Passive monitor controller design

[da goose]

Dear all,

All studios have at least one set of monitors which need easy volume control and switching between in and outputs and a mute, mono etc switches, specially when you are working computer based (DAW)
Some people use a small mixer for that which will color the sound, adds noise, adds distortion and so your monitors do not give the sound you really need.
Right now I’m in the middle of designing a 100% passive monitor controller with a 24 steps attenuator which is a LOT better!
My design is 100% passive which means no sound is going through any electronics, just relays for switching and hi precision resistor networks for a ‘what goes in comes out’ exactly the same way.

The prototype is done and I’m really satisfied with the result, it’s a 1 unit high 19 inch case with 3 stereo inputs and 2 monitor outputs and a separate subwoofer output, all switch-able from the panel.
Besides that it has a mono switch, left and right separated mute, a dim switch (instant 15db softer), input 3 has a extra 12 steps attenuator and it has a low noise, high output headphone out with volume control.
All the knobs are of really high quality and are illuminated with a 2 color led (red/green) so you see what is going on instantly just by looking at the switch.
The main attenuator is made in 24 steps going from 0db to -33db with 1.5db per step and it has a real BIG aluminum knob to tweak. Because of the 24 steps it’s really easy to use for a calibrated monitor setting (83db k-system) and it doesn’t have any stereoshift at all because of the precision resistors (less then 1%), standard pot meters have a real big disadvantage, they have a lot of resistance shifting (about 15%) between channels making it impossible for proper adjustment.
The box is completely balanced which is not really necessary for a monitor controller and makes it quite expensive but I thought it was worth it.

I designed this controller with my own studio situation in mind but now I want to make a version that will suit a lot of people and so I want to ask you some questions about what YOU search for in a monitor controller.
I would really appreciate it if you took some time thinking about what you would like to have in your studio for controlling your monitors, just use your imagination!
I have made a small list of items that will be on it but this could be changed depending on the results of my little ‘passivemonitorcontrollerinvestigation’

• 3 inputs with a 3 way switch (is 3 inputs enough?)
• 1 input with separate stepped attenuator (unbalanced)
• 2 switchable monitor outputs (could be more)
• Subwoofer output with in/out switch (stepped level control?)
• mono summing switch (check mono compatibility)
• dim switch (instant 20db dimming of the sound, is 20db ok?)
• Left/right separate mute switch (is just mute ok?)
• Headphone output with gaincontrol
• Main attenuator (goldplated Elma), 1.5db per step, 24 steps from 0db > -33db (could be any value, is -33db enough?)
• All outputs XLR connectors, 2x input XLR, 1x input 6.3mm jack (unbalanced)
• Should all input/outputs be balanced? (expensive because everything needs to be doubled L+-, R+-)
• What other options would you like to see?
• What color do you want it to be? (could be a custom color for fitting your studio’s design)
• What are you willing to pay for a box like this?

If you have any more questions, ideas or what ever, please feel free to contact me!

With kind regards,
Jeffrey de Gans

[rkwyent]

Looks great and it is a 1U design.
I would be interested in something like this.

However, it is my uderstanding that transformer based attenuators work best in regards to signal purity and gain control.

Just wondering what your thoughts are on that and if you considered a transformer type situation.

[da goose]

Some other info:

As you might have seen i used pretty thinck wire to the in and outputs and i used as short as possible wires. Also the audiorelays are as close to the connectors as possible and there are no audioconnections crossing the voltagelines to avoid humming and other noise.
In the final design i will also move the PSU section near the subwoofer output to the shielded section where the trafo is at, just to make sure.
The frontpanel is profeesionaly engraved at Schaeffer and so the letters wont come off or whatever.
The headphoneamp is done using a simple 5532 opamp stage but this will be done with a OPA opamp in the final version for even lower noise.
This version is build on veroboard (experimental printboard) but the final design will be on printed board.
When the unit is in 'normal' use (no muting, mono or whatever) all realys are in rest and so there is no voltage going through the relays, also to avoid noise.
The only relays in use are the input switching relays, they are switched on with the 3 steps frontpanel input switch.

