Re: Wrong cable... How bad is it?

Tom Lee

The solution depends on what problem you're trying to solve. If you need to meet the stringent flatness spec of the SG503, then you need the cable (or equivalent). Dennis has published an excellent analysis comparing off-the-shelf cables with the official Tek cable. To meet Tek's spec, you cannot assume that "50 ohm cable" is 50 ohm cable. Most are 52 ohms, nominally, and not controlled to that value with any precision.

The need for the special cable is caused by the SG503's sensing the amplitude at the front panel connector, not at the ultimate point of delivery. As such, any load mismatch will cause some non-unity VSWR and thus, some departure from flatness. Also, cable loss won't be corrected, either, but that's a secondary problem.

If you don't need the factory-level flatness, you can get by with short lengths of regular cable. The degradation in flatness will certainly be measurable, but may be acceptable (again, depending on your use case). Note that the SG503's sensitivity to cable characteristics is shared by virtually every other sig gen on the planet.

The SG504 replaces the need for expensive coax with the need for a special levelling head (which is almost never found accompanying the plug-in -- see the many efforts to homebrew them, such as the nice one by our own David Partridge). The 504 uses a different arrangement where the amplitude  is sensed at the point of delivery, rather than at the generator's front panel connector. If you use no (or negligible) additional cable, the amplitude will be controlled to exquisite precision by the levelling/leveling loop.

-- Cheers

Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070

On 2/8/2021 14:20, G Hopper wrote:
This is an interesting thread since I, like Stephen, have long assumed my
cables were ok. I'm no longer 100% sure and need to check them.

However, what REALLY caught my attention was Jon's comment "For use over 20
MHz eg with SG503 or SG504 or fast pulses like PG506, a calibrated
special 1/2 m Tektronix cable is preffered..."

Having one of those devices, and not having the special cable, what's the

Is the solution to the issue: getting/building a cable (with 50 ohm coax)
of exactly 1/2m length? or is it something else? There must be a solution
other than acquiring an original cable since Tektronix would have had a
specification and had cables made to that spec. I've got to believe that
it's something that mere mortals can reproduce. :-)

If it is, where can one gather the information?


On Mon, Feb 8, 2021 at 2:09 PM Jean-Paul <> wrote:

Stephen welcoming to the wonderful world of wired transmission of Signals,
researched since the earliest days days of Telegraph, telephone and vidéo.

First, câble Zo is important in a matched system eg 50 ohm cal gen and 50
ohm load.
For audio tests with 1 M Ohm , no difference.

As the frequency, and bandwidth go up, and rise times get short, the Zo
is more and more critical to obtaining low reflections, as is the cables
VSWR, attenuation and phase vs legnth and frequency.

Thus RG174/u is 50 ohm, but dia ~ 3 mm so rather high atten. db/ meter,
while RG58/u, is thicker and RG6/u very thick 12 mm so much mess atten per
m thoughts all are 50 Ohm.

For use over 20 MHz eg with SG503 or SG504 or fast pulses like PG506, a
calibrated special 1/2 m Tektronix cable is preffered . Generally the
shorter the cables the better for cal use.
Finally the wire and insulation quality and connectors precision are

We use the Tektronix and Genrad precision cables for critical applications
and avoid the cheap, poorly made Chinese clones.

Suggest that you consult the Belden Wire site papers and applications
notes, as well as the detailed specifications on whatever cables,
connectors and wire you use.

Bon courage


Join to automatically receive all group messages.