Cables connecting Generator to Oscilloscope BNC male to BNC male

[NHSA]

Hi,

Beginner here.

Is there a length limitation in regard to shortness of this cables ? e.g. do they need to have a certain minimum length ? Can I make then for example just 1 feet long ?

Can I use RG-174 for this cables ?

Thanks in advance for any tips.

regards Rainer

[NHSA]

Hi,

Question answered. Made a short cale ~ 1 feet and compared it with the one which was delivered with the Wave generator and saw no difference.

 

Rainer

[bob91343]

Assuming you terminate the cable in 50 Ohms, the only thing that changes with length is the delay and the amplitude.  The delay may be computed from the specs, usually somewhere around a nanosecond per foot, and the amplitude will change based on the listed attenuation per 100 feet in the cable specs.

RG-174 is fine.

[NHSA]

Assuming you terminate the cable in 50 Ohms, the only thing that changes with length is the delay and the amplitude.  The delay may be computed from the specs, usually somewhere around a nanosecond per foot, and the amplitude will change based on the listed attenuation per 100 feet in the cable specs.

RG-174 is fine.

Hi Bob,

Thank you very much. Your answer made me compare again the delivered cable and my two cables.

The delivered cable is a RG-58 an 104 cm long and my cable RG-174 is 36.5 cm long.

Comparing two equal Sine waves at 25 MHz I got a difference in timing of about 1.65 ns and comparing two of my cables I got 750 ps seconds difference and changing the Phase this is 7° in my cables.

This is when I connect output 1 from AWG to Input 1 and output 2 to input 2 but when I connect output 1 from AWG to input 2 and output 2 to input 1 I have no timing differnece at all   

Very interesting

Again Thanks

Rainer

[Damianos]

... ...
The delivered cable is a RG-58 an 104 cm long and my cable RG-174 is 36.5 cm long.

Comparing two equal Sine waves at 25 MHz I got a difference in timing of about 1.65 ns and comparing two of my cables I got 750 ps seconds difference and changing the Phase this is 7° in my cables.

This is when I connect output 1 from AWG to Input 1 and output 2 to input 2 but when I connect output 1 from AWG to input 2 and output 2 to input 1 I have no timing differnece at all   

Very interesting

Again Thanks

Rainer

The final result means that there is some error in the test.
Are the two channels of the generator synchronised? There may be a setting to do that.
Is the same valid for the oscilloscope? There may also be an adjustment for the time difference, between the channels.

For the generator side it's possible to use a tee adaptor to have the same signal, from one channel, if the cables are properly terminated at  the other end.

[ahbushnell]

The delay may be computed from the specs, usually somewhere around a nanosecond per foot, and the amplitude will change based on the listed attenuation per 100 feet in the cable specs.

RG-174 is fine.

Cable delay is more like 1.5 ns per foot.  Free space the speed of light is about 1 ns/ft.  Attenuation depends on the frequency.  Rise time will degrade with cable length.

[NHSA]

Hi Damianos and ahbushnell,

Thanks for your comments.

Yes there seem to be adjustments. As I said I am beginner and starting to learn the huge amount of settings an oscilloscope has as well as the wave generator.

Yesterday I found a so called " skew " with which I was able to also measure the difference which is 0.78 ns.

About sinchronyzing the channels there seems to be something. Need to play with that and as I said the possibilities are overwhelming. Lots of things to learn.

The oscilloscope s a Siglent SDS 1102CML+ and the wave generator is a Siglent SDG 2042X

Thanks and regards Rainer   

[jonpaul]

Bonjour Ranier NHSA:

We have studied cables and transmission for AES and SMPTE and presented papers and conferences in 1990s till 2014.

Differences are attenuation/ft, phase response, curve of atten vs freq, self capacitance.

All papramters are per uni distance eg db/100 ft or pF per ft.

