTG962E.html (https://www.uni-trend.com/html/product/NewProducts/UTG900E%20Series/UTG962E.html)
STEINBERG MR816 CSX SPECS
The MR816 is an analogue to digital, digital to analogue converter.
Operation Manual: https://download.steinberg.net/downloads_hardware/MR816/MR816_documentation/MR816_GettingStarted_en.pdf (https://download.steinberg.net/downloads_hardware/MR816/MR816_documentation/MR816_GettingStarted_en.pdf)
Specifications: https://www.steinberg.net/fileadmin/redaktion/PDF/Hardware/Technical_Specifications_MR816.pdf (https://www.steinberg.net/fileadmin/redaktion/PDF/Hardware/Technical_Specifications_MR816.pdf)
ANALOGUE INPUT
**************
MIC/LINE IN ....... XLR connector - balanced
Nominal Level: -60dBu to 10dBu
Maximum Level: +24dBu
Input Impedance: 3.5k Ohms
ANALOGUE OUTPUT
***************
LINE OUT ...... TRS connector - balanced
Nominal Level: +4dBu
Maximum Level: +18dBu
Input Impedance: 600 Ohms
HEADPHONE JACKS (2 of)
********************
Jack 1 & 2 ..... TRS connector - unbalanced - stereo
Nominal Level: 4mW + 4mW
Maximum Level: 25mW + 25mW
Input Impedance: 8 Ohms
Nominal Level: 12mW + 12mW
Maximum Level: 75mW + 75mW
Input Impedance: 40 Ohms
I think the Input Impedance quoted in the specs for the headphone outputs is for the input impedance of the headphones being used.
UNI-T UTG962E SPECS
The UTG962E is a Function/Arbitrary Waveform Generator.
Amplitude Range: < 10MHz 1mVpp to 10Vpp (50 Ohms)
User Manual:https://www.uni-trend.com/uploadfile/cloud/English%20manual/Benchtop%20Instrument/UTG900E%20English%20Manual.pdf (https://www.uni-trend.com/uploadfile/cloud/English%20manual/Benchtop%20Instrument/UTG900E%20English%20Manual.pdf)
MY QUESTIONS
I thought I could test the MR816 for distortion by inputting a sine wave and then looking at the output of the LINE OUTS and and also the output of the HEADPHONES with the oscilloscope set to FFT.
Connecting the Function Generator
******************************
1. The MR816 inputs are balanced (3 wires), so should I connect one wire of the Function Generator to HOT and the other wire to COLD and GROUND?
2. What amplitude should I use for the sine wave? Say, 0.894Vpp (ie -10dBV signal) or 3.472Vpp (ie +4dBu signal)
3. What frequency should I use for the sine wave? 1 kHz?
Connecting the Oscilloscope to the LINE OUT
***********************************
4. I will need to connect a load across the output (balanced - 3 wires). What should the size of the load be? Should it be a high value like 30K Ohms (This would match a bridging input) or should I use a 600 Ohm load?
5. The MR816 outputs are balanced (3 wires), so should I connect one wire of the load to HOT and the other wire to COLD and GROUND?
Connecting the Oscilloscope to the HEADPHONES
***************************************
6. I will need to connect a load across the headphone's output (unbalanced - 3 wires - stereo, so will do a test on left side then a test on right side). Would I use a 50 Ohm load for this?
Changing the Source Impedance
**************************
7. My Function Generator has an impedance of 50 Ohms. So my source impedance is 50 Ohms. If I needed the source impedance to be 150 Ohms, how would I do this?
Cheers
Lescro
-
aside from your questions is one important point: real distortion measurement require more than a scope and a signal source.
below a certain level of distortion, you're unable to detect them just with the scope monitoring the output signal; you'll need a notch filter which eliminates the injected signal and leaves the distortion and intermodulation untouched.
easiest possible way to achieve that is an active twin T notch filter with a very low noise opamp.
impedance matching is no issue in audio circuits; common output impedance is 600 ohms while input impedance is 10k, so you have voltage matching more or less out of the box; the 50 ohms of the signal source doesn't matter
for a balanced input you should inject a signal with 180° phase offset between hot and cold; most arbitrary generators with 2 signal outputs allow static adjustment of the phase between the 2 channels
caveat: if the interface supports phantom power, disable it before you burn the output drivers of your func-gen by accident.
