The Terminator !!!!

Terminator for Unused Digital Outputs

For decades it has been mandatory to use terminators in 10Base2 Ethernet networks that use coaxial cables and BNC connectors.
The purpose of the terminator is to absorb signals so that they do not reflect back down the line.

RF impedance theory says that if a transmission line is open circuited; it will have reflections at certain frequencies. Since the digital output signal has a wide range of signal content with regard to the frequency content, it will have some signal reflected. This reflected signal can back feed into the input of the digital output circuit (no circuit is perfect, and none have infinite input/output isolation) and then back into the rest of the digital circuitry, causing jitter.

A terminator used in Ethernets is a 50 ohm resistor is readily available to purchase. However, coaxial digital outputs of CDPs, DACs, DVD players, AV amps/receivers, etc require a 75 ohm resistor and these are very easy to make.

75 ohm resistors are just 1p each from places like Maplins (code G75R). Simply solder one across the terminals of a cheap phono plug. If the unused digital output has a BNC socket you will also need a BNC to Phono adapter (£2 Maplin FA11M).

 

I really must get round to doing this sometime... I'd hope my transport is well enough designed for it to make no difference but I ought to try it.

On the BNC front, it's probably better/cheaper to just buy a ready made 75Ohm terminator rather than using a phono plug and and adapter. Maplin don't seem too sell any but they shouldn't be hard to get hold of.

(Oh, and you might like to credit CAD for the picture of their product... The guy who runs CAD is a nice chap and would I'm sure appreciate it.)

 

I don't know why they should be functionally any better than DIY but apparently, there are a ready made 75 ohm terminators about.

RS Components : BNC End Terminations
1% 112-3154 £2.00 + VAT
0·1% 125-0470 £2.19 + VAT
0-1GHz 296-7949 £4.49 + VAT

 

I was really thinking of them being cheaper, but they may be better in terms of having one less junction for possible signal reflections I guess? I wouldn't get hung up about it mind.

Martin

 

The purpose of the load in 10 base2 is also for the collision detection, since the output from the NIC is a current not voltage and collisions [ie more than 1 talker] give a rise in the DC potential [there is a DC component to 10base2 output - this is why the centre pin is gold plated to avoid corrosion over time]. If you use 10base2 as an example, there is also a MINIMUM length of line, and between stations. I can't be bothered to dig out the maths but I doubt if the connector on the back of your CD player is a 'transmission line length' away from the PCB ie reflections are instantaneous not delayed as echo/standing waves. 75 ohm terminators are for radio, but I guess they would be happy for the hifi market

 

Use a resistor of around 470 ohm and you have terminators to fit into empty analogue inputs. It will stop RFI or EMI that might find it's way into the amp and reduce/eliminate hiss.

 

I hate to do this, but did you pay any attention to what I wrote to you on HFC??? Provided your amp doesn't have a problem with using a straight shorting plug to terminate an analogue input, this is going to be BETTER. Going straight to using a 470 OHm resistor on an analogue input is I think missing the point.

For those wondering about the suggestion of using a 470 Ohm resistor at all, here is a link posted by BlueMax over on HFC, which is helpful:

http://www.audioasylum.com/audio/tweaks/messages/1571.html

 

Martin
I saw that but thought that there is no signal in an unused input. I was hoping an expert at HFC forum will reply but did not. May be you could post your thoughts, in detail, here. Hopefully one of the resident experts will pick it up and answer.

 

OK, it would have been nice if you'd therefore phrased your last post as a question, rather than a statement of fact, but here's my argument:

There isn't a signal, no. Or rather the idea of what you're proposing is to prevent RFI becoming a signal on the input. If you short the input then the signal line is grounded so there won't be a signal. If you connect a resistor across the two this allows the signal line (i.e. where the centre pin of an RCA connects) to have a higher potential than ground*, hence an RFI signal could still be present.

*Just think V=IR, the point of a resistance is that it can have a potential difference across it.

 

My guess is that voltage and current of RFI is so small that the potential difference across the resistor will be miniscule.

Any electronics people on the forum who can throw some light on this?
Please.

 

Yes, but it's precisely this miniscule signal that you're trying to prevent!

 

What I was trying to say was that the potential difference across the resistor will be miniscule. Not the original RFI.

Best to leave this for an expert on the forum to comment.
Failing that Jon Risch should be able to answer at Audio Asylum BB.

 

I know.

My basic point was that if you short direct to earth, there is ZERO signal. If you have the resistor there, there is some. I haven't yet worked out how the size of this signal will compare with that from the open circuit input (off the top of my head I'm being thick and can't see how to do this...).

So... your earlier post which suggests that you should go straight to using a 470 Ohm resistor to terminate an analogue input to me encourages unnecessary effort, to achieve a potentially worse result. That is all I've been saying all along. As John Risch says, you should only bother with this if you have a problem with clicks/pops when switching through a shorted input. Something he also says is rare.

Do you see my point now?

 

Yes I see your point.
I somehow did not like the idea of shorting an input designed to have a load on it. Not being an expert I decided to err on the safe side.

Without expensive and sensitive measuring instruments, it will be difficult to know how the size of this signal (RFI) in the first place. Also, these interferences will vary from one location to the other.

It will certainly add more credibility to mods/tweeks if the inventor include some measurements to support. Thorsten made an attempt in his articles about DIY speaker cables. Businessmen like Russ Andrews know the importance of this. In his latest catalogue there some graphs to support his claims about his mains conditioners. Extortionately priced though. I have seen DIY versions on the net somewhere. Might have a go this winter.

 

I understand.

Yes, but it should be possible for a given amplitude RF signal, of a particular frequency, to calculate the signal this would cause open circuit and with a resistor across, and compare the two. That is calculate the relative amplitude of these two, which is what to some extent we were debating, not the absolute levels of either. This is beyond me at the moment though...

 

This is possible but I haven't got a clue what figures to use.
Russ Andrews sell these.

Doesn't state if they have resistors or not. But at £19.95p for 12 they might do.

He claim to do research work so should have suitable equipment.
Might be worth sending hinm an email asking what measurable results he got.

PS.
I did a simple test with volume control set at 80-100%, no CD playing and
selecting the unused input. With head close to the tweeter, I noticed a big reduction with the terminator in place.

 

I've always assumed the RA "shorties" shorted the inputs and so didn't have resistors in them, but I don't know. My bet would be no resistor as they'll be cheaper to make though

I tried listening to the 'signal' from unused inputs on my amp via my headphone rig, and basically there wasn't any noise as it stands, apart from what sounded like transformer noise right near the max. gain. Which was nice . So I won't be shorting any inputs just now. Oh, the one input where I did get significant hiss was on the CD input, being thrown out by my DAC (with nothing playing) Understandable though.

 

I don't know where the hiss at near max volume originates from but most of it does go down with the terminator in. If you have spare RCAs why not give it a try by shorting. No harm done.

 

The noise I got at near max vol. on my headphone amp (which is much higher gain than max on my pre I think) is an electrical hum, not a hiss, which I why I think it's a transformer thing. To give you an idea how loud I'm talking, I'd normally listen with my headphone amp set to about 7:30, and there's no noise at all till 3 or 4 o'clock. Also the constant hiss I get from my power amp on it's own is MUCH louder than anything coming from unterminated pre-amp inputs that I could hear with my headphone test, so I really don't think there's any point me bothering to short them. I will though get round to properly terminating the digital outs I have at some point.

 

All except Arcam amps, I have had this problem of sound level being too loud over about 9 O'clock. Sorted that out with DIY attenuators