Are there things that can be addressed in older circuits to modify them to deal better with the wireless routers and such?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

thor.zmt

Well-known member
Joined
Feb 3, 2020
Messages
776
Location
South East Asia (not China or SAR's)
This is split of from a different thread:

Reducing hum, noise and ground loops?

Are there things that can be addressed in older circuits to modify them to deal better with the wireless routers and such?

Add additional measures for RFI hardening.

My personal suggestion is make sure that XLR cables separate Pin 1 and shell and connect the screen to shell and run a dedicated wire for pin 1. In other words treat the shell and screen as "chassis", pin 1 as "audio ground" and Pin 2/3 as signal.

Personally my signal cables are CAT6 STP, 3 pairs of wires for pin 1, one pair for 2/3 and all shields are connected to shell.

If using Jack or RCA connectors etc, which lack a separate chassis connection make sure they are isolated and fit a binding post for shields and connect the shields ona wiretail with a stackable spade connector.

Inside equipment keep chassis and audio grounds separate. Depending on original design it is often useful to add small value capacitors around amplifier IC's (op-amp's etc.).

1) Place 10pF C0G/NP0 capacitor from -in to output and from +in to gnd on any op-amp anywhere in the circuit, IF the Op-Amp is unity gain stable (NE5532, TL07X and most common audio targeted and audio use units are unity gain stable).

This should be as close as possible to the IC. SMT capacitors are strongly recommended. If you do have a follower (- & out linked) or the +in is linked directly to GND the respective capacitors can be omitted.

Placing a 10nF C0G/NP0 capacitor between +V and -V, again SMD is preferred, improves the RFI resistance and stability and this capacitor should be below the Op-Amp with the smallest possible loop.

Make sure to also have 100nF from +V/-V to GND directly at the +V/-V Pins. Again SMT and C0G/NP0 is recommended, X7R SMD or TH film capacitors are acceptable.

Resistors or inductors in power supply lines need to be carefully considered, however placing SMD Ferrite Beads (330 Ohm @ 100MHz) in series with the supply lines per IC is definitely a good idea, especially if not making a 4-Layer or more design with contiguous ground and power planes. Again, put them close to the IC.

Adding all these capacitors dramatically helps suppress RF ingress.
Power supply lines on single layer PCB's or veroboard make good Antennae's. As do signal traces.

Do not allow any "floating" metal. Make sure all metal has a low resistance/impedance to either chassis (mostly) or audio ground.

All input and output connectors should have RFI mitigation measures:

For inputs using a "PI" filter with:

100p to chassis -> 330R Ferrite Bead -> 100pF to chassis

For outputs use "T" Filters with:

330R Ferrite Bead -> 100pF to chassis -> 330R Ferrite Bead

For low voltage power supplies using a "PI" filter with:

100nF to chassis -> 330R Ferrite Bead -> 100nF to chassis

Use this on each connection that is not bonded "hard" and directly to the chassis.

All components should be SMT.

For PCB mounted sockets, ferrite beads can be added by cutting traces and bridging the cut with a ferrite bead. 0402 and even 0603 capacitors may be too small to solder reliably by hand, so use 1206. Selfadheasive copper tape can be used to create a local image plane. If the PCB mounted sockets link ground and Chassis, cut or remove the links or exchange them for version that break the connection.

For situations like XLR Chassis sockets and jacks that use wires to PCB's, it may be best to make a small PCB that can be directly fitted on the Socket and contains SMD 0603 or smaller format capacitors (smaller is better) and preferably the largest physical size you can find for Ferrite beads to create filters that drain any RF noise into the chassis and prevent it from entering the signal path.

Loop areas matter and ideally the small PCB gets a solid image plane connected to the chassis.

High Voltage (Mains) connections require great care to be safe. Using an IEC mains inlet with build in RF Filter replacing a standard Plastic item is a good choice. If refitting equipment with a fixed power cable, use a powercon socket (or IEC if you enjoy the required dremelling and cutting and filing) and fit a separate mains EMC filter preferably as a canned unit with metal case for chassis mounting.

Large areas of plastic that lead to RF ingress can be covered with either self adhesive copper foil or with PCB material.

All connections after the RFI blocker should be made using shielded cable, shield = chassis and using separate audio ground where required.

Cat 6e STP (shielded twisted pair) network cable makes excellent internal wiring and is also excellent cable for signal cables.

Connect Chassis (and earth) together with audio ground at the power supply "star point" using 100nF & 100R and connect a pair of anti parallel diodes rated at 25A RMS current.

Usually a 25A rated diode bridge block rectifier can be used (link +/- together, the two ~ pin's are used to chassis on one side and ground on the other. The requirements for earth bond testing is to drop less than 2.5V with 25A test current.

Make sure that the diode bridge complies with the voltage drop limit and all other connections are low resistance. This makes sure you have no earth loops and the device remains electrically safety regulation compliant.

In cases where for example a large display window with display is present, creating a faraday cage from PCB material and adding ferrites to the cables can improve the RF ingress situation.

