Is agilent 1157A active probe is good for high speed digital signals ?

[electronic_guy]

Hi,

Does anyone of here have used agilent 1157A probe ? Is it a good probe to use for high speed signal integrity measurements ? Anything I should be aware of when choosing an active probe ?

Are there any better options that can be used instead of Agilent 1157A ? ( I need more that 2 GHz bandwidth). I'm choosing it for an Agilent Infiniium oscilloscope.

Please let me know of your ideas.

Thank you.

[MarkL]

...
Are there any better options that can be used instead of Agilent 1157A ? ( I need more that 2 GHz bandwidth). I'm choosing it for an Agilent Infiniium oscilloscope.

Please let me know of your ideas.

One alternative is to use a Tektronix P6249 4GHz active probe with either a N2744A TekProbe II adapter for your Agilent scope, or a Tektronix 1103 external power supply.  If you choose the N2744A, make sure your scope's firmware supports the P6249.

I have both of the above combinations and they work well with an Agilent MSO-X 3104A.  The N2744A is more convenient, but the 1103 allows the P6249 and other active Tek probes (with TekProbe II) to be used with any equipment (e.g., spectrum analyzers, non-Agilent/Tek scopes, etc.).

If you're buying any active probe used, I would add a recommendation to not to buy it "as-is" or without a return option.  It's easy to destroy the amplifier at the tip in these low voltage HF probes with over-voltage, and it's not repairable.  And also get the grounding accessories you're going to need with the probe, if at all possible.  It's usually going to cost a lot more to buy them separately afterwards.

Depending on your requirements and budget, you might also want to consider differential probe options.  Even with single-ended signals, they can provide a more accurate view of what your circuit is actually seeing since they reduce or eliminate probe ground effects.

[alm]

Agilent/Keysight is not very good about probe backwards compatibility, so it might be that you have problems connecting to more modern scopes. I haven't used this specific active probe, but I don't see a reason why it wouldn't work. Just be aware that using this probe may not give you 2 GHz bandwidth at the probe tip. Bandwidth is not a hard limit, but a gradual roll off. So the gradual roll off of the scope and probe will multiply. A 2 GHz probe on a 2GHz scope will not give 2 GHz bandwidth at the probe tip. Which is why the datasheet recommends the 1158A 4 GHz probe to give 2.25 GHz bandwidth at the probe tip with a 54846A scope.

Make sure you get the important accessories, like the special resistive tip and ground lead. In general active probes are fragile in electrical sense (they don't like over voltage at all) and expensive. If you can tolerate the high DC loading, you could consider a resistive divider probe like the HP/Agilent 54006A (and others like HP 10020A, Tek P6156/6158/6150). There's no electronics, so they are pretty universal. This should be cheaper, give you more bandwidth per buck and be more robust than active probes. The impedance will be more constant over its bandwidth than active probes, and impedance at GHz frequencies will be similar if not higher. But the DC and low frequency impedance is much lower. It's even possible to build such a probe yourself with decent performance. I also remember topics on this forum about it, but I can't find them right now.

I have both of the above combinations and they work well with an Agilent MSO-X 3104A.  The N2744A is more convenient, but the 1103 allows the P6249 and other active Tek probes (with TekProbe II) to be used with any equipment (e.g., spectrum analyzers, non-Agilent/Tek scopes, etc.).

I like the Tek 1103 too for versatility, but have you swept it up to 2 GHz? I haven't tried it (my fastest TekProbe active probe is probably 1 GHz, and I typically use it with slower differential probes), but I would be worried about skin loss and reflections introduced by the extra two BNC connectors in the signal path. I'd expect the Agilent/Keysight adapter to be better.

[MarkL]

...

I have both of the above combinations and they work well with an Agilent MSO-X 3104A.  The N2744A is more convenient, but the 1103 allows the P6249 and other active Tek probes (with TekProbe II) to be used with any equipment (e.g., spectrum analyzers, non-Agilent/Tek scopes, etc.).

I like the Tek 1103 too for versatility, but have you swept it up to 2 GHz? I haven't tried it (my fastest TekProbe active probe is probably 1 GHz, and I typically use it with slower differential probes), but I would be worried about skin loss and reflections introduced by the extra two BNC connectors in the signal path. I'd expect the Agilent/Keysight adapter to be better.

