Author Topic: > 1 GHz DIY differential probes  (Read 65572 times)

0 Members and 2 Guests are viewing this topic.

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26757
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #150 on: October 01, 2020, 08:58:09 pm »
A typical active probe has like 0.5pF capacitance, so anything you attempt to attach it somewhere will actually add a capacitance in the same magnitude or higher. I don't see how a purely resistive probe would help there. Plus, its low resistance makes it problematic for digital outputs which might as well create a burst clock only now and then.
Well, you should try the low-Z passive probes. These have very low capacitances and high bandwidth. I have a bunch of Tektronix P6156 probes ( http://w140.com/tekwiki/wiki/P6156 ) and these work very well (3.5GHz bandwidth with less than 1pf of capacitive loading!) but you'll need to use the 1:20 attenuator (or higher) to get an input resistance of at least 1k Ohm otherwise digital signals like LVDS will be loaded too much.
« Last Edit: October 01, 2020, 09:07:44 pm by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: > 1 GHz DIY differential probes
« Reply #151 on: October 01, 2020, 11:56:29 pm »
I expect you mean a 40Mb/s LVDS signal. The maximum frequency will be much higher, dependent solely on the transition time.
Well, one of the LVDFS channels is a 40MHz clock. So since when is frequency only a valid unit for sine waves? Of course the sharp edges have much higher frequency components. No need to mention that in this board, do we?

Yes, it is frequently (ho ho) necessary to mention it on this board - including in this case.

I mentioned the abbreviated version. For a longer version see https://entertaininghacks.wordpress.com/2018/05/08/digital-signal-integrity-and-bandwidth-signals-risetime-is-important-period-is-irrelevant/


Quote
It is unsurprising you see a distorted waveform, given that a *10 "10Mohm" impedance probe has a (capacitive) input impedance of <100ohms at the frequencies you are likely to see on an LVDS signal.
For a 11pF probe, a 400MHz component will "see" a resistance of ~36Ohm, while the DC components will "see" 10MOhm. How is it surprising that this distorts the waveform?
Anyway, have you ever actually tried to measure a 40MHz square wave with a passive probe?

Er, I noted distortion is unsurprising. It appears that you had not considered the way in which it will affect the waveform.

I have frequently (ho ho) measured such waveforms with passive probes over the decades.


Quote
A pair of resistive divider Z0 probes would be a better bet, since they have a much lower capacitance and a well defined input resistance of 500ohms (*10) or 1000ohms (*20). They are easy to make at home.
A typical active probe has like 0.5pF capacitance, so anything you attempt to attach it somewhere will actually add a capacitance in the same magnitude or higher. I don't see how a purely resistive probe would help there. Plus, its low resistance makes it problematic for digital outputs which might as well create a burst clock only now and then.

My resistive divider probes are "<0.7pF", so yes it does help considerably compared with the typical *10 passive probe you used[1].

This is nothing new; I have an "operating note" dated 1972 for my hp10020a resistive divider probes.

[1] Here's the context you chose to snip...
...Compared to the default 500MHz passive LeCroy probes, the difference is like night and day. I.e. with the passive probes, the signal looks like a distorted sine and with the active probes, it's a nearly perfectly rectangular signal....
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #152 on: October 02, 2020, 12:27:55 am »
Yeah, I really don't know if this brought up anything worthwhile. I can't day it did for me. Hope you feel better at least.
Trying is the first step towards failure - Homer J. Simpson
 

Online Mechatrommer

  • Super Contributor
  • ***
  • Posts: 11537
  • Country: my
  • reassessing directives...
Re: > 1 GHz DIY differential probes
« Reply #153 on: October 02, 2020, 06:25:06 am »
My resistive divider probes are "<0.7pF", so yes it does help considerably compared with the typical *10 passive probe you used[1].

This is nothing new; I have an "operating note" dated 1972 for my hp10020a resistive divider probes.

[1] Here's the context you chose to snip...
...Compared to the default 500MHz passive LeCroy probes, the difference is like night and day. I.e. with the passive probes, the signal looks like a distorted sine and with the active probes, it's a nearly perfectly rectangular signal....
while Z0 probe is the cheapest way, it doesnt help with high impedance nodes such as 1Kohm lines (regardless of frequencies low or high)... most logics will simply follow the distorted clock fine, but some other circuit will be screwed, maybe like comparator or serial comm/timing that needs strict synching... ymmv.

since at it, here is a 50ohm 40MHz clock line loading effect i probed with my diy 1/10X passive probe vs Rigol bundled probe (DS1054Z grade BW) last night just out of curiosity. 1st and 2nd picture is for reference only (unloaded signal on 1GSa/s 1 channel activated only vs 500MSa/s both channel activated) granted most of high frequency content of my 6GHz function (clock) Gen already absorbed by DSO capacitance and phase distortion (delay or synch) is not so obvious from this snapshots. anyway, for 1Kohm nodes such as usually found in mcu output, Z0 probe will not be suitable however slow the clock is, this is where active probe is more of a general purpose probe ie we dont need to keep track of using different probes...
« Last Edit: October 02, 2020, 06:31:53 am by Mechatrommer »
Nature: Evolution and the Illusion of Randomness (Stephen L. Talbott): Its now indisputable that... organisms “expertise” contextualizes its genome, and its nonsense to say that these powers are under the control of the genome being contextualized - Barbara McClintock
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: > 1 GHz DIY differential probes
« Reply #154 on: October 02, 2020, 08:03:50 am »
Yeah, I really don't know if this brought up anything worthwhile. I can't day it did for me. Hope you feel better at least.

There's an English proverb: "you can lead a horse to water, but you can't make it drink".

Perhaps it will avoid other readers being mislead.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline Noy

  • Frequent Contributor
  • **
  • Posts: 361
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #155 on: October 02, 2020, 01:03:44 pm »
I think i will respin the opa probe.
Even 4 active probes will be cheaper than one of the diffs here. I will try.

Plan is to use the wson package of the opa (maybe a bit better choice than sot23) and to add a sma connector + micro usb for power supply.

Is a LTC1044 useable for the OPA? With lots off caps + filtering? To get from 5V (maybe a powerbank / oszi output?) to +-5V for the opa..
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #156 on: October 02, 2020, 01:57:26 pm »
I used the LT1054 (which is an improved/beefier version of the LTC1044/ICL7660) to to create +/-5V from ~7V and the MC79M05/MC78M05 for stable +/-5V. There is quite a bit of supply voltage filtering done on the probe itself, so this should be OK. However, creating -5V from +5V without any step up converted seems a bit optimistic. The negative voltage created with such an inverting charge pump will not have the exact same amplitude and the voltage drop increases with the load current.
For the record: I converted the eagle schematics and PCB design to DipTrace some years ago if this would help.
« Last Edit: October 02, 2020, 02:00:44 pm by 0xdeadbeef »
Trying is the first step towards failure - Homer J. Simpson
 

Offline JohnGTopic starter

  • Frequent Contributor
  • **
  • Posts: 567
  • Country: us
Re: > 1 GHz DIY differential probes
« Reply #157 on: October 02, 2020, 02:08:09 pm »
I think i will respin the opa probe.

This OPA probe looks nice, and well documented, so thanks for pointing it out. The documentation looks especially good, but my German is preschool level, so I will need to run it through a translator.

However, keep in mind that two single-ended probes is not the same as a differential probe. Aside from the obvious, like using more scope channels (I fall in the camp where 4 is almost enough), it is very difficult to get good common-mode rejection at high frequency. Scope channels rarely have good accuracy, and if the cables and probes are not the same at all frequencies, this eats into input CM rejection. For lots of small-signal stuff, this might not matter, but if you are troubleshooting in an electrically noisy environment, it can make a big difference.

