Fifty Ohm Probes

[joeqsmith]

Large probe was very poor.  Smaller probe is not the best but much better.  The clamp and needle idea is not so great....

Swept to 2.7GHz.    Looks like the 0603 wedged in there is not so good.  Playing with 1206's the inductance was always a problem. 

BLUE 420MHz was 230 ohms
ORANGE 820MHz was 130 ohms

Transmission, ORANGE 1610MHz.   

[tggzzz]

BLUE 420MHz was 230 ohms
ORANGE 820MHz was 130 ohms

I would interpret that as indicating the tip capacitance is approximately 1.5pF.

For reference, here's a scan of the interesting bits of my hp10020 probe, 10:1, 500 ohms, 1.5GHz, <0.7pF, millimetre scale.

The metal part contains the 450ohm resistor and when inserted into the bayonet only the cable's screw thread is visible. The bayonet makes contact with the wide part of the metal shank (see the diagonal "score" marks) plus the end of metal part's narrow shank.

The metal part is light and there are visible external air gaps inside the plastic part - I interpret this as minimising material with Er>1 so as to minimise tip capacitance.

[Carrington]

In fact some Keysight probes use these non-Inductive resistors (MG660 and MG680): www.sider.com.hk/uploadfiles/files/mg.pdf

Why? Check this out, by Dr. Howard Johnson:

  How to Make A 1K-Ohm Probe: https://www.sigcon.com/Pubs/straight/probes.htm

Others non-inductive resistors:
   
   - https://www.vishay.com/docs/30203/rwrmil.pdf
   - https://www.vishay.com/docs/30204/rsns.pdf

[joeqsmith]

BLUE 420MHz was 230 ohms
ORANGE 820MHz was 130 ohms

I would interpret that as indicating the tip capacitance is approximately 1.5pF.

For reference, here's a scan of the interesting bits of my hp10020 probe, 10:1, 500 ohms, 1.5GHz, <0.7pF, millimetre scale.

The metal part contains the 450ohm resistor and when inserted into the bayonet only the cable's screw thread is visible. The bayonet makes contact with the wide part of the metal shank (see the diagonal "score" marks) plus the end of metal part's narrow shank.

The metal part is light and there are visible external air gaps inside the plastic part - I interpret this as minimising material with Er>1 so as to minimise tip capacitance.

I measured 1.7pf at 410MHz.     Should mention, the waveform is the same in both with same probe.  The reason risetime is so improved was higher BW input.   

I have built some disposable ones using two 1206s (about 1pf ea) in series, but the BW on these were around 1GHz.   Keep in mind that the standard ECL logic will have the following input capacitance per unit load:

MECL 10KH, 2.9pF
MECL 10K, 3.3pF
MECL III, 3.3pF

[tggzzz]

BLUE 420MHz was 230 ohms
ORANGE 820MHz was 130 ohms

I would interpret that as indicating the tip capacitance is approximately 1.5pF.

For reference, here's a scan of the interesting bits of my hp10020 probe, 10:1, 500 ohms, 1.5GHz, <0.7pF, millimetre scale.

The metal part contains the 450ohm resistor and when inserted into the bayonet only the cable's screw thread is visible. The bayonet makes contact with the wide part of the metal shank (see the diagonal "score" marks) plus the end of metal part's narrow shank.

The metal part is light and there are visible external air gaps inside the plastic part - I interpret this as minimising material with Er>1 so as to minimise tip capacitance.

I measured 1.7pf at 410MHz.     Should mention, the waveform is the same in both with same probe.  The reason risetime is so improved was higher BW input.   

I have built some disposable ones using two 1206s (about 1pf ea) in series, but the BW on these were around 1GHz.   Keep in mind that the standard ECL logic will have the following input capacitance per unit load:

MECL 10KH, 2.9pF
MECL 10K, 3.3pF
MECL III, 3.3pF

If n=needle, O=dielectric supporting the needle, r=resistor, then your current probe is effectively
nnnnOOOOOOOOrOOOOOOO
Would it be possible to remove some of the dielectric so that it is more like
nnnnOnnnnnnOrOnnnnnnO

I suspect that the HP10020's construction is more like that, and it might allow you to halve the capacitance associated with the dielectric.

I'm tempted to build one of these http://www.edn.com/design/test-and-measurement/4435414/Sub-picofarad-measurement-with-CMOS-inverters to help measure the tip capacitance.

[joeqsmith]

If n=needle, O=dielectric supporting the needle, r=resistor, then your current probe is effectively
nnnnOOOOOOOOrOOOOOOO
Would it be possible to remove some of the dielectric so that it is more like
nnnnOnnnnnnOrOnnnnnnO

I suspect that the HP10020's construction is more like that, and it might allow you to halve the capacitance associated with the dielectric.

I'm tempted to build one of these http://www.edn.com/design/test-and-measurement/4435414/Sub-picofarad-measurement-with-CMOS-inverters to help measure the tip capacitance.

To some degree, yes it could be made this way but I think the problem is going to be supporting the tip.   

[tggzzz]

You may find this relevant w.r.t. reducing the probe's tip capacitance, particularly figures 6 and 7.
http://m.eet.com/media/1149173/23518-46018.pdf

[grenert]

Here's an old EEVblog thread on making DIY 50 ohm probes:
https://www.eevblog.com/forum/projects/lo-z-probe/