Good design for a DIY active probe?

[Marco]

I've looked around the web, but they all seem a bit lacking. There is one which uses an opamp at the input, but that's just adding way way more input capacitance than necessary. The poor man's 1 GHz active probe is AC coupled and will be highly non-linear because of the small source resistor. The ancient one from Troubleshooting Analog Circuits still seems the best one over all. I can find no schematic at all for one with active DC offset correction, although that's a bit of luxury I could forego.

Anyone know of a better circuit? (Preferably with PCB layout.) I found these :
http://books.google.nl/books?id=3kY4-HYLqh0C&lpg=PP1&pg=PA16#v=onepage&q&f=false
http://elektrotanya.com/files/forum/2009/10/e04a036.pdf
http://welecw2000a.sourceforge.net/docs/Hardware/Aktiver_Tastkopf_mit_OPA659.pdf
http://oliverbetz.de/hit/hit_e.htm

[free_electron]

What are you looking for bandwidth wise ?
Just pick up a used 1151 or 1152 probe from agilent on ebay. You can find these for 100$. Build a supply to feed it the 3 or four voltages it needs and off you go. There was a topic here on the forum where someone did that. I posted pinouts and schematics there how to make the supply.

You are never going to beat the hybrids in those beasts....

[Marco]

1141 and 1152 you mean? Neither are up for anywhere near that at the moment. Tektronix P6201 looks more interesting actually, having BNC and inline 50 Ohm terminators build in for high impedance scopes (like my cheapo Owon DSO). Kinda hesitant at buying one off ebay from China though.

Beating them isn't necessary, couple 100 MHz bandwidth and mV level noise is acceptable.

[mazurov]

There are also Tektronix 6202 and 6045, the latter comes with power supply, see this -> http://www.ebay.com/itm/Tektronix-P6045-F-E-T-probe-package-uknown-instrument-/321057431273?pt=LH_DefaultDomain_0&hash=item4ac0839ee9

Both have higher bandwidth than your scope.

[muvideo]

Marco, less than one month ago I made the
same search and found the same references.
I needed only the lowest tip capacitance I could
afford and decent BW, I needed to spot only
the timing of the event, so didnt bother with
linearity. I liked Bob Peases' one but I built the
poorman's one, only because I had the few
components needed at hand, now I dont remember my
mosfet partnumber. Few minutes and it worked fine
for my task, it's so simple to build, but I did not
tried to characterize it, I would not know from where
to start

Fabio.

[Marco]

Mechatrommer : I want to mess around with FPGAs so anything over 5 volt ... less than a couple pF will not add significant load at the I/O speed I'm interested in.

Muvideo : did you build it dead bug style on some copper board?

[muvideo]

Muvideo : did you build it dead bug style on some copper board?

Perfboard, I used an old 0.6pF rod ceramic as input capacitor.

[mazurov]

This is another example of non-vendor specific active probe -> http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300848617979&ssPageName=ADME:B:SS:US:1123

Many old Tektonix active probes were equipped with mains power supply, either combined with amplifier box or standalone.

[nctnico]

Why not build a passive divider probe? I have build two and they work like a charm. I just can't find the website right now. Something with EMI RFI Guy but that doesn't Google properly.

[alm]

http://www.emcesd.com/1ghzprob.htm

[nctnico]

@Alm: Thats it exactly! Thanks for taking the trouble to find the link.

[Marco]

Also an option, though with 1K impedance you'd have to be a bit more thoughtful of when to use it than an active probe.

Someone made a good point in another thread though, the change in termination impedance when you parallel a traditional 1M scope input with 50 Ohm is by no means negligible at 100+ MHz. With the 15 pF my Owon supposedly has, termination impedance will already have dropped by nearly 10% at 100 MHz ... not a problem for a more expensive scope with a proper 50 Ohm input of course.

[nctnico]

1k Ohm may sound low but the input range of an active probe usually isn't much beyond +/- 15V. In some cases even much less. It depends on what you are measuring but you'll mostly find 50 or 75 Ohm signals in RF systems or 33 to 75 Ohm series terminated signals in digital systems. Compared to those impedances 1k Ohm is huge.

[alm]

1K impedance may seem low at DC, but what's the impedance of the so called high-impedance probes at high frequencies? A 100 MHz 10x high-Z probe might have an input spec of 10 Mohm // 12 pF. At 100 MHz, this will have an impedance of 133 ohm (I didn't even bother to take the 10 Mohm resistor into account). A commercial 20x resistive probe might have an input impedance of 1 kohm // 1.5 pF, or about 500 ohm at 100 MHz. Probe datasheets (for good probes at least, not the $5 eBay specials) will show an input impedance vs frequency graph. Note that the vertical scale is logarithmic. The input resistance is only relevant up to a few hundred kHz or so, after that it's the capacitance that matters.

A 500 MHz 10x high-Z probe will have an input capacitance of about 8 pF, at 500 MHz this probe will have an input impedance of 40 ohm. The 1 kohm // 1.5 pF probe will be about 175 ohm. There are also resistive probes with only 0.4 pF of input capacitance, which perform even better. A low-end commercial active probe might be 1 Mohm // 1 pF, or 320 ohm at 500 MHz. Not that much higher than the supposedly low impedance probe.

