Reccomendations on scope probes for RF

Hello, for a project I'm planning on starting in the next few months, I'll be building switching RF amplifiers at 128MHz, and it occurred to me that having good high bandwidth scope probes would be a big help, since I'll need to see waveforms at nodes which aren't 50ohm matched (the drain and gates of my transistors).

At my lab we have a Agilent DSO7054B 500MHz, 4ch scope. It's the kind which is compatible with a variety of active probes which are powered through the front panel. But one limitation I see in all of them is a limited input range +/-8V, which isn't enough for what I want to do (I'm expecting voltages up to 40V or more). And they don't seem to give any attenuator accessories. Seems like these probes are meant for high speed digital only.

So what feasible options are available? Are there third party probes with external amplifiers/power supplies which don't have this limitation? Can high quality passive probes give adequate performance? What I've noticed with using passive probes is that when I probe the same spot with two different channels, the waveforms will be different no matter how I do it (different phase and distortion), even if I carefully compensate them.

Thanks in advance.

It depends on how much money you want to spend. Also, I would have expected better performance out of your Agilent 'scope & probes.

I have a variety of different probes for RF. I have an active one, but it only takes +/-20V. I also have a passive Tektronix P6158 which is fine if you can live with the 1k load (it has 20:1 attenuation range). I also have a one of these https://www.howardelectronics.com/auburn/rfprobe.html (actually two - I broke one) which are similar to the Tektronix but 500 ohms. If you have lots of money there are probably other choices.

I tend to use the Tektronix probe most.

Keith

Agilent and Tektronix have also resistive 100:1 probes (Agilent part# is N2876A). But they have still limited voltage rating at higher frequencies, the same with standard passive probes. http://www.home.agilent.com/upload/cmc_upload/All/N2876-97000.pdf

Thanks for the replies.

A lower input resistance in the range of 1K-5Kohm would probably be fine for the probes, though I'd hope to have an effective input capacitance of ~1pF, which is why I was looking at active probes. I'm surprised that even passive probes have strict input voltage limits for higher frequencies... is that limit simply a function of heating in the attentuation network? Why does the max voltage roll off with frequency?

I'd recommend for you to make yourself a passive resistor probe. Use 1/8W (or even better 1/16W) resistors, value around 220Ohm. Put two of these resistors in series, minimizing the leads to no more than 1.5mm. After them solder a 50Ohm semi-rigid cable (like UT-141). This gives you a 500Ohm probe, that is 20db attenuation ratio. Probably 500Ohm won't have any effect in any circuit point, but put a DC-block after the semi-rigid cable (in front of the scope's 50Ohm input) not to feed the DC through the probe. Another key trick is to minimize the ground lead length. It should be about the same length as the two series resistors, that is about 15 mm. This kind of probe will have a flat response to about 3GHz, then it will rise. If you compensate the rise characteristic, you could have a flat response to 10GHz. I've made several and the material cost is about $10, but overall priceless, very good performance, rise time <50ps, estimated capacitance of about <0.2pF.

rfmw said:

I'd recommend for you to make yourself a passive resistor probe. Use 1/8W (or even better 1/16W) resistors, value around 220Ohm. Put two of these resistors in series, minimizing the leads to no more than 1.5mm. After them solder a 50Ohm semi-rigid cable (like UT-141). This gives you a 500Ohm probe, that is 20db attenuation ratio.

Click to expand...

How does this work out to 500ohm and 20dB with 220ohm attenuator resistors?

Probably 500Ohm won't have any effect in any circuit point, but put a DC-block after the semi-rigid cable (in front of the scope's 50Ohm input) not to feed the DC through the probe. Another key trick is to minimize the ground lead length. It should be about the same length as the two series resistors, that is about 15 mm. This kind of probe will have a flat response to about 3GHz, then it will rise. If you compensate the rise characteristic, you could have a flat response to 10GHz. I've made several and the material cost is about $10, but overall priceless, very good performance, rise time <50ps, estimated capacitance of about <0.2pF.

Click to expand...

Is it necessary for the whole thing to be rigid coax? That would be pretty inconvenient since then the whole thing would be rigidly fixed to the scope. Or could I just put a coax connector on a short length of rigid coax, then run flexible coax the rest of the way to the scope?

We do have some money to spend (I believe) so I was going to look at purchasing a solution before having to do a DIY hack.

I've been looking around a bit more, and I found this thing:
**broken link removed**
So it looks like it's similar to the concept described by rfmw; just a passive attenuator matched to 50ohms. Looks great, but I'm not sure what it's input range really is. It says "Max V in – 12.5 VRMS*1." which references the footnote "*1 Limited by scope input or 30 VRMS, whichever is less." Not sure exactly what that means... I'd expect it to depend on whether the attenuator is used, but it doesn't really say. Also, is uses an SMA connector, but for my frequency range I should be fine using a BNC adapter.

The 500ohm input resistance is kind of low though-probably enough to throw off the matching networks when probing gate voltages directly. Is there anything like this but with a higher input resistance (>1K)? If I used one of these things and just put an extra 500ohm resistor in series with the output, could I still expect good results (an extra 6dB of attenuation aside)?

Is there anything like this but with a higher input resistance (>1K)? If I used one of these things and just put an extra 500ohm resistor in series with the output, could I still expect good results (an extra 6dB of attenuation aside)?

Click to expand...

Please review post #3. 100:1 5K passive probes are available at least from Tektronix and Agilent. However, if you're not satisfied. with the RF RMS rating, you may need to design your own passive probes.

biff44 said:

https://www.google.com/url?sa=t&rct...s4zMCw&usg=AFQjCNE5RqywBdwn5A-uiqElxZDIkDQRaQ

Click to expand...

Yes I ran across this one too, but it also has a 500ohm input, and its input voltage is even less than the others (7Vrms continuous, 35V for 1second... but I assume all these probes can endure higher voltages for brief times).

FvM said:

Please review post #3. 100:1 5K passive probes are available at least from Tektronix and Agilent. However, if you're not satisfied. with the RF RMS rating, you may need to design your own passive probes.

Click to expand...

The N2876A does look good, but as noted before its voltage rating falls off with frequency (only around 12V at my frequency of interest). That seems bizarre, and I wonder why other similar probes don't specify similar behavior.

I'm open to adding on-board attenuation networks to extent the working voltage of the probes, but I don't think I'm willing to DIY the entire probe. That's why I asked about the feasibility of adding series resistance to the node of interest (like with a surface mount resistor) and probing the other side of the resistor. I'd like to get some feedback on the idea before asking my advisor to spend a couple grand on these things.

Also, is there any reason that one of the tektronix 50ohm probes would have trouble working with an agilent scope?

Thanks.