Author Topic: Need help with schematic and PCB for instrumentation (?) amplifier  (Read 1081 times)

0 Members and 1 Guest are viewing this topic.

Offline cooleb09Topic starter

  • Newbie
  • Posts: 4
  • Country: au
Hi   

I have seen a number of people making differential probes for their 'scope for significantly cheaper than the purchased ones, for use with  hobbyist level projects, including [this one][/https://xellers.wordpress.com/electronics/1ghz-active-differential-probe/] and [this one][/http://blog.weinigel.se/2016/02/26/ghz-differential-probe.html], and have decided I won't to make my own.   

However, unlike these projects I was hoping that I could attached some standard passive probes to the thing instead of relying on the card-edge/bodge wire thing that these designs have used.   

Additionally, I wanted something scaleable so that when I (hopefully soon) get a rigol ds1054z I can use it on all the channels.   

I have thrown together the schematic/pcb ([available here][/https://www.dropbox.com/sh/v9gne9ihuqbs8h6/AACF3gfhZ-c7xxGUBDKdh7m6a?dl=0]), but to be honest, it ain't pretty.   

Can anyone help me with getting this a bit more tidied up/organised/advice on component choice/overall design feasibility/etc?     
It would only need to be usable for a couple of MHz.   

Some thing that jump out at me are that the traces/layout between the channels aren't exactly the same - I couldn't find a way to do that in eagle.   

Also, I couldn't hide some parts of the PSU from the schematic in the PCB (although I can understand why this would be a limitation, is there a way to mark these parts as 'not on this board' in eagle?)   
I would be having the unit brought to mains-earth reference via the scope's ground terminals (to avoid ground loop issues), is this a bad idea?)   

Are there different OP-amps that might fit this project a bit better (perhaps less bandwidth, but cheaper and in a larger package?).   

Can I remove the input resistors and just rely on the probes input impedance (one of the designs from earlier noted that reducing their value could lead to some imporvement).   

Is this even vaguely feasible for a hobby project?
 

Offline Brumby

  • Supporter
  • ****
  • Posts: 12288
  • Country: au
Re: Need help with schematic and PCB for instrumentation (?) amplifier
« Reply #1 on: November 24, 2017, 05:57:34 am »
What things are you going to probe with these ... particularly the voltages involved?
 

Offline cooleb09Topic starter

  • Newbie
  • Posts: 4
  • Country: au
Re: Need help with schematic and PCB for instrumentation (?) amplifier
« Reply #2 on: November 24, 2017, 06:14:23 am »
Its mostly so that I can track voltages between arbitrary points in a circuit, moslty LV stuff (+/- 12V of gnd), I'm guessing I'm going to need a 10x probe though (rails on the amps are +/- 5).   
Some of my upcoming plans are making some servo drivers and an ESC/BLDC driver, but in general I think it would just be a useful thing to have (save needing to float things when debugging).
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16544
  • Country: us
  • DavidH
Re: Need help with schematic and PCB for instrumentation (?) amplifier
« Reply #3 on: November 24, 2017, 04:50:22 pm »
However, unlike these projects I was hoping that I could attached some standard passive probes to the thing instead of relying on the card-edge/bodge wire thing that these designs have used.

This is absolutely feasible however matching between the probes or lack thereof will limit performance.  The Tektronix 100 MHz 7A13 and 1 MHz 7A22 differential vertical amplifiers worked this way and used probes with trimmable DC and AC characteristics like the Tektronix P6055, P6055A, and P6135A.  The designer(s) of the 7A13 later founded Preamble and made external differential amplifiers which can be used with any oscilloscope.  LeCroy bought Preamble and still makes them along with trimmable probes.

If you design something like this, then the BNC inputs need to duplicate the performance of controlled capacitance 1 megohm inputs as found on an oscilloscope for the probes to work properly.  At high frequencies, this requires a pair of high input impedance buffers.  At low frequencies, a JFET or MOSFET differential feedback amplifier can be used.

If I were doing this, I would include input networks to trim the DC and AC characteristics of each probe however this is not a trivial undertaking.  Standard FR4 type circuit board substrates are not generally suitable for these types of circuits do to "hook" which explains why so many high voltage differential probes drift over time resulting in abysmal common mode rejection.
 
The following users thanked this post: cooleb09


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf