Of what quality does a function generator cable needs to be?

[Rene]

Hello,

I own a Siglent SDG1025 function generator (25MHz max frequency) and would like to buy a couple of BNC cables with different test leads attachments to use with the signal generator. As usual, prices for the BNC cables vary from very cheap to somewhat expensive.

Could someone tell me if the quality of the BNC cable really matters when it comes to function generator BNC cables? I am currently interested in buying a BNC to Mini Test Clip Cables (see attached Picture1).

If quality really matter, does anyone has any recommendation son what cable brand to buy?

Thanks.

[tautech]

If you are not pushing the frequency or drive capabilities to the max I doubt it will matter.

You might find some clues to what others use in:
https://www.eevblog.com/forum/testgear/show-us-your-square-wave/

Fast risetimes are probably the most demanding of signal integrity but for ordinary AWG use I wouldn't bother getting carried away. Something like you pictured should be fine.

[Rene]

Thanks tautech.

Hey, I noticed that there are a bunch of BNC cables that feature asymmetrical test lead length (see attached picture). There you can see that the black test lead is a lot longer than the red.

I am assuming that those type of BNC cables are a very bad idea for function generators running on high frequencies because of all the distortion different cable lengths may introduce.... but I can't help wondering if perhaps I am wrong and maybe those type of cables are actually meant to be used with function generators and offer some kind of benefit?

Thanks.

[Lightages]

With flying lead type cables you will not get accurate waveforms much over a few hundred kHz. You need properly impedance matched cables and termination to keep a proper waveform above that.

The quality of the cable will matter the higher the frequency you use and the longer the cable. Make sure you are using 50ohm cable and have proper 50ohm BNC connectors.

[tggzzz]

The only reasonable answer to your headline question is "they have to be good enough that they do not significantly affect the signal in my application". Hence you need to define your signal and your application.

At 25MHz I suspect that the transmission line qualities of the lead are less important that the shielding, capacitance and source resistance.

[dadler]

I like the stuff from probemaster: http://www.probemaster.com/index.php?cPath=5&osCsid=826ec36111782fa63edd2cdd996469e5

The cables are great. Some of the mini/microgrippers are of lower quality, but I have swapped most of mine out with pomona. Cheap too.

[TimFox]

Waveform reflection occurs at discontinuities between one characteristic impedance (50 ohms for typical coax) and another (higher for uncontrolled flying leads).  These reflections can become visible when twice the electrical length (in nanoseconds) of the cable is a substantial fraction of the rise time of the square wave.  Normal RG-58/U cable (solid polyethylene dielectric) has a velocity equal to 2/3 the speed of light, so 6 feet (180 cm) has an electrical length of 9 nsec.  Reflections can be reduced with appropriate termination.  The 6 inches (or so) of flying leads won't terminate properly, and has an electrical length of roughly 1 nsec.  The 2 inch (or so) mismatch in length between the red and black clip leads is for mechanical convenience when finding a ground point in the circuit:  at high frequencies the area inside the loop can be important.  What rise time is claimed for the square wave, and does the unit have a good square wave at its top frequency?

[tautech]

What rise time is claimed for the square wave, and does the unit have a good square wave at its top frequency?

<12nS
Here is an example I quickly grabbed:
https://www.eevblog.com/forum/testgear/show-us-your-square-wave/msg590757/#msg590757

[krivx]

Op hasn't mentioned square waves. If it's a sinusoidal test signal those leads tend to be fine for at least a few MHz.

[jpb]

As others have said, the most important thing is maintaining 50 ohms as far as possible. So using 50 ohm coax is important but also getting the load as close to 50 ohms as possible is also important. It is at the load end that things will go wrong most often - it is no good having a 50 ohm source connected via a 50 ohm coax to a circuit that looks nothing like 50 ohms.

In a practical sense, a short section to grabbers is probably ok - I use such an arrangement on my scope when testing OCXOs for instance and the results generally seem ok at least at 10MHz or so. Six inches of wire to the grabbers is perhaps longer than it need be, you could perhaps shorten them to a couple of inches.

Though I have quite a few expensive Pomono cables I also bought a cheap BNC to BNC cable and cut it in half and directly solder the cut end to prototype circuits. This works well except that the earth braid is quite thick and puts mechanical stress on the solder joint.

[TimFox]

As others have said, the most important thing is maintaining 50 ohms as far as possible. So using 50 ohm coax is important but also getting the load as close to 50 ohms as possible is also important. It is at the load end that things will go wrong most often - it is no good having a 50 ohm source connected via a 50 ohm coax to a circuit that looks nothing like 50 ohms.

Actually, if the source impedance is accurately 50 ohms, the coax is 50 ohms, and the termination is high impedance, this works reasonably well:  the reflection at the load termination restores the signal to the open-circuit value (twice the voltage into 50 ohms) and the reflection is absorbed in the source impedance and does not re-appear at the load.  This is equivalent to putting the high load impedance directly at the source output connector, but with the delay of the 50 ohm cable.
Double-termination (source and load) is more conservative, but so are suspenders and a belt.