Author Topic: DIY Differential probe again again - please help :-)  (Read 6006 times)

0 Members and 1 Guest are viewing this topic.

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
DIY Differential probe again again - please help :-)
« on: October 05, 2019, 03:37:51 am »
I have now read several super long treads about the subject on this great forum, but I simply need some help to know how to start and to find something that do work. There is a lot of ideas and discussions and diagrams all over but it always end up s an great start but never usable or something else that stop the project in it's tracks.

What I am hoping fore is a "between" board that I can couple between my scope and my standard probe.
500MHz no need for over 65V



The image is from this site where all components are listed but you have to pay for the article: https://www.elektormagazine.com/magazine/elektor-201409/27026
I have found this article: http://blog.weinigel.se/2016/02/26/ghz-differential-probe.html
And this wary long spec sheet for the THS3201 amplifier: https://www.farnell.com/datasheets/1447056.pdf
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #1 on: October 05, 2019, 05:58:52 am »
The image is from this site where all components are listed but you have to pay for the article
This is low ~60KHz bandwidth probe isolation adapter.
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1530
  • Country: gb
Re: DIY Differential probe again again - please help :-)
« Reply #2 on: October 05, 2019, 08:44:05 am »
A simple free approach (depends on scope) to differential probing is to
Use 2 channels on the scope
DO NOT use the ground clamp of the scope probes.
Use the math function of scope, subtract 1 channel from another.
Use one probe on one measurement point
Use the probe on the other measurement point.

The math function act like a differential function between 2 channels.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #3 on: October 05, 2019, 10:39:28 am »
The image is from this site where all components are listed but you have to pay for the article
This is low ~60KHz bandwidth probe isolation adapter.
Okay, thanks. I could not find that out just by looking at the image or components. If that had been a 500MHz would that have been enough? I mean, if all I want is to be able to measure where there is different grounds.
What about the second link, is that a usable design?
A simple free approach (depends on scope) to differential probing is to
Use 2 channels on the scope
DO NOT use the ground clamp of the scope probes.
Use the math function of scope, subtract 1 channel from another.
Use one probe on one measurement point
Use the probe on the other measurement point.

The math function act like a differential function between 2 channels.
Yes fine little trick but it have to be possible to build a probe instead of having to buy a probe or reducing my scope to half of the channels?

« Last Edit: October 05, 2019, 02:30:29 pm by FriedMule »
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #4 on: October 05, 2019, 02:35:04 pm »
Is it posible to make this Differential probe: http://blog.weinigel.se/2016/02/26/ghz-differential-probe.html to be able to handle about 65V?
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #5 on: October 05, 2019, 04:24:47 pm »
Is it posible to make this Differential probe: http://blog.weinigel.se/2016/02/26/ghz-differential-probe.html to be able to handle about 65V?
Yes indeed. Voltage divider at the input needed and maybe some circuit for offset error trimming needed. Note that schematics of final revision of that probe is not published. [edit] Github comment contains info about final modification which were done to achieve said performance which is great especially considering BOM cost of that thing.

Okay, thanks. I could not find that out just by looking at the image or components.
Main component is isolation amplifier, AMC1200. Bandwidth is specified in the datasheet: http://www.ti.com/lit/ds/symlink/amc1200.pdf

Quote
If that had been a 500MHz would that have been enough?
Even 10KHz BW probes have their use. What is your application that you are looking for 500MHz BW? What scope do you have, what is bandwidth of it?
« Last Edit: October 05, 2019, 04:38:59 pm by ogden »
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #6 on: October 05, 2019, 06:58:12 pm »
Is it posible to make this Differential probe: http://blog.weinigel.se/2016/02/26/ghz-differential-probe.html to be able to handle about 65V?
Yes indeed. Voltage divider at the input needed and maybe some circuit for offset error trimming needed. Note that schematics of final revision of that probe is not published. [edit] Github comment contains info about final modification which were done to achieve said performance which is great especially considering BOM cost of that thing.

