Weird scope effect

[cyberfuzzy]

Hi folks,

As I played around with the scope yesterday, I saw something weird.
The setup:
Waveform generator: sine-wave, 69 MHz, 5 Vpp, 50 ohms output.
1 m unshielded measuring line (not a 50 ohm cable) with crocodile clips at the end.
Two 10:1 probes connected to these crocodile clips.

When I touch the coaxial cable of the scope probes (or when I bring them together) I can see a phase-shift and an amplitude modulation on the scope. Watch the following video clip:



Does anyone know what's going on here? I know, there is no 50 ohm termination, but the potential of both probes should be the same at any time. And the coaxial probe lines are shielded.

@Dave: When you are able to replicate this issue, you could do a nice eevblog video of it, explaining what's going on here.

[PA0PBZ]

GND wire on the probes have different length.

[tom66]

I suspect the coax is not very good - poor shielding - and you are adding capacitance into the system and causing a phase shift.

[alsetalokin4017]

69 MHz isn't chopped liver -- at such high frequencies you need good probing techniques to get good results. Proper termination, proper shielding, matched line lengths, no finger-poking, etc.

[David Hess]

The long ground leads on the probes allows the ground impedance of the probes to change when you manipulate them.  This would not normally be a problem because ground would already be low impedance but in this case the long non-coaxial line from the signal generator allows coupling between the signal and ground.

Another way to think about this is that the unbalanced output from the signal generator is being partially balanced by the twin lead and the resulting ground signal is traveling up the inside of the coaxial shield on the probes *and* the outside where it reaches your hands and reflects.

As an experiment, minimize the ground lead length on the probes by using coaxial probe tip adapters.

[bjcuizon]

I agree with tom66, there is less or bad shielding on the coax. And that you are adding capacitance by touching it as you are also coupled to ground. Are those good quality probes? You might also want to watch this video about the effects of capacitance from the two minute guru:

Hope that helps.

[cyberfuzzy]

Thanks for your answers. I'm going to try it again with a shielded RG-58 from the wavegen.

Are those good quality probes?

I hope so. These are the standard 300 MHz probes (Rohde & Schwarz RT-ZP03) which are delivered with the scope (Rohde & Schwarz HMO724).

[bjcuizon]

Thanks for your answers. I'm going to try it again with a shielded RG-58 from the wavegen.

...And don't forget to terminate the coax. 

[Mr. Scram]

Hi folks,

As I played around with the scope yesterday, I saw something weird.
The setup:
Waveform generator: sine-wave, 69 MHz, 5 Vpp, 50 ohms output.
1 m unshielded measuring line (not a 50 ohm cable) with crocodile clips at the end.
Two 10:1 probes connected to these crocodile clips.

When I touch the coaxial cable of the scope probes (or when I bring them together) I can see a phase-shift and an amplitude modulation on the scope. Watch the following video clip:



Does anyone know what's going on here? I know, there is no 50 ohm termination, but the potential of both probes should be the same at any time. And the coaxial probe lines are shielded.

@Dave: When you are able to replicate this issue, you could do a nice eevblog video of it, explaining what's going on here.

What's the music?

[exe]

I sort of have the similar question. I noticed significant cross-talking on my micsig TO1104.
I was very upset about this, but after connecting my signal generator via a BNC cable it all was gone...

So, I think there is a big trap for young players: (cheap?) probes interfere a lot if their cables put together.

[Hydron]

These are not particularly cheap probes (R&S default offering by the look of it, same OEM also used by LeCroy and Pico); I'd lean more toward's David Hess's theory.

(unless if by cheap you mean the initial cable from the sig-gen, which is clearly inadequate)

[Damianos]

- Signal source of 69 MHz (wavelength a little more than 4 m), with unbalanced output.
- A balanced transmission line of 1 m (near a quarter of the wavelength) with unknown characteristic impedance, velocity factor or any other characteristic
- this connected to two high impedance cables with shield(!)
What we expect as a result?

Before we are blaming the cable's shields or the ground wires, let's start to think about standing waves and currents on the shields...
A search for "transmission lines" will shed some light.

However it is a chance for experiments!
Keep the wiring fixed and untouched, vary slowly the frequency and keep notes for the signals amplitude. Calculate the wavelength of the frequencies where you find maximums and minimums. Find the relation with the lengths of the wires. This may need to connect a resistor of a few hundreds Ohms at the end of what you call "measurement line".
Another "game" is to move a metallic thing (like a screwdriver) along the length of the wires, by observing the signal changes. This maybe is not much sensitive to give clear results.

Edit: removed blank space

[bjcuizon]

Dave and Alan already have a video about terminations and standing waves. Here it is:

[cyberfuzzy]

I continued playing around and I found the problem.

First I tried a shielded 50 ohms line from the waveform generator. No effect, same problem.

Then I connected the scope probes directly into the BNC connector of the waveform generator (60 MHz this time). Channel 1 of the generator was turned on, channel 2 was left off. Same problem. I tried this with four different scopes, many different probes and two different waveform generators:





So I thought it could only be the GND clip. For the next test I removed the GND clips of the probes. I put some aluminium foil into the BNC connector to connect the GND directly to the probes:



Now there was no way to disturb the transmission:

What's the music?

Sorry for the music. In my first video I wanted to remove the sound, because colleagues were talking. I couldn't find the function to just mute it. But I could choose a song for the video on youtube.

[kg4arn]

I agree with David Hess.  This is common mode on the probe shields.
I think the old Tek app notes recommend handling the probe to test for common mode.

[Hydron]

Probe to BNC adapters are your friend
e.g.

The R&S probes are likely compatible with this kit (made by the same OEM), which is an absolute bargain if you're within reach of RS:
http://uk.rs-online.com/web/p/oscilloscope-kits/8014925/

[exe]

£5.20 Price Each (I think VAT not included)... Damn, my probes are cheaper then that (PP-150). Anyway, RS Online does not work with individuals in Netherlands (neither Farnell, their local "partner" sinuss.nl is a rip-off).

May be there is something similar at tme.eu? Could someone point me out which of these should I buy (if any)? http://www.tme.eu/gb/katalog/#id_category=100365&search=bnc+tip&page=1&s_field=accuracy&s_order=DESC .

[Hydron]

For an accessory to a name brand probe 5.20 GBP (plus VAT) is indeed a bargain. The BNC adapters are not always compatible between brands unfortunately, but the ground clips often are (and name brand replacements of these simple items can be costly!).

For the OP with the R&S scope/probes this set also gives you a couple of spare tips as well as the adapter and ground clip so is well worth having even just as a spare. (I have a set of RT-ZP03 probes and this set is 100% compatible, though the hook clip thing is a slightly different style).

[David Hess]

I agree with David Hess.  This is common mode on the probe shields.
I think the old Tek app notes recommend handling the probe to test for common mode.

I think Jim Williams mentioned this someplace as well and it is the standard test I do when I suspect something is funny.  This is the same test I use when designing coaxial fed antennas to look for shield currents.

If a differential probe is not available, then sometimes this problem is amendable to wrapping the coaxial probe lead near the probe body through a common mode choke.

Probe to BNC adapters are your friend

These are very useful but most of the interesting devices do not have a BNC to plug them into.  What I do is solder a short coaxial pigtail with a BNC on one end into the circuit so the probe can then be connected without using the ground lead.  This defeats the purpose of using a low capacitance probe but works fine with low source impedances that will have no trouble driving the added capacitance.

It is too bad that printed circuit board to probe tip adapters are neither affordable nor available.  If I cannot use the coaxial pigtail described above, then I end up making a spring tip adapter out of bus bar wire as shown.