grapgic.png

 

Introduction

Having needed an audio cable recently for the lab, I was saddened to see the poor quality of some off-the-shelf pre-assembled cables. I didn’t realize poor audio cables were still a phenomenon in the 21st century. It got me wondering, how can we test and compare cables and complete cable assemblies?

 

It is an important question because many are forced to spend a small fortune on cables each time any hi-fi or audio/visual product is purchased from stores – the products do not come with the cables, and stores charge a premium for them because they know users won’t want to wait a further day or more to buy something cheaper and better online – especially at Christmas when the postal service is struggling.

 

All sorts of ‘techniques’ are used. As an example see the ‘gold-plated optical cable’. Incidentally the plating at a guess might cost just pennies or less, but a high premium is demanded for it – despite the fact that no property of gold is useful for optical signaling purposes. Gold has a high perceived value even if the quantity of it is tiny. The wording is always clever to leave a certain impression yet tell no lie.

 

The cables below may be really good. I have no idea, I have not tried them. Yet the prices are impressive. These screenshots are cropped and reduced in size for fair use as examples of price versus listed features.

pcworld-optical.jpg

pcworld-hdmi.jpg

maplin-audio.jpg

 

This post investigates how sometimes we can do better ourselves at a fraction of the cost of some pre-assembled cables – and we’re going to avoid anything non-persuasive (i.e. anything we can’t throw available test equipment at to prove or disprove), so no ‘directional’ conductors or oxygen-free copper or silver wires or Kevlar shielding allowed!. However, all comments and suggestions would be gratefully received – I know there are probably many practical considerations such as cable flexibility and durability which are of concern to users. Therefore the information here is more of use for a home environment until there is feedback from people on whether it is useful for studio or live performances too or what modifications they would prefer such as ultra-flexible cables. This post also examines how we can test off-the-shelf and home-made cables and see how good or bad they really are.

 

What is wrong with Off-The-Shelf?

The cable I inspected did not have great shielding (proven through tests – see later) consisting of wire strands providing perhaps 30% (this is a guess) of coverage of the centre conductor.

bought-cable-connector.jpg

 

In its favour, the cable outer insulation was very thick and maybe could withstand rough handling as long as the rough handling was no-where near the connectors.

label-on-the-cable.jpg

 

Furthermore it was a very low cost cable and probably would be adequate in a non-electrically-noisy environment. The connections would probably be fine if the cable was not mishandled.

 

What things would we want to see in good audio cable assemblies?

Everyone will have different requirements but from a general point of view these requirements would come out pretty high:

  • 100% coverage shielded cables, grounded, to minimise capacitive pickup and RF pickup
  • Two and three cores for flexibility. The two-cored cable could be used for mono or stereo applications; for mono use one of the cores would provide the audio signal and the other core would be used for the ground connection. The shield would be grounded at one end. For stereo use, the three core cable could be used, and the shield would be used as the ground connection at one end. For balanced audio use (e.g. with XLR connectors) then again either cable could be used.
  • Ideally a controlled pair cable for balanced audio applications, to reduce the effects of as many modes of noise pickup as possible

 

To provide protection from interference at low frequencies the best practice is to have the shield connected only at one end. When we hear mobile phone pickup on audio circuits, this is actually low frequency interference components that we are hearing. However we also want to minimise high frequency pickup up to about a few MHz just in case it mixes with signals in electronics. There is a ‘hybrid grounding scheme’ discussed in an Analog Devices paper (PDF) where one end has a capacitor to ground. Any ceramic 100nF capacitor would work admirably. A large ferrite may be slightly effective too but unlikely; the input impedance of audio circuits is far higher than the impedance from the ferrite and we might obtain less than a dB of benefit at best, and at audio frequencies there would be no benefit, it would just be a weight on the end of the cable. But worth experimenting with if there are issues caused by high frequency (of the order of MHz or higher) signals.

