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.

OPA2170: Can OPA2170 short-circuit output current limits be adjusted to custom levels during test process at TI?

Part Number: OPA2170
Other Parts Discussed in Thread: DAC60004, OPA2237, OPA170

Can OPA2170 short-circuit output current limits be adjusted to custom levels during test process at TI?

  • Hi Dave,

    Could you elaborate on your need a little bit? Do you have concern about the distribution of short-circuit output currents? I'm wondering if we'd be able to find you an alternate part which exhibits the specification you're looking for.

    That said, generally we don't provide this type of adjustment/customization and the parts are tested to meet the datasheet specs.

    If you'd like to take this conversation offline, please send me a friend request and we can discuss further.

    Thank you,
  • HI Paul. This is for analog output on precision scientific instrumentation. The opamp is driven by DAC60004 to generate 0-5V or 0-10V signal.

    I need a single supply(4-35V) opamp with low input offset(on the order of 1-2mV for worse case, but lower is better), whose input includes GND, and whose output can swing within a few millivolts of ground.

    I'm interested in short-circuit limits because the output of this opamp circuit is customer facing and be connected to GND or 32V. If the short-circuit current limit is low enough and the thermal resistance of the package is low enough, then the op-amp won't overheat. From my testing it seems that short-circuit current limit on sinking current is especially important.

    OPA2170 short-circuit sinking current is 20mA or less.

    If output is connected to 32V, and sinking 20mA, then 150 ohms used for RISO compensation dissipates 60mW and has 3V drop across it.

    OPA2170 dissipates(32V-3V)*20mA =  580mW. In a WSON package 580mW*71.5C/W = 41.47C temperature rise, which is respectable. 

    If the circuit-circuit sinking current limit were lower, it'd be better because the OPA2170 would dissipate less heat.

  • The OPA2170 tested well on the bench. The only thing that'd make it better would be if its output sank less current when short circuited to 32V.
    The OPA2237 looks like it might be promising. I haven't tested it. It has lower short-circuit sinking limit, but it's thermal characteristics aren't as good as the OPA2170 WSON.
  • Thanks for the additional info Dave. I'll share this with the team and get back to you with some thoughts. With the upcoming holidays, please allow for some delay in the reply.

  • Dave,

    R31 resistance value is the only way to limit current for 32V. The reason is the internal diode from U15A output to U15A pin8. This diode is just a diode and won't actively limit current.

    Try measuring pins 1 and 8 when 32V is applied to see that PIN1 > PIN 8
  • Ronald,
    D1 in the schematic limits the current through the internal diode from D15A output to U15A pin 8. When 32V is applied to pin 1, pin 8 tracks pin 1, but it's not destructive because of the low current and does not exceed the opamps voltage rating.

    I'm more specifically interested in limiting the current sinking ability of the op-amp, or finding a part similar to the OPA2170 with a lower short-circuit sinking limit.
  • Dave,

    Oh yes, the D1 diode. I didn't scroll up high enough the see that diode this morning.
    VCC and VOUT will increase in current limit as you have seen. Power dissipated is IOUT * VOUT so lowering VOUT and IOUT will be helpful.
    One other plan is to make output try to go high if output is greater than VCC. That would greatly reduce current.
    Do you have your circuit in Ti-Tina that you can upload?

  • Ron,
    How do you recommend making the output go high if the output is greater than VCC?
  • Dave,

    This modified circuit should provide full sink current until op amp V+ pin starts to rise,  at that point the output is suppose to go high so the output will not fight the external voltage anymore and op amps' V+ will rise with just a little current to run the PNP transistor.

    However, that doesn't model well. It won't solve if voltage starts at 32V but will solve for voltage ramping to 32V (DC transfer function). Also current doesn't drop and op amp's ground current flows in the wrong direction most of the time.  

    I will need to build this to prove it out.

    8688.opa2170 analog output(with high voltage input reversal).TSC

    So the worst case should be external voltage close to 12V and low current above that.

  • Ron,
    This looks very reasonable. I look forward to hearing about your results. I think it'll work.

    I found the stabilizing network in this blog post:

    I did step response and saw minimal overshoot as described in blog post. I think this means it's stable, but I'm not sure. How would I need to modify this TINA schematic to simulate for stability?

  • Dave,

    There are a few ways to check stability.

    For multiple feedback paths to inverting pin only, I break at inverting input and add C5 (expected input capacitance to ground )

    Phase margin is 74 degrees at 210 kHz and gain margin is 43dB at 8.2 MHz

    opa2170 bode (gainphase).TSC

  • Dave,

    Here is IOUT vs VOUT.  No pnp is white, with PNP is red.

    The red slope is a little higher than I expected. Alos input current is -3mA with VOUT at 32V. ~30V/10k = 3mA

    Here is V+ vs VOUT

    The PNP did not work with the switch open. This make sense because inverting > non-inverting with or without PNP.

    So, there is still room for improvement

  • Ron,
    The plots look good.

    What is the input current of -3mA you mentioned in your last post?
    I don't understand your comment regarding the switch and inverting vs non-inverting. The switch changes the gain of the amplifier. The pnp doesn't change the amplifier from non-inverting to inverting. What do you mean by this comment?

    Where do you think the steep slope comes from? Do you see any harm in using 1M and 24k instead of 100k and 2.4k resistors to lower power dissipation and current?

    Are these plots from simulations or actual hardware?

    Does phase margin of 74 degrees would mean the circuit is somewhere between critically damped(67 degrees) and totally overdamped(90degrees)?
  • Dave,

    Some of the slope does come from the feedback path. I used 2n3906 as the VCC diode [EB,C I should have left emitter open) ]at it leaked 2mA at 32V so using a real diode will save current that too.

    If PNP pulls non-inverting input high then that will fight input signal, that's why 3mA.
    If gain select switch is open then inverting input is almost same as OUTPUT force (32V). Non Inverting will be less voltage even with PNP turned on.
  • Dave,

    I increased resistors by 10X and now have better result for gain set to ~2. I added a NPN and diode to get good result for gain set to ~1.


    Here is the circuit as tested. (I simplified / modified some components for my testing)

    8688.opa2170 analog output(with high voltage input reversal) (2).TSC

  • Ron, what is the purpose of the npn? I see that it changes the gain of the opamp, but I don't see why this is desirable.
    Why is D1 added?
    I'll try building this myself. I'm hoping to gain your insight into the circuit.

    Have you noticed that TINA shows the OPA170 sinking 21mA and 20mA flowing into the ground terminal of the OPA170?
  • Dave,

    This is the third time that I have said that having a gain near 1, will result in IN- > IN+ so output will not go high, so there will be no reduction in output current. The diode keeps R9 and NPN from affecting the input voltage to IN+ (during normal operation). All of this will become clear after building the circuit. I have not looked deeply into the reason that it doesn't simulate as low current at 32V

    This was a fun project.