@Nelson:
http://www.audiosystemsgroup.com/Pin_1_Revisited.pdf
http://www.audiosystemsgroup.com/Pin_1_Revisited_Part_2.pdf
@sangram:
With TRS or TRS/XLR combo input jacks the S usally needs to be connected to the circuit GND reference, as those shall work with unbalanced cables also, where S is the refererence for the input signal. Which is a dilemma... it depends on the GND routing on the soundcard etc and its circuit design whether a problem will arise or not when large(!) GND currents (DC/AF/GF) are injected. In a benign environment (smaller home/studio or live sound setups) it will often work without problems.
Large GND currents are still a problem even with gear conforming to AES-48 because there is a risk that these currents couple into the two signal lines in a unequal fashion, depending on the quality of the cable. That's why GND-lift switches (at the receiving end) can be useful -- but they should not completely float the cable GND, rather use that disconnect/breaker network (to the chassis) to only damp the DC/LF amount of GND current on the cable for several reasons (one is to tie the signals level to the usable input common mode range in case of non-xfomer outputs). This again contradicts the requirement of no GND pigtails inside the equipment (unless the I/O section is a shielded compartment on itself).
In general, a balanced connection does not require to have signal symmetry wrt to a GND, rather impedance symmetry is the important thing, for longer cable runs. Still signal symmetry is a good thing to have but this is depending on the circuit details of the receiver.
Bottom line is, in a specific situation it depends on many factors whether there is potential for pin 1 problems... and mostly requires a specific solution.
- Klaus
http://www.audiosystemsgroup.com/Pin_1_Revisited.pdf
http://www.audiosystemsgroup.com/Pin_1_Revisited_Part_2.pdf
@sangram:
With TRS or TRS/XLR combo input jacks the S usally needs to be connected to the circuit GND reference, as those shall work with unbalanced cables also, where S is the refererence for the input signal. Which is a dilemma... it depends on the GND routing on the soundcard etc and its circuit design whether a problem will arise or not when large(!) GND currents (DC/AF/GF) are injected. In a benign environment (smaller home/studio or live sound setups) it will often work without problems.
Large GND currents are still a problem even with gear conforming to AES-48 because there is a risk that these currents couple into the two signal lines in a unequal fashion, depending on the quality of the cable. That's why GND-lift switches (at the receiving end) can be useful -- but they should not completely float the cable GND, rather use that disconnect/breaker network (to the chassis) to only damp the DC/LF amount of GND current on the cable for several reasons (one is to tie the signals level to the usable input common mode range in case of non-xfomer outputs). This again contradicts the requirement of no GND pigtails inside the equipment (unless the I/O section is a shielded compartment on itself).
In general, a balanced connection does not require to have signal symmetry wrt to a GND, rather impedance symmetry is the important thing, for longer cable runs. Still signal symmetry is a good thing to have but this is depending on the circuit details of the receiver.
Bottom line is, in a specific situation it depends on many factors whether there is potential for pin 1 problems... and mostly requires a specific solution.
- Klaus
Thanks for that, Klaus. I have read it before, but I forgot.
So all three previous references are talking in terms of 2
conductors and a shield. All I see out there in the industry is
3 conductors and a shield.
I appreciate where they're coming from, and this subject has
been hashed thoroughly on DIYAudio before. I remain happy
with chassis grounding the shield and using pin 1 for analog
ground.
So all three previous references are talking in terms of 2
conductors and a shield. All I see out there in the industry is
3 conductors and a shield.
I appreciate where they're coming from, and this subject has
been hashed thoroughly on DIYAudio before. I remain happy
with chassis grounding the shield and using pin 1 for analog
ground.
Thanks, Klaus, that was very helpful. I reference the S to signal ground personally, as anywhere else it causes some hum (I think the same as Mr. Pass, except I prefer to use 3-conductor shielded cable). It was being mentioned elsewhere on the boards that the standard interconnect methods are not always followed to the letter by all manufacturers, and it could be a problem in professional environments.
I reference the S to signal ground personally, as anywhere else it causes some hum (I think the same as Mr. Pass, except I prefer to use 3-conductor shielded cable). It was being mentioned elsewhere on the boards that the standard interconnect methods are not always followed to the letter by all manufacturers, and it could be a problem in professional environments.