It sounds great BTW! actually...it doens't sound at all

OK, now it's time for the pro's to have a look and comment!

BTW: I'm also thinking of designing a router/switcher for analog equipment, if you have any thoughts on that, please let me know!

[da goose]

Originally Posted by rkwyent

Looks great and it is a 1U design.
I would be interested in something like this.

However, it is my uderstanding that transformer based attenuators work best in regards to signal purity and gain control.

Just wondering what your thoughts are on that and if you considered a transformer type situation.

Wow... that is fast!

I haven't really thought about a trannybased attenuator actually, i like the sound of tranny's but i think for attenuation you don't want it it to sound at all but maybe i have to be in depth of the trannys more to make sure.
I also think it will be a lot more expensive to build and with the 24 steps switch it's also easier to go to a calibrated value.
It's also possible to make the values/steps more custombased because it's just a matter of using different values resistors, this version has 1.5db steps from 0db > -33db and when you turn on the dimswitch you can go lower if you want from about -20db > -53db. (didn't really measure that yet)
Do you have more info on trafo based attenuators for me? I'm interested!

[bob katz]

A very pretty thing you have built!

But why 1.5 dB steps? Why not 1 dB steps? 1.5 is fine for the casual user, but not the professional, in my opinion.

What are the internal impedances? How much does the last step vary when feeding amps of different impedance?

BK

[da goose]

Originally Posted by bob katz

A very pretty thing you have built!

But why 1.5 dB steps? Why not 1 dB steps? 1.5 is fine for the casual user, but not the professional, in my opinion.

What are the internal impedances? How much does the last step vary when feeding amps of different impedance?

BK

Thanks!

Right now i have chosen for 1.5db per step because it's a 24 steps switch and i wanted all steps to be equal and so it's from -33 > 0 right now.
I was also thinking of doing something like the first 8 steps in 3 to 6 db steps and the middle section (let's say, step 9 to 18) in 1db per step and the last section (to 0db) in bigger steps again.
In this way the middle section will be more precise the the rest, do you think this is a good idea or should i maybe just use the first steps in bigger steps and from, let's say step 9 to 24 in 1db steps?

Right now the attenuator is set to 10k in total, what do you mean with the last to vary the impendance? The resistance of the output stays at 10k all the time.
I think 10k is ok or do you think i should go for a different value?

The attenuator is a tricky thing i allready found out, the steps are to big or to small and stuff like that, i'm still searching for the perfect steps.
The thing is that a mastering engineer probably needs more precise steps then a mixing engineer who is switching to different levels (really loud to really soft) more and i'll try to make the box work for both mastering and mixing engineers which makes it extra hard i think.

Do you think the box should be balanced or will unbalanced do fine as well? I'm still a fan of balanced but what i understand from your book and comments from other people that unbalanced is better/cleaner.
For the box unbalanced will also be easier and cheaper to build, just a 2pole att switch instead of 4pole (just the 4pole elma was about €90 which is cheap BTW), a lot less relays, a lot less wires etc.

[MAzevedo]

A few things I would like to see:

-A phase flip on 1 channel (or both, but you don't have much faceplate left for two switches) for checking the Side information.

-An active gain stage on input 3. One thing I really need out of a monitor controller is the ability to easily do level-matched A/B between the flat and processed sounds, which generally means boosting the flat version 3-10dB. I guess I could calibrate the converter feeding input 3 10dB hot, but that wouldn't help when I am working with an analog source.

-A buffered meter output.

Regarding price, look at your competition. A Crane Song Avocet runs $2400 on VintageKing.com, and includes D/A conversion. You'd have to have a fairly low price point to make your controller + converters price competitive with something like that. (Though, I do like the rackmount controller better than having a remote, I hate having a big box cluttering up my desk.)

[just.sounds]

Trafo's=coloration. A passive type like this is the least coloring if the previous and the folowing stage have no problems with the impedance.

I build a design from the Elektor audio special (dutch magazine) Wich has also relays for the switching of the resistors 64 1dB steps remote control and a very clean 15dB amplifier.

http://www.just-sounds.nl/fototjes/relaixed.JPG

Works great but no sub out like yours.

