search

blinken / power-rail-probe

Low-noise oscilloscope power rail probe
Other
34 stars 1 forks source link

Increaseing the offset voltage. #1

Open smarley2 opened 3 months ago

smarley2 commented 3 months ago

Hi, on the next release could you consider increasing the offset voltage up to 60Vdc? I think would be necessary to change only the R5, right? Did you had a change to evaluate the noise on the real device and using different/better opamps?

blinken commented 3 months ago

Thanks for reaching out! Decreasing R5 to eg. 1k should work, though you'll start running in to the voltage ratings on other components. Input gain will change from -10 to -50. There may be impacts on the noise figure (but I haven't thought about it in detail), and there will be impact to the frequency response due to the higher bias voltage on C1 & friends.

I just received the first boards back for revision 1.0.1 (https://kicanvas.org/?github=https%3A%2F%2Fgithub.com%2Fblinken%2Fpower-rail-probe%2Fblob%2Fmain%2Flow-noise-power-probe.kicad_sch ) and I'll take a look at them (and this proposal) over the weekend. But there's no major reasons why it wouldn't work if you ensure the passive components are rated sufficiently well at the voltage & frequency range you need.

smarley2 commented 3 months ago

Ok, nice. Do you have a list of changes to the 1.0.1v? Because you mentioned some bug fix. Have you tried in the simulation a different opamp? Have you found the reason of the spike on the simulation? Did you tried ltspice instead of Tina?

Best regards

blinken commented 3 months ago

I just completed the characterisation today for v1.0.1, see https://github.com/blinken/power-rail-probe?tab=readme-ov-file#v101-3

I also tested changing R5 to 1k, and it does work well to increase the range of measurable voltages - I was able to measure my 30V power supply output quite comfortably @ 10mV range on the oscilloscope. However, with 50V AC-pass capacitors the frequency response is quite poor at ~30V DC bias (but I lack an accurate bias tee to test with a calibrated reference signal, so I was just eyeballing it).

I think if we switch to 100V-rated components for the AC pass capcacitors, then it should be ok to change R5 to ~3.3k and get the extra range from the device.

blinken commented 3 months ago

Just on opamps, the OPA322 was selected based on the following requirements

The last requirement in particular limits the selection quite a bit, but suggestions for better parts compared with the OPA322 are welcome.

smarley2 commented 3 months ago

Just on opamps, the OPA322 was selected based on the following requirements

  • Rail to rail output
  • Low noise
  • Output drive capability into 50 ohms at 2V

The last requirement in particular limits the selection quite a bit, but suggestions for better parts compared with the OPA322 are welcome.

On Kicad Sch you have two OPA332. But you are using the LMP7731 on the input and the OPA332 just to invert, right? Because if look TI you can find that LMP7731 is the less noise they have with rail to rail input.

Take a look on OPA323 to replace the OPA332, seems to be a better device.

smarley2 commented 3 months ago

Also, another point. Did you try to do a different approach like: Connect the AC cap directly to the output. Connect the OPA with 50R to the output with G=2. So you could have 1:1 probe, but I'm not sure how about the noise when you amplify with OPA. Could you try to simulate it and see the impact?

blinken commented 3 months ago

Take a look at this article https://circuitcellar.com/research-design-hub/projects/building-a-power-rail-probe/ which is a similar design, and covers your questions/the design decisions in detail (much better than I can here!)

On Fri, 2 Aug 2024, 08:48 smarley2, @.***> wrote:

Also, another point. Did you try to do a different approach like: Connect the AC cap directly to the output. Connect the OPA with 50R to the output with G=2. So you could have 1:1 probe, but I'm not sure how about the noise when you amplify with OPA. Could you try to simulate it and see the impact?

— Reply to this email directly, view it on GitHub https://github.com/blinken/power-rail-probe/issues/1#issuecomment-2264781495, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAB5ETHE4TIFUA52R2UYQ3TZPM2TDAVCNFSM6AAAAABKXQC5IKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENRUG44DCNBZGU . You are receiving this because you modified the open/close state.Message ID: @.***>