Okay, thanks. I could not find that out just by looking at the image or components.
Main component is isolation amplifier, AMC1200. Bandwidth is specified in the datasheet: http://www.ti.com/lit/ds/symlink/amc1200.pdf

Quote
If that had been a 500MHz would that have been enough?
Even 10KHz BW probes have their use. What is your application that you are looking for 500MHz BW? What scope do you have, what is bandwidth of it?
I am currently trying to play with this circuit, and hope tinker with related projects:-) http://www.foxdelta.com/projects/pm1.htm
The reason for the voltage is that I hope to use it for more then just that one project, and I know I will work with projects in the 60V range
So a probe where I can just probe when I need to, and not have to build an other later for that voltage.

May I ask if you please could guide / draw something that could be used to allow 65V?
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #7 on: October 05, 2019, 07:20:50 pm »
May I ask if you please could guide / draw something that could be used to allow 65V?

Indeed you may ask. I already did tell you - you need voltage divider. You may skip offset trimming at the beginning. Question "could you give me solution on silver plate" is one I would simply answer "definitely no". So please do your research and ask good question which is worth to answer.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #8 on: October 05, 2019, 07:52:34 pm »
May I ask if you please could guide / draw something that could be used to allow 65V?

Indeed you may ask. I already did tell you - you need voltage divider. You may skip offset trimming at the beginning. Question "could you give me solution on silver plate" is one I would simply answer "definitely no". So please do your research and ask good question which is worth to answer.
LOL sorry if I sounded like I did ask for a silver plate and you build it to:-)
What I men was a drawing ehm... is it called a block diagram? On how you would mount it, I presume a simple voltage divider with two resistors? 1:2 / 1:10 or? But before the offset trimming? and then the probe diagram?
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #9 on: October 05, 2019, 08:05:38 pm »
What I men was a drawing ehm... is it called a block diagram?
Yes, I got it. You want diagram? Then draw it.

Quote
I presume a simple voltage divider with two resistors? 1:2 / 1:10 or?
Who knows. You want me to give exact divider ratio? This is what I call providing something on silver plate. First you need to know input voltage range of the existing design, gain and output range as well.  After that you will start to get clue about  voltage divider. BTW simple, two resistor divider won't work on 500MHz probe. It (divider) shall be frequency-compensated.

BTW you expect me to answer your questions, yet you ignored to answer mine - about your scope and it's bandwidth.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #10 on: October 05, 2019, 08:33:30 pm »
What I men was a drawing ehm... is it called a block diagram?
Yes, I got it. You want diagram? Then draw it.

Quote
I presume a simple voltage divider with two resistors? 1:2 / 1:10 or?
Who knows. You want me to give exact divider ratio? This is what I call providing something on silver plate. First you need to know input voltage range of the existing design, gain and output range as well.  After that you will start to get clue about  voltage divider. BTW simple, two resistor divider won't work on 500MHz probe. It (divider) shall be frequency-compensated.

BTW you expect me to answer your questions, yet you ignored to answer mine - about your scope and it's bandwidth.
Sorry trust me I am not asking for any silver platter, every time it sounds like that, please blame my missing language skills! There is no fun in electronic if you either buy it finished or make someone do it for you, but also no fun in getting lost way over your head:-)
Sorry forgot about your scope question: it's a 2-Channel 500MHz LeCroy LT342L
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #11 on: October 05, 2019, 09:27:30 pm »
Sorry forgot about your scope question: it's a 2-Channel 500MHz LeCroy LT342L

Good. Now we at least know that you possibly/actually may need what you ask for. Sorry for my seemingly non-forthcoming attitude as well - we are tired of those who come and expect somebody else do their homework. When you show that you are seeking answers yourself and need just guidance or even better - confirmation that you are on right track, you will get increasingly better feedback from others.

