• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 48 subscribers
  • Views 171 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

OPA551: OPA551 overheat quickly

Desmond Gong
Prodigy 10 points

Part Number: OPA551

The schematic is attached

This design is a interface between a HIL device and our BMS PCBA. It suppose to amplify HIL outout of ~3V by the gains designed in each CT amplifier circuit. 

We noticed a relative warm tempurature(55~60 degree C) when the board is running idle(means only a 55V power supplay is connected). 

The question we have is that, is this design feasible for our purpose? When we connect BMS to this section, it draws over 50W of power and over heat very quickly.

  • 0
    TI__Mastermind 18656 points

    Hey Desmond, 

    Thanks for reaching out. I removed your schematic from the public forum, I have a local copy I will refer to now. 

    These power amps can certainly be somewhat power hungry, though the overall temperature depends very much on board thermals and general PCB layout. 

    Is the 55~60C with all CT amplifiers on without any drive current into the BMS?

    What is the load impedance of the net CT_"X"? 

    Assuming the power amp is not driving any current, we will se an Iq of about 7-10mA per device. With 15 of these operating at 55V supply, that is ~8W idle. 

    A bit more information on the load will greatly help in allowing me to run sims on the circuit. 

    Thanks!

    Best,

    Jacob

  • 0
    Prodigy 10 points

    Thanks for the help Jacob, 

    To answer your question
    1. Is the 55~60C with all CT amplifiers on without any drive current into the BMS?
    Yes, this condition is CT amplifiers on, both input and output are disconnected

    2. What is the load impedance of the net CT_"X"? 

    The CT_x net is connected to the cell voltage input of the L9963E battery monitor IC through the AFE filtering network.
    I do not know the exact effective input impedance seen by OPA551.
    According to the L9963E datasheet, the Cx inputs are cell voltage measurement inputs with very low leakage current, so I originally expected the load to be very high impedance.
    However, during testing I observed that connecting CT14 (42V equivalent stack voltage) increases board power by approximately 1W, while connecting CT1 (3V) increases power by approximately 3W. Therefore I suspect the actual load seen by OPA551 may not be purely high impedance in my implementation.
  • +1 in reply to Desmond Gong
    TI__Mastermind 18656 points

    Hi Desmond, 

    Thanks for the details here!

    Are the OPA551 devices especially close together? It is still somewhat surprising to hear that Iq burn is causing a ~30C increase in temperature. Do you have a layout picture or set of files I can look at?

    I do have some theories for what may be going on, though it is not certain.

    I spent a bit more time with the circuit, I now understand that we are using this ISO amp to do current shunt measurement. The OPA55x is being used to appropriately scale each cell voltage by the corresponding gain set in each channel. 

    I do not think that the OPA is necessarily configured wrong, though there is a characteristic of the design that may introduce issues to the OPA. 

    If we consider the connection of C384 and C388, we are loading the op amp output with 1.1uF between another low impedance load. This large cap load will likely cause the op amp to destabilize and oscillate. Op amps in oscillation tend to burn more current than a stable device. 

    Here is a stability test in the configuration of CT14:

    CT1 is even lower phase margin, maybe eluding to why CT1 is burning more current than CT14. 

    Regardless, could we either try removing the caps off of one of the channels, or alternatively probing the CTx node with an scope? I suspect we will see oscillation. 

    Thanks,

    Jacob