Hypex NC252MP Balanced Volume control, am I doing this right?

[Bartholomeus]

First of all I want to thank the ASR community for all the great readings I had the past few months, it helped me a lot making discissions for my new audio gear without all the fuss.

My audio setup is as follows:

Source: PC (USB)
DAC: Topping D50 III
Power amp: Audiophonics AP300-S250NC (Hypex NC-252MP)
Speakers: Bowers & Wilkins 707 S2

What bothers me is the fact I only have digital volume control, especially after reading some cases where the DAC went on to 100% volume. I’ve had a similar issue in the past with an older integrated amp so I want to put a potentiometer (ALPS RK27 10KOhm Quad channel) in between the DAC and Poweramp.

Of which I thought to be a simple implementation, turned out to be a headache for the past week.
I think I’ve worked it out how to implement the potentiometer but I’m still stuck at some points of which I hope you guys can help me.

Current wiring: (ALPS RK27 10KOhm Quad channel)

XLR Pin #1 to Chassis
XLR “side” pin to shield
XLR Pin #2 & #3 to Potentiometer IN
Potentiometer OUT to PCB pin #1, #2 and #15, #16.
Potentiometer GND to PCB bolt
Shields connected without touching the Potentiometer.

I did a short test run with some loose wires and It all seemed to work correctly.

My questions:
1. Is the GND from the potentiometer to the PCB correct?
2. What do I do with the floating end of the shield near the PCB? Do I just cut it?
(Of what I’ve read the original way Audiophonics connected it to the PCB Pin #3 and #14 would be
wrong)
3. Are there any other things I have to take in consideration?

This is my first time doing something like this with an amplifier so I hope you guys can point me in the right direction.

I’ve added a drawing of the current idea.

(Edit 1: Potentiometer is a ALPS RK27 10KOhm. The amplifier does have a on/off switch but i would prefer a hard limit on the volume)

 

[mcdn]

Hi @Bartholomeus !

First of all, digital volume control is better in almost every way than analogue:

• Better linearity (because the maths is exact)
• Less crosstalk (because there is no capacitative coupling)
• Better dynamic range (because it's usually done at 24 or 32 bit resolution)
• Better impedance matching (because there is no increase in output impedance)

The NC252MP has an input impedance of 47kOhm, while the RK27 pot has an impedance of 100kOhm. It will work, but it will have a very strange relationship between pot position and volume. An active preamp/volume control would be a better fit.

But what problem are you really trying to solve? Just the risk of full volume output from the DAC? If so then you have a power button on the amp, or you could put a switch instead of the pot?

 

[Bartholomeus]

Hi @Bartholomeus !

First of all, digital volume control is better in almost every way than analogue:

• Better linearity (because the maths is exact)
• Less crosstalk (because there is no capacitative coupling)
• Better dynamic range (because it's usually done at 24 or 32 bit resolution)
• Better impedance matching (because there is no increase in output impedance)

The NC252MP has an input impedance of 47kOhm, while the RK27 pot has an impedance of 100kOhm. It will work, but it will have a very strange relationship between pot position and volume. An active preamp/volume control would be a better fit.

But what problem are you really trying to solve? Just the risk of full volume output from the DAC? If so then you have a power button on the amp, or you could put a switch instead of the pot?

Thank you for the replay @mcdn !

I forgot to add some of the points you made in my post, I will add them.

• The pot is a 10KOhm, does this change anything for the better regarding the relationship between the pot and the input impedance of the amplifier?
• There is a switch on the amplifier itself to turn it on/off, but i don't want to take the risk i'm to slow to shut it off.

Yes, that’s exactly what I’m trying to do. I use my setup only in nearfield, mostly gaming and listening to music. Because I have tinnitus in one ear I’m trying to minimize the risk of further damage to my hearing may the DAC ever malfunction.

My idea was to add a potentiometer so I have a variable hard limiter on the amplifier. I’m aware a potentiometer is not the best way to go and was thinking of replacing it with something like a Khozmo stepped attenuator in the future.

The reason I got this idea was to minimize the amount of boxes on my desk. But if a passive potentiometer or stepped attenuator is in all ways inferior to a active pre-amp and defeats the purpose of the great sound of the Hypex Ncore I willing to look at that option.

 

[voodooless]

It's not just the channel imbalance that a pot will give, especially at low levels. It's also an imbalance of the balanced signal itself, because of the inaccuracies in the pots, leading to a less effective balanced connection. You may experience more noise, but it can also introduce distortions. So, I think volume controls in a balanced domain are usually not the best idea.

