Author Topic: > 1 GHz DIY differential probes  (Read 29537 times)

0 Members and 1 Guest are viewing this topic.

Online Marco

  • Super Contributor
  • ***
  • Posts: 4768
  • Country: nl
Re: > 1 GHz DIY differential probes
« Reply #50 on: October 16, 2018, 10:14:08 am »
High resistance at low frequency, high enough impedance to still be useful at high frequency. So still useful for general purpose probing in a pinch.

The really high frequency differential probes all try to be constant impedance, just high enough to not fuck a 50 Ohm line too much ... and that's all they are useful for.
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 6801
  • Country: 00
  • +++ ATH1
Re: > 1 GHz DIY differential probes
« Reply #51 on: October 16, 2018, 11:32:57 am »
can we make 100fF?

Femto farad ? I'm guessing its not that easy, example of a 3 decades old, HP 54701A 2.5GHz active probe (not a differential probe) has 0.6pF (typical) with input resistance of 100K Ohm.


Online Rerouter

  • Super Contributor
  • ***
  • Posts: 4624
  • Country: au
  • Question Everything... Except This Statement
Re: > 1 GHz DIY differential probes
« Reply #52 on: October 16, 2018, 11:44:20 am »
If you want crazy low input capacitance, things get complex, but I would guess we can reduce what is currently there and iterate towards the goal, anything under 0.5pF already appears to be crazy level money,

Most of the schematics I have glimpsed in the past from reverse engineered crazy high bandwidth probes seemed to use an input network to cancel out the apparent capacitance. and pretty much no protections at all.
 

Offline Gerhard_dk4xp

  • Regular Contributor
  • *
  • Posts: 115
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #53 on: October 16, 2018, 11:51:31 am »
I absolutely LOVE these HP probes, also the 1152A successor!
The 0.6pF make the series resonance of the ground lead much less important
and you can probe around without fear of burning sth. @ less than 50V.

regards,
Gerhard
« Last Edit: October 16, 2018, 11:56:20 am by Gerhard_dk4xp »
 

Offline Mechatrommer

  • Super Contributor
  • ***
  • Posts: 9731
  • Country: my
  • reassessing directives...
Re: > 1 GHz DIY differential probes
« Reply #54 on: October 16, 2018, 12:12:28 pm »
can we make 100fF?
Femto farad ? I'm guessing its not that easy, example of a 3 decades old, HP 54701A 2.5GHz active probe (not a differential probe) has 0.6pF (typical) with input resistance of 100K Ohm.
sorry i screwed up (msg deleted) i didnt see the low-f respond (1st attachment). so 1pF seems feasible? i got 0806 smd 1pF here ;D (2nd attachment) they both load the 500 ohm node by a bit anyway @ 1GHz ???

anything under 0.5pF already appears to be crazy level money,
how about pcb printed capacitor? or diy like 2 plates separated by a piece of paper? is it wrong to be unacademical?
« Last Edit: October 16, 2018, 12:15:37 pm by Mechatrommer »
It's extremely difficult to start life.. one features of nature.. physical laws are mathematical theory of great beauty... You may wonder Why? our knowledge shows that nature is so constructed. We simply have to accept it. One could describe the situation by saying that... (Paul Dirac)
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19753
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #55 on: October 16, 2018, 05:22:05 pm »
Or from a few bits of coax and a combiner, e.g. http://emcesd.com/pdf/cd94scr.pdf

More expensive than a bunch of BFU660F's though. How about something simple like this? Is the simulator really giving me that bad an impression of how it would work with a tight layout and some =<0204 components? R1=R2=0 is only useful for say 10s of mV of input, probably want to bias the bases to ~4V with opamps instead of resistor dividers, might want to clamp R5-R8 with Schottkys.