Just my $0.02,
John
"Reality is that which, when you quit believing in it, doesn't go away." Philip K. Dick (RIP).
 

Offline Noy

  • Frequent Contributor
  • **
  • Posts: 361
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #158 on: October 02, 2020, 03:02:26 pm »
I have a PCB from this project:
https://xellers.wordpress.com/electronics/1ghz-active-differential-probe/

Still in my "build in future" box. But currently i do not find the little bag with the already purchased parts.. Maybe i will order them again with the other parts for the Opa Probe... Then i have both.
Its not a 1.4G but i think enough for my scope.
Maybe i will rework this design (4 layer, jlcpcb, 0402 parts, micro USB) together with the opa design.


Found same nice parts: LM27762 (+ pre boost converter?? Or only round about +-4.5V for the opa) / TPS65133
« Last Edit: October 03, 2020, 08:20:19 pm by Noy »
 

Offline teomondo

  • Newbie
  • Posts: 5
  • Country: it
Re: > 1 GHz DIY differential probes
« Reply #159 on: October 30, 2020, 06:02:14 am »
Dear ntcnico
Is DIP1400 1.4GHz differential probe currently avaliable?
I have an Operating manual dated 2019 any new version avaliable?
What about the price of DIP1400 1.4GHz differential probe and how I can buy it.
Thanks for your answer
Roberto
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26757
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #160 on: October 30, 2020, 10:36:27 am »
Dear ntcnico
Is DIP1400 1.4GHz differential probe currently avaliable?
I have an Operating manual dated 2019 any new version avaliable?
What about the price of DIP1400 1.4GHz differential probe and how I can buy it.
The probes are a standard product so usually in stock; the manual hasn't changed. The price is 125 euro ex. VAT and ex. shipping. Please PM me for more details.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Noy

  • Frequent Contributor
  • **
  • Posts: 361
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #161 on: November 14, 2020, 10:58:23 am »
Your datasheet states 2k ohm impedance at DC between the inputs.
Do you have a chart where the input impedance is shown over the full bandwith?
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26757
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #162 on: November 15, 2020, 04:33:53 pm »
Your datasheet states 2k ohm impedance at DC between the inputs.
Do you have a chart where the input impedance is shown over the full bandwith?
No. But 2k parallel with 1pf is likely a good worst case assumption.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: JohnG

Offline JohnGTopic starter

  • Frequent Contributor
  • **
  • Posts: 567
  • Country: us
Re: > 1 GHz DIY differential probes
« Reply #163 on: April 13, 2021, 12:38:22 am »
Nico's probe does what it says. Even nicer, I can power it right from one of my scope's USB ports. I am a happy customer! Already solved a problem for me.

Nico, if you design one with higher input impedance, I would likely get one of those, too... Just sayin' ;)

Cheers,
John
"Reality is that which, when you quit believing in it, doesn't go away." Philip K. Dick (RIP).
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: > 1 GHz DIY differential probes
« Reply #164 on: April 13, 2021, 01:57:14 pm »
Nico,

Are these differential probes still available?

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26757
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #165 on: April 13, 2021, 05:17:10 pm »
Nico,

Are these differential probes still available?
Yes but I'm currently out of stock; a new batch is being produced.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline beardedhustler

  • Newbie
  • Posts: 4
  • Country: no
Re: > 1 GHz DIY differential probes
« Reply #166 on: June 04, 2021, 11:52:09 am »
Sorry for hijacking the thread, but since it's a bit related and I eventually will look into the topic (differential design) of this thread as well I hope it's ok.