All the data is available at the cable manufacturers website, eg

https://www.belden.com/products/cable/coax-triax-cable/50-ohm-coax-cable

In general the attenuation increases and capacitance/ft increase as diameter is reduced, eg sequence RG-6/U>>RG59/U>>RG198/U

Cable delay depends on the dielectric constant of the insulator.

 cables with teflon or polyethelene, the parameter is 0.75..0.85

https://www.electronics-notes.com/articles/antennas-propagation/rf-feeders-transmission-lines/coaxial-cable-velocity-factor.php

For short cables in the lab the difference will be th frequency response and fast edges ....
eg a cheap Chinese cable may be much worse that a fine Pomona and the Tektronix or HP CAL cables much better.

Usual lengths are 0.5 and 1m
As the cables are so easy to buy, we seldom make them up.
Termination for best frequency response requires special connectors and crimp tools.
A fine example is Belden 1694A (75 Ohm) which goes to 6 GHz.

So it depends on what you are oding, eg looking at audio 20- kHz it may not matter, calibrating of a 1 GHz scope transits big difference.

Indeed for pS pulsers NO cables are used, and the pulse gen (Leo Bodnar) has a BNC (Female) direct to the scope BNC.

By the way, the connector on the scope is a BNC MALE and the cable usually has BNC female.

Its the bayonet shell and not the center conductor pin that determines the connector type.

Bon Chance,

Jon

[tautech]

Hi Damianos and ahbushnell,

Thanks for your comments.

Yes there seem to be adjustments. As I said I am beginner and starting to learn the huge amount of settings an oscilloscope has as well as the wave generator.

Yesterday I found a so called " skew " with which I was able to also measure the difference which is 0.78 ns.

About sinchronyzing the channels there seems to be something. Need to play with that and as I said the possibilities are overwhelming. Lots of things to learn.

The oscilloscope s a Siglent SDS 1102CML+ and the wave generator is a Siglent SDG 2042X

Thanks and regards Rainer   

Signal propagation delay is very relevant when using different cables or probes and quite noticeable when using voltage and current probes hence the deskew feature to null propagation delays from measurements.

If say for example you were working with custom flyback coils and MOSFET ON/OFF timing to maximize output without coil saturation you would need to null any propagation delays from your measurement system to obtain the most accurate results.
Deskew fixtures are perfect for this and also help to set correct attenuation for current probes and this is commonly done with scopes that have Power Analysis functionality so that all measurements related to current vs voltage timing can be precise for accurate measurement results.

[David Hess]

I usually make my cables 1, 2, and 5 nanoseconds long (8.2 inches per nanosecond with Teflon) so I know how much skew is produced and I can compensate it easily if necessary.

[jonpaul]

Check the specs of the Sigelent equipment, the scope and gen may have more internal skew and jitter than you think.

Best test is the 40 pS Leo Bodnar pulser.

J

[NHSA]

I usually make my cables 1, 2, and 5 nanoseconds long (8.2 inches per nanosecond with Teflon) so I know how much skew is produced and I can compensate it easily if necessary.

Hi David,

That is interesting. From where to where is that measurement ?

Connector tip to connector tip ... cable length when cutting ... shield length after peeling off the outer shielding insulation ... length of internal center wire insulation ...

Yesterday I measured my skew again and it is stable 780 ps or 0.78 ns. Do not remember if both cables are the same. Will test it today   

regards rainer

[jonpaul]

Rainier rebonjour,

The connectors and ends add little to the delay.

Its completely the cable length and velocity factor.

Normally it is measured with connectors in place.

Beware that the accuracy and resolution of such a time measurement is dependent on the risetime, jitter   of the gen and scope.

Jon

[NHSA]

Rainier rebonjour,

The connectors and ends add little to the delay.

Its completely the cable length and velocity factor.

Normally it is measured with connectors in place.

Beware that the accuracy and resolution of such a time measurement is dependent on the risetime, jitter   of the gen and scope.