For example you just blindly guessed two values of divider out of thin air w/o showing that you invested your time finding it (divider ratio). Do you think that seeing that I want to invest my time in return to solve problem which is not actually mine? Hint: not at all.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #12 on: October 05, 2019, 09:43:06 pm »
Sorry forgot about your scope question: it's a 2-Channel 500MHz LeCroy LT342L

Good. Now we at least know that you possibly/actually may need what you ask for. Sorry for my seemingly non-forthcoming attitude as well - we are tired of those who come and expect somebody else do their homework. When you show that you are seeking answers yourself and need just guidance or even better - confirmation that you are on right track, you will get increasingly better feedback from others.

For example you just blindly guessed two values of divider out of thin air w/o showing that you invested your time finding it (divider ratio). Do you think that seeing that I want to invest my time in return to solve problem which is not actually mine? Hint: not at all.
You are completely correct, I have, on an other subject and forum, med many who did expect everything on a silver... hmm no say gold platter and was annoyed that no one did jump at the chance. So I do understand your reaction. :-)
My 1:2 / 1:10 was as to ask about if you did mean something like a 1:2 or 1:10, not thinking of some certain number but only if you meant a:b resistor thingy.
I know that it would be stupid of me to start recherching the wrong place, so I'd rather ask again and maybe sound stupid, then waste your or my time:-)

EDIT: By the way, I do really appreciate that you tell me how my bad English is being read! It helps me so much and I have to try to be better to formulate what I want to say, so that my text really do express what I mean. So many many thanks!! :-)
« Last Edit: October 05, 2019, 10:04:24 pm by FriedMule »
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #13 on: October 05, 2019, 10:20:31 pm »
My 1:2 / 1:10 was as to ask about if you did mean something like a 1:2 or 1:10, not thinking of some certain number but only if you meant a:b resistor thingy.
Yes. Two resistors. Like common 1:10 scope probe. Thou to reach 65 volts with that design, most likely other ratio needed. That's your homework to investigate and try to think of ratio and actual value of those.

Quote
EDIT: By the way, I do really appreciate that you tell me how my bad English is being read!
I think you are doing fine.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #14 on: October 05, 2019, 11:13:31 pm »
My 1:2 / 1:10 was as to ask about if you did mean something like a 1:2 or 1:10, not thinking of some certain number but only if you meant a:b resistor thingy.
Yes. Two resistors. Like common 1:10 scope probe. Thou to reach 65 volts with that design, most likely other ratio needed. That's your homework to investigate and try to think of ratio and actual value of those.

Quote
EDIT: By the way, I do really appreciate that you tell me how my bad English is being read!
I think you are doing fine.
Yes now comes the fun part, first do I have to find out the max voltage the probe can handle safely, then the devider, I hope:-)
Thank you so much for helping me in so many ways!

EDIT: oh forgot, and this may I need on a silver platter:-)
I know that the author has only tried the B variant, but the variants are way over my head, is it possible for me, in any way, to decide what variant is best for my usage or can you see any obvious candidate?
« Last Edit: October 05, 2019, 11:25:30 pm by FriedMule »
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline ogden

  • Super Contributor
  • ***
  • Posts: 3731
  • Country: lv
Re: DIY Differential probe again again - please help :-)
« Reply #15 on: October 06, 2019, 05:37:27 am »
One author actually finished, B
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: DIY Differential probe again again - please help :-)
« Reply #16 on: October 06, 2019, 12:27:16 pm »
Option B is just a standard instrumentation amplifier (IA) circuit made out of 3 individual op-amps. You can buy an IA as a single IC but it won't have a very high gain*bandwidth product - maybe 10MHz at unity gain. Using individual high speed (and looks like current-feedback here) op-amps makes it easier - though not easy - to achieve an arbitrarily high bandwidth.

Note that the voltage divider at each input of a high bandwidth differential probe needs to be precise both at DC, which is easy - just choose the correct resistor values - and across the entire frequency range, which is deceptively difficult to pull off, especially if you need >2:1 attenuation and >10MHz bandwidth.