PS. that Deltasense probe has nice specs, large voltage range and input impedance.
I don't think this configuration will work very well (probably due to the miller effect). I did find a fast opamp and I'm going to try a classic difference amplifier approach. After all the reason to use a differential (instrumentation) amplifier is to have extremely high input impedances on both inputs but in this case a few kilo Ohms is already OK. Perhaps a discrete solution specifically designed to act as a differential probe can be pushed to a higher bandwidth especially if it doesn't need to amplify. DC offset and temperature stability will be challenging nuts to crack though.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 4768
  • Country: nl
Re: > 1 GHz DIY differential probes
« Reply #56 on: October 16, 2018, 05:37:55 pm »
There's not a whole lot of amplification in that circuit, 1.5 or so since the emitter resistance of RF transistors is rather high, so Miller capacitance isn't a big deal. The simulator says the circuit has >400 Ohm of input impedance at 1 GHz. When you increase R1&R2 the Miller capacitance is reduced further still.
« Last Edit: October 16, 2018, 05:43:39 pm by Marco »
 

Online dietert1

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: de
    • CADT Homepage
Re: > 1 GHz DIY differential probes
« Reply #57 on: October 18, 2018, 07:14:15 am »
Your difference stage (and whatever amplifier you make) will have an input capacitance of about 0.5 to 1 pF, which is much better than the usual passive probe and may be good enough for a 1 GHz probe.
Yet if you want the 0.04 pF or 0.02 pF that cerebus proposed above and that Mr. Rosenkränzer implements in his probe, there will be a capacitive divider in front of the amplifier. This is also mentioned in the Keysight video. The difference stage always looses a factor 4, because you drop one of two output ports and you want to drive the line to the scope with 50 Ohm output impedance. In total you will have a factor 1:50 or 1:100.
I think you really want some gain. If you put emitter followers in front of your difference stage this will get you roughly a factor 5 (BFU66F Cbe/ Ccb). Those who want DC to GHz might use THS4302 gain blocks.
By the way, i received a LMH3401 evaluation board. That is an excellent difference stage, like +/- 1 dB up to 2.9 GHz. Image is boring.
 

Offline dzseki

  • Frequent Contributor
  • **
  • Posts: 389
  • Country: hu
Re: > 1 GHz DIY differential probes
« Reply #58 on: October 18, 2018, 08:43:38 am »
can we make 100fF?

Femto farad ? I'm guessing its not that easy, example of a 3 decades old, HP 54701A 2.5GHz active probe (not a differential probe) has 0.6pF (typical) with input resistance of 100K Ohm.



I have the HP 1120A which is only 500MHz active probe (not differential), it has 100kOhm||3pF input impedance in 1:1 mode and 1MOhm|| <1pF when using a divider (10:1 or 100:1).
The point is this is a quite old stuff, appeared in one HP Journal from 1969 I think, the service manual do actually contains all the schematics, even the probe's hybrid, it might worth a look.
HP 1720A scope with HP 1120A probe, EMG 12563 pulse generator, EMG 1257 function generator, MEV TR-1660C bench multimeter
 

Online dietert1

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: de
    • CADT Homepage
Re: > 1 GHz DIY differential probes
« Reply #59 on: October 18, 2018, 09:29:22 am »
In my opinion that old technology is completely obsolete and its analysis a waste of time. With the THS4302  gain block i mentioned you can make a high impedance 1:1 probe with 0.4 pF input capacitance and a bandwidth of 2.4 GHz.
« Last Edit: October 18, 2018, 09:33:27 am by dietert1 »
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19753
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #60 on: October 18, 2018, 10:04:18 am »
By the way, i received a LMH3401 evaluation board. That is an excellent difference stage, like +/- 1 dB up to 2.9 GHz. Image is boring.
That sounds promising. Did you test single-ended in and single-ended out? I did some testing with an LMH6703 (dead-bug style) and that looks promising as well albeit with less bandwidth. At least there is no difference between reversing the inputs.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online dietert1

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: de
    • CADT Homepage
Re: > 1 GHz DIY differential probes
« Reply #61 on: October 18, 2018, 11:25:27 am »
Yes, the LMH3401 was tested with the HP 8560A again, so it is single ended in (TG) and single ended out. The curves from both inputs to the same output appear flat +/- 1 dB up to at least 2.9 GHz. I would not trust my setup to better than a dB.

