10:1 attenuation probe with SMA connector (instead of probe tip)

[Sparky]

Hi friends,

I'm measuing a signal exposed on an SMA connector with a 10:1 attenuation probe (for reduced loading of the DUT). My current setup (see photo attached) uses Keysight 2873A (500MHz BW probe) with a probe-tip to BNC adapter and then a BNC to SMA adapter.

The setup works but isn't as robust as I'd like.

I searched for a better solution and the closest I've found is this "BNC To BNC 10X Attenuation Cable":
https://probemaster.com/6143-bnc-to-bnc-10x-attenuation-cable/

This looks like it would work though I still need to use an adapter.

I'm wondering how difficult modifying an existing Keysight probe would be:
- Use an existing Keysight 10:1 probe so the probe attenuation detection and probe compensation works
- Cut off the probe completely!
- Solder/crimp an SMA connector to the probe cable

I'm unsure what trouble I'm getting myself into by removing the probe itself; I guess inside the probe handle is the 9Meg ohm resistor in parallel with a capacitor and I might need to replicate that...

Am I overlooking simpler ways to achieve what I want? If I use a 20dB attenuator with BNC connectors I'll loose the convenience of the probe detection by the scope -- trying to keep that -- and also I don't know how probe compenation is performed in such a case.

Thanks for any insights!

[joeqsmith]

Assuming the boards output is 50 ohm, I would just use SMA to SMA cable and set the DSO's input to 50 ohms. 

[TheoB]

You mention loading is an issue, so you need a X10 probe. That still has 15pF as load. Above 20MHz you can better use a resistive divider into 50 Ohm coax cable. That keeps a constant 500 Ohm load up to high frequencies. But you have to manually set the attenuator factor in the scope.

[joeqsmith]

I assumed when they brought up the 10X probe, they were not aware of the loading effects these types of probes have.    Still if its a 50 ohm source, using 50 ohms all the way though seems like the logical choice. 

I have not seen an SMA style resistive probe before.   Do they make such a product?

[Sparky]

hi @joeqsmith and @TheoB -- thanks for replying with thoughts and insights! Much appreciated! Let me update with some further details.

Assuming the boards output is 50 ohm, I would just use SMA to SMA cable and set the DSO's input to 50 ohms. 

Unfortunately the output of the DUT is not designed for a 50ohm load.

And apologies: I should have been explicit about the input termination.
- The input is 1Mohm // 13pF (Keysight CX1151A Passive Probe Interface Adapter)

You mention loading is an issue, so you need a X10 probe. That still has 15pF as load...

Yes, I want the DUT to see a high impedance so I thought to use a 10:1 probe instead of 1:1 probe, but I did not consider the capacitive load that is there.

To add more detail: I'm not selecting a 10:1 probe for the voltage attenuation; the voltage being measured will be <10V. A 1:1 probe would be better for less noise, but such probes have considerably less bandwidth.

...Above 20MHz you can better use a resistive divider into 50 Ohm coax cable. That keeps a constant 500 Ohm load up to high frequencies. But you have to manually set the attenuator factor in the scope.

This sounds interesting as a solution! 500MHz bandwidth would be sufficient.

The attenuation can be configured.

Given the input termination (1Mohm) I could make a 2:1 cable from coax with a 1M resistor in series with the "probe tip" (SMA connector).

How to address probe compenation?

I have not seen an SMA style resistive probe before.   Do they make such a product?

Neither have I; it does not seem to be a thing.

Ultimately I'd like a solution that's more robust than the slide on adapters; avoid the probe tip being damaged and less clumsy than multiple adapters.

It sounds like a custom cable from coax might be the way to go.

[TheoB]

What I meant with a resistive divider is that the scope is 50 Ohm (add a 50 Ohm termination if the scope does not have that) and feed in with a 50 Ohm coax. At the dut side inject into the 50 Ohm coax with 450 Ohm. The dut now sees a constant 500 load and the scope sees a 50/500 part of the signal (10x attenuation). If your dut cannot handle 500 Ohm at DC, add a blocking capacitor (450 Ohm + 100nF gives 3.2kHz high pass). At the dut side you can also implement a source follower to make an impedance transformation from high to 50 Ohm. I searched for the probe adapter and found that you probably use it with a "Device Current Waveform Analyzer". If so are you measuring current by a small burden voltage over a shunt resistor?

[joeqsmith]

Maybe you can modify the board to place an 0603/0402 in series with the SMA.  Then use coax back to the scope.   Odd they would use a 50 ohm connector but not provide a proper driver for it.   What is the source?

If you are looking for a low cost active probe or details on typical resistive probes, try here: 
https://www.eevblog.com/forum/testgear/12-ghz-active-probe-project/msg5006290/#msg5006290

[Sparky]

What I meant with a resistive divider is that the scope is 50 Ohm (add a 50 Ohm termination if the scope does not have that) and feed in with a 50 Ohm coax. At the dut side inject into the 50 Ohm coax with 450 Ohm. The dut now sees a constant 500 load and the scope sees a 50/500 part of the signal (10x attenuation). If your dut cannot handle 500 Ohm at DC, add a blocking capacitor (450 Ohm + 100nF gives 3.2kHz high pass). At the dut side you can also implement a source follower to make an impedance transformation from high to 50 Ohm. I searched for the probe adapter and found that you probably use it with a "Device Current Waveform Analyzer". If so are you measuring current by a small burden voltage over a shunt resistor?

Thanks for the details of the resistor divider idea -- now I understand it better. The 500ohm load would still be too high, though.

You are correct -- the probe adapter is used with a CX3300 Device Current Waveform Analyzer, and various current sensor heads having bandwidths up to 200MHz; the effective bandwidth of the system can be around 140MHz. The CX1151A probe adapter (bandwidth 300MHz) allows to measure voltage on a separate channel. The recommended probe to use with the CX1151A is N2843A 500MHz 10:1 passive probe. I attached a picture of a typical setup with sensor head showing "voltage monitoring" output (the SMA connector I attach to). With the setup I can measure voltage and current simultaneously.

My goal is to improve performance and robustness of the voltage monitoring setup by replacing the probe assembly (N2873A probe I have on hand + multiple (slip on) adapters) with a single cable solution: a BNC connector on one end, SMA on the other end, a bandwidth 300-500MHz, minimal loading of the DUT (a low-power low voltage battery operated device), and compatible with the CX1151A interface (1Mohm // 13pF).

Maybe you can modify the board to place an 0603/0402 in series with the SMA.  Then use coax back to the scope.   Odd they would use a 50 ohm connector but not provide a proper driver for it.   What is the source?

If you are looking for a low cost active probe or details on typical resistive probes, try here:
https://www.eevblog.com/forum/testgear/12-ghz-active-probe-project/msg5006290/#msg5006290

Perhaps a single piece of coax with BNC and SMA connectors on it will be fine. But, is it OK to ignore probe skew and not have any compensation? I'm not sure if I'm over thinking it...

Thanks for the link -- I'll have to read up on more details! It's interesting this somewhat simple task has a lot of details

[tdunning]

It's a bit late to reply, but I just designed what you are looking for:

https://github.com/tdunning/sma-scope-probe

This repo has PCB layouts and a link to a parts list.