0
\$\begingroup\$

I have no answer for this strange opamp behavior. The circuit shown is from a typical nonisolated ,Non grid-tie inverter product (imagine H bridge Inverter). which has a common -ve high power circuit and isolated low-voltage 24V and 3.3V. DC Bus and Battery share the same common -ve.

  • The battery voltage sensing opamp shows wrong output i.e 3.1V saturated voltage. whereas the DC Bus voltage opamp works well in all condition. Upon inspection of voltages i observe below voltage, a) Following are the reading across various nodes of OPAMP (Opamp LMV931MGX/NOPB):
  • Bat- R104 to GND= -122
  • BAT+ R108 to GND= -16.9
  • Battery voltage 105.3
  • BAT U25 -ve IP (3) (R105)= 0.372V (note +ve sign)
  • BAT U25 +ve IP (1)(R110)= -0.385V (note -Ve sign)
    1. DC link sensing U23 gain 0.004799
    2. DC link sensing U23 Pin 3 (R95)=1.622
    3. DC link sensing U23 Pin 1 (R102)=1.622

My question :

  1. Why is the voltage on inverting and non-inverting pin are not the same? how can i make it work properly? also note that i assume Opamp is saturating because it is seeing (-ve) -0.0.385V at its +ve non inverting pin. How can i fix this? Note the 3.3V is isolated from HV (battery and DC bus) voltage it is reading differentially.
  2. Both circuits are same in nature DC Bus sensing opamp has no issue and Battery sensing opamp throws wrong reading. i have tried replacing opamp with equivalent part# but still, the issue is the same. I don't think this is noise related issue as i didn't see any sign of noise.

two Differential opamp sensing configuration, One Battery voltage sensing (100V) Other DC Bus voltage sensing 400V

Edit: We have AC sensing in the same way as this. now noticed that when the AC sensing is in circuit we see this DC battery wrong reading issue , when we have AC sensing out of circuit. battery sensing circuit works fine (remember its non isolated H bridge inverter)

\$\endgroup\$
  • \$\begingroup\$ What is bandwidth of your scope? Also, make a 1 inch by 1 inch loop, and go looking for magnetic fields. \$\endgroup\$ – analogsystemsrf Nov 16 '18 at 4:10
  • \$\begingroup\$ Your op-amp (+) input needs a ref voltage equal to 1/2 Vcc. Your input common mode range is centered around that ref voltage. \$\endgroup\$ – Sparky256 Nov 16 '18 at 4:15
  • \$\begingroup\$ You just miss ground connection on R110 lower pin ;) \$\endgroup\$ – carloc Nov 16 '18 at 11:33
0
\$\begingroup\$

The op-amp is specified operate with a common mode voltage range up to 200 mV beyond the supply rails. Less than that over temperature. -0.382 V is beyond that. Can you keep the bat+ input closer to the 0 V of your isolated supply?

\$\endgroup\$
  • \$\begingroup\$ I am aware that the opamp is seeing -Ve 0.38 voltage and hence saturating. but my question is why is it going -ve. why both pins of opamp are not same. i have other circuit connected to same high power -ve bus and it is reading perfectly. \$\endgroup\$ – Busy_Life Nov 16 '18 at 14:57
  • \$\begingroup\$ It is going negative because the only thing driving it is -16.94 V via resistive divider R108-R110. In the other circuit VDC must be positive. \$\endgroup\$ – Steve Hubbard Nov 16 '18 at 21:15
  • \$\begingroup\$ Yes but why only this circuit is going -ve (-16.94V). where as other circuit is also connected to same -ve bus . Also see my new discovery under Edit \$\endgroup\$ – Busy_Life Nov 16 '18 at 22:37
0
\$\begingroup\$

Your op amp won't work correctly unless the inputs have voltages within the supply range of the op amp. Your stream of input resistors keeps the voltage difference from being overwhelming, but it's still outside of the range. I assume R110 is connected to ground...maybe if you pull it to 3.3V you can keep it in the input range of the op amp.

\$\endgroup\$
  • \$\begingroup\$ Yes R110 is connected to ground. I have added 1.5V offset to R110 (by disconnecting its other end from ground and connecting to 1.5V. that didnt help. \$\endgroup\$ – Busy_Life Nov 16 '18 at 15:00

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.