 

[mcdn]

10k is definitely better. If you've already got the pot then use it, there will be no audible issue since the D50 has a low output impedance (50-100ohms) and good current delivery. Some 1% resistors and a switch would be technically better but not audibly better.

 

[retroflex]

(Of what I’ve read the original way Audiophonics connected it to the PCB Pin #3 and #14 would be
wrong)

Where did you read this? Seems dubious to me, but maybe there's a good reason. I would remove the PCB screw connection, ground the potentiometer to the shields of the corresponding channels, and connect the shields to pins #3 and #14. But I'm not an expert.

 

[Sokel]

It's not just the channel imbalance that a pot will give, especially at low levels. It's also an imbalance of the balanced signal itself, because of the inaccuracies in the pots, leading to a less effective balanced connection. You may experience more noise, but it can also introduce distortions. So, I think volume controls in a balanced domain are usually not the best idea.

It depends what pot one uses.
Alps are nice ones,here's one of them,4 channels at 100mV (pot at about 10 o' clock )

Note the scale,and the small slope difference down low is filter's resistor tolerance not the pot's (didn't give it the attention I gave to the hipass measuring 20 resistors
)

 

[voodooless]

It depends what pot one uses.
Alps are nice ones,here's one of them,4 channels at 100mV (pot at about 10 o' clock )

Note the scale,and the small slope difference down low is filter's resistor tolerance not the pot's (didn't give it the attention I gave to the hipass measuring 20 resistors
)

View attachment 382944

Yeah, I already don't like that. Even these small differences will affect the CMRR.

 

[Sokel]

Yeah, I already don't like that. Even these small differences will affect the CMRR.

On the other hand I don't like DAC's VC SNR at 100mV.
Compromises I guess.

(you'll make me match them down there,I hate measuring resistors! )

Edit:You mean the balanced circuit CMRR?
No,that has nothing to do with this,what you see is the output of the whole device (a two-to-four channel x-over),these channels go straight to amps as you see them.
When I say filters,I mean the x-over ones.

 

[SSS]

First of all I want to thank the ASR community for all the great readings I had the past few months, it helped me a lot making discissions for my new audio gear without all the fuss.

My audio setup is as follows:

Source: PC (USB)
DAC: Topping D50 III
Power amp: Audiophonics AP300-S250NC (Hypex NC-252MP)
Speakers: Bowers & Wilkins 707 S2

What bothers me is the fact I only have digital volume control, especially after reading some cases where the DAC went on to 100% volume. I’ve had a similar issue in the past with an older integrated amp so I want to put a potentiometer (ALPS RK27 10KOhm Quad channel) in between the DAC and Poweramp.

Of which I thought to be a simple implementation, turned out to be a headache for the past week.
I think I’ve worked it out how to implement the potentiometer but I’m still stuck at some points of which I hope you guys can help me.

Current wiring: (ALPS RK27 10KOhm Quad channel)

XLR Pin #1 to Chassis
XLR “side” pin to shield
XLR Pin #2 & #3 to Potentiometer IN
Potentiometer OUT to PCB pin #1, #2 and #15, #16.
Potentiometer GND to PCB bolt
Shields connected without touching the Potentiometer.

I did a short test run with some loose wires and It all seemed to work correctly.

My questions:
1. Is the GND from the potentiometer to the PCB correct?
2. What do I do with the floating end of the shield near the PCB? Do I just cut it?
(Of what I’ve read the original way Audiophonics connected it to the PCB Pin #3 and #14 would be
wrong)
3. Are there any other things I have to take in consideration?

This is my first time doing something like this with an amplifier so I hope you guys can point me in the right direction.

I’ve added a drawing of the current idea.

(Edit 1: Potentiometer is a ALPS RK27 10KOhm. The amplifier does have a on/off switch but i would prefer a hard limit on the volume)

If the current potentiometer wiring works without hum and/or noise then you may keep it this way. It is somewhat a deviation from true balanced connection since the potentiometer signal ground is tied to the PCB ground although this works. A clean way would be to have a floating above ground potentiometer connection. It can be done but also deteriorates the true balanced way since different resistor values in the balanced connection. But also may make no problem. The hint that digital volume control is better in channel balance and channel seperation is true. But ALPS potentiometers were all time used also in hi-end amplfiers and these work well. I prefer analog preamps from the 1980ies which work still perfectly and have also these potentiometers. I hate to be dependent on digital control only which my audiolab DAC has and I don't use it and let it stay at 0 dBFS.