A device refitted thusly and used with a correct cable that does not mix audio ground and chassis/screen and guards conscientiously against RF Ingress can often improve sound quality and SNR dramatically.

Add internal power supply modifications and possibly topology improvements to lower thermal and active device noise and you have a device suitable for employment in the 21st century and in systems with 120dB SNR.

Thor
 
Last edited:
This is split of from a different thread:

Reducing hum, noise and ground loops?



Add additional measures for RFI hardening.

My personal suggestion is make sure that XLR cables separate Pin 1 and shell and connect the screen to shell and run a dedicated wire for pin 1. In other words treat the shell and screen as "chassis", pin 1 as "audio ground" and Pin 2/3 as signal.

For a balanced input I'm not seeing an advantage in connecting audio ground from the source device ?

Personally my signal cables are CAT6 STP, 3 pairs of wires for pin 1, one pair for 2/3 and all shields are connected to shell.

Sounds nice although I'd question the symmetry depending on which pair used for Signal. But this may not be significant tbh.
I don't have STP CAT cable to look at cross section etc.
Meanwhile the pro audio world runs on Screened 2-Core (or starquad in some cases) so I don't think it'll catch on !
But for DIY all good I guess.

Do not allow any "floating" metal. Make sure all metal has a low resistance/impedance to either chassis (mostly) or audio ground.
+1

Loop areas matter and ideally the small PCB gets a solid image plane connected to the chassis.
+1

Large areas of plastic that lead to RF ingress can be covered with either self adhesive copper foil or with PCB material.
+1

In cases where for example a large display window with display is present, creating a faraday cage from PCB material and adding ferrites to the cables can improve the RF ingress situation.
+1
Depending on where you are in the design / build process it might be worth looking at display windows with a degree of screening built in.
eg (general info only - no personal link)
EMC Display RFI Shielding Laminated Window and Screen
 
For a balanced input I'm not seeing an advantage in connecting audio ground from the source device ?

That depends on the precise design of electronics.

If we have a classic balanced system with a transformer coupled output and input, then absolutely no. Even just one transformer will do.

Alternatively we could use AD and DA and digital isolators.

If we have modern systems using electronically balanced I/O and we are looking for 120dB SNR/DNR, definitely.

Sounds nice although I'd question the symmetry depending on which pair used for Signal. But this may not be significant tbh.

It is not significant, as each pair has an individual foil screen.

I don't have STP CAT cable to look at cross section etc.

Cat6e STP uses standard 24 gauge / 0.2mm^2 conductors.

Each pair has an individual foil screen. There is an overall foil screen and woven shield.

Meanwhile the pro audio world runs on Screened 2-Core (or starquad in some cases) so I don't think it'll catch on !

I know. And it's a loss IMNSHO.

But for DIY and one's own studio, why not use an improved option?

Thor
 
Last edited:
It is not significant, as each pair has an individual foil screen.
Good Point. I had missed that in my thinking.

Cat6e STP uses standard 24 gauge / 0.2mm^2 conductors.

Each pair has an individual foil screen. There is an overall foil screen and woven shield.
Understood. I'll try and get my hands on some. It's not that common. At least here (UK) and I don't want to get a 100m drum !

I know. And it's a loss IMNSHO.

But for DIY and one's own studio, why not use an improved option?

Can't argue with that. Hence I also said "But for DIY all good I guess." :)
 
Understood. I'll try and get my hands on some. It's not that common. At least here (UK) and I don't want to get a 100m drum !

I do basically buy boxes. If you have a studio to wire, with patch cables etc, a few 10m microphone cables etc an 100m box is not that much.

The cable is also great for AES-EBU and probably will do fine for USB 2.0 high speed as well. So it's really universal cable.

Thor
 
Last edited:
If you are designing new PCB's you may want to treat the audio board as an RF device, using board level coaxial connectors, shielding, feed thru caps. MCX connectors are OK, not too big, easy to work with. These can be placed on opposite side of shields to ease routing in and out of shields. Coax RG402, RG405 provides good shielding efficiency, but RD-316 crimp cable is likely sufficient. Connector type has many available cable options. Those board steel/tin/nickel shields with a frame and a snap on cover works good. The RF in the environment is going not going to improve or make life easier on audio equipment. Try leaving a cellphone next to a device for laughs. 5G in my area uses spectrum from 630MHz to 3.8GHz.
 
Most of the CAT6 cable I've seen is rather stiff. What types are flexible enough to lay flat on a studio floor?

I have seen one cable used in a telco equipment room that was very pliable but it was unmarked and unidentifiable.
 
Flexible network cable is stranded and shouldn't exceed a length of 50 cm. It's simply called "patch cable"

Neither UTP nor patch cable should be on the floor. It's light enough to fly it.
 
Most of the CAT6 cable I've seen is rather stiff. What types are flexible enough to lay flat on a studio floor?

You mean as microphone cable? Cat 6 lays flat enough, with a little more effort, but it needs to be uncoiled properly and will never be as limp PVC stranded Microphone cable.

If you must have this, you can use starquad with one pair as pin 1 instead.

Thor
 

Latest posts

Back
Top