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.

Compensation of voltage drop using OPA549 (high current operation)

Other Parts Discussed in Thread: OPA549

Hi everyone,

I'm using OPA549 to compensate voltage drop due to internal and external resistance of a current source (Ri in the schematic, around 20 mohm) based on the circuit schematic below. The current passing through the load can be up to 10 A, so I chose OPA549 (the op-amp with very high output/sink current). The circuit kind of worked (Voltage different between A and B kept almost 0 and unchanged, this is necessary for my application) for low currents (<0.6A), but it would not work at higher current, which might be due to thermal shutdown according to the datasheet. I'm getting a better heat sink, but by passing current of 10A (or 8A continuously), the power dissipation would be more than 40 W (I'm using a +/- 5V power supply), and it is difficult to find heat sink so powerful and with acceptable price. So I was wondering if I can get some help here about the circuit improvement. Are there better methods for voltage compensation? If not, is there any good method to mitigate the heat dissipation problem? Thanks a lot!

3252.Voltage compensation using OPA 549.tif

More information about how the circuit works can be found from http://www.circuit-fantasia.com/circuit_stories/inventing_circuits/ser_nr_comp/ser_neg_res_comp.htm

  • Shawn,

    I'm not sure I totally understand your application but I believe I know the issue. I believe it relates to the output voltage swing capablity of the OPA549 under high current conditions. Check out the the performance graph in the upper-right of page 6 of the data sheet. The output of the OPA549 can come within approximately 2.8V of the positive power supply voltage when sourcing 2A at 25'C. The output voltage swing capability decreases (increased voltage from the rail voltage) as the load current is increased. It's not clear from your schematic, current and voltage conditions, what the voltage would be at the output terminal of the OPA549. Higher supply voltage would likely solve the problem but power dissipation may increase significantly. It depends on the resistance of the load and voltage on the current source.

    I hope this is helpful in diagnoising the problem.

    Regards, Bruce

  • Dear Bruce,

    Thank you so much for your prompt response. The voltage of point A in my schematic is expected to be equal to point B (internal ground) with the circuit design, so the Vout would vary from 0 to -0.4 V (40 mohm*10A) relative to point B, and my measurement results confirmed this (in the order of mV with I up to 0.6A). The supply voltage in my application is +/- 5V, so |Vsupply| - |Vout | would vary from 4.6 to 5 V, will this cause problem?

    I have tried higher supply voltage (+/- 15 V) and the maximum current was only 0.3 A, which I believe was due to thermal shut down (higher power dissipation with same current), so I turned to +/-5 V supply. I was considering further decrease the supply voltage (say +/-4 V) to reduce power dissipation,  but the power dissipation may remain very high at high current. Do you have any information about very powerful heat sink for OPA 549?

    Best regards!

    Shawn

  • Shawn,

    It appears that your output voltage is not so close to the supplies so as to cause an issue.

    I believe the problem may be due to where you have connected the ref and Ilim pins. We generally show these connected to ground in a dual supply situation.  That would be "B" in your circuit. But there is a limitation that these pins be 8V or more below the positive supply voltage. There is a section of the specification table that deals with this just above the power supply sectioin. As a result, you should connect the ref and Ilim pins to the -5V power supply.

    Let me know if this helps.

    Regards, Bruce.

  • Hi Bruce,

    You're right, I connected the Ref and Ilim pins to internal ground. I will change the connection to -5 power supply. Thank you for the advice.

    Regards!

    Shawn

  • Hi Bruce,

    I  reconnected the Ref and Ilim pins to -5 power supply, but there was no significant improvement. The stable max current was still 0.6A. When I increased the current to 0.7 A, it worked only for several seconds. I also tried 0.8 A and 0.9 A, it stopped working even in shorter time (around 2 seconds). I was wondering if the heat sink I'm using is not good enough even for so low current. I ordered one heat sink with much larger cooling area from DigiKey, but unfortunately it did not fit the size of OPA549, and it is not easy find a suitable one. Do you have any recommendation about heat sink for OPA549? Thank you very much!

    Regards!

    Shawn

  • Shawn,

    I'm sorry, I forgot to address the heat sink issue.

    If you get proper operation for a few second, then not, the issue is sure to be heat related. You have a substantial power dissipation and need some serious heat sinking. You have nearly 5V across the conducting output transistor so the dissipated power is approximately 5V times your output current. 5W will heat the OPA549 very rapidly without a heat sink. You can experiment with a simple piece of flat 0.090" (2.3mm) thick aluminum. A 4" x 5" piece should have a thermal resistance of about 10'C/W, or so. That would allow you some time to get quick tests at higher current.

    For continuous operation you will probably need a heat sink that provides a total thermal resistance of closer to 3'C/W.  The junction-to-case thermal resistance is approximately 1.4'C/W so you need a heat sink in the range of 1.5'C/W or less. This would typically be a finned aluminum heat sink with a thickness of approximately 7mm or more in the mounting area. The total heat sink would need to be approximately 8cm x 12cm or equivalent area (just guessing from some experience). If you need to insulate the case of the OPA549 from the heat sink, it can add another 0.8'C/W, or so, requiring an even larger heat sink. (I assume you are using the tab-mount package and it is electrically connected to V- supply.)

    The data sheet has some recommendations on heat sinking. Wakefield is a well-established manufacturer of heat sinks and you can get some good ideas from looking at their products and specifications. The rating of heat sink thermal resistance is not a precise science. Your needs will depend on many factors such as max power dissipation, duty cycle, max ambient temperature, mechanical mounting condition, ventilation, etc.

    Hope this helps.

    Regards, Bruce.

  • Hi Bruce,

    Thank you very much for the analysis and advice on heat sink. I agree  that the power dissipation would be around 5V*I. I will try to get a better heat sink and see if it works.

    In addition to the method shown in my schematic, do you happen to know any other method/advice about compensating resistive voltage drop using OP-AMP or other TI products? I'm concerned that I cannot reach 10 A operation even with very good heat sink. As you know, the power dissipation by OPA549 would be around 50 W at 10 A with the present method, which is very challenging to dissipate.

    Best regards!

    Shawn

  • Shawn,

    I'm not exactly clear on the big picture of your problem and solution. Assuming that you implement the same basic correction loop but making your own op amp, here are some possibilities:

    1. Use a standard op amp to drive power MOSFET buffers (source followers). Power them from +/-1V power supplies. Note that if you are using linear power supplies, this just moves the dissipation to the power supply.
    2. Use a switching "Class D" audio power amplifier. I'm not well connected with this product line so I can't be of much help. There is an audio amplifier forum that might help.
    3. Build your own class D amplifier. This would be much easier if your correction is unipolar in nature as it is a switching buck regulator controlled by an op amp.

    These would all require a higher level of creativity and resourcefulness on your part. You seem like a clever guy. Go for it. :)

    Regards, Bruce.

  • Hi Bruce,

    Thank you very much for your advice. My knowledge about electronics is pretty limited, so it will take some time for me to figure out those approaches. Hoping I can figure them out, :).

    It is not convenient to tell too many details about my problem. Basically, I need to compensate voltage drops across a number of resistive elements in series with some other elements (resistors and current sources), and each resistive voltage drop needs to be compensated individually, which make it more complicated.

    Best regards!

    Shawn