• 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!

Question: Behringer UMC 204, Distortion on main out

Lilith

Active Member
Joined
Mar 28, 2021
Messages
267
Likes
79
Hi,

I'm just fiddling with my UIMC204 interface to see how strong the distortion is on its outputs and how it depends on the Main Volume.
I connected one channel of the outpout (which normally goes to my monitor controller) to the left input. I set the input gain so that it it does not clip at the input.
The signal of the left input I checked with a spectrum analyzer playing a sine tome at 1 kHz. I don't know the absolute levels, but what I get is this:
V4XoeGq.png


In red you see the signal from the PC and in cyan the signal that goes from the interface output into its input. There's significant amount of distortion. I don't know if this is comparable with Amirs measurements: https://www.audiosciencereview.com/...ringer-umc204-hd-audio-interface-review.9856/ but it seems worse to me.
The main volume in this case is at 100% (~5 o clock). When I dial the volume back a bit to ~2 o clock it looks better, but I'm loosing 12 dB of volume, which is not acceptable. Does anyone know if the unit is really that s h i t t y?

tABH0rY.png
 
OP
Lilith

Lilith

Active Member
Joined
Mar 28, 2021
Messages
267
Likes
79
and with REW... it's between ~0.5 - 0.8 % THD (0.8% at maximum output volume). :rolleyes:
This is when the UMC 204 is connected to an active USB Hub:

TAof39m.png


And this is when the hub is not powered. > more noise, more distortion.

40SspM2.png
 
Last edited:

Lalaland

Member
Joined
Oct 5, 2022
Messages
22
Likes
41
Location
Norway
I am going through the exact same test as you present on my UMC204, witch lead me to your thread. Found it was a good opportunity for me to register here and not just lurk around.
I got about the same results as you present, although I have been able to improve results a bit by fiddling with levels. I'll see if I can post a few measurements later, if you are interested. There is info out there about possibly improving distortion by correct use of TRS to RCA plugs and cables, but I have not tested this yet
 
OP
Lilith

Lilith

Active Member
Joined
Mar 28, 2021
Messages
267
Likes
79
I am going through the exact same test as you present on my UMC204, witch lead me to your thread. Found it was a good opportunity for me to register here and not just lurk around.
I got about the same results as you present, although I have been able to improve results a bit by fiddling with levels. I'll see if I can post a few measurements later, if you are interested. There is info out there about possibly improving distortion by correct use of TRS to RCA plugs and cables, but I have not tested this yet
I was mainly interested in the distortion of the TRS outputs. I don't hear any distortions with sine tones so I guess it musst be better than the measurement (which also includes distortion of the inputs.)
 

Lalaland

Member
Joined
Oct 5, 2022
Messages
22
Likes
41
Location
Norway
Hi again, sorry I was going to give you a response but whole family been having "fun" with the flu or something like it.
Anyways, I believe what you posted a picture of is mostly distortion on the front inputs - that is to say, the input preamps.
Here are TRS insert measurment that bypass those preamps.

Not perfect, but I don't think output is too bad. Measured at main all the way up (only 717mVrms)
Behringer Umc204HD Distortion.jpg
 

pLudio

Member
Forum Donor
Joined
Dec 8, 2020
Messages
95
Likes
86
Location
Sweden
Hi again, sorry I was going to give you a response but whole family been having "fun" with the flu or something like it.
Anyways, I believe what you posted a picture of is mostly distortion on the front inputs - that is to say, the input preamps.
Here are TRS insert measurment that bypass those preamps.

Not perfect, but I don't think output is too bad. Measured at main all the way up (only 717mVrms)
I measure 1.09 Vrms with a 0 dBFS sine on both RCA and TRS Main Out. Did you forget to set Mix to full PB?

Used a cable where RCA pin goes to TRS ring and sleeve to sleeve and took some loopback measurements.

Clipping:
file.php


Not clipping:
file.php


file.php

file.php


file.php
 

pLudio

Member
Forum Donor
Joined
Dec 8, 2020
Messages
95
Likes
86
Location
Sweden
Measurements indicate that the Clip LED turns off slightly before it stops clipping so a little extra margin is needed, and you want a bigger margin anyway considering the harmonic distortion.

This time I grabbed an XLR cable and attached TRS adapters on both ends and measured Main Out TRS at max vol with Mix at max PB (1.09 Vrms) inserted at the front with the Midas pre-amp. Line mode, no Pad. I adjusted the pre-amp gain to get the same input levels as last time. Gain was set to about 1 o'clock.

Clipping like last time:
file.php


Not clipping but still ******:
file.php


file.php


file.php


file.php
 
Last edited:

pLudio

Member
Forum Donor
Joined
Dec 8, 2020
Messages
95
Likes
86
Location
Sweden
So I removed the TRS adapter in the front and connected the XLR directly in the combo jack and got a much higher input signal. What's going on here?

