0
\$\begingroup\$

I tried to search something about this but found only a small amount of information.

What confuses me is that all isolation amplifiers, optocouplers etc. I could find usually work at much lower frequencies than hundreds of MHz. So it is a mystery to me how the oscilloscopes work.

For example, here is some photos of Hantek DSO1152S https://www.eevblog.com/forum/testgear/hantek-dso1152s-wont-boot-after-firmware-update/100/ and DSO1122S http://www.davidefabbri.net/files/foto/SANY0116.JPG

It is very difficult to understand from those photos how the isolation works. I only could find about DC-DC converters that seemingly used there https://lib.chipdip.ru/503/DOC001503388.pdf

\$\endgroup\$
3
\$\begingroup\$

Generally you're going to find a transformer in the signal path. Some methods use an optocoupler plus a transformer to handle DC-high frequency, and some modulate the input (which can then go down to DC), pass it through a transformer then demodulate after isolation.

The photos you linked appear to show a transformer and optoisolators on the bottom of the board, with a canned cheap DC-DC on the top of the board to supply power. Probably not the best isolation.

\$\endgroup\$
  • \$\begingroup\$ It would appear the DC/DC brick can isolate to 1500V (Mornsun WRA series). Out of curiosity, what would you say that this isolation scheme is lacking? \$\endgroup\$ – Platytude Jan 13 at 5:06
  • 2
    \$\begingroup\$ Pretty low isolation impedance. There's 85pF in the DC-DC alone and more from that blue safety rated capacitor (and the transformer, and the opto and stray). Possibly a fair bit of noise current from the DC-DC too, so again if the isolated "ground" in the signal you are measuring isn't low impedance to earth you may see artifacts. \$\endgroup\$ – Spehro Pefhany Jan 13 at 5:17
  • 3
    \$\begingroup\$ I can speak from experience that those DC/DC bricks usually aren't the quietest. I've had more than one EMI headache caused by them. \$\endgroup\$ – Platytude Jan 13 at 5:27
  • \$\begingroup\$ @Spehro Pefhany Oh, didn't expect transformer (because I thought transformers usually have relatively narrow frequency range). Thank you. I wonder, what kind of modulation is usually used in that case. Maybe FM? Also, those (DSO1xxxS) are probably among the cheapest isolated oscilloscopes out there, so... \$\endgroup\$ – Dmitry Jan 13 at 14:14
  • \$\begingroup\$ Btw, here is approximate capacitance between channels (between "grounds" of channels) of Fluke 199C. Something around 29 pF. imgur.com/a/10HpMDF I know, not a good way of measuring low values of capacitance. Just a rough estimation. \$\endgroup\$ – Dmitry Jan 13 at 14:16
0
\$\begingroup\$

The isolation is in the internal DC-DC converter for each channel.

The float voltage between Multimeter and [earth]Ground 1000V may be conservative due to external AC-DC isolation.

In the case of SANY0116.jpg ( cropped below ), the isolation is in the DC-DC converter provided by a MORNSUN hybrid IC for each of the two channels with the creepage distance around the analog circuit and >=2mm air gap between channels.

enter image description here

  •  Wide input voltage range (2:1)
  •  High efficiency up to 86%
  •  Isolation voltage 1.5KVDC
  •  Output short circuit protection (self-recovery)
  •  Operating temperature range: -40℃ to +85℃

Ref enter image description here

This 2W version is obsoleted by a 3W version.

\$\endgroup\$
  • \$\begingroup\$ This is the isolation for power. What about the data? Form low-resolution pictures it is unclear, which side the ADCs are located, but there is some interface. The interface is on the other side of this board. I deleted my answer because it is unclear whether the entire front end with ADC is isolated, or there is some analog path running up to 200 MHz. That is the question. \$\endgroup\$ – Ale..chenski Jan 13 at 18:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.