 

[voodooless]

Edit:You mean the balanced circuit CMRR?
No,that has nothing to do with this,what you see is the output of the whole device (a two-to-four channel x-over),these channels go straight to amps as you see them.
When I say filters,I mean the x-over ones.

Yes, I understand, but it shows why I have a problem with this setup as a volume control.

 

[Sokel]

Yes, I understand, but it shows why I have a problem with this setup as a volume control.

No,it doesn't.
The absolute level of the two channel input are identical to the last mV for a very broad voltage range,you can see that.
Forget what this cheap xover does,we're talking about the pot's channel imbalance here.

The overall output of the pot is as near as it gets between it's two (balanced) output channels before x-over's input.
I know cause I have measured it a gazillion times straight at the output with the DMM.

 

[Bartholomeus]

Thank you for all the input, I really appreciate it!

@voodooless
I did not think of the deviation of the balanced signal when I started this little project. I assumed putting a pot in the amplifier would add some noise and/or channel imbalance but that would be okay if it solved my issue.

How would you go around this issue? If I understand correctly the only option to keep the balanced signal intact would be to add an active pre-amp, is that correct?

@mcdn
So you suggest instead of using a pot to add some fixed resistors to keep the signal and channel balance from deviating. Would a stepped attenuator give me the same results in this case? I do lean towards a variable solution in case I change speakers.

@Sokel
So the minimal difference in the balanced signal and channel balance would not be an audible issue?

I did measure the resistance of the pot and it showed some deviation:
DMM on 20K.

knob on 12 o clock:
Pot 1: 7.78
Pot 2: 7.74
Pot 3: 7.77
Pot 4: 7.66

knob at 10 o clock:
Pot 1: 8.76
Pot 2: 8.68
Pot 3: 8.71
Pot 4: 8.61

Although I’m too inexperienced with this to tell if this would be a audible issue.

@SSS
Yes, it works without audible hum or noise as far as I can hear of course. How would I achieve a floating above ground connection to the pot?
I do like the small steps digital volume control offers, but fully agree with you regarding the dependency on digital control only. One of the reasons I bought the ALPS was indeed older hifi equipment that uses it and the good reputation it has.

@retroflex
I came to this conclusion after reading a lot about the Pin #1 issue, the papers from Hypex itself states that the shield should not be connected to Pin #3 and #14 because it’s not part of the audio signal and could only add unwanted noise into the signal. I can fully agree with you that I’m also absolutely not an expert on this stuff!

 

[SSS]

Assuming that the XLR source signal is balanced (no connection to what I call ground in the source) and the Amplifier inputs are also balanced. Then it could be done floating above ground as follows for one channel: XLR input pin-2 connected to the full-volume position pin of the potentiometer. XLR input pin-3 connected to the potentiometer pin zero-volume position. This pin-3 connection goes also to one pin of the amp input. The potentiometer slider (middle pin) goes to the other amp input. Thus you need for one stereo channel only one pot out of the four and also for the second stereo channel only one pot. No ground connections from the pot connectors. But the pot case can be grounded which may be anyway when the mounting position is metal. If the balaced wires are shielded then the shield may go to the pcb ground or metal case whichever creates less noise or hum. By the way, on some vintage studio gear this was solved with a three potentiometer circuit for one channel in order to keep input and output resistance within certain limits.

 

[ElJaimito]

Hi @Bartholomeus !

First of all, digital volume control is better in almost every way than analogue:

• Better linearity (because the maths is exact)
• Less crosstalk (because there is no capacitative coupling)
• Better dynamic range (because it's usually done at 24 or 32 bit resolution)
• Better impedance matching (because there is no increase in output impedance)

The NC252MP has an input impedance of 47kOhm, while the RK27 pot has an impedance of 100kOhm. It will work, but it will have a very strange relationship between pot position and volume. An active preamp/volume control would be a better fit.

But what problem are you really trying to solve? Just the risk of full volume output from the DAC? If so then you have a power button on the amp, or you could put a switch instead of the pot?

Might I add better channel tracking, which for me is the most important of all!

 

[Sokel]

@Sokel
So the minimal difference in the balanced signal and channel balance would not be an audible issue?