With the Midas pre-amp set to minimum(!) gain and the output signal lowered even further I get:
file.php
 
Last edited:

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,338
Likes
3,278
Location
.de
So I removed the TRS adapter in the front and connected the XLR directly in the combo jack and got a much higher input signal. What's going on here?
In a combo input, it is typical for the TRS portion to have extra attenuation to accept line level (so the jacks are not directly connected). Often it's just some extra series resistance that's forming a voltage divider with existing mic input impedance, e.g. 2x 22k / 3k.

The preamps in these Behringers have a reputation for generating a lot of H2, and that certainly seems to be the case when cranking it. Not exactly hi-fi, this. Gain at 1 o'clock isn't even what I'd consider very high.
 

Lalaland

Member
Joined
Oct 5, 2022
Messages
22
Likes
41
Location
Norway
I think I worded myself poorly in that last post, as to why my output was only 0.7Vrms. Indeed you are right, max output is just above 1Vrms. I turned down output to get away from clipping the input(Inserts at the back). I just did a test to verify, and I get about the same results you are presenting.

After tinkering and testing this device, I am mostly not very happy and am probably gonna change it for something better at some point. I really don't like that anaemic output and the fact that they are not even properly balanced
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,338
Likes
3,278
Location
.de
It's about all you can get from a bargain basement design running on 5 V single-supply - at least in years gone by, this interface used to be extremely cheap for something this feature-packed. (Literally HALF the price of a Scarlett 2i4 2nd gen, or even less.) At least the mic preamps have an actually acceptable noise floor for dynamic mics, which seems to involve a bit of a hack by itself, as the NJM2122 opamp they used is not normally stable below 30 dB of gain (it was probably developed as a mic preamp for portable tape / DAT / MD recorders).

Apparently the old Behringer ADA8000 ADC/DAC has similar distortion issues with its input amps (the ADA8200 seems substantially better). It's a good thing I researched that - it seems to be using some ADC and DAC chips that I had never even heard of (AL1101 and AL1201 by Wavefront Semiconductor, ex Alesis Semiconductor). You live and learn.
 

pLudio

Member
Forum Donor
Joined
Dec 8, 2020
Messages
95
Likes
86
Location
Sweden
So the Line/Inst button doesn't do anything for XLR while it increases the signal by 23 dB for TRS when in instrument position. The Pad button attenuates the signal by 20 dB for both XLR and TRS.

I tried to measure the Midas pre-amp at it's best but I'm sure it can be slightly improved. -12 dBFS in signal looked pretty good when measuring the inserts so I aimed for that.

Main Out TRS to XLR, which is always in instrument mode:
file.php


file.php


file.php


I pressed the Pad button now to attenuate the signal by 20 dB and keep the noise floor lower:
file.php


file.php


file.php


file.php


file.php


file.php


file.php


file.php
 

Lalaland

Member
Joined
Oct 5, 2022
Messages
22
Likes
41
Location
Norway
So the Line/Inst button doesn't do anything for XLR while it increases the signal by 23 dB for TRS when in instrument position. The Pad button attenuates the signal by 20 dB for both XLR and TRS
I'm not sure what you will end up using this interface for, but I believe you should be careful with the settings for inputs. Inst(Instrument) setting is as you point out much more sensitive, and is for guitars and keyboards etc. with a very low output signal. A line level input signal will probably cause some damage if used with Instrument setting
Apparently the old Behringer ADA8000 ADC/DAC has similar distortion issues with its input amps (the ADA8200 seems substantially better). It's a good thing I researched that - it seems to be using some ADC and DAC chips that I had never even heard of (AL1101 and AL1201 by Wavefront Semiconductor, ex Alesis Semiconductor). You live and learn.
Do you think there's potential in the ADA8200? I have seen measurements which seems quite good for what it costs. I'm kind of uncertain, It seems to cheap to be good
 
OP
Lilith

Lilith

Active Member
Joined
Mar 28, 2021
Messages
267
Likes
79
I'm not sure what you will end up using this interface for, but I believe you should be careful with the settings for inputs. Inst(Instrument) setting is as you point out much more sensitive, and is for guitars and keyboards etc. with a very low output signal. A line level input signal will probably cause some damage if used with Instrument setting

Do you think there's potential in the ADA8200? I have seen measurements which seems quite good for what it costs. I'm kind of uncertain, It seems to cheap to be good

I doubt there will be a damage when having the "wrong" setting, otherwise the unit sucks.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,338
Likes
3,278
Location
.de
Main Out TRS to XLR, which is always in instrument mode:
It looks like the preamp is really running out of loop gain at high gain settings. I don't dare imagine what happens at, say, 7 kHz.

