3E Audio 260-2-29A : TPA3255 / HA3588 CoilCraft / PFFB Fully differential.

[daniboun]

Interesting thx )

We see the shield is connected to the chassis.

 

[UAN]

Сбалансированный интерфейс обычно имеет три провода:

1. Эталон (обычно заземление)
2. Сигнал (горячий)
3. Инверсный (холодный)

Как видите, вы оставили кабель отключенным (заземленным)

View attachment 311451

Симметричный сигнал использует для передачи три проводника: два информационных (прямой и обратный) и один заземление.
На фото разъема:
красный – прямой +,
черный - обратный -,
желтый - земля.
Соответственно, желтый цвет должен быть подключен к контакту заземления розетки TRS.
На этом старом фото желтый провод не подключен, но контакт заземления розетки TRS физически подключен к корпусу, то есть к массе усилителя.

 

[daniboun]

A balanced signal uses three conductors for transmission: two information (direct and inverse) and one ground.
In the photo of the connector:
red - straight +,
black - inverse -,
yellow - earth.
Accordingly, yellow must be connected to the ground pin of the TRS socket.
In this old photo, the yellow wire is not connected, but the ground contact of the TRS socket is physically connected to the case, that is, to the amplifier ground.

Recommended practice should be in a simple way for balanced xlr plug

 

[MCH]

Recommended practice should be in a simple way for balanced xlr plug


View attachment 311514

Capacitor between XLR shell and ground pin

The topic isn't new, but still seems to be contested. Here and here it is advised to add 100pF capacitor in a way that both pin #1 and shell are connected. Is this what's normally recommended or connection between the cable and pin should be severed? Also, what is the optimal capacitor...

www.audiosciencereview.com

 

[daniboun]

@S=klogW

I have a practical case to test, since you have the E1DA Cosmos ADC )
can you measure the impact of these two scenarios with your 3E Audio module?

 

[antcollinet]

@S=klogW

I have a practical case to test, since you have the E1DA Cosmos ADC )
can you measure the impact of these two scenarios with your 3E Audio module?

View attachment 311525

I believe in the AES method there should also be a single connection somewhere between the module/signal ground and chassis.

 

[S=klogW]

@S=klogW

I have a practical case to test, since you have the E1DA Cosmos ADC )
can you measure the impact of these two scenarios with your 3E Audio module?

View attachment 311525

Yes, I'll set that up and run the tests tomorrow, plus I'll try the third option suggested by @antcollinet : single module/signal ground connection only to avoid ground loop circulating currents.

 

[daniboun]

Yes, I'll set that up and run the tests tomorrow, plus I'll try the third option suggested by @antcollinet : single module/signal ground connection only to avoid ground loop circulating currents.

Many many thanks ))) to help I found this for best practice :

https://audiosciencereview.com/forum/index.php?threads/xlr-input-to-amp-module-connections.15547/

You re the best amigo )))) yeah

 

[antcollinet]

Yes, I'll set that up and run the tests tomorrow, plus I'll try the third option suggested by @antcollinet : single module/signal ground connection only to avoid ground loop circulating currents.

Just to be clear, the single ground from the module to chassis - is in addtion to pin 1 to chassis. Here is a picture I found from a (admittedly draft for comment) version of AES48

 

[daniboun]

Just to be clear, the single ground from the module to chassis - is in addtion to pin 1 to chassis. Here is a picture I found from a (admittedly draft for comment) version of AES48
View attachment 311555

That is exactly what Bruno Putzeys described.

 

[antcollinet]

That is exactly what Bruno Putzeys described.

View attachment 311563

Yep - I wonder why

 

[S=klogW]

Interesting thx )

We see the shield is connected to the chassis.
View attachment 311509

I've always thought this diagram very instructive. One of the things that it is saying that if you connect the shield wire at both ends of the cable to the two chassis then you are essentially making a new, full-coverage Faraday screened cage which protects all the electrical circuits and signal wires seen above from the effects of Radio Frequency Interference (RFI) getting in. Rather like two spacecraft connected by an air-tight umbilical connection not letting any air out of either ship.

Note also that both the input and output circuits show just one connection to their respective chassis. In this case the connections have been drawn to the 'bottom' of each chassis, but they can also be drawn to pin 1 of each XLR conn. with the short looped wire from pin1 to its chassis making a nearly equivalent circuit. The difference in performance between the two options comes down to which scheme gives the lowest impedance to circulating ground currents.