 

For connectors the following properties would be good to see:

  • Ruggedness
  • Fully shielded
  • Easy to solder terminals
  • Conductors secured with more than just solder
  • A cable clamp
  • Strain relief

 

A look at some Connectors

Neutrik connectors are considered good and reliable, so I made it a priority to try out some of these.

neutrik-an.jpg

 

I did not go for the most expensive one but instead picked a model that is mentioned on various Internet forums; the NP2XNP2X. I wasn’t entirely happy with it but it’s definitely a very good product – I suspect this particular model is also popular because it provides lots of space for soldering. It had a trough-like shape for soldering the centre connection, and just a curved surface for soldering the screen. Another point is the slightly unusual shell found on Neutrik connectors. The large metal shell actually fits from the ‘business’ end of the connector, not the solder terminal side. The smaller plastic threaded part is pushed over the cable before soldering the connections. The cable clamp was achieved by a separate plastic piece. It was quite a novel attachment and it may be quite effective. I just didn’t like that nothing is secure until the final part of the assembly is screwed. Up till that point, the wires are not secured (solder alone is not really secure). Still, I can see why this is a popular connector.

 

I also tried a lower-cost REAN branded one from Neutrik too, part code NYS204NYS204.

rean-an.jpg

 

The NYS204 has a more conventional design. The cable clamp is integrated into the main connector assembly. It would not be ideal for very low level signals in a noisy environment because it had an unscreened plastic shell. In its favour it is very low cost.

 

amphenol-an.jpg

Next, I examined an Amphenol right-angle connector part number ACPM-RB Amphenol right-angle connector part number ACPM-RB. It has a multi-part metal shell. Once the curved cap is placed on the right angle part of the connector, the back shell can then be screwed on and it will secure the curved cap. It was also great that it had solder terminals with holes; the sleeve terminal has a partially knocked-out hole, leaving a hook. The integrated cable clamp has three prongs which seems like a good idea. A plastic sleeve is also supplied to help with preventing shorts to the metal shell. Finally the metal backshell also has a small piece of plastic strain relief to keep the wire straight for a short distance beyond the metal shell. Overall, it is a very high quality part.

 

switchcraft-an.jpg

However my favourite was one by Switchcraft, model 297Switchcraft, model 297. It had the integrated cable clamp, solder terminals with holes, and an integrated insulator to protect against shorts from the tip or ring wires to the sleeve. There was no hole for the sleeve terminal and because of the integrated insulator the sleeve wire would have to be folded over and soldered on the other side.  It has many other great features however. All of the solder terminals are bent in an angle such that there is adequate room for bringing the wires from the outer side. The construction of the plug uses some type of glass fibre for acting as an insulator unlike plastic seen on other connectors. The solder terminals are nickel-silver electro-tinned for great ease of soldering. Furthermore the metal used for the solder terminals  and integrated cable clamp is thicker than some other connectors had. There is no plastic strain relief but this doesn’t matter when built into a complete cable assembly (see later).

 

A look at some Cables

In order to have results I can use in future, I tried to only use easily available cables with a known part number. Cable is expensive (copper is expensive) so I wanted to standardise on just a few types of cables that would be useful for audio and non-audio applications. The cables that were selected were Belden 9533Belden 9533 and Belden 9841NHBelden 9841NH.

 

The Belden 9533Belden 9533 is a 4.1mm thin cable with three cores and a foil shield with an associated bare wire (called a drain wire) and is primarily intended for RS-232 use. The 9841NH9841NH has two cores and a braid and foil shield. It is thicker (5.9mm) and physically tougher. It is overkill for home audio applications but I liked it because it could also be used for non-audio applications. It is primarily intended for RS-485 signalling.

belden-9533-4mm-foil-removed-annotated.jpg

 

It should be noted that these cables are actually even better constructed than their datasheet description : ) The 9533 cable is not a controlled impedance cable but nevertheless the three conductors are tightly held together in a consistent intimate configuration in their foil shield as seen below. Also the foil shield is nicely folded on the outer edge for very effective 100% coverage.