Jon

Hi Jon,

Thank you. Just made a test with my AWG by changing the cables position and measuring the skew and it remains at 0.6 ns and this seems to be a time delay of the AWG. Both channels of the AWG are Phase locked.

My cables have a length difference of 3mm.   

Interesting   

Rainer

[bob91343]

While I feel I know the basics of transmission lines, I found that I learned a lot from my nanoVNA.  I watch the display change as I modify the cables for length, impedance, discontinuities, terminations, and more.  The Smith chart display is a wonderful learning tool.  A wide frequency spectrum is visible all at once, and you can even see the effect of adapters.  For only $50 it has been a marvelous investment in my education.

[rf-loop]

I usually make my cables 1, 2, and 5 nanoseconds long (8.2 inches per nanosecond with Teflon) so I know how much skew is produced and I can compensate it easily if necessary.

Hi David,

That is interesting. From where to where is that measurement ?

Connector tip to connector tip ... cable length when cutting ... shield length after peeling off the outer shielding insulation ... length of internal center wire insulation ...

Yesterday I measured my skew again and it is stable 780 ps or 0.78 ns. Do not remember if both cables are the same. Will test it today   

regards rainer

Cable delay is naturally from signal pathway length aka contact point to contact point. If measure example length fron BNC tip end to BNC tip end then there is "big" nearly 20ps error.

User can also easy demonstrate it. Connect BNC so that just contact and can see signal, look delay to reference (take example long average from  edge to edge delay... after then slowly push connector fully in and  there can see roughly 9ps difference (+/- something depending connectors quality etc)
(if user have instruments where this kind of time delays can detect and measure.
SDS1000CML+ is not suitable for this - at all. All is there far below noise and jitter and resolution, if I remember right max res on display is 50ps(?) (one horizontal pixel with 2.5ns/div) also it have old school analog pathway trigger system with higher trigger than more modern technology ones)
SDS1000CML+  specified skev between channels is <500ps. )


Just for example, measured Siglent SDG6000X channels time skev between channels outputs connectors.
Using 120MHz sqr and 500MHz sine. In these cases under 25ps including all possible estimated max errors and inaccuracies sum in test. (all my measured skev under 15ps)
Because I do not have perfect matched cables naturally I made all combinations cross check and calculus also for eliminate scope skev between channels and cable delay differences. Naturally we all know also that BNC connectors are just horrible shit if want really perfect and stable and repeatable signal travel time when we start talk about pico seconds. Just example semicrap RG316/U what I used in this time have nominal time 4.856ps/mm and also it is some variations. Now think BNC connector... where is point where signal connects, in pin and in GND... after then, just keep fun. Of course if we talk only nano seconds then problem with BNC and can just forget.
Naturally SDG6000X is not SDG2000X and at this time I can not check it.




 

[David Hess]

I usually make my cables 1, 2, and 5 nanoseconds long (8.2 inches per nanosecond with Teflon) so I know how much skew is produced and I can compensate it easily if necessary.

That is interesting. From where to where is that measurement?

Connector tip to connector tip ... cable length when cutting ... shield length after peeling off the outer shielding insulation ... length of internal center wire insulation ...

The acid test is when I add the cable to an existing line, how much is the delay increased?  So in practice that has to include the delay of the female-to-female adapter but measuring from connector end to connector end is close enough.

Even a 100 MHz oscilloscope can see 1, 2, or 5 nanoseconds of delay, but cannot see the difference between my various "2 nanosecond" cables.  My sampling oscilloscope could of course.

[tkamiya]

For me, it really depends on why I'm making the connection.  I may just use random length of RG58 and such or use phase stable cable at specific length.  I don't like RG174.  It's too fragile.

[David Hess]

For me, it really depends on why I'm making the connection.  I may just use random length of RG58 and such or use phase stable cable at specific length.  I don't like RG174.  It's too fragile.

I agree about RG-174 being awfully fragile.  RG-316 which is the Teflon version is somewhat better.  It also helps to use crimp connectors instead of clamp connectors.