Measure fast differential signals (100MHz) using instrumentation amp

[crutschow]

I presume, DC common mode won't require differential measurement at all, just a trace offset or AC mode. I was expecting high frequent CM signals, but the OP didn't tell exactly.
............................

If it's a true differential signal (opposite polarity signals on each line) then it still needs a differential measurement, regardless of any common-mode noise.

- - - Updated - - -

Thanks that's what i'll do

For good performance at those frequencies you will need a careful layout of the circuit on a board with a ground plane. Use short direct leads for all the high frequency signals, and 0.1uF decoupling caps with leads as short as possible directly from the IC power pins to the ground plane.

 

[D.A.(Tony)Stewart]

Choose a gain of 6dB for all stages for 500MHz signal with 1.8GHz GBW chip and slew rate spec will limit output level.

This may yield 5.5 dB gain per stage. Ground plane with 50 Ohm striptline to 50R out into coax , terminated with 50R at scope is required.

 

[levnu]

i have another implementation question:
i to make a PCB w/ the THS3202 configuration from above .... but i have only one around 6V power supply ... i need to provide also the -6V thus i though about
TI inverter:
https://www.ti.com/lit/ds/symlink/tps60400.pdf

- But this inverter can provide only 60mA max current and i'm not sure if i'll be able to drive the 3 THS3202 op amps .... what will be the -6V current consumption by the op amps?
- How can i get this info from the data sheet ... what parameter?

https://www.ti.com/lit/ds/symlink/ths3202.pdf

 

[FvM]

Maximum quiescent current is specified as 20 mA per amplifier. But you have to consider also the current flowing into the feedback resistors and more important, into the output load. All-in-all, the 60 mA charge-pump inverter is probably too weak.

 

[levnu]

this i was afraid off ... what can i do ? maybe place few inverters from the same 6V supply?

- - - Updated - - -

i have found next TI DC to DC inverter do you think it can be good enough??
http://www.ti.com/lit/ds/symlink/tps63700.pdf

 

[FvM]

I have found next TI DC to DC inverter do you think it can be good enough??

Something like this can work. TPS63700 isn't specified for +6V input, however. Good bypassing and input + output chokes for the switching converter are suggested to reduce noise, also a shielded inductor and good layout.

I wonder if you wouldn't better start with a conventional linear regulated power supply.

 

[levnu]

I can make a 5V input to the TPS63700 from the 6V from my Power supply ...
My main issue is how to provide the Output of -5V & 320mA ....... that's why i choose TPS63700
Is it possible to input 5V and to get a -5V output w/ 320mA output current from a linear regulated power supply?

another consideration is that i want it to have a small pcb that has only one ~6V Power Supply input and from this 6V Power Supply i would like to provide all currents and required voltages in this circuit

Basically it suppose to be a cheap & simple differential probe.

 

[Borber]

Agilent has separate power supply for their differential probes. Why such solution is not good for you? How are you going to solder TPS63700?

 

[levnu]

regular diff probes are too expansive for me .. and i don't have such ... i want to imitate the diff probe on my own custom PCB
what do you mean how i'm gonna solder tps63700..?? i'm gonna do a small PCB and gonna send it to PCB assembly...according to the datasheet

 

[D.A.(Tony)Stewart]

I think your approach is too myopic for this design.

Always start with detailed input and output requirements and a packaging concept. 100MHz BW is insufficient as a spec because BW changes with slew rate and signal amplitude.

Commercial Diff. Probes have high impedance , very low capacitance and variable high gain and high bandwidth for the CMRR with a 50 Ω output.


Since Diff Probes have a wide application range, we have no idea of your objective.

Modern "simplified" specs look like this.

KEY PERFORMANCE SPECIFICATIONS

<100 ps rise time (P7350 guaranteed)
5 GHz to 3.5 GHz bandwidths (typical)
Low input capacitance: <0.3 pF differential (P7350 typical)
≥60 dB (1000:1) Common Mode Rejection Ratio (CMRR) at 1 MHz (P7330, P6330 typical)
KEY FEATURES

High-bandwidth, low-noise differential probing
Small probe head allows easy probing of SMDs
APPLICATIONS

Communications (Gigabit Ethernet, Fibre Channel, InfiniBand)
Semiconductor characterization and validation
PCI Express
Serial ATA
IEEE 1394
USB 2.0 (The P7350 is not recommended for USB 2.0 compliance testing)
MIPI® D-PHY
DDR

 

[FvM]

You'll hardly find a commercial active probe that has a switched mode power supply integrated with the amplifier. Modern oscilloscopes have +/-5 and +/-12V supply pins in the probe interface for it, older Tektronix scopes had a Lemo connector providing +/-15 V.

Of course you can't use a linear power supply to convert +6 to -6V or similar. But you can have a transformer supply, rectifier, filter capacitors and linear regulators.

Making a several 100 MHz BW amplfier with good waveform reproduction is demanding enough, you don't need to make your live hard (and put the amplifier performance at risk) by adding specific power supply features.