We have a Signatone probing station with (passive) "micropositioners" (probes) that we use to probe on bonding pads on various chips and read data with some reader attached. Coaxial cables from probes to reader are ~1m, so signal integrity isn't always great when the frequency increases, thus the need to add an active stage/probe. For now we've mostly dealt with various SPI flash chips, eMMC chips in 1-bit bus width mode, up to ~36MHz clk speed (I got enough probes that I can go for the 4-bit bus width mode which supports clk speeds up to 200MHz if I want, but it's not always beneficial to spend time and energy trying to place six probes within the small area where the bonding pads are located and at the same time not break any bonding wire between the pads and the controller, etc, so usually I stick to using a 1-bit bus width..).

Due to the possibility of using the HS200 mode for eMMCs I think a 1GHz BW is an ok goal to aim for here, and since eMMCs use single ended signals I don't have the need for a differential probe (for now - that might change later on if we decide to try probing UFS chips, but that's another story - we'll cross that bridge when we get to it...). Another important thing to mention is that the goal here is to make the pcb that we can custom fit on the micropositioners in some way to utilize what we already have and thus buying a general purpose probe isn't an option for this now.

I've started on a few single-stage designs trying out various op-amps from TI (LMH3401, LHM6702, OPA855/OPA858), and getting a BW > 1GHz seems quite straight forward with the op-amps mentioned, but I'm a bit unsure on the pitfalls when it comes to designing a active "probe". I guess trying to get as high input impedance / low input capacitance should be a goal as well. Anything else to consider other than trying to impedance match output from op-amp to 50Ohm (coax cable) to maximize power transfer?

It's been a few years since I did design and back then it was mainly high power RF PAs with different things to consider during the design process. I've only used TI's spice simulator for now, and although it's a nice tool I guess it has its limitations. I tried importing some of the spice models they offer on their web page into Keysight ADS, and although I get almost identical results for similar simulations I'm not sure if e.g. the models are valid for other simulations, e.g. S-parameter simulations, which for some devices result in S11>0dB for a wide frequency range (and in my experience from the RF PA design days that's a big no-no).

Suggestions to components, literature, etc, is greatly appreciated!
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 19281
  • Country: gb
  • Numbers, not adjectives
    • Having fun doing more, with less
Re: > 1 GHz DIY differential probes
« Reply #167 on: June 04, 2021, 03:13:43 pm »
Sorry for hijacking the thread, but since it's a bit related and I eventually will look into the topic (differential design) of this thread as well I hope it's ok.

You would probably get better responses if you started a new thread with your questions. Plus, if the responses are useful, other people are more likely to find them and benefit from them.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline beardedhustler

  • Newbie
  • Posts: 4
  • Country: no
Re: > 1 GHz DIY differential probes
« Reply #168 on: June 07, 2021, 06:19:41 am »
Sorry for hijacking the thread, but since it's a bit related and I eventually will look into the topic (differential design) of this thread as well I hope it's ok.

You would probably get better responses if you started a new thread with your questions. Plus, if the responses are useful, other people are more likely to find them and benefit from them.

You're probably right; I'll make a new thread about it.

Edit: new thread up at https://www.eevblog.com/forum/projects/input-to-single-ended-probe-designs/

Any suggestions are greatly appreciated!
« Last Edit: June 07, 2021, 06:59:29 am by beardedhustler »
 

Offline LooseJunkHater

  • Frequent Contributor
  • **
  • Posts: 283
  • Country: de
Re: > 1 GHz DIY differential probes
« Reply #169 on: December 24, 2023, 08:01:37 pm »
Nico,

Are these differential probes still available?
Yes but I'm currently out of stock; a new batch is being produced.

Do you have any intention of making some high voltage (500v+) differential probes?
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 26757
  • Country: nl
    • NCT Developments
Re: > 1 GHz DIY differential probes
« Reply #170 on: December 25, 2023, 03:59:33 pm »
Nico,

Are these differential probes still available?
Yes but I'm currently out of stock; a new batch is being produced.

Do you have any intention of making some high voltage (500v+) differential probes?
Not at this moment. There are enough high voltage diff. probes on the market for reasonable prices. And such a product should also have a CAT rating which adds safety testing as an extra development cost.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf