1
\$\begingroup\$

I need to precision buffer a DC (or slow changing) voltage of around 2.5V using an opamp in non-inverting unity gain mode and driven from a single supply using no resistors. Supply voltage is 5V. Is it enough just to wire it as described? Opamp chosen is OPA2188.

Potential problems I can see are PSU noise being introduced, which may mean special power supply decoupling. Any others?

\$\endgroup\$
1
\$\begingroup\$

Common mode voltage range could be an issue:

enter image description here

With power supply at 5V you are left with a max common mode voltage of 3.5V. In non-inverting configuration the inverting input will sit at about 2.5V so you only have 1V headroom. If your supply lowers just a bit you risk malfunctioning. And even if it stays stiff at 5V, note that those specs are given with an \$R_L=10k\Omega\$ (see above the table). So if you use the opamp as a buffer your load could have lower resistance and this could impact the max common mode voltage tolerated by the opamp.

\$\endgroup\$
  • \$\begingroup\$ Thanks - good point. I will probably change it to ADA4528-2 \$\endgroup\$ – Dirk Bruere Aug 3 '15 at 11:14
  • \$\begingroup\$ RE: "it is not stated with which accuracy that (V+)-1.5V is guarenteed, nor the manufacturing spread is specified". The value given is a maximum, not a typical. From a vendor like TI, it should be trustworthy. Likely a typical part would work with a slightly higher input CM voltage, and the value given is intended to account for manufacturing variation. From a company like TI, this should be a 6-sigma limit, meaning no more than about 1 in a billion parts should fail the spec. \$\endgroup\$ – The Photon Aug 3 '15 at 16:39
  • \$\begingroup\$ @ThePhoton Thanks for the useful information, especially the 6σ threshold (BTW, is there some industrial standard, or some other document where these info are collected, or is it common practitioners' knowledge?). So I stand corrected for that part. I'll edit my answer. \$\endgroup\$ – Lorenzo Donati Aug 3 '15 at 19:50
  • \$\begingroup\$ The 6-sigma number is part of "Six Sigma" quality management. It's not necessarily a promise to customers but something they should do to avoid waste, returns and unhappy customers. Correction: note in the wiki article on Six Sigma that 1 ppb defect rate is actually overly optimisitic. \$\endgroup\$ – The Photon Aug 3 '15 at 20:24
  • \$\begingroup\$ FYI, I found this document which says "TI manufacturing uses process capability measurements as a key component of process monitoring and control with a goal to achieve values of Cp > 2.00 and Cpk > 1.67." Wiki article on process capability index tells how this relates to test margins. \$\endgroup\$ – The Photon Aug 3 '15 at 21:03

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.