Oscilloscope probe on my Rigol D815 (50 ohm) spectrum analyzer?

[jrs45]

I really like the spectrum analyzer, and have used it for EMI precompliance via field probes and LISN.

But I'd really love to be able to use this to look at the spectra of voltage waveforms, basically using just an oscilloscope probe.

I don't think i can connect the probe directly, since it's a 50ohm input.  Is there a buffer or active probe of some sort available for a reasonable price that will allow this kind of use, safely?  I'd pay a couple hundred dollars for such a thing.  Really I only care about <1MHz (though 10MHz+ would be nice).  I'd be interested in up to around 50V most of the time with a 10:1 probe, and maybe 500V with a 100:1 scope probe.

I have a really old HP spectrum analyzer with a 1Meg input that works great with a probe, but it's a beast and not as functional.  Why can't I do this easily with the D815?

[Relaxe]

I am currently at the same place as you:

I have a spectrum, I want to hook directly to a signal, with a scope probe.
Goal is to see the spectral content of switching noise on SMPS switching net, in the 10MHz to 200MHz range.

I have no clue how to do that safely.

About the impedance: there is passive 50 Ohm passive probes out there, so I guess it's "problem solved" for this.

Now, what else is required/recommended? A DC-Block and an attenuator? Would that suffice? Do we really need to pass the signal trough a LISN?

Thanks to all the senior member of this board for constantly educating us!

[Relaxe]

For the probe itself:
CalTest CT4204 or CT4206.
~200 USD at Digikey.

And this thread seems very helpfull:`
https://www.eevblog.com/forum/testgear/fifty-ohm-probes/

[David Hess]

What you are looking for is a high impedance buffer:

https://www.mathews-engineering.com/store/p1/High_Impedance_Buffer_Amplifier.html

It does what an oscilloscope does internally to convert the 1 megohm impedance of a high impedance probe to a low impedance.  They are commonly used to allow oscilloscope probes to be used with 50 ohm instruments like spectrum analyzers.  An active probe essentially does the same thing.

[Relaxe]

What you are looking for is a high impedance buffer:

This is EXACTLY what I was looking for.
Thanks a lot!

[jrs45]

What you are looking for is a high impedance buffer:

https://www.mathews-engineering.com/store/p1/High_Impedance_Buffer_Amplifier.html

It does what an oscilloscope does internally to convert the 1 megohm impedance of a high impedance probe to a low impedance.  They are commonly used to allow oscilloscope probes to be used with 50 ohm instruments like spectrum analyzers.  An active probe essentially does the same thing.

Awesome, thanks, the buffer is what I want.  Know of any cheaper ones off hand?  I don't need so much bandwidth.  I'll search around.

Active probes aren't meant for 50ohm input usually are they? 

[ogden]

Goal is to see the spectral content of switching noise on SMPS switching net, in the 10MHz to 200MHz range.

I have no clue how to do that safely.

Get 50Ohm DC block and put it on your spectrum analyzer "always on" unless you are sure you are working with circuit that does not output DC. Having DC block on your SA input terminal you just connect 50ohm cable directly to supply output.

What you are looking for is a high impedance buffer:

This is EXACTLY what I was looking for.

No. You need DC block. Hi-impedance buffer is needed for applications where you want to probe RF circuit (supposedly) without affecting it's function. Honestly for such applications you need j-fet probe instead of that useless and overpriced buffer. For DC supplies you just connect your SA using mentioned DC block or plain capacitor if you are on a thin budget. [edit] Purists may add series 50 ohm resistor at the supply end

[jrs45]

So a DC block (looks like a HPF going to 100kHz), and a piece of 50 ohm coax, that's it?

If I can guarantee there's <50VDC, looks like I don't even need that. 

The port says it has a limit of 50VDC and 20dBm on the 50 ohm input, so that would imply around 2Vrms limit if I did my calculations correctly.

This also assumes that 50 ohms wouldn't alter the circuit, which it certainly will in my case.

I suppose I could homebrew a divider with high impedance into 50 ohms, but it would be really nice if this were already built nicely. 

For example, something I can plug a x1 or x10 (or x100) probe into, scales appropriately, and feeds the 50ohm input safely even if I accidentally probe something I shouldn't.  Seems like such a thing should be <$200 if we only need a few hundred MHz?  I'm surprised it doesn't seem to exist.

Goal is to see the spectral content of switching noise on SMPS switching net, in the 10MHz to 200MHz range.

I have no clue how to do that safely.

Get 50Ohm DC block and put it on your spectrum analyzer "always on" unless you are sure you are working with circuit that does not output DC. Having DC block on your SA input terminal you just connect 50ohm cable directly to supply output.

What you are looking for is a high impedance buffer:

This is EXACTLY what I was looking for.

