2
\$\begingroup\$

I am trying to set up lab equipment to measure my soon-to-come sample, but I couldn't get my mind around one thing, impedance termination. I will tell you what I have, what I want to get, and my concerns. My probes are already connected to coaxial wires that are terminated by trixial connectors. However, my oscilloscope/function generator/etc all have BNC connectors. So i need to get a triaxial to BNC adapter. my concern is, do I get a 50 ohm or 75 ohm adapter. What is it based on? The function generator? The coaxial wires I already have? how do I know if it is terminated by a 50 ohm or 75 ohm? (I cannot get in contact with the person who ordered them, nor i have their datasheets, they are just there and I have to figure it out with what I have).

I understand that they nature of my measurements are also important. But I am not asking if I have to change the wires I already have, I want to work with them, and then maybe change them latter to what better fits my measurements. Most of my measurements are DC. so does it matter at this point? I do measure a square wave of about 300 MHz, is this considered high frequency?

\$\endgroup\$
  • 1
    \$\begingroup\$ RF : 50 ohms. Video : 75 ohms. \$\endgroup\$ – Brian Drummond Jul 25 '15 at 21:25
  • \$\begingroup\$ Triax? The only context I've used triax in is low-leakage current/voltage measurements with a SMU. Probes, as in a probe station? \$\endgroup\$ – W5VO Jul 26 '15 at 5:06
  • \$\begingroup\$ Yes, the system is now connected to SMUs. I have some business with it, but after that i need to use the probes with external equipment like function generator and oscilloscope \$\endgroup\$ – himura Jul 26 '15 at 7:50
5
\$\begingroup\$

Most normal BNC terminated lab equipment uses 50 Ohm. 75 Ohm is mainly used for cable tv and related equipment.

Can you read the number printed on the side of the cables? For example RG58is 50 Ohm but RG59 is 75 ohm.

At low frequencies it does not matter although the amplitude may indicate incorrectly if you terminate in a different impedance from which it was designed. For example if you use a function generator that was designed for 50 ohm and terminate it in 75 ohm the amplitude will 20% higher than intended (e.g. if you set it to 1V p-p output you will actually get 1.2V).

What constitutes low frequency depends upon the length of the cable - if the time delay in the cable is more than say 10% of a cycle of the highest frequency component then you may see ringing or other artifacts. The propagation speed in a cable will be about 65% of the speed of light or about 1.5ns per foot. For your 300MHz which has a wavelength of ~2 feet any impedance mismatch of more than a couple of inches could cause problems.

\$\endgroup\$
4
\$\begingroup\$

300 MHz is a high frequency for normal oscilloscopes. This usually requires a good probe, and a scope with a termination in matching impedance, preferably internal, but possibly external, as close as possible to the scope. If you want to see the square wave as a nice square wave you will have to have equipment that has an upper frequency limit of at least 1.5GHz (5 harmonics of a 300MHz sinewave), and if you need to see details of the square wave I would go for 3GHz. (a 4GHz oscilloscope can cost 40 000 USD, it is serious equipment)

The triax cables seem to be for broadcast/telecom, and I searched online and it seems to be 75 ohm only, so the triax cables must be terminated with 75 ohm as far as I know.

To make a proper measurement at 300MHz there should only be one characteristic impedance. only 50 ohm or only 75 ohm. If you combine different transmission line impedances you will degrade the signal and create reflections.

50 ohm is very common, and all non-broadcast coax I've seen has been 50 ohm. If you have to use the 75 ohm cable (which i think the triax is) you should terminate it properly on the scope end, with a 75 ohm termination, and any extra cables/connectors/etc should have that same characteristic impedance.

\$\endgroup\$
  • \$\begingroup\$ Yes, I have a 4GHz Oscilloscope. Thank you for your reply. So if my tirax cable is in fact 75 ohms, then I should just get 75 adapters. But what about the oscilloscope? does it matter? \$\endgroup\$ – himura Jul 26 '15 at 7:54
  • \$\begingroup\$ This is the cable I have: newark.com/trompeter/trc-50-1/triaxial-cable-36awg-50-ohm-per/…. SO I should just get 50 ohms adapters? The oscilloscope/function generator has nothing to do with it? \$\endgroup\$ – himura Jul 26 '15 at 8:00
  • \$\begingroup\$ The important thing is that you have the right characteristic impedance all the way from your device under test (d.u.t.) and a matching termination at the scope end. Your scope should have a 50 ohm termination option, or it should be default. \$\endgroup\$ – user55924 Jul 27 '15 at 6:03
  • \$\begingroup\$ yes it has 50 ohm termination. Thank you fro your input! \$\endgroup\$ – himura Jul 27 '15 at 9:51
1
\$\begingroup\$

I think you've correctly identified that you have 50Ω triax, and I've only seen 50Ω test equipment and hardware. You can probably go ahead and get 50Ω adapters. That being said, let me mention a few gotcha's I've found with triax.

  • There are 2-lug connectors (like regular BNC), and 3-lug connectors.
  • Not all adapters drop the same conductor! I have some adapters that drop the inner guard braid, and some that drop the outer shield braid.
  • Triax on probe stations are often configured for low-frequency and DC measurements (e.g. a SMU). You will probably have little difficulty hooking up a power supply, but getting accurate 300MHz measurements will be difficult depending on what you are trying to measure.
  • The triax probes I have used are all single needle, meaning that only the center conductor is connected to your circuit. If this is the case, you will have a lousy ground connection (and the shield won't be a direct ground. RF probes have 3 contacts in a GND-SIG-GND connection
\$\endgroup\$
  • \$\begingroup\$ Thank you for your reply! very insightful! I have noticed the 20lug/3-lug thing. I want to ask you, it is not possible to simply connect the 2-lug triax connector to a BNC connector, right? As for your second point, I do not really understand what do you mean by dropping the inner and outer guard. And as in your case, yes I believe that my probe station is configured for DC measurements, however I do not get why hooking up a power supply will be difficult. Finally, yes mine are only the center conductor. I was thinking of manually connecting the outer parts with wires to the ground. \$\endgroup\$ – himura Jul 27 '15 at 9:49
  • \$\begingroup\$ I was advised to get an active prove for AC measurements (the 300 MHz). Won't that help in the floating ground issue? \$\endgroup\$ – himura Jul 27 '15 at 9:50
  • \$\begingroup\$ @himura I said you should have little difficulty, meaning it will be easy. My triax is 3 lug, so I can't give direct experience. You might consider trying it since you probably have some regular BNC equipment, but I suspect it will not fit. An active probe is a good idea, especially if it isn't a 50 ohm output, but you will still need to make a decent ground connection. We have used Picoprobe active probes, but primarily for reducing capacitive loading on output pads. \$\endgroup\$ – W5VO Jul 27 '15 at 13:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.