Looks like one of the inexpensive Klark Teknik mic boosters (same parent company) may be useful when you want highest fidelity out of a dynamic mic. Or just don't turn up the gain much beyond 3 o'clock and compensate for the modest levels in post. Preamp noise floor comes up to about 22 dB above ADC noise at full gain, so you can pretty much back off by 20 dB and still have better EIN than a Steinberg UR22II/C.

Do you think there's potential in the ADA8200? I have seen measurements which seems quite good for what it costs. I'm kind of uncertain, It seems to cheap to be good
It seems to be a "does what it says on the box" kind of deal. Input dynamic range at 103 dB(A) is about in line with what other decent CS4270-based interfaces get, while the output side at 104.5 dB(A) appears to fall slightly short of the values around 106 dB(A) that seem possible otherwise (clearly the DAC side of this codec is slightly beating its spec). Also, sample rates are 44/48 kHz only, since S/MUX is not supported.

It's a "gets the job done" device with "gets the job done" converters and digital filters. If you just need some (less critical) extra 48 kHz input channels in the home studio and know how to set your levels halfway sensibly, sure, why not. People have recorded on worse equipment. It's obviously very much bread and butter and not a Focusrite Scarlett OctoPre or an Audient ASP800, but those cost twice or >3 times as much respectively.

(Speaking of Behringer, the piece that I'd like to see tested in detail the most would have to be the UMC1820. I expect a generic CS4272 grave littered with NJM4580s, but the question is whether the Focusrite equivalent is that much better.)
 

Lalaland

Member
Joined
Oct 5, 2022
Messages
22
Likes
41
Location
Norway
It seems to be a "does what it says on the box" kind of deal. Input dynamic range at 103 dB(A) is about in line with what other decent CS4270-based interfaces get, while the output side at 104.5 dB(A) appears to fall slightly short of the values around 106 dB(A) that seem possible otherwise (clearly the DAC side of this codec is slightly beating its spec). Also, sample rates are 44/48 kHz only, since S/MUX is not supported.

It's a "gets the job done" device with "gets the job done" converters and digital filters. If you just need some (less critical) extra 48 kHz input channels in the home studio and know how to set your levels halfway sensibly, sure, why not. People have recorded on worse equipment. It's obviously very much bread and butter and not a Focusrite Scarlett OctoPre or an Audient ASP800, but those cost twice or >3 times as much respectively.

(Speaking of Behringer, the piece that I'd like to see tested in detail the most would have to be the UMC1820. I expect a generic CS4272 grave littered with NJM4580s, but the question is whether the Focusrite equivalent is that much better.)
I'm not so concerned with ADC, as for my use it's multichannel DAC/active speaker duty. I guess it's not much use as I need something that measures better than my MiniDSP 4x10HD if I was to upgrade. I would prefer to move DSP duties to some software solution at some point
 

pLudio

Member
Forum Donor
Joined
Dec 8, 2020
Messages
95
Likes
86
Location
Sweden
It looks like the preamp is really running out of loop gain at high gain settings. I don't dare imagine what happens at, say, 7 kHz.
While the TRS outputs on the back are fake balanced, the inputs on the fronts are properly balanced so I connected a Topping D10 Balanced to XLR (instrument mode), pushed Pad, set the amplification to 5 o'clock, -53 dBFS out from the D10B, ca -12 dBFS in, 7 kHz + 13 kHz. Looks surprisingly good, no funny 20 kHz IMD peak sticking out. Much cleaner signal source than loopback.
D10B_instrument+pad_7+13_5o.png
 
Last edited:

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,338
Likes
3,278
Location
.de
Huh. Interesting. So perhaps the input is somewhat susceptible to common-mode distortion then? The dominant H2 should have been a giveaway. That would be good news for a number of (but not all) microphones.

But that makes no sense, CM distortion would improve with the lower input levels seen at higher input gain. Weird.

(I almost can't believe it would be the output. That being said, a H2 at -72 dBFS at full input gain would be at -123 dBFS relative to the output side if monitor gain was set to align 0 dBFS levels at minimum input gain. It is not at all uncommon to see such remaining low-level nonlinearity in old converter designs, and the CS4272 DAC can trace its lineage to the 1999 CS4392 after all. This is mostly an issue for early multibit delta-sigma DACs though, including the CS4398 which performs at like 20 bits effective. This one is only like a 2-bit modulator or something though, otherwise it wouldn't require such an obscene amount of noise shaping. I guess you could switch the generator to 16-bit output with dithering and see whether that makes any difference.)
 
Last edited:
Top Bottom