Built a Differential O-Scope Amp

I've been in need of a differential o-scope probe and due to their price, the cheapest I found was $350ish, I decided to build one as a learning exercise. Now I know that the bandwidth and the CMRR of this home-built diff probe is not up to par with a professionally manufactured one, but it does work for what I'm using it for. The amplifier box currently accepts 2 10X scope probes and in combination with the internal circuitry of the diff-amp the attenuation is 100x. I know this isn't ideal and I have 2 100x probes ordered and I will swap out the input resistors and caps on my diff amp to give a much higher input impedance and it will be proper 100x attenuation. The amp box also has a grounding post that is tied to the scope's ground so that I am able to take measurements from completely floating circuits(something battery powered) by connecting the grounds together, then the 2 probes can each be connected to points other than ground. If the device-under-test is already earth-referenced(ie: connected to the mains), then the grounding post on the diff-amp is left unconnected and each of the two probes may still be connected to points away from ground/earth. Attached is the schematic, the board I designed for the amp, and a picture of the output on my scope. The device creating the input is a TL494 PWM PSU controller powered from a 16VDC isolated supply. The ground post of my diff amp is tied to the 0V/GND rail of the isolated supply giving it an earth reference and then the two probes are each connected to an output of the TL494. Previously I was only able to measure 1 output at a time and was unable to see the dead-time between the two outputs where as now you can clearly see on the scope's screen that I am able to easily see the dead-time between the outputs.

The amp is powered by 2 9V batteries giving a +9V/-9V supply voltage to the opamp.

View attachment Differential Sch.pdf
View attachment Differential Board.pdf

Congratulations! To better evaluate your design, I would wish to see input and output side-by-side, also some frequency information. One good specification of such amplifier is the DC to maximum frequency response, etc.

The TL074 Op-Amp I used isn't very fast, only 13V/uS, so the frequency response isn't that high. I do have some THS3201 amps here that have a slew-rate of 6700V/uS so I'm going to try redesigning the circuit to see if I can improve the speed and response of the circuit. My scope only reads up to 5MHz and most of the stuff I work on only runs at 25-50KHz so I'm not too worried about the slower response.

Well although I have the THS3201 amps on hand, they are SMD components and I am unable to get the simulation model to work properly in MultiSim. I have found another op-amp with a high slew-rate that also has the advantage of having the same pin-out of the TL074, plus Multisim has the model for it built-in.

According to the datasheet, the LT1365CN is a voltage feedback op-amp with a slew-rate comparable to current feedback amps. It has a slew-rate of 1000V/uS and a quoted 70MHz gain bandwidth. In simulation I have ran it up to 5MHz without a significant difference in rise/fall-time between the input and output, but with a square-wave input there is a bit of ringing when the pulse goes high and when the pulse goes low. I'm currently trying to figure out what is causing this as it works fine with a sine-wave input(s).

GeorgesWelding said:

..................................

....................... In simulation I have ran it up to 5MHz without a significant difference in rise/fall-time between the input and output, but with a square-wave input there is a bit of ringing when the pulse goes high and when the pulse goes low. I'm currently trying to figure out what is causing this as it works fine with a sine-wave input(s).

Click to expand...

It may be ringing due to stray summing junction capacitance. Try a few pF to a few tens of pF of capacitance across R2 and R8. You need to experimentally determine the correct value to minimize the ringing without significantly affecting the frequency response. Whatever you select, each capacitor should have the same value as any difference will reduce the high frequency common-mode rejection.

GeorgesWelding said:

Well although I have the THS3201 amps on hand, they are SMD components and I am unable to get the simulation model to work properly in MultiSim.

Click to expand...

I don't know the SPICE syntax used by Multisim, but try the attached model. I had to edit the TI model to get it to work with SIMetrix.

Keith

I've updated the schematic to include the components that are inside each of the 10X scope probes that I have hooked to the amplifier box.

View attachment Differential Sch 2.pdf

LT1365CN Op-Amp. Output Ringing.