DIY designs will be worse, but I still expect for example the EMCESD design to beat a high-Z probe at input capacitance.

[muvideo]

Hello, dont know if I understand all of the arguments exsposed before,
but sometimes all is needed is not huge BW but simply a low tip capacitance,
decent BW, division ratio not too high for decent vertical resolution on
low amplitude signals.

For example I built the mosfet active probe because I wanted
to check the sampling window of the internal ADC of a microcontroller.
The internal S/H of the ADC is modeled as 8pF capacitor with a 1k
resistor in series and is not discharged between conversions.
The procedure was to charge the SH capacitance with a slow
conversion on a steady known DC, and then start a conversion
on a pin where there was a resistor, the probe measured the pulse
on this load resistor. The pin, pad and pcb have their own
capacitance.

In this case if the probe capacitance is low enough and the
resistive loading is high enough I can easily see  sampler's "gate"
opening (this is easy), but also closing.
The 10:1 probe with it's capacitance was uneffective.
The active probe was showing the pulse nicely and I could
choose the right loading resistor to discharge the pin, a good
compromise was 100kOhm.

Could a test like this be done better with a passive probe?

Thanks,
Fabio.

[nctnico]

An active probe with a low input capacitance is always better than a passive divider probe because a passive divider probe has an attenuation of 20 times or more (like 40 times) so you'll see more noise from the scope itself.

[alm]

I would expect the high DC loading of resistive probes to be an issue here.

The main issue with resistive divider probes is the low DC input impedance. It requires test points with a fairly low output impedance. This is not usually an issue for fast digital circuits, which is why Howard Johnson advocates resistive divider probes. 10x resistive probes are also common, but DC resistance is down to 500 ohm. High frequency impedance is similar to the 20x probes, though. Many FET probes are also 10x.

An issue with active probes is that the input impedance varies over many orders of magnitude over its bandwidth. This can introduce distortion, depending on what the source looks like. A resistive source in combination with a capacitive probe will form a low-pass filter, slowing down rise times. The input impedance of resistive probes is, as the name implies, much more resisitive and constant over its bandwidth.

[signality]

Way too late to be of any use for the original post but in case anyone else is looking for the same thing ...

Here's a schematic and PCB design for a DIY version of Bob Peases's original jfet input high impedance active scope probe adapter:

http://easyeda.com/project_view_Bob-Pease-s-High-Impedance-Active-Scope-Probe-PCB-Design_g2QNnljhf.htm

with a simulation:

http://easyeda.com/project_view_Bob-Pease-s-High-Impedance-Active-Scope-Probe-simulations_gsmEyQ8qj.htm

There's also the outline of an idea for a similar (but maybe lower frequency) probe using an opamp:

http://easyeda.com/project_view_High-impedance-probe-using-opamp_gmtnFzRL3.htm

Enjoy.

 

I've looked around the web, but they all seem a bit lacking. There is one which uses an opamp at the input, but that's just adding way way more input capacitance than necessary. The poor man's 1 GHz active probe is AC coupled and will be highly non-linear because of the small source resistor. The ancient one from Troubleshooting Analog Circuits still seems the best one over all. I can find no schematic at all for one with active DC offset correction, although that's a bit of luxury I could forego.

Anyone know of a better circuit? (Preferably with PCB layout.) I found these :
http://books.google.nl/books?id=3kY4-HYLqh0C&lpg=PP1&pg=PA16#v=onepage&q&f=false
http://elektrotanya.com/files/forum/2009/10/e04a036.pdf
http://welecw2000a.sourceforge.net/docs/Hardware/Aktiver_Tastkopf_mit_OPA659.pdf
http://oliverbetz.de/hit/hit_e.htm

[Alex Eisenhut]

I've looked around the web, but they all seem a bit lacking. There is one which uses an opamp at the input, but that's just adding way way more input capacitance than necessary. The poor man's 1 GHz active probe is AC coupled and will be highly non-linear because of the small source resistor. The ancient one from Troubleshooting Analog Circuits still seems the best one over all. I can find no schematic at all for one with active DC offset correction, although that's a bit of luxury I could forego.

Anyone know of a better circuit? (Preferably with PCB layout.) I found these :
http://books.google.nl/books?id=3kY4-HYLqh0C&lpg=PP1&pg=PA16#v=onepage&q&f=false
http://elektrotanya.com/files/forum/2009/10/e04a036.pdf
http://welecw2000a.sourceforge.net/docs/Hardware/Aktiver_Tastkopf_mit_OPA659.pdf
http://oliverbetz.de/hit/hit_e.htm

Tektronix P6032.

Oh wait, I'll bet you wanted it solid state? https://www.eevblog.com/forum/Smileys/default/smiley_laughing.gif

[branadic]

Please refer to:

http://welecw2000a.sourceforge.net/docs/Hardware/Aktiver_Tastkopf_mit_OPA659.pdf

and the additional thread:

http://www.mikrocontroller.net/topic/188227

[David Hess]

The service manuals for the old Tektronix P6202A (500MHz 10M 2.0pF) and P6201 (900 MHz 1M 1.5pF) active probes have schematics which show what is involved in adding offset to the simple designs Bob Pease and Jim Williams wrote about.