This circuit it's just 3 resistors and 2 capacitors, it will be pretty small, so I think this one will fit.
I also didn't know those protruding connectors existed, good to know
Thank you
Help Design an unbalanced consumer Line level to Balanced interface
- Thread starter Whoops
- Start date
Help Support GroupDIY Audio Forum:
This circuit it's just 3 resistors and 2 capacitors, it will be pretty small, so I think this one will fit.
I also didn't know those protruding connectors existed, good to know
Thank you
I know this is a very old thread, but I just ran across it today. This is a classic case of a design de-railed by the widespread, but absolutely false, idea that balanced lines require equal and opposite signal voltages. I've been harping about this for 30 years but I guess the internet is much better at misinformation (even many textbooks) than I am at myth-busting!
Anyway, this is an easy problem to solve without phantom power and without active parts - it takes only 3 resistors in the circuit sketch attached and can (should actually) be built inside the handle of a TS plug! The resistor values shown are based on the following design targets:
1. Instrument output impedance, Zo: 0 to 1 kΩ
2. Load seen by instrument output: approximately 10 kΩ, depending slightly on preamp Zi
3. Mic preamp input impedance, Zi: 2 kΩ minimum
4. Signal attenuation, Av: 20 dB for Zi of 10 kΩ or more; 26 dB for Zi of 2 kΩ
5, Differential-mode output impedance: 1.82 kΩ
6. Common-mode output impedance balance: ±9 Ω (including 1% resistor tolerances and effect of instrument Zo)
7. Phantom voltage tolerance: Phantom voltage, +48 V applied via 6.8 kΩ resistors in the preamp, will produce a common-mode voltage less than 7 V on the balanced line. This voltage will be further attenuated at the instrument output by the series 9.1 kΩ resistor. This may cause increased distortion at the instrument output, but I can't image any damage being done. Of course, phantom power should be turned off whenever possible.
Because this circuit is not galvanically-isolated, there is always the possibility of ground currents flowing from instrument to mic preamp through the shield connections and cable. Cables using foil and drain-wire construction are very vulnerable to shield-current-induced-noise or SCIN as described in papers by the late Neil Muncy, myself, and Jim Brown. This problem doesn't usually reveal itself because microphones are not grounded at the their end of the cable. But here, if the instrument is grounded (by its own power cord or another piece of gear its connected to), small power-line related currents will flow in the shield, exciting the SCIN mechanism (and producing buzz in the signal received at the mic preamp).
This circuit compromises loading on the instrument output and its output impedance to achieve the 20 to 26 dB attenuation. If more attenuation is desirable, simply adjust the 909 Ω and 1 kΩ resistor values downward proportionally, i.e., by the same percentage. Adjusting values upward for less attenuation is not recommended since the output impedance of the device will also rise (undesirable, especially with very long cables where it could cause high-frequency roll-off.
Bill Whitlock
AES Life Fellow (and balanced interface evangelist!)
Anyway, this is an easy problem to solve without phantom power and without active parts - it takes only 3 resistors in the circuit sketch attached and can (should actually) be built inside the handle of a TS plug! The resistor values shown are based on the following design targets:
1. Instrument output impedance, Zo: 0 to 1 kΩ
2. Load seen by instrument output: approximately 10 kΩ, depending slightly on preamp Zi
3. Mic preamp input impedance, Zi: 2 kΩ minimum
4. Signal attenuation, Av: 20 dB for Zi of 10 kΩ or more; 26 dB for Zi of 2 kΩ
5, Differential-mode output impedance: 1.82 kΩ
6. Common-mode output impedance balance: ±9 Ω (including 1% resistor tolerances and effect of instrument Zo)
7. Phantom voltage tolerance: Phantom voltage, +48 V applied via 6.8 kΩ resistors in the preamp, will produce a common-mode voltage less than 7 V on the balanced line. This voltage will be further attenuated at the instrument output by the series 9.1 kΩ resistor. This may cause increased distortion at the instrument output, but I can't image any damage being done. Of course, phantom power should be turned off whenever possible.
Because this circuit is not galvanically-isolated, there is always the possibility of ground currents flowing from instrument to mic preamp through the shield connections and cable. Cables using foil and drain-wire construction are very vulnerable to shield-current-induced-noise or SCIN as described in papers by the late Neil Muncy, myself, and Jim Brown. This problem doesn't usually reveal itself because microphones are not grounded at the their end of the cable. But here, if the instrument is grounded (by its own power cord or another piece of gear its connected to), small power-line related currents will flow in the shield, exciting the SCIN mechanism (and producing buzz in the signal received at the mic preamp).
This circuit compromises loading on the instrument output and its output impedance to achieve the 20 to 26 dB attenuation. If more attenuation is desirable, simply adjust the 909 Ω and 1 kΩ resistor values downward proportionally, i.e., by the same percentage. Adjusting values upward for less attenuation is not recommended since the output impedance of the device will also rise (undesirable, especially with very long cables where it could cause high-frequency roll-off.
Bill Whitlock
AES Life Fellow (and balanced interface evangelist!)
Attachments
I think it's important to point out a flaw here. As much as I love the little extruded boxes sized for Neutrik XLRs, putting a phone jack on the box may introduce another issue. Assume that there's a non-trivial voltage difference between the instrument and the mic preamp. If this current flows in an unbalanced cable, i.e., instrument output to the input of this box, you've created the classic coupling mechanism (common-impedance coupling) that makes RCA interconnects so awful. That's why I recommend putting the resistors in my circuit in the TS plug handle, keeping shield currents out of an unbalanced cable shield. Granted, you've used the ring contact to provide a separate path for the signal reference (a good thing) but now you're depending on the common-mode rejection in the mic preamp to suppress it (and you've added loose-tolerance electrolytics to degrade that CMRR). On a related issue, I don't believe the capacitors are necessary as I explained in my post yesterday (Oct 20), where the instrument output stage sees only +7 V or under 1 mA worst-case. If coupling caps are included, because there's no ground reference for common-mode voltage, the instrument output will still see a rather long spike of 48 V when the XLR is plugged into a phantom powered mic pre. If you deem coupling caps necessary, I'd make them considerably larger to not only extend LF response and reduce phase distortion of bass. I'd suggest 47 µF to extend LF -3 dB point down to about 0.7 Hz. As has been previously explained in this forum, long signal chains using coupling capacitors chosen for -3 dB at even 10 Hz will affect timbre of things like kick drums adversely. Also, putting a jack on the adapter will allow (perhaps encourage) musicians to put unbalanced TS extension cords on the instrument output to reach the box, which will add common-impedance coupling to the problems (it's why RCA cables are so awful). Putting the adapter in a TS plug means the cable coming out of it is already balanced mic level and can be treated as a normal mic.
MidnightArrakis
Well-known member
So.....putting your schematic into a TS phone-jack housing big enough to contain a 47uf electrolytic cap will do the trick here??? Sounds like a winner!!!
/
Last edited:
Technically, no. If you add an impedance to one balanced line, you need to put exactly the same impedance in the other to maintain balance. Bear in mind that these capacitors will be rather large if they're to withstand 48 V (likely 50 or 63 V rated). I guess I still don't understand what they would allegedly accomplish. Are we trying to make the interface work perfectly with phantom on? Would anyone intentionally turn phantom power on for a dynamic mic?My intent was simply to protect the instrument output from 48 V and up to 7 mA of current. The circuit as I drew it reduces this to a harmless 7 V and less than 1 mA. And 3 resistors can easily fit in the TS barrel. I doubt that the pair of electrolytics would fit with a shoehorn!
MidnightArrakis
Well-known member
[Would anyone intentionally turn phantom power on for a dynamic mic?] -- Intentionally? Possibly if they are mentally deranged!!! Accidentally??? Some "newbie" audio dude who doesn't know any better and who thinks that -- ALL -- of the mic-input channels "need" phantom-power could just easily flick all of the phantom-power channel switches to "ON" and think nothing of it!!! You never know.....you know???