I posted a previous thread about the home-brew differential o-scope amplifier/probe that I built using a TL074 Op-Amp and although it works great, the bandwidth is pretty much limited to around 100KHz before the rise/fall-time of the op-amp starts to become an issue(measuring a square-wave). I have a few THS3201 current-feedback op-amps that have a massive slew-rate but I am unable to get the spice model to function properly in Multisim and I've never used a CFB Op-Amp before so I'm a little lost at how they work. Also they are SMD components with a fine pin-pitch and I don't have the tools to solder such a thing. Que the LT1365CN. It is a voltage-feedback op-amp with a slew-rate comparable to a CFB op-amp(1000V/uS). The pin-out of the LT1365CN is exactly the same as my TL074 so I thought I could just swap the two out and everything would be golden. I don't have the LT1365CN chip ordered yet so I decided to plug it into the simulation in Multisim and while it did improve the bandwidth of the differential amplifier by leaps and bounds, when I give the amp a square-wave input I am getting some ringing on the output of the amplifier which is absolutely unacceptable. I've tried everything that I can think of, which wasn't much, and I cannot get the ringing to stop. I've attached the schematic and shot of the scope output from Multisim so you can get an idea of what I'm working with here. Attenuation of the input stage is ~X100 so with a 200Vp-p input the output is only 2Vp-p.

Orange and Green traces are the two inputs measured at the input pins of the op-amp. The magenta trace is the output of the amplifier circuit which would normally be fed into my scope's input.

View attachment Differential Probe #6.pdf

Re: LT1365CN Op-Amp. Output Ringing.

Have you tried a small capacitor across your feedback resistor? Have you tried changing your resistor values?

Not sure what you mean by "200V p-p input", but it sure sounds scary. Do you have a schematic?

Nice project, congratulations!

With an improved opamp and high precision (0.1%) resistors, your basic circuit will provide very good results at a low cost.

Optionally, I would increase the battery voltage to +/- 12volt, for a wider common mode range. You can use a pair of A123 cells.

One question, though...if U1b and U1c are configured as voltage followers, why did you add R3, R4?

Come to think of it you're right. They really serve no purpose at all. This was a tweaked design from elsewhere on the internet and the original design used a CFB op-amp where mine are VFB. I've never used a CFB op-amp so I'm not sure if that was the reason they were in the original design or not. Honestly I'm not sure how I missed that when I redid this design. Good catch!

Re: LT1365CN Op-Amp. Output Ringing.

Schematic is in the PDF file linked below the image. Ignore R3 & R4, somebody just pointed out in my first thread about this amplifier that those serve no purpose since the input stages of the amp are voltage followers and they have since been removed. Not sure how I missed this when I was building this thing...

The input stage of this amplifier consists of a 9M-Ohm o-scope probe terminated into a 100K-Ohm input stage of the amp(this resistor divider is in parallel with a capacitor divider) which gives approximately 100X attenuation of the input signal before it reaches the op-amps. Therefore if I put 200V peak-peak input or +/-100V across the two input terminals the output of the amp is only 2Vp-p or +/-1V.

It works fine with the TL074 but when swapping to this higher slew-rate op-amp, I'm getting the ringing. I'll try your suggestions though.

- - - Updated - - -

Ok, since my threads have now been merged...

The schematic in post# 8 using the LT1365, I have removed resistors R3 & R4. This did away with 1/2 of the ringing. The other half of the ringing was my bad. I guess I forgot to set the rise-time of my input signal that Multisim was generating and it was set at 1pS. I increased this to 50nS rise/fall-time on the input signal and the ringing is gone. Really the ringing remains non-existent down to about 25nS rise/fall-time of the input signal. With the rise time set below 25nS the ringing is present, but is less intense and only lasts about 10-20nS depending on the rise/fall-time of the input signal. This is completely acceptable for what I'm doing with this amplifier circuitry especially when the rise/fall-times of the circuits I will be measuring is usually on order of 75-500nS at BEST.