No. You need DC block. Hi-impedance buffer is needed for applications where you want to probe RF circuit (supposedly) without affecting it's function. Honestly for such applications you need j-fet probe instead of that useless and overpriced buffer. For DC supplies you just connect your SA using mentioned DC block or plain capacitor if you are on a thin budget. [edit] Purists may add series 50 ohm resistor at the supply end

[Elasia]

Just curious.. what is the collective's opinion on this home brew fet probe?

https://www.instructables.com/id/DIY-1GHz-Active-Probe-for-Under-20/

[edpalmer42]

HP has an input pod that's used on a few of their models.  You'd have to do some reverse engineering & customization, but it would provide a 300 MHz unity-gain buffer.  Model number is 54003A .  Input impedance is 1 Mohm // 10 pf.

[ogden]

So a DC block (looks like a HPF going to 100kHz), and a piece of 50 ohm coax, that's it?

Yes. This is all you need to measure power supplies. Supply won't notice 50 Ohm load unless you work on extremely low power supplies.  Further reading: AN-1144. Obviously you can't poke around RF circuit with 50 Ohm cable. You need J-FET RF probe for that. You can make one yourself. Fancy scope probe adapter for 50 Ohm devices is pointless waste of money. I have feeling that I said it already, right? [edit] Illustration attached, from said AD appnote AN-1144.

[Elasia]

I need a good set of j probes for my home lab and cant justify hundreds for just a single probe

Think im going to take that design i ran into and improve it using actual mfg'd boards from pcbway advanced material.. they are pretty affordable

I got altium 20 as well so can do proper impedance path / layer stack calculations the lazy way / improve the design hopefully

I'll post it to hackaday but probly wont get to this project till maybe late June at the soonest

[jrs45]

Thanks.  This would be for much more than power supplies.

I'd much rather spend $100 for a properly built, tested, characterized and reliable FET probe than spend a day hand building one.

Does a commercial version of this probe exist?  I'm really surprised this isn't more common.  Wouldn't it be nice to easily and safely look at voltage spectra on a cheap spectrum analyzer?

[Elasia]

Thanks.  This would be for much more than power supplies.

I'd much rather spend $100 for a properly built, tested, characterized and reliable FET probe than spend a day hand building one.

Does a commercial version of this probe exist?  I'm really surprised this isn't more common.  Wouldn't it be nice to easily and safely look at voltage spectra on a cheap spectrum analyzer?

You can find used ones littered all over ebay if you search fet probe

But... its used, who the hell knows what you are going to get / condition until you can quantify it yourself

[M0HZH]

FFT math on a modern oscilloscope is also an option, if you're just looking for spectrum composition of PSU outputs and such.

As for the idea of running a straight 50ohm cable w. separation cap to the spectrum analyzer input, keep in mind spectrum analyzers front ends are very delicate and generally designed to take no more than +20dBm or so; that's a bit over 6 Vpp into 50ohm. I would never do it without a buffer circuit with solid protection.

[james_s]

If you've got a DC blocker then there's no way you're going to get over 6Vpp, at least not unless there's something seriously wrong with that power supply. If one is worried there are also attentuators that can be placed in line with the input. I have some that I picked up recently that I use any time I'm not 100% sure of the power level of whatever I'm trying to look at.

[jrs45]

To be clear, this is not (just) for a power supply.  So some flexibility would be nice.

Yes, there are FET probes on ebay, but like you said, there is a risk that it's damaged or doesn't work to spec, and there is no good way to know.

The FFT on a scope is pure garbage compared to a SA.  It's difficult to use and not very useful.

There seems to be a market opportunity here, and I'm stunned it hasn't been filled.  Really, I'd spend $200-$300 for a proper solution:

1.  Connect a regular oscilloscope probe to my spectrum analyzer.  Or some simple similar probe
2.  Include DC block and output overvoltage protection (a cap and diode?)
3.  100-500MHz bandwidth (maybe?)
4.  Possibly include some scaling (this isn't even needed)

How is this not a thing that exists?  My ancient, beastly HP SA even has a separate 1Meg input for a probe.  (But it only goes to like 40Mhz, is very slow, etc.)

[ogden]

1.  Connect a regular oscilloscope probe to my spectrum analyzer.  Or some simple similar probe
2.  Include DC block and output overvoltage protection (a cap and diode?)
3.  100-500MHz bandwidth (maybe?)
4.  Possibly include some scaling (this isn't even needed)

How is this not a thing that exists?

When you *actually* use scope probe for hi-freq (> 100MHz) RF work - you will see why. First part of the homework - calculate impedance of your 100MHz probe at 100MHz, see how it will impact impedance of 50Ohm transmission line (at given freq) you will try to probe.