I did measure the resistance of the pot and it showed some deviation:
DMM on 20K.

knob on 12 o clock:
Pot 1: 7.78
Pot 2: 7.74
Pot 3: 7.77
Pot 4: 7.66

knob at 10 o clock:
Pot 1: 8.76
Pot 2: 8.68
Pot 3: 8.71
Pot 4: 8.61

Although I’m too inexperienced with this to tell if this would be a audible issue.

Resistance is good but you better measure it with signals to see where you stand and which channels to use (mix) for each channel for better CMRR,the closer,the better.
Measuring the end result won't be the worst idea either.

 

[Bartholomeus]

Assuming that the XLR source signal is balanced (no connection to what I call ground in the source) and the Amplifier inputs are also balanced. Then it could be done floating above ground as follows for one channel: XLR input pin-2 connected to the full-volume position pin of the potentiometer. XLR input pin-3 connected to the potentiometer pin zero-volume position. This pin-3 connection goes also to one pin of the amp input. The potentiometer slider (middle pin) goes to the other amp input. Thus you need for one stereo channel only one pot out of the four and also for the second stereo channel only one pot. No ground connections from the pot connectors. But the pot case can be grounded which may be anyway when the mounting position is metal. If the balaced wires are shielded then the shield may go to the pcb ground or metal case whichever creates less noise or hum. By the way, on some vintage studio gear this was solved with a three potentiometer circuit for one channel in order to keep input and output resistance within certain limits.

That is a verry interesting take, I did not found that possibility anywhere in my search. Thank you for the detailed description. It somehow feels counterintuitive to connect the pot this way but I will definitely look this further up and test it to see how it works. The pot will indeed be mounted on a metal plate that is connected to the case.

 

[Bartholomeus]

Resistance is good but you better measure it with signals to see where you stand and which channels to use (mix) for each channel for better CMRR,the closer,the better.
Measuring the end result won't be the worst idea either.

Thank you for the tip, I will definitely measure the pot with an active signal to see which channels match best. I will post the results I find this weekend.

May I ask, how did you ground the pot?

 

[SSS]

That is a verry interesting take, I did not found that possibility anywhere in my search. Thank you for the detailed description. It somehow feels counterintuitive to connect the pot this way but I will definitely look this further up and test it to see how it works. The pot will indeed be mounted on a metal plate that is connected to the case.

Yes, you can only try and measure the signal passthru result and hum/noise in order to decide what the best in your setup. One advantage of my proposal is that you only have to match two potentionmeters for channel matching over various volume settings.

 

[Ropeburn]

I've thought about a solution for this scenario as well, and thought of doing it like this:

A simple 2x XLR in and out box (aluminium for shielding), with a 4 pole 3 position rotary switch acting as a 3-step attenuator, equivalent to a 4-gang 20 kOhm potentiometer, but much better balance between gangs and better longevity. Circuit:

For one balanced channel. Ro = source output impedance, RL = destination input impedance (=load). Switch numbers are positions, A/B are poles. Above E6 resistor values result in ~0, -7.2, and -13 dB attenuation. To be adapted to desired attenuation levels.

Using 0.1% metalfilm resistors, absolute worst case deviation is +-0.01 dB between balanced sides and stereo channels, assuming 1V input attenuated as above. I assumed the absolute worst possible theoretical constellation for this, with all resistors reaching maximum tolerance into the worst possible direction within the circuit. Real life conditions can be expected to reach half that deviation or less.

Switch type:

Datasheet says, contact resistance <1 mOhm, isolation resistance 25 GOhm. Seems good enough, and it's cheap. I've used these before in coarse 8-bit digital signal switching applications (hamfisted live jamming on modified digital drum machines, drunk and sweaty switching like a madman), and they are sturdy, satisfying to the hands and seemingly last forever.

Built into a nice little box, the whole thing is effectively a stereo, balanced 2-step defeatable attenuator (="pad") that's relatively simple to build and very cheap. Small cases with 4x XLR mounting holes are 20-30 moneys, good XLR sockets are maybe 2-3 moneys each, plus a few cm of wiring.

Practical advantages: a little "problem solver box" that can be used with any source and destination that meet standard line in/out impedances. No need to drill into your nicely built amplifier. Three attenuation levels so you still use your source for volume control, but can run it "hot" in any listening situation, to take almost full advantage of its SNR. Easily and precisely adaptable attenuation to personal use scenario using E12 or E24 resistor values. Much easier and cheaper to build that a finely stepped attenuator, and still solves the original problem while eliminating the disadvantages of a (costy) potentiometer.

Am I missing something crucial? Please do tell, guys, if there's a glaring oversight in the concept. Are the principles sound?