[squeegybug]

Originally Posted by just.sounds

Trafo's=coloration.

Absolutely not true, for the intended application of volume control. In fact,
exactly the opposite, TVC is the *most* transparent approach available.
Much more stable than switched-resistor controls or potentiometers.

And balanced operation is easy if desired, since it is inherently available
in the design.

Jeffrey, that is a nice looking product. I might have some ideas to add.
When I saw this thread today I first thought 'use TVC', then saw rwk
mentioned it as well. One thing though about the transformers
(or autoformers if you prefer) -- all the ones I have seen
(Intact, S&B, Promitheus, Sowter, etc.) are too big to fit into a 1RU chassis.
For my needs, I would actually prefer 2U anyway, to allow more room for some
additional switches I would like to see, more on that later when I get a chance.

Steve

[bob katz]

Originally Posted by da goose

Thanks!

Right now i have chosen for 1.5db per step because it's a 24 steps switch and i wanted all steps to be equal and so it's from -33 > 0 right now.

I think 24 steps will do fine at 1 dB/step. I once made one where the last step was another 10 or 15 dB for a dim position. You could also do a range switch or a dim/mute switch so you don't waste any of those precious 24 1 dB steps.

In this way the middle section will be more precise the the rest, do you think this is a good idea or should i maybe just use the first steps in bigger steps and from, let's say step 9 to 24 in 1db steps?

Unfortunately, for a modern day mastering studio that might be doing wide dynamic range material today and Snoop Dogg tomorrow, you need at least 16 steps of 1 dB each.

Right now the attenuator is set to 10k in total, what do you mean with the last to vary the impendance? The resistance of the output stays at 10k all the time.
I think 10k is ok or do you think i should go for a different value?

10K is very dangerous from the point of view of capacitive loss in the cabling to the amplifier.

The attenuator is a tricky thing i allready found out, the steps are to big or to small and stuff like that, i'm still searching for the perfect steps.
The thing is that a mastering engineer probably needs more precise steps then a mixing engineer who is switching to different levels (really loud to really soft) more and i'll try to make the box work for both mastering and mixing engineers which makes it extra hard i think.

A lot of mixing engineers would disagree with you. While it is true that they may change gains a lot and play loudly and softly a lot, I'm trying to teach mixing engineers that being within 1 dB of a "standard monitor gain" that they can repeat from day to day can really help them in their mixing consistency.

Do you think the box should be balanced or will unbalanced do fine as well? I'm still a fan of balanced but what i understand from your book and comments from other people that unbalanced is better/cleaner.

Unbal CAN be better/cleaner if the number of active gain stages in the chain is less... But if the sources are extremely well designed balanced sources and the loads are equivalent, then a balanced unit can perform (theoretically) better. I'm not sitting on the fence here, we'd have to see specific schematics. Then there's the ground loop thing but a well-designed single purpose studio should be able to have that part of their act together.

Good luck!


BK

[Paul Gold]

Is the main attenutor constucted like a potentiometer? If so the input impedence would be 10K (reflected back to the source). The output impedence would vary between 10k and whatever it is at 33dB of attenuation. You would need a constant impedence attenuator like a bridged T to keep input and output impedence constant.

[dcollins]

Originally Posted by squeegybug

Absolutely not true, for the intended application of volume control. In fact,
exactly the opposite, TVC is the *most* transparent approach available.
Much more stable than switched-resistor controls or potentiometers.

I don't think most people would agree with these statements!


DC

[Paul Gold]

I think the way to construct the attenuator would be to have a fixed loss of 12-18dB. Gain at the power amplifier is usually plentifull. If your potentiometer (attenuator) is 10K and you have a fixed loss of 18dB (at the top of the throw) then your maximum output impedence is approximately 1k25. That should be low enough not to bother anything. It might be a good idea to hand match resistors per channel to maximize CMRR.

[da goose]

Wow! That is a lot of input guys!
Have to check all of it later on, i allready have some nice ideas right now which i will implement.

Thanks!
Get back soon :)

[just.sounds]

Originally Posted by squeegybug

Absolutely not true, for the intended application of volume control. In fact,
exactly the opposite, TVC is the *most* transparent approach available.
Much more stable than switched-resistor controls or potentiometers.

And balanced operation is easy if desired, since it is inherently available
in the design.
Steve

Okay maybe it is possible to make a colorless trafo but i have not encountered one.
I heard the Sowter and they are to my ears not as transparant as a resistor attenuator.

But what i do believe is that some analog stages react better to a trafo attenuator than a resistor based one.

But i'm up for the challenge. If you can provide a sample of a piece of music raw and with a trafo attenuator in a loop with a much used converter like a digi192 or Lynx aurora clocked internal. I can do the same with my resistor network (wich is fully balanced) with buffer amp (and maybe without if you would really like) and we can hear the difference!

Schematic

RelaiXed -- Balanced pre-amplifier

[mischa janisch]

Originally Posted by MAzevedo

A few things I would like to see:

-A phase flip on 1 channel (or both, but you don't have much faceplate left for two switches) for checking the Side information.

-An active gain stage on input 3. One thing I really need out of a monitor controller is the ability to easily do level-matched A/B between the flat and processed sounds, which generally means boosting the flat version 3-10dB. I guess I could calibrate the converter feeding input 3 10dB hot, but that wouldn't help when I am working with an analog source.

-A buffered meter output.

Regarding price, look at your competition. A Crane Song Avocet runs $2400 on VintageKing.com, and includes D/A conversion. You'd have to have a fairly low price point to make your controller + converters price competitive with something like that. (Though, I do like the rackmount controller better than having a remote, I hate having a big box cluttering up my desk.)

This is exactly what I've been thinking - I use the phase phase flip quite a bit to check the side information... also the buffered meter output is a must for me. I'm just not sure about an active gain stage - I think the extra stepped attenuator for one input in his design is ok for me (if it's not more than 1db steps over a range of 10-12db) to bring the mastered version down in level... soundwise probably superior and cheaper to build.

On the other hand the Avocet IS the one to beat in terms of price vs performance...

[bkuijt]

Originally Posted by MAzevedo

A few things I would like to see:

-A phase flip on 1 channel (or both, but you don't have much faceplate left for two switches) for checking the Side information.

-An active gain stage on input 3. One thing I really need out of a monitor controller is the ability to easily do level-matched A/B between the flat and processed sounds, which generally means boosting the flat version 3-10dB. I guess I could calibrate the converter feeding input 3 10dB hot, but that wouldn't help when I am working with an analog source.

-A buffered meter output.

Regarding price, look at your competition. A Crane Song Avocet runs $2400 on VintageKing.com, and includes D/A conversion. You'd have to have a fairly low price point to make your controller + converters price competitive with something like that. (Though, I do like the rackmount controller better than having a remote, I hate having a big box cluttering up my desk.)

Your design looks really nice and solid.

I agree with the above points.
Also I would prefer 1 db steps on the master.
It would be important to have small steps or continuous attenuation on input 3 to have precise level matching possible.

Next to the Avocet your competition would be the Coleman monitor controllers.
The Coleman M3PH MKII goes for £650,- ex vat at KMR Audio.

keep us posted with your progress!
Bests, Bastiaan

[da goose]

OK, here's some new ideas i have with help from you guys.

The attenuator:
The attenuator will have 1db steps instead of 1.5db steps, so the range will be from 0db > -23/24db (depends on steps, mine has 23 instead of 24steps).
To get more range i will tune the dim switch so that it starts at -24db > -48db.
I'm also seriously thinking of going for a T ladder type so that the impandance will remain the same all the time. (is 10k ok or should i go for perhaps 25k?)
There is one big disadvantage, to go balanced you need a 8 pole switch instead of 4, i'm not even sure if it's easy to get, i've seen 'm but i think they where not elma switches.
Unbalanced is not so much of a problem, you need 4 poles but what i understood from most of you guys is that unbalanced is fine and maybe even better.
Maybe it's a better idea to go realy based for the volume control, i can use a 1 pole 24steps switch then, have to figure that out how many relays i would need.
What do you think about this idea?