So, if I scratch away any paint and oxidisation from the chassis before making the ground connection using serrated-toothed washers and use Sanchem NO-OX-ID protective grease and I make a nice low impedance soldered, thick copper braided short connection to XLR pin 1, then I don't expect to see a significant difference between the 'Right' and 'Wrong' diagrams i.e. the quality of the measured result will depend on the quality of the electrical and mechanical connections that are made in the pursuit of the best/lowest ground impedance to ensure lowest ground circulating currents.

 

[antcollinet]

I've always thought this diagram very instructive. One of the things that it is saying that if you connect the shield wire at both ends of the cable to the two chassis then you are essentially making a new, full-coverage Faraday screened cage which protects all the electrical circuits and signal wires seen above from the effects of Radio Frequency Interference (RFI) getting in. Rather like two spacecraft connected by an air-tight umbilical connection not letting any air out of either ship.

Note also that both the input and output circuits show just one connection to their respective chassis. In this case the connections have been drawn to the 'bottom' of each chassis, but they can also be drawn to pin 1 of each XLR conn. with the short looped wire from pin1 to its chassis making a nearly equivalent circuit. The difference in performance between the two options comes down to which scheme gives the lowest impedance to circulating ground currents.

So, if I scratch away any paint and oxidisation from the chassis before making the ground connection using serrated-toothed washers and use Sanchem NO-OX-ID protective grease and I make a nice low impedance soldered, thick copper braided short connection to XLR pin 1, then I don't expect to see a significant difference between the 'Right' and 'Wrong' diagrams i.e. the quality of the measured result will depend on the quality of the electrical and mechanical connections that are made in the pursuit of the best/lowest ground impedance to ensure lowest ground circulating currents.

View attachment 311575

No, there is a massive difference.

In the 'right' case, all the common mode noise between devices is conducted through the chassis and goes no-where near the analogue electronics.

In the 'wrong' case, you connect the screen to the analogue electronics ground plane. Now all the common mode noise between devices is conducted through that ground plane, and will cause interference to any single ended signals tracked across the PCB of the circuits.

 

[daniboun]

I've always thought this diagram very instructive. One of the things that it is saying that if you connect the shield wire at both ends of the cable to the two chassis then you are essentially making a new, full-coverage Faraday screened cage which protects all the electrical circuits and signal wires seen above from the effects of Radio Frequency Interference (RFI) getting in. Rather like two spacecraft connected by an air-tight umbilical connection not letting any air out of either ship.

Note also that both the input and output circuits show just one connection to their respective chassis. In this case the connections have been drawn to the 'bottom' of each chassis, but they can also be drawn to pin 1 of each XLR conn. with the short looped wire from pin1 to its chassis making a nearly equivalent circuit. The difference in performance between the two options comes down to which scheme gives the lowest impedance to circulating ground currents.

So, if I scratch away any paint and oxidisation from the chassis before making the ground connection using serrated-toothed washers and use Sanchem NO-OX-ID protective grease and I make a nice low impedance soldered, thick copper braided short connection to XLR pin 1, then I don't expect to see a significant difference between the 'Right' and 'Wrong' diagrams i.e. the quality of the measured result will depend on the quality of the electrical and mechanical connections that are made in the pursuit of the best/lowest ground impedance to ensure lowest ground circulating currents.

That is a good resume )
I just wonder if the JST connectors and cables supplied by 3E audio can follow the requirements of the AES method..... we are impatiently awaiting the measurements)

 

[daniboun]

BTW want to quote this

 

[MCH]

I've always thought this diagram very instructive. One of the things that it is saying that if you connect the shield wire at both ends of the cable to the two chassis then you are essentially making a new, full-coverage Faraday screened cage which protects all the electrical circuits and signal wires seen above from the effects of Radio Frequency Interference (RFI) getting in. Rather like two spacecraft connected by an air-tight umbilical connection not letting any air out of either ship.

Note also that both the input and output circuits show just one connection to their respective chassis. In this case the connections have been drawn to the 'bottom' of each chassis, but they can also be drawn to pin 1 of each XLR conn. with the short looped wire from pin1 to its chassis making a nearly equivalent circuit. The difference in performance between the two options comes down to which scheme gives the lowest impedance to circulating ground currents.