You have a good point.  I haven't swept it up to 2GHz, and my MSO-X 3104A only goes up to 1GHz.

However, if we assume the probe behaves the same in both scenarios, then we're only talking about the added effects of a BNC jumper from the 1103 to the scope.  I don't have a VNA to do a detailed analysis, but I *can* sweep it on a 8595E spectrum analyzer with a tracking generator.

The following is using Pomona 2249-C-24 jumpers, which is a RG58C/U 50R 24" BNC jumper of reasonable quality.  Admittedly better ones are available.

The SA has N to BNC adapters on it.  The SA display is normalized using one 24" BNC jumper from TG output to SA input.

Below is a sweep where the TG is connected to the probe input on the 1103 through the normalized cable, and the 1103 output is connected with a second 24" jumper to the SA input.  So, this sweep represents one run of 24" BNC jumper.  It's about -1.3dB at 2GHz, which I would consider tolerable.  Using simple extrapolation, at 4GHz it should end up being around -3dB which is also the BW of the P6249.

Because I had it laying around, the second sweep is using a no-name N to SMA jumper, with an SMA to BNC adapter on the 1103 end.  Better, but I'm guessing it's probably better coax.

I also checked the N2744A by itself, inserted at the SA input, and it is hardly even noticeable.  I didn't bother to attach the sweep here.

An interesting project is a mini TekProbe II power supply, like this one:

  https://www.eevblog.com/forum/circuit-studio/example-project-tektronix-tekprobe-adapterbreakout

It has an absolutely minimal transmission line like the N2744A, but is not tied to a specific scope manufacturer.  I didn't get the impression the OP was looking for a sub-project.

[ADT123]

At higher bandwidths also consider a low impedance passive probe for digital signals as they can end up cheaper and better due to probe input capacitance. 

The 1157A probe has 0.8pf of input capacitance so whilst it has a DC input impedance of 100K, its impedance at 2GHz will be 99 ohms.

A low impedance passive probe like the PicoConnect 914 has a DC impedance of 500 ohm (x10 50ohm probe) but its tip capacitance is only 0.3pf giving a capacitive load of 265 ohms at 2GHz.  The parallel value (500 & 265 ohms) is 273 ohms.

These may both seem really low but fast digital signals usually have drivers with low output impedance and often 50 or 100 ohm terminations so whilst probes attenuate the signals it may well still be OK.  The above is also a bit of a simplification - the amplitude of the waveform may be attenuated more by the passive probe due to the 500R DC load (eg if the impedance of the signal is 50R) but the waveform shape (edges / overshoot etc) will be better preserved by the lower capacitance passive probe. 

https://www.picotech.com/accessories/passive-low-impedance-probes/picoconnect-914-4-ghz-10-dc-coupled-probe

[joeqsmith]

I wonder how laxmux's home made active probes are moving along.   Sounded like they were getting close but haven't posted in a while.   

https://www.eevblog.com/forum/testgear/12-ghz-active-probe-project/msg4973839/#msg4973839

[electronic_guy]

Thank you all for the suggestions and answers.

What are the safety measures that usually an active probe needs ?

I’m thinking of using a normal passive probe to measure the voltage first and then use the active probe to ensure that it doesn’t go beyond the limits of the active probe. Would that be a good move or is it a overkill to buy a passive probe of 1GHz or so for that.

What about the overshoots and undershoots that might be on the voltages. Would that destroy an active probe. Im being hypochondriac about this now. 

[tggzzz]

No harm in being cautious Amateurs think about how things might work, professionals think about how things might fail.

ADT123 makes a very sensible points. Those are emphasised in "The Trouble with Oscilloscope Probes and an “off-piste” design for microwave and gigabit application" https://www.picotech.com/download/manuals/picoconnect-900-series-the-trouble-with-probes-paper.pdf That also nods to the practicality of a homebrewed probe, which is obviously cheap and can be tailored to the probing point's dimensions.

The PicoConnect 194 capacitance of 0.3pF is impressive, and the price isn't bad either (£499).

You can sometimes find used 1.5GHz HP10020A probes on fleabay; I have four and I'm not selling them