1
\$\begingroup\$

I have a transimpedance amplifier circuit that takes 100kHz small current and gives the voltage output proportional to current amplitude. Everything works properly when nearby HV circuit is off. When it is on, the amplifier gives changing and sometimes weird results. I enclosed both in ungrounded metal boxes (should I ground the metal box around the HV?), but still problems arise. What else I should do? My circuit is on a board and wires are long, also they are not soldered (just test board). I know for RF, one should use short wires, do I also need coaxial wires between?

Overall if anyone has a good guide how to work out circuit boards with HV and RF around, please let me know.

Thanks!

\$\endgroup\$
1
  • \$\begingroup\$ I don't know why the OP left this open , why it was bumped? \$\endgroup\$ May 26, 2017 at 21:47

1 Answer 1

1
\$\begingroup\$
  • All input and return signal paths must couple equally to stray noise to cancel. This is normally done with shielded twisted pairs and for lower capacitance on cable using a common mode signal to drive the shield over twisted pairs, called an Active guard method.
  • Ensure the CM signal is within TIA input limits and better performance is obtained using a ferrite balun sleeve or SMT twin coil CM choke with differential cap LPF.
  • No open loops in sensor wires are permitted as this creates a loop antenna.

    • More details on your exact signal being measured, the interfering signal, and the physical layout are needed to give a better answer.
\$\endgroup\$
9
  • \$\begingroup\$ I have an optical input system which operates in the rough bandwidth the OP is discussing and the currents are also small. Sometimes there is no escaping the use of a Faraday cage over the front end -- soldering down a metal can over the entire front end part -- with appropriate board layout underneath, as well. Guard rings around the sensitive inputs. Power supply analog vs digital isolation with beads and caps, too. \$\endgroup\$
    – jonk
    Oct 11, 2016 at 16:52
  • \$\begingroup\$ yes the best Optical sensors have built in shields , while the feedback R lowers the load impedance while reducing gain at the same time. HV electric fields create large CM fields to signal pair can be a challenge for long cables requiring deeper understanding of coupling on cable. \$\endgroup\$ Oct 11, 2016 at 17:00
  • \$\begingroup\$ Thanks Tony and Jonk. In my case, I have to have open sensor circuit since it should work as antenna. Oscillator is connected with + potential to one electrode that electric emits signal into the liquid, the receiving electrode picks up the signal depending on ionic content of the liquid so again I need to have an open loop. I put the whole circuit (-oscillator) inside the Faraday cage hoping that would eliminate need for deliberate design adaptations. Maybe my cage is not good enough? Power supply is analog but I am not sure how to decouple it from digital part or if that is the problem. \$\endgroup\$
    – Biosensing
    Oct 11, 2016 at 19:25
  • \$\begingroup\$ Consider a differential C bridge arrangement instead of single ended TIA current amp and measure impedance change near a balanced bridge impedance. Consider a PLL to filter out noise. \$\endgroup\$ Oct 11, 2016 at 19:29
  • \$\begingroup\$ Thanks Tony, I like the bridge idea, I might try that. I am not sure how to do PLL or that it would work with my, probably distorted shape and small current. \$\endgroup\$
    – Biosensing
    Oct 11, 2016 at 19:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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