So, if I scratch away any paint and oxidisation from the chassis before making the ground connection using serrated-toothed washers and use Sanchem NO-OX-ID protective grease and I make a nice low impedance soldered, thick copper braided short connection to XLR pin 1, then I don't expect to see a significant difference between the 'Right' and 'Wrong' diagrams i.e. the quality of the measured result will depend on the quality of the electrical and mechanical connections that are made in the pursuit of the best/lowest ground impedance to ensure lowest ground circulating currents.

View attachment 311575

See section 4.3:

https://www.hypex.nl/media/c8/3d/26/1648383450/Application%20note%20Legacy%20pin%201%20problems.pdf

 

[S=klogW]

No, there is a massive difference.

In the 'right' case, all the common mode noise between devices is conducted through the chassis and goes no-where near the analogue electronics.

In the 'wrong' case, you connect the screen to the analogue electronics ground plane. Now all the common mode noise between devices is conducted through that ground plane, and will cause interference to any single ended signals tracked across the PCB of the circuits.

I agree that would be a significant difference. So maybe, I'm not interpreting the 'Wrong' diagram correctly. Is the 'Wrong' diagram saying that the chassis only gets a connection to the protective/safety earth (yellow/green wire to the mains plug) and the chassis doesn't get a connection to the (usually labelled) Vminus output of the psu?

 

[antcollinet]

I agree that would be a significant difference. So maybe, I'm not interpreting the 'Wrong' diagram correctly. Is the 'Wrong' diagram saying that the chassis only gets a connection to the protective/safety earth (yellow/green wire to the mains plug) and the chassis doesn't get a connection to the (usually labelled) Vminus output of the psu?

No, in the wrong case, you are connecting the shield of the xlr interconnect to the ground of your analogue circuit.

In this case, the ground loop current conducts through your analogue ground - probably back to earth through the chassis connection - or on though an another analogue interconnect to the next device in the chain. That ground loop current causes noise pickup in analogue circuits in your device, as shown in the link posted above by @MCH

Bear in mind your PSU may well be a bipolar +/- supply - in which case you definately don't want to connect v- to the chassis.

You *do* connect one point of your analogue ground to the chassis - since this is only a single connection to the "system shield", then there is no current in that connection. It just serves to set a reference point, and ensures the common mode voltage difference between devices isn't too large for the common mode rejection of the input diff amps. This analog ground might also be the v- of a single ended supply or the midpoint of a bipolar +/- supply. But this is design dependent.

 

[S=klogW]

No, in the wrong case, you are connecting the shield of the xlr interconnect to the ground of your analogue circuit.

In this case, the ground loop current conducts through your analogue ground - probably back to earth through the chassis connection - or on though an another analogue interconnect to the next device in the chain. That ground loop current causes noise pickup in analogue circuits in your device, as shown in the link posted above by @MCH

Bear in mind your PSU may well be a bipolar +/- supply - in which case you definately don't want to connect v- to the chassis.

You *do* connect one point of your analogue ground to the chassis - since this is only a single connection to the "system shield", then there is no current in that connection. It just serves to set a reference point, and ensures the common mode voltage difference between devices isn't too large for the common mode rejection of the input diff amps. This analog ground might also be the v- of a single ended supply or the midpoint of a bipolar +/- supply. But this is design dependent.

Let me try again then to get 'Wrong' right . Do I use exactly Figure 24 for the 'Right' setup, but in addition, I make a connection from pin 1 of the input XLR chassis connector, mounted on the 3eAudio's chassis, to the ground plane/ground pin of the 3eAudio board? Is this how to do it 'Wrong'?

 

[antcollinet]

Let me try again then to get 'Wrong' right . Do I use exactly Figure 24 for the 'Right' setup, but in addition, I make a connection from pin 1 of the input XLR chassis connector, mounted on the 3eAudio's chassis, to the ground plane/ground pin of the 3eAudio board? Is this how to do it 'Wrong'?

View attachment 311593

No, because that would mean multiple connections from the board to the chassis, which could circulate current (one from each XLR). You want one single (short) connection from a ground point on the board to the chassis. (Represented by the connection to the bottom horizontal line of the chassis in figure 24). This should be directly from a connection point on the ground plane.

The AES drawing above shows this should be a star point on the chassis for the incoming (mains) earth, the PSU 0V reference, and the signal 0V reference.