Audient Sumo Summing Mixer - causes to uneven 16 channels

[canidoit]

Hi,

I was having a look at the internals of the Audient Sumo and trying to find a way of calibrating this unit so that all 16 channels signal would be equal, as I noticed the signal is not outputting it equally. Some signals were between 0.1db - 0.8db differences.

I am presuming the Component side board is where this issue is happening but it is all resistors, capacitors, ICs, etc.? Am I correct? I was hoping there were like trimpots on each channel so I can adjust the levels.

How would I go about troubleshooting this - would I use a multimeter and touch ground with the negative and on the positive touch a component end on the left side of the board and check what the voltage is? Do I look at the voltage difference and where it is different is where the issue lies?

How does a technician fix the uneveness issue, what usually causes this?

Can I insert trimpots somewhere on this board and make this adjustable?

Does anyone know where I can obtain a service manual for this unit?

(image from kingkorg post)

 

[Brian Roth]

A 0.1 dB level difference is waaay beyond what you need to worry about. 0.8 dB....maybe.

Audient is still in business, and they are a decent company. Contact them to see if they have any service info available.

https://audient.com/
Bri

 

[canidoit]

Hi Brian,

I think accumulatively it is affecting the result, as the center is off when I listen to the summed audio.

I am reading up on potentiometers and apparently they add distortion or affect the sound quality of the signal so I am wondering whether this is proportional to the range of the potentiometer? So if I add a trimpot that reduces the signal from 0-2db, would this be low enough not to affect the audio quality?

I am thinking of possibly snipping those resistors legs at the end of the Component Side boards and inserting a trimpot(0-2db reduction) between it. Looks like all 16 inputs are all funnelling in from the DB25 input connectors to those resistors. Anyone think this is a good idea? To confirm if this is audio, I would connect the positive of an earphone to one of those resistor legs and the negative to the chassis and see if I hear audio?


From what I recall when I was making -10db pad XLR cables, I think 10K or 100K resistor is -10db reduction on the hot line level audio signal? So possibly use a trimpot that is 2K?

 

[Ike Zimbel]

I would do some frequency/amplitude sweeps before starting to carve up your circuit board. In my experience, the single most common cause of imaging issues is capacitor failure. I was just in an Audient desk a few months ago to replace a bunch of electrolytics that had gone bad from heat build up.

 

[zamproject]

Looks like all 16 inputs are all funnelling in from the DB25 input connectors to those resistors

Nope ?!? this look much more like 8 SYMetrical line input from one DB25
there is probably the same layout duplicated on the PBC bellow for another 8 line in

Best
Zam

 

[JohnRoberts]

when troubleshooting a multichannel sku like that you can compare channels to each other . Identifiy the one that is low (or hi?) and then dig deeper into the outlier channel to see what is different of changed.

JR

 

[Newmarket]

Hi Brian,

I think accumulatively it is affecting the result, as the center is off when I listen to the summed audio.

Not understanding what that means. It sounds like a panning issue rather than a channel level thing ?

I am reading up on potentiometers and apparently they add distortion or affect the sound quality of the signal

Where did you read up on potentiometers causing distortion ?
No they don't (well anything can if you take it far enough but in reality they don't for these purposes) unless they are seriously broken eg oxidised to the extent that a low signal level cuts out. But your initial issue was level but now you're suggesting distortion ?
The problems that occur with potentiometers are scratchiness whilst being moved and accuracy of the taper particularly if not a linear taper.

 

[canidoit]

Hi Newmarket,

It is a level thing, as I have done test tone checks on individual channels and they fluctuate between 0-0.8db differences. I ran a mix as 8 stereo stems without calibrating the feed to the Sumo using software, and the center image was off noticeably. When I calibrated using software so that it fed the Sumo equally, the center image was center.

There are a couple of posts online that I read discussing distortion from potentiometers. Doesn't adding resistors to an audio signal path add some form of distortion?

Thank you all for your replies!

 

[JohnRoberts]

Hi Newmarket,

It is a level thing, as I have done test tone checks on individual channels and they fluctuate between 0-0.8db differences. I ran a mix as 8 stereo stems without calibrating the feed to the Sumo using software, and the center image was off noticeably. When I calibrated using software so that it fed the Sumo equally, the center image was center.

Identify the channel that is high/low amplitude and dig deeper into that channel to find the specific cause.

There are a couple of posts online that I read discussing distortion from potentiometers.

In general NO. Hypothetically a dirty wiper contact could introduce some nonlinearity but I have never seen (heard) that happen. I have seen dirty wipers, but not distortion from one.

Doesn't adding resistors to an audio signal path add some form of distortion?

Again in general NO. I have seen some extreme cases where a cheap resistor introduced distortion (from voltage coefficient) in a high power audio amp feedback network. That resistor was exposed to > 100V. At typical audio signal levels even cheap resistors are linear.

Thank you all for your replies!

If you have an expensive enough test bench you can find nonlinearities in everything. Some adventurous experimenters had to build their own very low distortion test equipment (like ultra clean sine wave sources) to measure distortion levels in passive components.