Active stage for input 3:
I'm thinking about this idea with a range of about +6db and -6db with a 12 steps attenuator for the AB comparsion.
It might even be a better idea to make the AB a pushbutton just as the other swithes on the panel.

Phase/polarity flip:
I had this idea from the beginning actually, don't ask me why i didn't implement it.
The only thing is that if i go for a unbalanced version it's quite hard (not even possible i think) to flip polarity without going active, which i don't want so that makes it quite hard.

I might even go for 2 versions of the controller, one balanced with a serial type attenuator like the one i build with a polarity flip button and another version as described above, T ladder and unbalanced.
The only thing i'm thinking of right now is the price it's going to be, the idea was a bit to get somewhere in between the avocet and cheapo controllers so that you have quality without paying TOO much.

What would you expect the price to be for a piece of equipment like this?

What about the idea of getting it custombuild? perhaps you want 0.5db steps att or you want 1 output but more inputs or you want a blue frontpanel etc.

[MAzevedo]

Originally Posted by mischa janisch

This is exactly what I've been thinking - I use the phase phase flip quite a bit to check the side information... also the buffered meter output is a must for me. I'm just not sure about an active gain stage - I think the extra stepped attenuator for one input in his design is ok for me (if it's not more than 1db steps over a range of 10-12db) to bring the mastered version down in level... soundwise probably superior and cheaper to build.

On the other hand the Avocet IS the one to beat in terms of price vs performance...

I like to keep my monitor position constant as a loudness reference, if I were to use input 3 for the eq'd version, I would have to turn up the monitors to match the attenuation on my EQ. Not that it would be the end of the world or anything, it just ends up being a bit more fiddling about than I would like.

[da goose]

BTW, i'm thinking of starting the production somewhere this summer, i will be moving houses (and so my studio) in june/july and from that place i have the possibility to set up some serious production (yep, all handbuild)
Right now it's hard to build stuff because i simply don't have room for it and my wife is complaining about the mess it makes
In my new house if have a BIG place for soldering, drilling, screwing, screaming and build more then one machine at a time.

shhhttt.. i also have plans on doing a passive tube eq.. i've got my prototype running here as well...

[bkuijt]

Something I forgot to mention in my post before:

I would like push buttons for source selection, not a rotary switch.
When comparing 1 and 3 it's doesn't work if you have to go past 2 all the time...
So that would be my custom demand

But... for now I am still satisfied with what I have (Emm Meitner Switchman).

bests, Bastiaan

[jamsmith]

All I can say is that while some design consideration could have been better thought out as has been pointed out here, the quality of your work and attention to details is outstanding. The beauty of the construction is admirable and make me with my many years with a soldering iron in hand envious!

A+, 5 stars!

[Paul Gold]

Originally Posted by da goose

OI'm also seriously thinking of going for a T ladder type so that the impandance will remain the same all the time. (is 10k ok or should i go for perhaps 25k?)

The problem is that you want a high input impedence and a low output impedence. This is impossible with a pad unless there is loss. A 10K output impedence is not good, a 25k output impedence is a disaster. The general rule is a 1:10 source to load impedence ratio. Most modern input impedences are 10k and up. So you really want a 1k output impedence. You can get away with higher but as it increases driving the capacitive cable load becomes more difficult. The problem with a 1k input impedence is that many things won't drive that cleanly. It's a trade off. Fitting passives into an active world isn't easy.

[squeegybug]

Originally Posted by just.sounds

But what i do believe is that some analog stages react better to a trafo attenuator than a resistor based one.

But i'm up for the challenge. If you can provide a sample of a piece of music raw and with a trafo attenuator in a loop with a much used converter like a digi192 or Lynx aurora clocked internal. I can do the same with my resistor network (wich is fully balanced) with buffer amp (and maybe without if you would really like) and we can hear the difference!

You likely know this, but just to clarify a bit for the thread's sake (and sorry Jeffrey for the slight detour) -- these TVCs are simply used for voltage conversion and impedance transformation, without power/heat losses. Easy loads on source/target combination. Each tap is wired to a separate switch terminal. There is no hum, no ground loop noise. A black background. Completely open and fast sound.