4mm-jacket-removed.jpg

 

The 9841NH 9841NH also has a similar folded foil shield but it is also covered with a 90% coverage braid surrounding the foil shield making the cable better screened at lower frequencies and providing physical resiliency.

belden-9841NH-6mm-jacket-removed.jpg

 

Also the 9841NH does have a controlled impedance pair of wires and filler material is used to keep the wires in formation. We don’t actually care about the impedance, but we care that the two wires are consistently intimate with a twist, so that any noise signal that is picked up will be common; it will therefore be eliminated in a balanced audio configuration. The 9841NH would be superb for balanced audio applications as a result.

belden-9841NH-6mm-all-disassembled-annotated.jpg

 

In summary it is very interesting that both of these cables should perform very well yet are low cost and available in single metre lengths if desired; no need to buy a huge roll!

 

Building Audio Cable Assemblies

Engineering practice dictates that solder alone should not secure a connection, wire breaks should not short against other pins, the cable should be clamped and cable strain relief should be provided to allow the cable to flex without strain near a connector. One particular difficulty is holding the connector and wire together while soldering. Some websites incorrectly state that it is ok to add solder to the connector and wire separately and then re-melt solder on the connector and then add the wire while the solder is molten in an attempt to make it easier to form the join.

 

A clamp can solve the problem. Any clamp with soft inserts (e.g. pieces of wood) could be used. The one below is all aluminium and has a V-shaped groove so the part doesn’t move too much.

connector-clamped.jpg

 

Another issue is that the solder might not ‘stick’ to the large metal terminals. The solution is to use a high-power soldering iron and a large bit (3mm or slightly more) to get heat to the join. A small bit of flux paste will result in an excellent join. The photos below show an example walk-though assembling an audio cable. Don’t forget to pass the connector backshell onto the cable first.

wires-stripped2-annotated.jpg

 

When the braid shield needs soldering there is always the problem of how to deal with the pigtail. One nice method is to use solder sleeves but there is not enough room to use them in a typical audio connector. Instead it is easiest to use a small piece of heatshrink tubing to insulate a small portion of it. Twist the braid into shape, place the heatshrink tube into position and then tin the end; the tube will shrink due to the conducted heat. The aim is to just tin the end of the pigtail, not the entire length.

wire-tinned-annotated.jpg

 

Similarly any other wires should be just very lightly tinned; no more. Once the wires are inserted through the holes they can be folded back and then a small bit of flux pasteflux paste applied before soldering.

flux-applied-annotated.jpg

 

The join is complete when the solder fills the hole entirely and the entire wire end is covered with solder and has a fillet all around the side that was soldered.

soldered-tip-residue-annotated.jpg

 

With fresh clean connectors it is best to use No-Clean fluxNo-Clean flux. It will leave a residue but it can be ignored or (if you prefer) wiped off (using flux remover and a brush or cleaning wipe) as shown here:

soldered-tip-cleaned.jpg

 

I don’t have a specialist tool for the cable clamp but a low-cost crimp terminal toollow-cost crimp terminal tool will probably work very well. Note that ideally you want the plastic outer insulation to extend one half the diameter of the cable from the end of the clamp if possible with the specific connector in question; it is not always easy with your first connector especially since they are so space constrained and mine is a little low here.

clamping-action.jpg

 

By opening and closing the jaws of the crimp tool and rotating the assembly it is possible to get a neat clamp all around.

all-soldered.jpg

 

Heatshrink tube can be used as insulation where possible, but audio connectors do not have a lot of space. An alternative option could be to use Kapton tape Kapton tape (use a pair of tweezers to pull it through) wrapped around individual conductors. It is very thin so will fit in a tight space, and it won’t leak a sticky residue.

kapton-applied.jpg

 

The end result will allow the connector backshell to screw on comfortably. At this point there should be a lot of confidence that the wires will not have broken during the backshell assembly (due to the cable clamp securing the heavy cable successfully) nor shorted anywhere. However it should be verified immediately with a multimeter.

 

Next some heatshrink tubing (16mm16mm and 12mm12mm are useful; these particular ones are adhesive lined and have 4:1 shrink ratio) can be placed around a portion of the connector backshell and the cable. 4:1 ratio tubing is really ideal for this. Adhesive lined is excellent. It can be purchased in short lengths good for at least 10 complete cable assemblies so it is not expensive. Start at one end (the connector shell end) and make sure that it shrinks evenly toward the cable end, but briefly stop before it touches the cable. Let it cool a bit, partially unscrew the connector backshell, squeeze a bit of thread lock adhesivethread lock adhesive and then tighten the backshell. Then finish the heatshrinking and then retest. The end result will be very rugged and should provide years of good service. For repair the end can be snipped off, the heatshrink cut off, and the threadlock is semi-permanent; it can be disassembled with force.

heat-shrinked-backshell.jpg

 

Testing Cable Assemblies

Some basic tests are obvious; the usual tests for shorts and intermittent connections can be done with a multimeter. I also wanted to check how effective the cable was at shielding from unwanted signals however. There are different ways that unwanted signals can be picked up. Capacitive coupling can occur if the cable is very close to another cable for example. Another source can be whenever there is changing flux nearby, i.e. inductively coupled noise. Radio signals may also be picked up; these should be inaudible but there is a possibility that mixing or demodulation may inadvertently occur in the attached electronic equipment. One test method would be to deliberately have a noise source nearby and see how much is picked up by the test cable. I decided to generate a square wave type of signal at high current and place it close to the test cable. I used an IRFZ44 MOSFETIRFZ44 MOSFET and a 48V DC supply; the load was a 25 ohm resistor (two 50 ohm, power resistors in parallel) so that almost 2A was being switched. I didn’t have an alternative power MOSFET handy, so I have to be careful not to exceed the safe operating area (SOA) of the MOSFET. The heart of it is shown below. The source signal into the MOSFET was an 8V 500usec pulse every 10msec; this would result in a 100Hz tone along with harmonics.

testbed-speaker.png

 

The circuit was just quickly hacked together as a test - it wasn't pretty but it functioned. Note that at this level of power (close to 100W) you do want to make sure nothing accidentally shorts. I have a habit to place new experiments inside a translucent box the first time they are powered up, if they are high powered circuits. I don't want something accidentally bursting uncontained.

mosfet-cct-built.jpg

 

I coupled the test cable with a few twists over 10cm or so as a quick experiment. With the off-the-shelf cable I could hear a very loud 100Hz sound through the amplifier and speaker. I then swapped out the cable for the cable I had assembled and repeated the test – no 100Hz sound was audible with the ear right against the speaker.

 

This experiment could be refined with a higher frequency and spectrum analyzer however for the initial tests I wanted to use easily available test equipment. The simple test was enough to validate that home-made cable assemblies could vastly out-perform some commercial cable assemblies. Sticking with simplicity another far better way to test this is to use a radio receiver. It is one of the cheapest yet most sensitive devices in a lab. It can detect femto-watts (10^-15 W) of signal power. As a rough analogy of power ratios consider listening to a pair of headphones 20 km away from your ears

testbed-radio.png

 

Note that most radio receivers will not pick up RF signals so low in frequency, but there are plenty of harmonics that could be received. I upped the pulse rate to 1 every millisecond (and shortened it to 100usec). With the cable connected to the radio I could easily hear a harmonic with the radio tuned to 20kHz with the commercial cable. Loud and clear! I swapped to the home-made cable and there was no signal received (just radio background noise). This is a convincing result that the shielding is exceedingly effective even at these relatively low frequencies.

 

Summary and Next Steps

A lot of topics were quickly covered in this post; cable requirements, connector requirements, how to assemble good audio cables and how to test them. It can be seen that it is entirely practical to hand-assemble audio cables that will out-perform many commercial ones, and confirm their behaviour with low-cost test tools. It was very impressive to see such outstanding results from the radio receiver method using the model 9533 and 9841NH cables that were tested.

 

After having tried these simple tests it was interesting to find an IPC Test Method (PDF) which uses a fairly similar method of checking cable shielding. It too uses a source cable with a test signal, this time from an RF signal generator. The cable under test is mounted parallel to the test signal cable at a fixed distance. The received signal is measured with a field intensity meter. With more time or if I was checking many cables it would be worthwhile building up such a dedicated test rig. I may try to run similar tests on other types of cables, for non-audio applications.

 

finished-cable-heatshrinked.jpg