JR

 

[Newmarket]

If you have an expensive enough test bench you can find nonlinearities in everything. Some adventurous experimenters had to build their own very low distortion test equipment (like ultra clean sine wave sources) to measure distortion levels in passive components.

+1

 

[Newmarket]

Hi Newmarket,

It is a level thing, as I have done test tone checks on individual channels and they fluctuate between 0-0.8db differences. I ran a mix as 8 stereo stems without calibrating the feed to the Sumo using software, and the center image was off noticeably. When I calibrated using software so that it fed the Sumo equally, the center image was center.

I see. I think. Are you using the potentiometers around the 0dB point ?

There are a couple of posts online that I read discussing distortion from potentiometers. Doesn't adding resistors to an audio signal path add some form of distortion?

No. With the caveats that JR has stated. Voltage coefficient is a thing but it's tiny tiny in this context.
Capacitor voltage coefficient is more of a thing using eg X7R dielectric rather than NP0. But even then it's very small.
And not related to a level difference.
Where did you read those posts online ? To put it very mildly - the internet is not always a source of reliable information.

 

[ccaudle]

Some signals were between 0.1db - 0.8db differences

What measurement tools are you using? Are you sure your tools are accurate enough to measure 0.1dB consistently?

Next step, are you sure that the audio outputs of whatever device you are using are consistent to within 0.1dB? Fixing a problem of poor output calibration by modifying the summer seems like fixing the problem in the wrong place to me.

Are all the inputs coming from the same device? If not, then a few tenths of a dB could be caused by differing levels and even differing output impedance of different devices.

I was hoping there were like trimpots on each channel so I can adjust the levels.

Trimpots would require adjustment in the factory (extra time and expense), and are in general not as stable as precision resistors. I would expect a device like that to just use 1% or tighter resistor values, that should get within 0.1dB matching (assuming identical signal levels, and identical output impedance of the devices driving all the channels).

would I use a multimeter

Keep in mind that multimeters have restricted frequency range, so you need to choose your signals to make sure you stay in the most accurate range of the meter. You will have to check your meter manual to verify the accuracy and frequency range limitations in AC voltage mode.

touch ground with the negative and on the positive touch a component end

Assuming you have a battery powered meter, you would measure between two components, because you are dealing with balanced signals (which are referenced from signal leg to signal leg, not from signal leg to ground). Assuming that the marketing info is correct and the Sumo keeps balanced signal path all the way through. A schematic would help to verify.

@kingkorg did you ever make a schematic of the entire Sumo? I saw on an old post that you had traced out the monitor control section, but did not reply to the last post which asked if you had a schematic of the entire unit.

apparently they add distortion or affect the sound quality of the signal

I think you have been reading too much nonsense if you are worried about distortion of trimpots and resistors. Metal film resistors are the most linear components in use. In comparison to metal film resistors trimpots may have measurable distortion, but just toss out any info you have that makes claims about distortion but doesn't provide graphs comparing distortion of different devices. State of the art distortion measurements can get down to nearly -160dB referenced to full scale, so anyone who starts talking about distortion of resistors but isn't providing data, and also discussing audibility concerns (i.e. even if you can measure the distortion it is not possible to hear distortion from a decent resistor) then that source is not worth your time.

But that brings up another question, namely why are you even concerned with distortion? The only reason to use an analog summing mixer of that style is to add distortion because you like the sound of the particular kind of distortion it makes. If you are really concerned about distortion then mix in the DAW. Mixing is in the simplest terms adding numbers together. Unless the DAW you are using is really incompetently designed it will be able to add numbers together without errors.

So to summarize:
Verify your tools are appropriate.
Verify your input signals are matched.
Trace out the circuitry so you know the likely places inaccuracy could creep in.
After you understand the circuitry then you can compare channel-to-channel between two worst case channels to see where the difference in level first occurs.

 

[Newmarket]

Let's wait for someone to advocate bulk metal foil over metal film now...

 

[ccaudle]

bulk metal foil over metal film

Metal foil actually have a non-obvious distortion mechanism that causes higher distortion than metal film.
This link is directly to that discussion, general resistor discussion starts at around 18:30. The whole video is really good (one of the co-founders of Audio Precision presenting to undergraduates at Oregon State):
Discussion of metal foil resistors in low THD design presentation

"Low frequency thermal modulation distortion in the range of 10-200 Hz is typically much higher than that implied by its ultra-low TCR specification.
This phenomenon is currently unexplained...."

As I said, only listen to someone who can provide data to back up the claims.

 

[kingkorg]

I have the schematic, but not the permission to share it. I acquired it through regular path, just ask Audient for the schematic.

There are some essential issues with this topic. If you want "quality" you just don't run the sound through additional external unit such as summing mixer. You run it through the unit to get whatever it's bringing to the table.

Reducing "distortion" on this unit is just bollocks, this unit is all about adding harmonic content. It does miracles in this respect when pushed, just use audio interface with enough juice to push it properly. I love it, and use it all the time.

