Impedances and transmission line effects...

[pwdrive]

S is the ending end: sending end impedance 600 ohms.  R is the receiving end, terminating impedance 600 ohms.

Nothing at L/SHE except a frame to join the cables S -> L/SHE and L/SHE -> R.  L/SHE -> E is the stub, added overnight by accident.

There is an equaliser at R which contains reactive components, but such that there is no in-band reactive component to the impedance "seen" by the line from L/SHE.

If you can explain the observed effects without invoking transmission line concepts, I'd be fascinated to know. Much cleverer people than me have tried unsuccessfully (doesn't mean it can't be done, of course).

Maybe yourself or Mark can explain it to me using transmission line concepts, I'm not 100% up to speed with transmission line concepts but I would be interested in learning more.

Lawrence.

[Diabolical Artificer]

Blimey, opened up a can of worms with my question. : ) I've followed most of it, though transmission line theory is a bit heavy going.If I have it right though in answer to my original query, I can use any combination of BNC test leads for AF use as long as they don't have old style thick pin types and the length of the cable isn't too long.

The only longish cable's I have for use at home are about 2M long and are of decent quality Klotz balanced twisted pair. I havn't experienced any HF roll off or other issues, but would'nt expect to.

I also have some 10M mic cables which I havn't used yet, so I'll keep reading this thread with interest. Thanks for all you input.

Andy.

[ppppenguin]

50m+ of cable at audio will never cause any problems due to transmission line effects.

I use a lot of BNC cablles etc for video. I don't worry about BNC mating problems. If I get a bad contact I simply scrap the offending connector or cable.

[Terry]

 I don't worry about BNC mating problems. If I get a bad contact I simply scrap the offending connector or cable.

Yes, Jeffrey, but as you denied all knowledge of the old thick pin/thin pin connectors I've mentioned here and elsewhere in the past, presumably all of your BNC connectors are of the modern mix'n'match variety ...

But then, if you are using good quality connectors, bad contacts shouldn't happen, surely ...? 

[ppppenguin]

My BNCs are a wild and wonderful assortment of all ages. Mostly decent professional grade parts but a good scattering of cheap/nasty. Every now and then a crimped plug falls off the cable, means either that it wasn't done right in the first place or has been heavily abused over the years.

[Mark Hennessy]

Blimey, opened up a can of worms with my question. : ) I've followed most of it, though transmission line theory is a bit heavy going.If I have it right though in answer to my original query, I can use any combination of BNC test leads for AF use as long as they don't have old style thick pin types and the length of the cable isn't too long.

The only longish cable's I have for use at home are about 2M long and are of decent quality Klotz balanced twisted pair. I havn't experienced any HF roll off or other issues, but would'nt expect to.

I also have some 10M mic cables which I havn't used yet, so I'll keep reading this thread with interest. Thanks for all you input.

Andy.

Hi Andy,

Post #2 says it all really. Don't worry about cable type or transmission line theory - it's not an issue at audio frequencies in a workshop. But do watch for loading effects

I'll say a few more things about it, as it's the biggest problem I see every day in my job:

In a workshop environment, the problem occurs most often when a signal source of some type includes a level meter (or some other form of output level indication, such as markings on a dial).  Usually, these readouts are calibrated on the assumption the output is correctly loaded, so the output voltage will be higher than the meter says if you don't.

So if you don't take the time to measure the actual output with a 'scope, then you could be somewhat confused! Typically, if you're measuring the gain of an amplifier, it will appear to have more gain than you expect, but that's only because the signal level is higher than the meter on the generator suggests.

The function generators I use most at home are uncalibrated - just a simple rotary control - so I know that I have to measure their output directly. I actually prefer that.

Also, my preferred generators are 50 ohms, so loading effects are less of an issue for typical audio applications. So if they did have an accurate meter, the error between reported and actual would be minimal to insignificant. For example, 50 ohms feeding 10k amounts to a loss of 0.04dB. But 600 ohms feeding 10k is half a dB, which might be significant.

Today, a lot of generators are microprocessor based, so you adjust the output level by typing in numbers or turning a rotary encoder, reading the value from the screen. So again, what do these numbers mean? If you ask for 1V, that might be 1V before the output impedance (a.k.a. "EMF"), or 1V at the output terminals, when correctly terminated. Again, just to reiterate, if it's the former case, then for low-Z out going into hi-Z in, these numbers are virtually identical in most situations. But if it's the latter, the actual output could be double the expected value.

Believe me, this subject catches out a lot of people in many walks of life.

This is a good overview of the subject: https://www.youtube.com/watch?v=tClE8s6RZdg

That video mentions 'scope probes as another "trap". And another thing to watch for is RMS verses peak to peak. Signal generators tend to quote volts RMS, but function generators usually quote peak to peak. Of course, there are variations. Some allow you to change the readout type; many don't. Obviously, on an analogue 'scope, pk-pk is what we measure. A digital 'scope will offer measurement options for either (and a lot more besides). A DMM might measure RMS, in which case it'll usually say "True RMS" in big letters on the front. Or it might be "Average-dection, RMS reading", which is fine for measuring sine waves, but non-sinusoidal waveforms will give different numbers compared to a "True RMS" meter. And DMMs have much less bandwidth than a 'scope*, so knowing the upper limit of your meter is really important.

In fact, knowing your gear is the most important rule in any workshop

* Just in case someone wants to pipe up at the back, yes, there are exceptions. E.g., I have an old Fluke 8920A which has a bandwidth of 20MHz, and 'scopes with less bandwidth exist.

Finally, if you're curious about learning more about transmission line theory, this video is another good practical demonstration of what can happen: https://www.youtube.com/watch?v=PPgxFd97taY

Hope this helps,

Mark

[ppppenguin]

Mark's cautionary tales should serve as a warning to us all when we're trying to measure something. From the simple EMF vs PD problems he's described to a systematic 0.05dB (yes, a twentieth of a dB and it really mattered) error between a client's measurement and mine on professional video equipment. Took a lot of work to troubleshoot that one. It's relevant here because it turned out to be a return loss problem on some (very expensive) testgear. Fixed by a 6dB pad. But it meant that my client had been sending out all their video routers with a systematic 0.05dB gain error for some years. Since the published spec was for 0.1dB error max, this was using up a lot of the tolerance.

[Diabolical Artificer]

Thanks for the links and info Mark. My sig gen an old Advance H1 doesn't have a meter and the volts pot is ish, so I always measure the input to an amp say on the input RCA with a Fluke DMM. Same goes for measuring OP power.

I'm lucky to have an old valve B&K AC millivolt meter that measures in RMS, Peak, P-P, Average and dB's but to be honest I rarely need complete precision but it's good to know how to obtain accurate measurements.

Andy.