Post Go back to editing

Why would the current limit fail to work?

Category: Hardware
Product Number: LT3091

I'm using the LT3091 to generate a negative voltage, and I'm using its current-limiting function to limit the current between 1 mA and 2 mA. Through the use of 10 Mega-ohm resistors (10 Mega-ohms for 2 mA, 20 Mega-Ohms for 1 mA) and a SPDT switch, I can switch between the 1 mA and 2 mA limit. However, in testing, it's not limiting the current to those set limits. Rather, when set to 1 mA, it outputs 7 mA before the output voltage drops while it outputs 8 mA when the switch is set to 2 mA. Why would this be occurring?

For more information, I'm using the LT8333 to generate -32 volts (as seen at the input of the LT3091), and the set output voltage is -30 volts. The voltage differential should be enough to handle the current I'm hoping for. There is a 1 Mega-ohm load to act as the minimum required load for proper voltage regulation.

The LT3091 itself is the 14-lead DFN package. Could there be issues with that package to cause it to behave that way? Would the 16-lead TSSOP package function any better?

schematic_simplified.pdf

  • Hi Jason

    Can you send your schematic? Also how did you calculate the 10 and 20 Mohm resisters.

     

  •   Hi John. I've attached my schematic in the original post. I calculated the resistors based on the datasheet for the LT3091. In it, it says that the programming scale factor is 20kΩ • A. For example, a 40k resistor between ILIM and IN programs the current limit to 500mA. Using that, I came up with the following equation:

    RESISTANCE = (PGRM SCALE)/CURRENT

    EX: [20kΩ • A]/[0.5] = 40, 000 ohms (or 40 kOhms).

    Thus, for 1 mA and 2 mA respectively, I calculated that to be 20 Mega-ohms and 10 Mega-ohms. In my physical design, I used 0603 resistors tailored to 0.1% tolerance. 

    1307.schematic_simplified.pdf

  • Hi 
    The LT3091 can regulate an output from 1 mA to 1.5A  load. The datasheet describes the Programmable Current Limit for I out greater than 100mA, with a accuracy of +/- 10 percent.


    The regulator is not guaranteed to work below a 1mA load. 


    To increase precision the datasheet recommends the current limiting resister be Kelvin connected.


    I looked for other parts that would be better suited but they only go to -20V. 
     

  •  Thanks, but what I'm trying to do is limit the max current to 1 or 2 mA. Where in the datasheet does it say that the programmable current limit can only limit as low as 100 mA? Also, how would you kelvin connect the limiting resistor?

  • See page 4 of the datasheet under Programmable Current Limit. As a sanity check 10 - 20 M ohms is same as an open circuit. 

     

    The Kelvin layout is not as critical but for precision it may contribute better results. This is a layout implementation and more important for small resisters but it does contribute to over all performance. Its best to consult a layout expert on this implementation.