[_Wim_]

Thanks.  This would be for much more than power supplies.

I'd much rather spend $100 for a properly built, tested, characterized and reliable FET probe than spend a day hand building one.

Does a commercial version of this probe exist?  I'm really surprised this isn't more common.  Wouldn't it be nice to easily and safely look at voltage spectra on a cheap spectrum analyzer?

You can also buy those DIY probes on Ebay directly (look for RF probe, not FET probe), no need to build yourself. 

Example:
https://www.benl.ebay.be/itm/RF-Active-Probe-0-1-1500-MHz-1-5-GHz-analyzer-oscilloscope-RF-cable-included/332393676082?hash=item4d6434fd32:g:Q3QAAOSwIipZzBvz

I have a couple of these, and they work quite ok. They are based on an old elektor article.

[jrs45]

Let's see, call the probe 30pF, so impedance is 1/(jwC) = 53 ohms.  Yes, of course this will influence a 50 ohm transmission line.

But I'm not ever measuring 50 ohm transmission lines.  I'm measuring various voltage waveforms, including from power rails, various test waveforms (<100kHz) and such, and would like to see their harmonics and noise characteristics cleanly, as one would do on a scope.  It doesn't have to be an actual scope probe, just something comparable.

I don't see why this is such an odd thing to want, and why the market hasn't created anything like this for a reasonable price.

The homebrew FET probe looks like the closest thing available.  If ebay hobbyists are selling them for $15, why isn't someone selling a professional version with proper characterization and protection for ~$100?

When you *actually* use scope probe for hi-freq (> 100MHz) RF work - you will see why. First part of the homework - calculate impedance of your 100MHz probe at 100MHz, see how it will impact impedance of 50Ohm transmission line (at given freq) you will try to probe.

[Elasia]

Thanks.  This would be for much more than power supplies.

I'd much rather spend $100 for a properly built, tested, characterized and reliable FET probe than spend a day hand building one.

Does a commercial version of this probe exist?  I'm really surprised this isn't more common.  Wouldn't it be nice to easily and safely look at voltage spectra on a cheap spectrum analyzer?

You can also buy those DIY probes on Ebay directly (look for RF probe, not FET probe), no need to build yourself. 

Example:
https://www.benl.ebay.be/itm/RF-Active-Probe-0-1-1500-MHz-1-5-GHz-analyzer-oscilloscope-RF-cable-included/332393676082?hash=item4d6434fd32:g:Q3QAAOSwIipZzBvz

I have a couple of these, and they work quite ok. They are based on an old elektor article.

Hah look at that.. i knew someone had to be making them.. looks like a good few are with several different designs

Look pretty sloppy some of em.. get what you pay for

[ogden]

Let's see, call the probe 10pF, so impedance is 1/(jwC) = 160 ohms.

Here you go. 50 Ohms becomes 38.10 Ohms even with exceptional 10pF scope probe. Ask any RF engineer - 38Ohms is good match or not, do they want you to connect 10pF in parallel to any part of RF circuit or not?

I'm also not ever measuring 50 ohm transmission lines.

Emphasis was on 50Ohms. OK - what exactly you want to measure, if anything?

[Elasia]

To be clear, this is not (just) for a power supply.  So some flexibility would be nice.

Yes, there are FET probes on ebay, but like you said, there is a risk that it's damaged or doesn't work to spec, and there is no good way to know.

The FFT on a scope is pure garbage compared to a SA.  It's difficult to use and not very useful.

There seems to be a market opportunity here, and I'm stunned it hasn't been filled.  Really, I'd spend $200-$300 for a proper solution:

1.  Connect a regular oscilloscope probe to my spectrum analyzer.  Or some simple similar probe
2.  Include DC block and output overvoltage protection (a cap and diode?)
3.  100-500MHz bandwidth (maybe?)
4.  Possibly include some scaling (this isn't even needed)

How is this not a thing that exists?  My ancient, beastly HP SA even has a separate 1Meg input for a probe.  (But it only goes to like 40Mhz, is very slow, etc.)

AuburnP-20A
https://www.tequipment.net/AuburnP-20A.html

AuburnP-20B
https://www.tequipment.net/AuburnP-20B/

There you go

[jrs45]

Emphasis was on 50Ohms. OK - what exactly you want to measure, if anything?

Like I said, various supply rails, voltage waveforms (from opamps), and similar.  Pretty conventional stuff, not RF per se, so no 50 ohm transmission lines.

Nothing sensitive to loading from a conventional scope probe, of course.

[jrs45]

AuburnP-20A
https://www.tequipment.net/AuburnP-20A.html

AuburnP-20B
https://www.tequipment.net/AuburnP-20B/

Excellent, thank you, this looks about right for my use case (though a commercial version of the FET probe also looks attractive).