Some times Jim Brown's papers don't have a good correlation between title and subject! Try:
www.audiosystemsgroup.com/AES-RFI-SF08.pdf
www.audiosystemsgroup.com/Pin_1_Revisited.pdf
www.audiosystemsgroup.com/Pin_1_Revisited_Part_2.pdf
www.audiosystemsgroup.com/AESPaperNYPin1-ASGWeb.pdf
The best Bill Whitlock paper at Jensen Transformer is:
http://www.jensen-transformers.com/as/as085.pdf
Bill Whitlock also write's lots of magazine articles, I have a list of almost 60 papers. Also a long audio presentation at a London AES meeting.
Bill Whitlock is now into IC's and is working with THAT Corp. see:
www.thatcorp.com/appnotes.html
http://www.thatcorp.com/appnotes.html
Thanks for the valuable input Klaus.
From my observations on existing devices, you may find the TRS sleeve pin (or XLR pin 1) connected either to the device shield or to the device signal ground, whether the device is balanced or unbalanced...
Assuming a transformer-less balanced receiver which may be connected to balanced TRS or unbalanced sources TS, I would connect the sleeve pin to the chassis by default and provide a ground lift. Since there will be a switch, do you see a benefit of providing a 3-way switch to handle
- sleeve to chassis
- sleeve lift (sleeve unconnected or connected to chassis thru R?, RC?)
- sleeve to signal ground?
The same question may be raised about the output. Assuming a balanced driver with floating output which may be connected to balanced TRS or unbalanced TS devices, would there be a benefit of providing a 3-way switch to select between sleeve to chassis, lift or to signal ground?
Sorry I did not find this question earlier.
A standard XLR 3 pin connector has four connections. Pin 1 is the shield. Pin 2 is now hot or signal it did not used to be! So be careful when working with older gear especially Altec. It also is not next to pin 1 so pay attention. Pin 3 is signal return which may be ground in some cases.
There is also a case connection. This may be connected to the shield at the send end only. That would be the male in a cable or the female in a chassis. If you want RF protection use a 1 nf also known as .001 uf or 1000 pf capacitor between the shell and ground. The capacitor can be used at both ends. I do not connect the shield to the case!
If you are sending from an unbalanced source some folks ground pin 3. A better way is to use a 100 ohm resistor between pin 3 and ground at the send end. That puts approximately the same impedance in both input paths so you get some balancing action.
When using 3 conductors and an isolated shield best practice depends on the frequencies of interest! I put both the third wire and the shield on pin 1 of the send end and only the third wire at the receive end. If I expect frequencies higher than normal then the shield gets connected at both ends Think AES digital!
An important issue occurs when using transformers for balanced use as opposed to electronic differential inputs. A transformer is only connected to pins 2 and 3 even if there is a center tap. The center tap should not be connected to pin 1. When you have a single transformer either at the send or receive end it will not work if either pin 2 or 3 is open at the other end.
However a differential or electronic input will pass signal even if only pin 2 or 3 is connected. When going unbalanced into a differential input lowest noise occurs when the unused pin is connected to what is called ground but is signal return. Very rarely this does result in an instability. That is why I prefer to connect it to signal return via a 100 ohm resistor at the send end. Less noise and some balancing action to reduce low frequency external interference.
If you have an unbalanced input that is fed from an electronic differential source you can leave the unused side open or have a resistor (either 600 or 10 K ohms) go to ground. Keep in mind that without the resistor the input will mysteriously stop working if the source is changed to a transformer balanced source!
When you have a transformer source it really wants to see the proper termination resistance. This is usually 600 ohms. If you have an electronically balanced source they most often work better with a 10K or higher load. So the universal termination is a 620 ohm resistor in series a 10 nf to 50 nf (.01 to .05 uf) capacitor. This will load the transformer source at the higher frequencies where it is most needed and keep it from sounding "Harsh." It also leaves the electronic source alone for most of it's work and only loads it at higher frequencies where the is less musical energy although more slew rate current is required.
A very handy test is to plug your equipment into an outlet by itself. Measure from the earthed safety ground to the audio shield or case. Reverse the AC cord and read it again. The proper connection is whichever gives you the lowest reading. If you have a 3 prong plug you can test for this with a ground lift adapter that has been filed down to allow it to be put in backwards. If it turns out that the best connection is not the way the equipment's plug is wired you may wish to consider swapping the transformer's primary AC connections. Do not try to permanently use an adapter as this will defeat the power switch and fuse under some fault conditions.