-
Thank you HB9EVI for your reply.
I just need to clarify some things.
1. To look at the signal from the LINE OUTS that are balanced, would I place a 600 Ohm load across HOT and GROUND and measure across HOT and GROUND?
2. I would still use the FFT function on the oscilloscope, but have the oscilloscope connected to the notch filter output. The notch filter input would be from HOT and GROUND of the LINE OUT.
3. Are there any notch filter ICs or active notch filter circuits you would recommend?
Thanks
Retsel
-
hi there
like I said, in audio circuitry you don't work with impedance matching like in rf circuits; very likely your line out driver could drive 600ohms, but not necessarily; if you see need for a termination, use a 10k resistor each between hot-gnd and cold-gnd, but it's not really needed
the FFT function of most DSOs is rather a toy function than a measurement tool; much better results you get with a higher quality sound card and a piece of software; I use a E-mu 0204 which allows quite reliable measurements in a dynamic range of about 85dB; if you want more, it gets very fast expensive.
an all-in-one notch filter circuit I don't know about; best way is to carefully match the Cs and Rs from a bigger pile and use a very low distortion opamp; if I remember right, I used a OPA2134 for my twin T notch. I followed this guide here: https://sound-au.com/project52.htm
the trickies part is to determine the precise resonance frequency of the filter setting, but that's a one-time work only
-
Hi HB9EVI
Thanks for the clarification.
I have a RME Fireface UCX Audio Interface. So what you are saying is that I could use it with the appropriate software to determine the THD. I have heard of ARTA software (but never used it.) Is ARTA a good package?
Cheers
Retsel
-
hi there
in general these audio interfaces are quite decent interfaces with good technical parameters; I don't know your Steinberg, but I have the mentioned Creative, an older 2ch Tascam and a newer 4ch Behringer; they all do quite fine work; I mostly use the Creative E-mu to measure my home brew PAs; as linux only user I cannot say much about the windows world; i use jack with jaaa and japa
-
Connecting the Function Generator
******************************
1. The MR816 inputs are balanced (3 wires), so should I connect one wire of the Function Generator to HOT and the other wire to COLD and GROUND?
2. What amplitude should I use for the sine wave? Say, 0.894Vpp (ie -10dBV signal) or 3.472Vpp (ie +4dBu signal)
3. What frequency should I use for the sine wave? 1 kHz?
I'd do 775mV RMS sine at 1KHz (0dBu). That should produce a -18dBFS signal in the digital domain, at least in pro gear. Start by establishing that, and verify your gain structure. This voltage should come out at the same level on the line outputs. (If you're looking at this on the scope, beware that the waveform will show more than 775mV, since the scope presents Peak on the graticule, not RMS. The formula to convert is Vpeak = Vrms * sqrt(2). Further, you'll see peak-to-peak, meaning you need to multiply with 2 as well. )
If the voltage out is the same as in, in unity gain mode, then there might be trouble, if it still sounds bad.
Connection-wise I'd start with XLR pins 2 and 3, BNC center going to 2, and playing with connecting 1 and 3 together.
Connecting the Oscilloscope to the LINE OUT
***********************************
4. I will need to connect a load across the output (balanced - 3 wires). What should the size of the load be? Should it be a high value like 30K Ohms (This would match a bridging input) or should I use a 600 Ohm load?
5. The MR816 outputs are balanced (3 wires), so should I connect one wire of the load to HOT and the other wire to COLD and GROUND?
High impedance load -- say 10K or so. XLR to BNC as above.
Connecting the Oscilloscope to the HEADPHONES
***************************************
6. I will need to connect a load across the headphone's output (unbalanced - 3 wires - stereo, so will do a test on left side then a test on right side). Would I use a 50 Ohm load for this?
600Ω should be about right.
Changing the Source Impedance
**************************
7. My Function Generator has an impedance of 50 Ohms. So my source impedance is 50 Ohms. If I needed the source impedance to be 150 Ohms, how would I do this?
I'd not bother. As others wrote, this is not at transmission line frequencies; but lower. 50Ω source will do fine.
Oh, and what you actually need is an analog distorsion analyser, like a Tek AA501A. :D