I am designing a headphone amplifier for use of live sound in ear monitoring. A mixing board will send two balanced signals for stereo operation, and the headphone amplifier will receive these two XLR connectors. How should I connect the pin 1 of the two receptacles on the headphone amplifier?
I am working with the assumption that all shields are connected to the mixer chassis. I am also assuming that the mixer earth may not be the same as the earth on stage.
I am working with the assumption that all shields are connected to the mixer chassis. I am also assuming that the mixer earth may not be the same as the earth on stage.
I've read articles and posts on the matter. The simplification of this information seems to exclude multiple conductor scenarios.
This would create a chassis ground loop. Is this a problem?
Nope, no issue at all.
The current flows in the chassis and not in the (internal) single ended reference (Unless you screw that up by connecting it to chassis at more then one point).
Now there is a proviso to all this, google "shield current induced noise" for the residual issue, which is seldom a problem in practise.
Also see Tony Walderons stuff on "designing interference free audio systems", worth the time.
Regards, Dan.
The current flows in the chassis and not in the (internal) single ended reference (Unless you screw that up by connecting it to chassis at more then one point).
Now there is a proviso to all this, google "shield current induced noise" for the residual issue, which is seldom a problem in practise.
Also see Tony Walderons stuff on "designing interference free audio systems", worth the time.
Regards, Dan.
From my reading of all the pin 1 problem papers/reports:
Pin1 always goes to chassis at the socket entry.
Preferably OUTSIDE the Chassis, rather than a pigtail looped inside the Chassis.
Pin1 is NOT part of the Signal.
The two signal pins 2 & 3 are taken to the signal processing In/Out of the circuit.
Pin1 always goes to chassis at the socket entry.
Preferably OUTSIDE the Chassis, rather than a pigtail looped inside the Chassis.
Pin1 is NOT part of the Signal.
The two signal pins 2 & 3 are taken to the signal processing In/Out of the circuit.
Great, makes sense. Typically would you make this one connection between audio and chassis grounds near the input, or is it mostly a trial and error type thing?
I use Oshpark to prototype my designs. They are great, but they take about a month to get the pcbs in my hands. I'd like to iron out as many details as possible before the testing stage.
I use Oshpark to prototype my designs. They are great, but they take about a month to get the pcbs in my hands. I'd like to iron out as many details as possible before the testing stage.
Think of induced voltages in the signal path instead of worrying about all apparent ground loops which are often innocuous.
Not really.
Pin 1 is in fact a functional not safety earth, it serves two purposes:
1: It helps to control the common mode voltage.
2: It provides continuity between the cable shield and the chassis, important to keeping RF out.
It is not a safety ground connection (That goes via the power connection), as Ze may be far too high to handle the fault currents.
Regards, Dan.
Pin 1 is in fact a functional not safety earth, it serves two purposes:
1: It helps to control the common mode voltage.
2: It provides continuity between the cable shield and the chassis, important to keeping RF out.
It is not a safety ground connection (That goes via the power connection), as Ze may be far too high to handle the fault currents.
Regards, Dan.
Nope. As others have pointed out, the beauty of balanced interconnections is that they reject ground loops. The shield is not part of the signal path, so it doesn't matter if a bit of ground current flows in it.
And H.Ott tells us that if "a bit of" shield current becomes a big shield current, then one can fit an extra Chassis to Chassis cable to lower the voltage between Chassis. The extra connection forms a loop and that loop does not affect the signals.
In the original 'Pin 1 Problem' paper, Neil Muncy wrote.
2.8 Cable Construction versus Shield-Current- Induced Noise
To put this issue into perspective, a survey was conducted using a simple two-device equipment interface arranged to emulate a typical ASP system installation. The test setup, shown in Fig. 2, was modeled after a similar one described by Perkins [20, fig. 18]. The noise current through the shield in all tests was adjusted to approxiamtely100mA to simulate worst-case shield currents actually encountered in operating real-world systems.
2.8 Cable Construction versus Shield-Current- Induced Noise
To put this issue into perspective, a survey was conducted using a simple two-device equipment interface arranged to emulate a typical ASP system installation. The test setup, shown in Fig. 2, was modeled after a similar one described by Perkins [20, fig. 18]. The noise current through the shield in all tests was adjusted to approxiamtely100mA to simulate worst-case shield currents actually encountered in operating real-world systems.
Indeed. Look at it as creating a tunnel for the signal between the two chassis. It is as if you build one very large chassis
Jan
Unfortunately many of these rules were devised when balanced in/outs were always transformers. These have true galvanic isolation (you could have huge pin 1 currents and not effect the signal). Modern equipment rarely uses transformer in/outs so there is a large grey area. There are many variations of electronic balanced in/outs and each has there own pros and cons. Blanket statements don't apply anymore. One of the best references are the white papers by bill Whitlock of Jensen transformers.
A studio I used to work in had a Neve v 48 Chanel mixer (a $400k pro mixer with balanced +4dbm in/outs but they were electronic). The inputs constantly developed ground hums. We tried various wiring schemes that didn't work. The only way to clean them up was to install transformers. My point, not all balanced I/O is equal, far from it.
Not really, if you do it right you can still have huge pin one currents without problems.
The thing with modern input stages is that common mode voltage is rather more restricted then it was when all the inputs were transformer balanced, and the common mode ipedance is usually somewhat lower.
Both of these things actually argue for getting the pin one connection to chassis to be as low an impedance as possible so that the two cases more closely approximate the same potential.
In fact these rules were only properly documented recently (AES 48 is the document describing the right way to do this) and it very much assumes sand based input and output.
Regards, Dan.
The thing with modern input stages is that common mode voltage is rather more restricted then it was when all the inputs were transformer balanced, and the common mode ipedance is usually somewhat lower.
Both of these things actually argue for getting the pin one connection to chassis to be as low an impedance as possible so that the two cases more closely approximate the same potential.
In fact these rules were only properly documented recently (AES 48 is the document describing the right way to do this) and it very much assumes sand based input and output.
Regards, Dan.
When Neil Muncy wrote the Pin1 Problem paper (this is what AES 48 is based on) he tested what happens when noise current is forced through the shield circuit. When the chassis XLR connectors are correctly installed, the noise current is not a problem.
He wrote:
6.3 System Test 1: I/O Shields Terminated at Chassis Entrance Point
The generic mixer shown in Fig. 9 was installed as device 2 in the test system shown in Fig. 8.
Using the procedure outlined in Section 6.1, a 60-Hz square-wave test current of approximately 100 mA was injected into the ground loop. In the first test, all I/O cable shields were terminated directly to chassis ground at the point of entry, as shown in Fig. 8. The noise current flowing in the resulting ground loop travels around each chassis to the equipment ground conductor without sharing audio signal ground conductors anywhere in the system. Regardless of the input used, the degradation of the system dynamic range was too small to measure, even with all input faders wide open and all unused inputs unterminated.
He wrote:
6.3 System Test 1: I/O Shields Terminated at Chassis Entrance Point
The generic mixer shown in Fig. 9 was installed as device 2 in the test system shown in Fig. 8.
Using the procedure outlined in Section 6.1, a 60-Hz square-wave test current of approximately 100 mA was injected into the ground loop. In the first test, all I/O cable shields were terminated directly to chassis ground at the point of entry, as shown in Fig. 8. The noise current flowing in the resulting ground loop travels around each chassis to the equipment ground conductor without sharing audio signal ground conductors anywhere in the system. Regardless of the input used, the degradation of the system dynamic range was too small to measure, even with all input faders wide open and all unused inputs unterminated.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.