You want something like 500MHz bandwidth which implies the probe having a -3dB bandwidth that is at least Pi times higher (and ideally closer to 10x). This is just enormously difficult to do and commercially available active differential probes with >2GHz bandwidth have 5 digit prices, and I don't mean in Yen or Yuan. You need to set your sights much lower or else buy a probe outright, or, perhaps, copy a successful design *exactly*.



 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #17 on: October 06, 2019, 01:05:34 pm »
Option B is just a standard instrumentation amplifier (IA) circuit made out of 3 individual op-amps. You can buy an IA as a single IC but it won't have a very high gain*bandwidth product - maybe 10MHz at unity gain. Using individual high speed (and looks like current-feedback here) op-amps makes it easier - though not easy - to achieve an arbitrarily high bandwidth.

Note that the voltage divider at each input of a high bandwidth differential probe needs to be precise both at DC, which is easy - just choose the correct resistor values - and across the entire frequency range, which is deceptively difficult to pull off, especially if you need >2:1 attenuation and >10MHz bandwidth.

You want something like 500MHz bandwidth which implies the probe having a -3dB bandwidth that is at least Pi times higher (and ideally closer to 10x). This is just enormously difficult to do and commercially available active differential probes with >2GHz bandwidth have 5 digit prices, and I don't mean in Yen or Yuan. You need to set your sights much lower or else buy a probe outright, or, perhaps, copy a successful design *exactly*.
Okay yes it makes a bit harder, I have tried to find sub 1GHz probe designs and I have also found some, but I do not know how good they are. If you look at the design I have chosen here, it should work fine up to about 800MHz but only in super low voltage. The design you can buy are for the most for high voltage and yes costs 4 times my scope.
I have read every comment in this tread: https://www.eevblog.com/forum/testgear/gt-1-ghz-diy-differential-probes/
Found other probes like this: https://xellers.wordpress.com/electronics/1ghz-active-differential-probe/

But my "wall" is simply that I do not have enough knowledge to build my own, so I have to, as you say, copy a design.
But untill now, I have not found a design that your great guy's says do work.
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: DIY Differential probe again again - please help :-)
« Reply #18 on: October 06, 2019, 01:51:25 pm »
Okay yes it makes a bit harder, I have tried to find sub 1GHz probe designs and I have also found some, but I do not know how good they are. If you look at the design I have chosen here, it should work fine up to about 800MHz but only in super low voltage. The design you can buy are for the most for high voltage and yes costs 4 times my scope.

I (and ogden) have already told you this is an enormously difficult task, and your only hope - hell, my only hope! - would be to copy someone else's design exactly. I mean, the same components, board layout, physical construction, every last detail.

This is not the sort of problem a beginner is likely to run into, nor even the sort of problem any but RF and high-speed digital design engineers will ever have to contend with, frankly, so I have to ask why you think you need a >500MHz differential probe? Earlier you wrote this:

...If that had been a 500MHz would that have been enough? I mean, if all I want is to be able to measure where there is different grounds....

But that isn't an answer. Suffice it to say that there are precious few instances where an RF or high speed digital signal is not supplied by an isolated power supply, so there isn't a safety requirement for needing a differential or isolated probe. There are plenty of examples of high speed differential digital signal standards such as Ethernet, LVDS, HDMI, SCSI, etc., but, again, not the sort of thing the typical EE will ever work on at the signal level, much less a beginner or hobbyist.

So what, exactly, do you intend to probe with an oscilloscope that has convinced you you need a really expensive differential probe good for (much greater than) 500MHz?
 

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7549
  • Country: 00
  • +++ ATH1
Re: DIY Differential probe again again - please help :-)
« Reply #19 on: October 06, 2019, 02:00:52 pm »
A 500MHz differential probe capable of < 65Volt ... cheap ... DIY & relatively easy to build ... just nope.

Suggesting to change plan.