[Are we trying to make the interface work perfectly with phantom on?] -- Intentionally as "normal"? NO!!! But, as in possibly "accidentally"??? Why not???
[doubt that the pair of electrolytics would fit with a shoehorn!] -- After reading your original response very, very early this morning, I set about to search for a "large-barreled 1/4" phone-plug" connectors and I came across one that -- might -- work that is made by AMPHENOL in Australia. I have already contacted them and have requested some additional mechanical detail information and also a 3D CAD-model of this connector to see if I can "make things fit" inside of it.
I will also need to take a look at the mechanical dimensions of various Surface-Mount electrolytic capacitors in order to see if they just might fit as well. If I can determine that 3 type 1206 resistors (or, type 0805 if necessary) and 2 SMT electrolytic capacitors can squeeze into the available space within this AMPHENOL connector, then I could design a teeny-tiny PCB to place all of the parts onto while connecting to the TS terminals of the phone-plug and the TRS connections of a mic cable. By the way.....where would the second (other) capacitor go between? I -- think -- I know where, but I am not totally certain.
[to maintain balance] -- The MOODY BLUES were big on "Balance" in many of their songs!!!
/
Brian Roth
Well-known member
Many of the "cheap" desks have only a single switch to turn on/off phantom for the entire desk....not on module by module switches.
Shrug....just a fact.
Bri
Shrug....just a fact.
Bri
Well, in my defense, even if you use a pair of electrolytics at the adapter's output terminals, the instrument will still see the 7 V pulse (current-limited to <1 mA) at the moment the adapter's output is plugged into a phantom-powered mic input, all that changes is that the pulse would likely last only a fraction of a second. But, this voltage and current level is a power delivered on less than 0.007 W. I can't imagine it damaging even 3.3 V CMOS logic, let alone the output stage of an op-amp line driver - even if applied for an indefinite period. I guess it's clear that I don't think the capacitors add anything worthwhile to the design. But I don't want to kick a dead horse either.
BTW, Moody Blues is one of my all-time favorite groups. It's too bad the noise floor on many of their early recordings was so poor. Little did I know in the sixties that I'd later throw myself headlong into "The Question of Balance" in real life. My efforts to bust the myth of equal-and-opposite signals defining balance often makes me feel like Sisyphus!!
BTW, Moody Blues is one of my all-time favorite groups. It's too bad the noise floor on many of their early recordings was so poor. Little did I know in the sixties that I'd later throw myself headlong into "The Question of Balance" in real life. My efforts to bust the myth of equal-and-opposite signals defining balance often makes me feel like Sisyphus!!
Attachments
The important thing in this case is that the circuit can have phantom power protection so in case phantom power is ON by mistake, of it's turned on accidentally that the output of the Keyboard is not toasted
Not intentionally, by accidental yes.
In live sound sometimes the time for setup up is so short and the stress is so much that accidents happen a lot.
So it's important that the circuit protects the source device of the possibility of phantom power being turned on on that channel.
Hi Bill,
I really got confused with your post.
The circuit Midinight Arrakis showed was the circuit you drew and showed to us on post 44,
it's actually your circuit, the first one you posted in this thread:
https://groupdiy.com/threads/help-d...evel-to-balanced-interface.82656/post-1069113
This was what you drew:
So MidnightArrakis drew a schematic with your circuit:
The 4,7uf value for the capacitors was chosen by you.
You don't agree with those any longer?
I'm really having problems understanding why the circuit couldn't be in an extruded box, that will stay on the floor like a DI box.
The normal way is to use an unbalanced cable (normally 2 meters) from the Keyboard output to the input on the DI Box that is on the floor under the keyboard.
So those 2 meters of unbalanced cable would be there with a DI box,
so what's the problem with replacing the standard DI Box with an Extruded box with your circuit?
Sorry if I'm missing something but I'm really not getting it.
Sorry, my confusion about the caps. There's my tendency to always recommend coupling caps go large. But using a plug, not a jack, as the input to the circuit is very deliberate and important. It discourages use of a 2-conductor extension cord and placement of the box at the mic pre end of a long unbalanced cable. Doing so defeats the purpose because then there would be a long run where signal and ground currents will mix (common-impedance coupling) adding noise. The box belongs as close as possible (perhaps a 1 meter cable to let the box rest on the floor) to the device driving it.
That will not happen with my boxes Bill,
I want your circuit inside a small extruded box, it will be used on a stage in Live sound, it will stay next to the keyboard or drums machine on the floor, exactly like a DI would. So the cable that connects the instrument to the box will be 1 meter to 2 meters, anyway will be exactly the same size of unbalanced cable like if a Di box was used.
Then from the Box output a Long XLR cable (2 conductors + shield) will take the signal from next to the keyboard (going around the stage) to the console stage box , normally in the side of the stage.
So the longer cable will always be the XLR cable and not the unbalanced cable.
My doubt now is that you posted 2 circuits for this, and I'm not sure now which one I should use, if Version 1 or Version 2
So your version 1 was this one:
And your version 2 it's this one:
So, Which one to use?
Thank you
Both do about the same thing, except the first one also does DC protection of the source.
However it must be noted that they are both somewhat sensitive to source impedance. Good CMRR requests a very low source impedance.
However it must be noted that they are both somewhat sensitive to source impedance. Good CMRR requests a very low source impedance.
The capacitor coupled version has it's HPF tuned for <10Hz but use of 20% tolerance capacitors could (very) slightly degrade CMRR, FWIW.
JR
JR
I have many 4,7uf capacitors, I could measure all of them and choose 2 very closely matched in terms of capacitance value,
would that help?
Thanks
Yes....
Not a huge problem, but I'm just doing the typical web forum dance of sharing TMI.
JR
Yes, but remember, the fact that the output impedance is not zero has a much bigger effect.
Zout=100ohms results in CMRR=40dB in the whole audio bandwidth.
In comparison, having one cap at +10% and the other at -10% results in CMRR 43dB@50Hz, 53dB@150Hz, 57dB@250Hz.
Last edited:
MidnightArrakis
Well-known member
>> Version 2:
/
When I posted the first version, I catered to the paranoia about phantom power damaging something. It can't with only 7 volts open circuit and 1 mA short circuit available at the instrument without any capacitors.
Regarding taking this circuit outside of a phone plug barrel and adding a cable between the instrument output and the circuit, there's a very important reason not to do that. It's important to have NO unbalanced paths ANYWHERE in the signal path. This circuit is NOT a direct replacement for a direct box. If you put an unbalanced cable between instrument and direct box input, you can easily stop the current flow in the shield (how the buzz couples) by flipping a ground-lift switch! There is no such option here. Please build it the way I intended - put it in the barrel of a 2-C phone plug with a shielded-twisted-pair cable of any length exiting it. You can easily fit 3 resistors in the barrel. And extend the reach of it's output cable by adding as much XLR extension as you need. Build it any other way and you'll likely have noise issues in certain system setups - I'll feel no obligation to help.
Regarding taking this circuit outside of a phone plug barrel and adding a cable between the instrument output and the circuit, there's a very important reason not to do that. It's important to have NO unbalanced paths ANYWHERE in the signal path. This circuit is NOT a direct replacement for a direct box. If you put an unbalanced cable between instrument and direct box input, you can easily stop the current flow in the shield (how the buzz couples) by flipping a ground-lift switch! There is no such option here. Please build it the way I intended - put it in the barrel of a 2-C phone plug with a shielded-twisted-pair cable of any length exiting it. You can easily fit 3 resistors in the barrel. And extend the reach of it's output cable by adding as much XLR extension as you need. Build it any other way and you'll likely have noise issues in certain system setups - I'll feel no obligation to help.
