• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 47 subscribers
  • Views 130 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.

Original question:

THS3121: Improved driver for peltier (TEC)?

OPA2675: TEC driver has different results in Tina-TI and circuit

Daniel Ka
Intellectual 607 points
Part Number: OPA2675
Other Parts Discussed in Thread: TINA-TI

Tool/software:

Hello,

about two months ago, I asked for an alternative operational amplifier in order to match current driving requirements of my application. There, the current feedback OPA2675 was recommended. That was a helpful suggestion.

So I built a driver for a small TEC (thermo electric cooler aka. peltier element) based on the OPA2675 with the TEC connected to the two outputs of the opamp ("bridged configuration").

The simulation helped me to figure out the correct values for the feedback network so that the opamp provides the power to the TEC with the control voltage UControl = 0 - 3.3 V (which is provided by a DAC). I included Voffset in the circuit in order to shift the control voltage but I noticed that my application does not need that and thus I set R1 = 1 TOhm.

To compare the circuit with the simulation, I used a 1.5 Ohm resistor instead of the real TEC element and took some measurements: For example at  UControl = 171 mV, I measure UPeltier = -372 mV (across R5) and a current of 233 mA. At UControl = 3,2 V, I measure 101 mV and a current of -63 mA through R5. The bridge is balanced (no current through R5), when UControl = UBias = Vref/2 = 2.55V. 

I noticed that the simulation agrees with the circuit only if the feedback resistor Rf is set to half the value of the circuit. In other words, in the circuit I have Rf = 510 Ohm whereas in the simulation I have to set Rf = 260 Ohm in order to make simulation match the values measured on the bench. 

Any idea why this is?

Please note that I have attached the Tina-TI simulation below so you have it handy in case you want to look at it.

Thanks and have a nice weekend.
Daniel

52-4TI BridgeInverted OPA2675 Vcc=6.5V Uc(max)=3.3V.TSC

  • 0
    TI__Genius 16575 points

    Hi Daniel,

    I was able to look over the tina sim and I cannot think of anything that could be causing this. I am curious, in the real circuit do you have Voffset or is this not connected anymore? Also, when probing around and collecting more data does this also match simulation very closely? Or is there some other issue that can be happening where it is a coincidence that the circuit works when adjusting Rf? Are the other components the correct values as well? As well as the biasing circuitry on the non-inverting input?

    Best Regards,

    Ignacio

  • +1 in reply to Ignacio Vazquez Lam1
    Intellectual 607 points

    Hello  ,

    Thank you for taking care of my request.

    In checked the real circuit: I disconnected Voffset by removing resistor R1, RG is 2700 Ohm and all feedback resistors are 510 Ohm. U_Bias is at 2.55V.

    This is the data that I collected compared to the simulated values:

    Scrutinizing my setup, I noticed that my amps meter adds an input burden of 2 mV/mA. ‍Rolling eyes  In other words, the amps meter increases the resistance of the load from 1.5 Ohm to 3.5 Ohm. Therefore the real circuit has to provide about twice the gain compared to the simulation. I unintentionally "corrected" this difference by reducing the gain of the simulation by halving Rf.

    Does that make sense to you?

  • 0 in reply to Daniel Ka
    TI__Genius 16575 points

    Hi Daniel,

    I have not experienced this myself, but I can see what you found to be very plausible. It is adding an offset to one side of the resistor (where you're measuring from) and is therefore reducing the voltage difference across the resistor and therefore reduces the current through the resistor. It is reducing what is expected and why reducing the simulation gain gives the same result. You are dealing with such a small load that millivolts of offset will affect your results which could be what was happening in this case. If so, this was a great catch!

    Best Regards,

    Ignacio