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #20 on: October 06, 2019, 03:29:50 pm »
You are all a great bunch of guys!! :-)
Okay I did not know that my project was so outrageous, if it is a futile goal for me right now, then let be refrace my question then. What would be a doable probe attachment that can handle up to 65V with isolated ground and ???MHz?
Functioning but not necessary looking like this:

Could I use either: http://blog.weinigel.se/2016/02/26/ghz-differential-probe.html
Or: https://xellers.wordpress.com/electronics/1ghz-active-differential-probe/
As template, they both have circuit design and component list.
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: DIY Differential probe again again - please help :-)
« Reply #21 on: October 06, 2019, 03:41:39 pm »
...
Could I use either...

You want us to spend considerable time assessing projects for you but still haven't answered the most basic question which has been asked several times already: What Are You Trying To Measure?
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #22 on: October 06, 2019, 03:53:11 pm »
...
Could I use either...

You want us to spend considerable time assessing projects for you but still haven't answered the most basic question which has been asked several times already: What Are You Trying To Measure?
Sorry if I have not explained that in a good enough manner, I did try at my comment number 6: https://www.eevblog.com/forum/beginners/diy-differential-probe-again-again-please-help-)/msg2725412/#msg2725412
But please ask any question I may not have answered.
« Last Edit: October 06, 2019, 03:55:28 pm by FriedMule »
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 7044
  • Country: nl
Re: DIY Differential probe again again - please help :-)
« Reply #23 on: October 06, 2019, 04:57:34 pm »
Coax is inherently asymmetric, that makes a differential measurement pretty much impossible at high frequencies with a single oscilloscope probe. A 500 MHz amplifier which can directly take +/- 65 V on the inputs without dividers is implausible too, so you need dividers.

So you need either a small probe with 2 pins directly going to dividers and then a differential amplifier. Or two divider probes going to a differential amplifier. For the divider probes you can have oscilloscope probes, or coax divider probes.

The only existing design which somewhat meets your requirements is Smith's balanced coax probe using a power combiner. But it can't work down to DC, in fact you probably need to add blocking capacitors so the common mode doesn't blow the combiner. You could hybridize this, by adding two extra wires and using a normal low frequency differential probe (ie. high impedance dividers with an instrumentation amplifier) for the low frequencies and Smith's probe for the high frequencies.
 

Offline FriedMuleTopic starter

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: dk
  • Can make even the simplest task look imposible.
Re: DIY Differential probe again again - please help :-)
« Reply #24 on: October 06, 2019, 05:23:37 pm »
Coax is inherently asymmetric, that makes a differential measurement pretty much impossible at high frequencies with a single oscilloscope probe. A 500 MHz amplifier which can directly take +/- 65 V on the inputs without dividers is implausible too, so you need dividers.

So you need either a small probe with 2 pins directly going to dividers and then a differential amplifier. Or two divider probes going to a differential amplifier. For the divider probes you can have oscilloscope probes, or coax divider probes.

The only existing design which somewhat meets your requirements is Smith's balanced coax probe using a power combiner. But it can't work down to DC, in fact you probably need to add blocking capacitors so the common mode doesn't blow the combiner. You could hybridize this, by adding two extra wires and using a normal low frequency differential probe (i.e. high impedance dividers with an instrumentation amplifier) for the low frequencies and Smith's probe for the high frequencies.
Okay, thanks for the link, I hav not read it all yet but can see that it do not contain any diagram, so it may be great to read but maybe also hard for me to copy from, like the other say I have to, to get a little chance for something useful. :-)

hmm... What if I use my standard 10:1 probe infront of the probe circuit? That will give about 6.5V and then make a divider that I have to calculate (I know that it may not be easy) and maybe end up with a 10-100MHz. I shall not pretend that I know what a coaxial divider is, but that is only fun to find out:-)
Even if I appear online is it not necessary so, my computer is on 24/7 even if I am not on.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf