Coax (BNC) cable impedance question - for level not clock

Blue Jinn

Blue Jinn

Rider of the ARPocalypse
I have a signal generator with BNC outputs, and a couple of AC voltmeters with only BNC connectors. I've a mix of RG58 (53.5Ω), RG59(73Ω), and might even have RG62 (93Ω) . Does it matter?

I've no idea what the output impedance of the signal generator a Wavetek 186. I've managed to hook it up to both a Fluke DVM and the HP, (they're not tube but solid state) to get a .316v sine wave, and it seems as long as the voltage of the sine wave at various freq's is correct that would be the the focus. I've a mix of RG58 (53.5Ω), RG59(73Ω), and might even have RG62 (93Ω). Does it matter for this purpose?

NOTE: My understanding is for clock & SP/DIF you need 75Ω. Second question, is RG59 close enough?
 
In radio frequency circuits, the correct impedance is vital, but in audio frequency use, im not sure any of my test gear could detect the difference. In digital audio, the key feature is not voltage, its not rounding off the edges of square waves. I would guess your generator is 50 ohm output, most are. I would suspect that unless you need the sort of accuracy that means your meters go back twice a year to the manufacturer for recalibration you just wont notice anything if the interconnect has DC continuity. A lot of test generators have banana plug outputs and bnc converters!
 
It's the wavelength innit! Given by, in theory c/f where c is speed of light and f the highest frequency of interest. So for 20kHz that is 15,000 mtrs. Signals* don't quite go at c in cables and it depends on the cable but you can see that 20kHz is going to be well over 10km!

At audio frequencies then the impedance of a cable does not matter for other than very large distances. For video and S/PDIF it can. Both should use the correct connectors and cable, usually 75R and be terminated at the end of the line. Otherwise video gets "reflections" repeat vertical lines and S/PDIF can drop out and crackle. (in practice I have never had a problem with S/PDIF using RCA plugs which are not 75 Ohms)

But, some signal generators, old ones mainly, have an output attenuator designed to deliver a constant, 600 Ohm source impedance and that will only deliver the correct voltage into 600 Ohms. Best to check.

We are blessed now with super accurate digital AC meters for peanuts but, most have a limited frequency response. Many are a dB or two down at 2(yes TWO!) kHz an some as low as 1kHz. Even the "true rms" types cannot be trusted past 400Hz or so. A spot check at 100Hz and 10kHz will tell you the unvarnished.

*Note, SIGNALS, it is the WAVE that travels in the cable not "electricity". You can learn more by studying Mr Maxwell's famous equations but don't ask ME questions...WELL past my pay grade!

Dave.
 
Blue Jinn said:
NOTE: My understanding is for clock & SP/DIF you need 75Ω. Second question, is RG59 close enough?
Click to expand...
First question: RTFM. See what cable is recommended for your equipment.
Second question: Yes, it matters. Close enough is not good enough. Not so much the impedance of the cable, but the termination from point to point.
--But, some signal generators, old ones mainly, have an output attenuator designed to deliver a constant, 600 Ohm source impedance and that will only deliver the correct voltage into 600 Ohms. Best to check-- @ecc83 has it right. The "impedance" needs to be equal at both ends.

Further, the "x ohm" label has little to do with the resistance of the cable. It has more to do with the "impedance" of the cable. If you look at the differences in coax cable(s), look at the ratio of the conductor to the dielectric. That is what really affects the signal across the line.

With respect to a "word clock", the purpose of the word clock is to ensure that the A/D and D/A converters clock at the exact same time. If you're deep enough into that level of sync, you need to have equals on both ends.

IMHO: Using a meter is not the right tool. You should really put a scope on it and measure the leading adges of both clocks and check that they are indeed synced, in-phase, and not noisy. .
 
In my experience, RG 59 is 75 ohms and is what most plumbling is done with. And I do it for a living
 
@Track Rat is right on. Just don't confuse it with RG-58, which is 50 ohms.
@Slouching Raymond is also correct, for short distances.
 
I have been smiling a little on this topic. Signal generator, audio voltmeter, lots of discussion over different specifications of coaxial cable. Then i look at the huge pile of audio cables in the studio that i have not put away. Apart from one that is turquoise, which I know the impedance of, the rest are going to be totally random impedance, connected and selected because they are short, or long, or just different colour from another. They all work totally invisibly. We have on screen EQ curves and displays that show exactly what the frequency response and levels are. I am pretty certain nobody has a cable they use that sounds different, because if you had one, it is faulty, or deliberately made to be a filter. We also know that high impedance sources and long cables creates a great filter. If you are making a test rig, you just have to remember the basic science that could impinge on your results, and factor them in. On your workbench, individual cables with banana plugs would work just as well. If your source is 20m away from your measurements, then you consider what the interconnect is also doing, as it becomes part of the ‘equipment’ under test. Like doing PAT tests with long power cables, the fail comes from the total things being tested. The cable resistance gets identified as a fault, when it isn’t. You adjust the test result based on your knowledge.
 
You must log in or register to reply here.
Back
Top