0.8db difference? If you are so picky, just compensate before the input. When you mix you do it through the summing mixer. You don't mix something in the box and just blindly run it through a summing mixer in the end, that gets you nowhere.

Would you mod an SSL or Neve to achieve the same?

The only mod i did to this unit is compressor sidechain low cut to prevent pumping with low end heavy content. This unit does wonders when used properly, needs no "improvement" whatsoever.

 

[Newmarket]

Metal foil actually have a non-obvious distortion mechanism that causes higher distortion than metal film.
This link is directly to that discussion, general resistor discussion starts at around 18:30. The whole video is really good (one of the co-founders of Audio Precision presenting to undergraduates at Oregon State):
Discussion of metal foil resistors in low THD design presentation

"Low frequency thermal modulation distortion in the range of 10-200 Hz is typically much higher than that implied by its ultra-low TCR specification.
This phenomenon is currently unexplained...."

As I said, only listen to someone who can provide data to back up the claims.

Okay. I was being light hearted in that comment because it's something that I've seen voiced in various 'audiophile' opinions but never been convinced myself.
In that context it's probably useful to say that I have been involved in the design of extreme precise scientific instrumentation and never needed to consider metal foil rather than metal/thin film technologies.
I'll take a look at the link when I have time. Esp as I spent some time with an AP2 in the last century.

 

[canidoit]

I have the schematic, but not the permission to share it. I acquired it through regular path, just ask Audient for the schematic.

There are some essential issues with this topic. If you want "quality" you just don't run the sound through additional external unit such as summing mixer. You run it through the unit to get whatever it's bringing to the table.

Reducing "distortion" on this unit is just bollocks, this unit is all about adding harmonic content. It does miracles in this respect when pushed, just use audio interface with enough juice to push it properly. I love it, and use it all the time.

0.8db difference? If you are so picky, just compensate before the input. When you mix you do it through the summing mixer. You don't mix something in the box and just blindly run it through a summing mixer in the end, that gets you nowhere.

Would you mod an SSL or Neve to achieve the same?

The only mod i did to this unit is compressor sidechain low cut to prevent pumping with low end heavy content. This unit does wonders when used properly, needs no "improvement" whatsoever.

Hi kingkorg, I am currently adjusting before the Sumo input using the DAW but my concern was that if I decided to insert outboard on multiple Sumo inputs.
DAW Stem 1 > SB4001 > A > Sumo input 1-2
DAW Stem 2 > API2500 > B > Sumo input 3-4
So I would need to somehow insert a device that can shift the balance on A and B so that the Sumo receives it compensating for its level irregularities.

I would prefer not to feed an adjusted signal to the outboard, as that might just emphasize the issue, makes the issue more complex or alter how the outboard behaves when it does not receive a centered stereo signal.

Just the shift in center is what bothers me, as I know, this causes a shift in how I have positioned all the elements and it's levels in my mix. I know the right way is to mix into the Sumo from the start, as you are already making decisions compensating for the 16 channels of unevenness, but I I have a system happening where I only add the summing on a completed stereo mix for experimentation and creativity.

I probably would try to fix my SSL/Neve if the center was shifted and it had limited controls where you would not be able to adjust the levels on the left and right signal so that it comes out dead center on a stereo signal.

 

[kingkorg]

I understand the issue. But i honestly don't think you should obsess that much, it's a negligible difference. You should have no issues acquiring the schemo from the Audient though. If you eliminate all of those quirks, you end up with the same exact thing you put into it, so it defies the purpose. I don't think your api2500 has matched channels to that extent either.

 

[gamcneer]

Hi,

I was having a look at the internals of the Audient Sumo and trying to find a way of calibrating this unit so that all 16 channels signal would be equal, as I noticed the signal is not outputting it equally. Some signals were between 0.1db - 0.8db differences.

I am presuming the Component side board is where this issue is happening but it is all resistors, capacitors, ICs, etc.? Am I correct? I was hoping there were like trimpots on each channel so I can adjust the levels.

How would I go about troubleshooting this - would I use a multimeter and touch ground with the negative and on the positive touch a component end on the left side of the board and check what the voltage is? Do I look at the voltage difference and where it is different is where the issue lies?

How does a technician fix the uneveness issue, what usually causes this?

Can I insert trimpots somewhere on this board and make this adjustable?

Does anyone know where I can obtain a service manual for this unit?

(image from kingkorg post)

 

[emrr]

The output of every converter set I’ve looked at has had similar gain variation, same with consoles. Anywhere there’s a gain or pan pot, there’s likely similar, in anything analog. Gain steps on really old stuff is usually 2db, fancier stuff 1.5db.

Hell, your average 1/4” plastic jack patchbay will introduce slight level variations. One of the ways I monitor patchbay connection condition is to record and compare tones through all channels, so I have a benchmark to compare against. Easy to spot jacks that aren’t making a great connection as they age/fatigue. Then there’s gain variation on converter inputs too….