In contrast to resistors/pots which vary power by changing both voltage and current via losing electrical energy to heat. And causing impedance mismatches and frequency response variations along the way. I see you have a buffer amp, and think that is probably a good idea.

I don't design them but these guys and others do:

Stevens & Billington
Intact
Promitheus
Sowter

There is starting to be lots of information out there on transformer volume controls, several high end audio designers are using them in their 'preamps' now. I bought a fairly basic lower cost single-ended model, easily sounds better (or rather, doesn't sound) than the active/resistor/pot devices I have used.

But you are right, as with any passive attenuator, the specific characteristics of the source and target, plus cable construction and length, will determine just how well *any* passive devices will work in a given system. With my Lavry DA going to Bel Canto power amp with Teflon Belden cables it is an easy match. As it likely is with most modern converters and amps.

I have recorded loops with other resistor/pot controllers and been planning to do it for this new TVC. But the only ADDA converter I have now is Lavry Blue so don't know if you can get access to one of those. And it ideally should be done using broad spectrum generic sustained tones at various levels. But music can probably serve

Steve

[just.sounds]

Hmm a lavry will be hard to get hold of here in the Netherlands. I could get acces to a Apogee or a Prism maybe DCS but i don't know anyone with a lavry.

But maybe Da Goose is going to try this?

I'm very curious!!

Da Goose if you stay with resistors maybe you can look at DACT or Goldpoint attenuators.
I think that would save production costs

[Bob Yordan]

Hiya

It looks cool.

I use something simular passive etc and very cheap from Pro Audio (Australian company).

I wonder why did you add a stepped output level knob??

Shouldn't the calibration apply to the level of the final file/mix etc?
The volumes of mixes that shall be mastered are mostly at different levels.

Is it necesarry to know the exact output level of the control room monitoring system during a session?

I turn up and down the control room monitor level a lot of times during a session. To adapt it to speaking level (talking with customer) etc and to adapt it to different speaker pairs.

Best wishes

[dpdemo]

Why not go with a 32 position attenuator, then develop a log-taper control? You could then provide fine, 1 dB steps in the critical range and have large attenuation steps at the bottom of the range (gives the dimming Mr. Katz suggested).

I designed something for a phono preamp years ago that had this - I happened to use a Shallco switch (big! thing) and matched 1% resistances. Zero crosstalk, near-perfect attenuation tracking L to R. Mine was a pot, as opposed to a ladder since I had an amp stage behind it.

I still have the design notes buried somewhere if anyone is interested. Unfortunately, the design software was written on a now-obsolete and ancient HP lab computer...

ps - that is one very pretty unit!

[Thomas W. Bethe]

How about an external power supply to get all the power supply mess out of the 1U chassis? I have always liked external power supplies for audio equipment especially where the signal has to stay as pristine as possible. Just an idea...

[da goose]

Originally Posted by Thomas W. Bethe

How about an external power supply to get all the power supply mess out of the 1U chassis? I have always liked external power supplies for audio equipment especially where the signal has to stay as pristine as possible. Just an idea...

Hi Tom,

I get your point about the external PSU but actually... i don't think it makes such a big deal anyway.
The DC voltage has to come into the box anyway and the external DC cable crosses the signal cables externally, tends to break soon, the PSU has to be somewhere in the rack (i hate that personally) and then you have AC and DC in the area of your cables.
I use really good shielded toroidal transformers and as you can see on the picture, i have extra shielding in the PSU section (in the final version there will be just DC going straight to the board.

About the 32 steps attenuator, i never seen these actually, did you see 'm anywhere?
I'm using my prototype for some time now (24 steps) and so far it's more then enough, i made the fixed point at 12 o clock right now so i have enough steps up and down and it's easy to go back to the centre this way.

But maybe someone else has some other ideas about that?

I'm still working on it right now, i think by the end of the summer i will have some new prototypes working here.
I'm about to get the key to my new house and that is priority number one right now.
I'm also think about a router for fast passive swithcing between sources, what would you like to see in a router? any fresh ideas might come in handy

[George Necola]

it's already developed..
Dangerous MONITOR