• State
  • Locked Locked
  • Replies 2 replies
  • Subscribers 49 subscribers
  • Views 316 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

OPA838: non-linear temperature drift - transimpedance amplifier

Jan Bock
Intellectual 2750 points
Part Number: OPA838
Other Parts Discussed in Thread: OPA837

Hello Team, 

my customer has designed the following TIA and observe an non linear temp. drift issue: 

The supply voltage of the OPA838 is 0V and 3.3V, the VREF_OFFSET is 1.45V, so that we can use the TIA in both "directions" around the reference voltage despite unipolar voltage supply.

The requirements for the OPA838 are very high in that its amplitude linearity should be very constant, both over input current and temperature.
Unfortunately, we see in measurements that non-linearities show up here. We modulate a current source with the input signal with non-linear characteristic as input signal (X-axis of the following graph) and record the output signal of the TIA (Y-axis of the graph) at different temperatures. Here it can be seen that the curves intersect at an output amplitude of about 300mV at the two measurement temperatures.
1. could this be due to a temperature-dependent amplification?
2. possibly in combination with the small and large signal behavior of the OPA838?
3. is there a recommendation which OPAMP is even better here?

Regarding the table: the amplitude is formed as difference of "High - Min", the last column shows then the difference of the amplitudes at different temperatures. Here you can see in the range of 0.2 - 0.9 (input current source) a maximum difference of about 4.8mV. An always equal difference would not be a problem, the "intersection" of these curves (nonlinear effect) is what gives us problems.

Input current source [V]

50°C

25°C

High [V]

Min [V]

Amplitude [V]

High [V]

Min [V]

Amplitude [V]

Difference (25°C - 50°C)[V]

0,20

1,4308

1,4282

0,0026

1,4338

1,4312

0,0026

-2,2204E-16

0,30

1,4308

1,4263

0,0045

1,4338

1,4296

0,0042

-0,0003

0,40

1,4307

1,4224

0,0083

1,4338

1,4261

0,0077

-0,0006

0,50

1,4306

1,4112

0,0194

1,4346

1,4155

0,0191

-0,0003

0,60

1,4315

1,3763

0,0552

1,4346

1,3837

0,0509

-0,0043

0,70

1,4337

1,3209

0,1128

1,4357

1,3277

0,108

-0,0048

0,80

1,4337

1,2524

0,1813

1,4357

1,2563

0,1794

-0,0019

0,90

1,4337

1,1801

0,2536

1,4257

1,1808

0,2449

-0,0087

1,00

1,4337

1,1038

0,3299

1,4392

1,1038

0,3354

0,0055

1,10

1,4353

1,0271

0,4082

1,4392

1,0233

0,4159

0,0077

1,20

1,4353

0,948

0,4873

1,4392

0,94

0,4992

0,0119

1,30

1,4353

0,833

0,6023

1,4392

0,826

0,6132

0,0109

I hope you can help!

Thanks

Jan

  • 0
    Guru 48395 points

    Well Jan, 1st we have to address stability issues. when you say source current, what is its capacitance. the OPA838 is not unity gain stable, if you have pushed it into oscillations that might be your issue, the OPA837 is the drop in unity gain stable version. but it too may need some phase margin work. 

    If your source has very low C, you can stabilize the OPA838 circuit with a cap to ground onthe inverting input, min stable gain of say about 6V/V, so 5*160pF on the inverting node to gnd as 800pF. That would be easy to try. 

    Also, this seems a DC measurement? might as well match source Z for this bipolar input VFA. on the V+ bias network, make the 100ohm equal to the feedback R. Also, sometimes these very high speed input stages don't like 1uF source C - might need to insert like a 10ohm in series with that to the V+ input. 

    The OPA838 is a very power effiicient decomp VFA, the transistors are still >10GHz Ft so it can be a bit touchy. 

  • 0
    Guru 140200 points

    Hi Jan,

    never connect the scope or voltmeter directly to the output of an OPAmp. Always insert a suited isolation resistor.

    Jan Bock said:
    The supply voltage of the OPA838 is 0V and 3.3V, the VREF_OFFSET is 1.45V, so that we can use the TIA in both "directions" around the reference voltage despite unipolar voltage supply.

    What is "VREF_OFFSET" during in all these measurements?

    What is C101? A tantal with temperature dependent leakage current?

    I would replace C101 by a good quality 100nF X7R.

    By the way, improper soldering and washing can also make issues...

    Jan Bock said:
    Unfortunately, we see in measurements that non-linearities show up here. We modulate a current source with the input signal with non-linear characteristic as input signal (X-axis of the following graph) and record the output signal of the TIA (Y-axis of the graph) at different temperatures.

    What do you mean by an "input signal with non-linear characteristic"?

    What is your input signal? Can you show a scope plot? (For the scope plot you may run the current signal into a simple ground referenced resistor.)

    What is the output signal when you choose a pure DC input current?

    What is the output signal when you choose a pure AC input current?

    Divide and conquer Relaxed

    Kai