For 1/4" phone plugs you can treat the tip as hot or pin 2, the ring as return or pin 3 and the shell as the shield or pin 1. Here the case and shield are forced to be tied together. The tip is of course left when used as unbalanced stereo.
The last issue is with electronic balanced inputs. These work well as long as you stay within their linear range. An opamp powered from 12 volt rails cannot fix a balanced line with 15 volts of noise on it! The US NEC (electrical code) allows 3 ohms of safety ground resistance so it is not unusual in large scale sound systems to see 3 volts between the neutral and ground. If you have two different AC sources thsi can be 6 volts RMS or 16.9 volts P-P. That is where transformers are at their best!
A standard XLR 3 pin connector has four connections. Pin 1 is the shield. Pin 2 is now hot or signal it did not used to be! So be careful when working with older gear especially Altec. It also is not next to pin 1 so pay attention. Pin 3 is signal return which may be ground in some cases.
There is also a case connection. This may be connected to the shield at the send end only. That would be the male in a cable or the female in a chassis. If you want RF protection use a 1 nf also known as .001 uf or 1000 pf capacitor between the shell and ground. The capacitor can be used at both ends. I do not connect the shield to the case!
If you are sending from an unbalanced source some folks ground pin 3. A better way is to use a 100 ohm resistor between pin 3 and ground at the send end. That puts approximately the same impedance in both input paths so you get some balancing action.
When using 3 conductors and an isolated shield best practice depends on the frequencies of interest! I put both the third wire and the shield on pin 1 of the send end and only the third wire at the receive end. If I expect frequencies higher than normal then the shield gets connected at both ends Think AES digital!
An important issue occurs when using transformers for balanced use as opposed to electronic differential inputs. A transformer is only connected to pins 2 and 3 even if there is a center tap. The center tap should not be connected to pin 1. When you have a single transformer either at the send or receive end it will not work if either pin 2 or 3 is open at the other end.
However a differential or electronic input will pass signal even if only pin 2 or 3 is connected. When going unbalanced into a differential input lowest noise occurs when the unused pin is connected to what is called ground but is signal return. Very rarely this does result in an instability. That is why I prefer to connect it to signal return via a 100 ohm resistor at the send end. Less noise and some balancing action to reduce low frequency external interference.
If you have an unbalanced input that is fed from an electronic differential source you can leave the unused side open or have a resistor (either 600 or 10 K ohms) go to ground. Keep in mind that without the resistor the input will mysteriously stop working if the source is changed to a transformer balanced source!
When you have a transformer source it really wants to see the proper termination resistance. This is usually 600 ohms. If you have an electronically balanced source they most often work better with a 10K or higher load. So the universal termination is a 620 ohm resistor in series a 10 nf to 50 nf (.01 to .05 uf) capacitor. This will load the transformer source at the higher frequencies where it is most needed and keep it from sounding "Harsh." It also leaves the electronic source alone for most of it's work and only loads it at higher frequencies where the is less musical energy although more slew rate current is required.
A very handy test is to plug your equipment into an outlet by itself. Measure from the earthed safety ground to the audio shield or case. Reverse the AC cord and read it again. The proper connection is whichever gives you the lowest reading. If you have a 3 prong plug you can test for this with a ground lift adapter that has been filed down to allow it to be put in backwards. If it turns out that the best connection is not the way the equipment's plug is wired you may wish to consider swapping the transformer's primary AC connections. Do not try to permanently use an adapter as this will defeat the power switch and fuse under some fault conditions.
For 1/4" phone plugs you can treat the tip as hot or pin 2, the ring as return or pin 3 and the shell as the shield or pin 1. Here the case and shield are forced to be tied together. The tip is of course left when used as unbalanced stereo.
The last issue is with electronic balanced inputs. These work well as long as you stay within their linear range. An opamp powered from 12 volt rails cannot fix a balanced line with 15 volts of noise on it! The US NEC (electrical code) allows 3 ohms of safety ground resistance so it is not unusual in large scale sound systems to see 3 volts between the neutral and ground. If you have two different AC sources thsi can be 6 volts RMS or 16.9 volts P-P. That is where transformers are at their best!
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.