In the meantime i experimented a little with the ADA4927 probe from Mr. Rosenkränzer. I modified the difference stage to the standard setup with 4 301R resistors (inputs = Gnd). The 2K5 0603 input resistors were replaced each by a 2x 620R 0402 resistor pair. I also removed two pads and brought the coax cable closer to the amp.
Now it works better, but the asymmetry of the ADA4927 still exists. With our WR64Xi and a Lattice FPGA measured risetimes are 345 psec on indirect input, 306 psec on direct input. The direct trace still shows more overshoot/ringing at 1.4 GHz .
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19753
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #62 on: October 18, 2018, 12:04:35 pm »
OK. I think I'm going to re-design my board with an LMH3401 then instead of the LMH6703.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 11250
  • Country: us
  • DavidH
Re: > 1 GHz DIY differential probes
« Reply #63 on: October 20, 2018, 02:35:21 am »
Bob Pease's probe with 0.29 picofarads of input capacitance is shown below.  The printed circuit board stiffener adds another 0.08 picofarads but drilling holes in it reduces this to 0.06 picofarads of added capacitance.

Old designs like the Tektronix P6202 (which is what I would start with if I wanted to design a general purpose probe) use an input divider to reduce the capacitance of the input device, 2.0 picofarads in this case.
« Last Edit: October 20, 2018, 02:39:01 am by David Hess »
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 4876
  • Country: gb
Re: > 1 GHz DIY differential probes
« Reply #64 on: October 20, 2018, 08:07:01 am »
Both designs are well and good in themselves, but don't forget that the title contains >1GHz and differential in its title.

Both are single ended, the JFETs in Bob's design run out of steam at ~200MHz and the P6202 is a 500MHz design.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Online dietert1

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: de
    • CADT Homepage
Re: > 1 GHz DIY differential probes
« Reply #65 on: October 20, 2018, 08:44:31 am »
Thanks for posting the probe schematics. We cannot emphasize enough the advantage of using a low capacitance active probe, even if its bandwidth is "only" 500 MHz instead of 1 GHz and even if it's not differential. When you try to make a probe for time domain measurements, you want the gain curve to level off smoothly at the upper bandwidth limit. Otherwise you will see ringing, like with audio brickwall filters. That may be a reason to prefer a discrete design since in general it will have less components. For example the proposed transistor difference stage levels off smoothly.
Anyway i would recommend to have a look at the "contemporary" TI proposals for a 2 GHz scope frontend with LMH5401 and LMH6401. Using that kind of building block is roughly how Mr. Rosenkränzer arrived at a working DIY type GHz differential probe.
Really, in 2018 we should have DSOs with builtin probe deembedding including some handy calibration scheme.
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19753
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #66 on: October 20, 2018, 12:03:03 pm »
At some point you just can't get away with a discrete solution especially on readily available PCB materials. The parasitics will kill the circuit. Meanwhile I've updated my design with the LMH3401 to see how that works. The datasheet for the LMH3401 says that TI choose to have the feedback resistors on the chip itself to avoid the parasitic capacitance of the circuit board. Unfortunately these are 200 Ohm and for a 1:20 differential attenuation this means a differential input impedance of only 2k Ohm. It is still OK-ish because on a typical differential pair terminated with 100 Ohm it will result in an amplitude error of 5% due to loading the signal. Still seeing is believing. The simulation shows some peaking at higher frequencies (several GHz) so hopefully the loss of FR4 takes care of it.
« Last Edit: October 20, 2018, 12:44:21 pm by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online dietert1

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: de
    • CADT Homepage
Re: > 1 GHz DIY differential probes
« Reply #67 on: October 21, 2018, 09:17:32 am »
I bought some ATF-35143 FETs to make buffer input stages. That works +/- 1 dB at least up to 2.9 GHz with a 51K/10K input divider made from 0805 SMD resistors. Then the parasitic capacitance of the 0805 resistor roughly compensates the FET input capacitance. Low frequency noise is less then a mV. Attached schematic describes a test setup. With smaller parts and a careful layout a 10:1 divider should compensate. I will try to drive the LMH3401EVM inputs with two of these input stages.
« Last Edit: October 21, 2018, 09:22:40 am by dietert1 »
 
The following users thanked this post: 2N3055

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19753
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #68 on: October 21, 2018, 10:37:40 am »
That is interesting! Are these FETs available as a matched pair? I think that will greatly improve the CMMR and DC offset matching.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Gerhard_dk4xp

  • Regular Contributor
  • *
  • Posts: 115
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #69 on: October 21, 2018, 10:57:20 am »
Discontinued.
No matched pairs, if that word is allowed at all in the context of FETs.
Maybe one could enforce equal currents with a mirror.

possible replacements:
https://www.digikey.de/product-detail/de/CE3521M4-C2/CE3521M4-C2CT-ND/6165474/?itemSeq=265991607     >

https://www.digikey.de/product-detail/de/CE3520K3-C1/CE3520K3-C1CT-ND/6165473/?itemSeq=265991735     >

<   https://www.digikey.de/product-detail/de/cel/CE3514M4-C2/CE3514M4-C2CT-ND/6165472     >


Also interesting:

<    https://www.digikey.de/product-detail/de/analog-devices-inc/ADL5565ACPZ-R7/ADL5565ACPZ-R7TR-ND/2773701     >

regards,
Gerhard


« Last Edit: October 21, 2018, 11:19:21 am by Gerhard_dk4xp »
 

Offline Carrington

  • Super Contributor
  • ***
  • Posts: 1201
  • Country: es
Re: > 1 GHz DIY differential probes
« Reply #70 on: October 21, 2018, 11:07:01 am »
I just wanted to say that this is getting more and more interesting, and I think that something pretty good is going to come out of here.  :-+
My English can be pretty bad, so suggestions are welcome. ;)
Space Weather.
Lightning & Thunderstorms in Real Time.
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 4768
  • Country: nl
Re: > 1 GHz DIY differential probes
« Reply #71 on: October 21, 2018, 07:38:56 pm »
That is interesting! Are these FETs available as a matched pair? I think that will greatly improve the CMMR and DC offset matching.
How close can you match the capacitance of the resistors though? Seems to me the solder blobs would throw it off.
 

Online 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1482
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #72 on: October 21, 2018, 10:56:24 pm »
I'm aware it's neither differential nor >1GHz, and well, it's only available in German, but the most elaborate take on a DIY active probe I stumbled over in the last years was this one:
http://welecw2000a.sourceforge.net/docs/Hardware/Aktiver_Tastkopf_mit_OPA659.pdf
The probe uses an OPA659 with a trimmable capacitive divider. The main developer and author of this PDF "branadic" is also active in this forum btw.
Building some of these is one of multiple projects on my "stack".

As a side note: I also bought some of the probes by Alfred Rosenkraenzer (PCB version). When I bought the first one, I was well aware that the bandwidth sounded a bit optimistic. Still, my main intention was to measure LVDS signals in the 20-40MHz range and the probes worked excellently for that. I.e. where the passive 500MHz probes just showed a sine shaped signal, I got a clear rectangular signal with steep edges. I already used them for 5V signals as well though (e.g. 100ns CS pulses) where they are also far superior to 500MHz passive probes. Admittedly, the scoped used was a 600MHz Lecroy so I don't really care if the probes really have a much higher bandwidth or not. And yes, looking at the differential amplitude, the probes show a 20:1 attenuation, I guess the 10:1 value was meant for each input. In the meantime, the description in the eBay offers was also changed as far as I can tell.
Trying is the first step towards failure - Homer J. Simpson
 

Online dietert1

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: de
    • CADT Homepage
Re: > 1 GHz DIY differential probes
« Reply #73 on: October 22, 2018, 09:20:56 am »
Today i received a THS4302 evaluation module. Compared to the OPA659 the THS4302 has more bandwidth and some gain.
The EVM gain curve  measured by our HP 8560A shows 7.8 dB at low frequencies (G=5 gives 14 dB - 6 dB output line attenuation). 3 dB bandwidth is roughly 1.8 GHz.  Up to 2.9 GHz it levels off smoothly, so this will result in low ringing.
The evaluation module has a terminated strip line as input. A probe needs a different design. THS4302 input capacitance is specified as < 1 pF, so if you put a 10:1 divider in front, you may get a 0.2 pF high impedance probe.
 

Offline Hydron

  • Frequent Contributor
  • **
  • Posts: 462
  • Country: gb
Re: > 1 GHz DIY differential probes
« Reply #74 on: October 22, 2018, 09:39:30 am »
Looks like there's also a gain-of-10 version of the THS4302 too (the THS4303) if it's of interest.

Thanks for all the work being done in this thread - sounds like the final result could be pretty handy.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf