LM317: Any knowledge on Protection Circuit internal operating details

Part Number: LM317

Tool/software:

HI,

We're using an LM317 as a current regulator (45mA) to drive a current loop communications circuit for petrol pumps. The current in the loop is simply turned on and off with optocouplers and transistors in the master and slave devices. Baud rate is 4800 so bit width ~200uS.

We are having problems with some newer slave devices where occasional data bits are malformed when the slave (re)enables the current. It appears that the regulator does not supply current immediately in these instances.

Our thinking on this is that the newer slave devices have a faster switching rate than older devices and are triggering the protection circuitry in the LM317. We do not have any control over the slave devices nor much in-depth knowledge of their circuitry.

We can mitigate the problem by adding some parallel capacitance to the circuit or by bleeding some current so that the regulator is never completely turned off.

I have not been successful in finding any information about the working of the protection circuit.

Is there a protection component that may be triggered by fast rising current? Is going from 0 current to 45mA a problem in some other way for the protection circuit? Are there other aspects of the protection circuit or LM317 in general that may be at play here?

Thanks in advance for any comments anyone may be able to give.

Excuse the noise due to an unreferenced probe.
First trace shows good comms, second trace shows failure in bit transition from high to low.

  • Hi Carl,

    I am reviewing this request and I will get back to you by end of this week.

    Best Regards,

    Ishaan

  • Carl,

    Could you please share a schematic of your circuit?

  • Sure, here's the circuit.

    The circuit is powered from a DC/DC circuit for isolation but the problem has been isolated to the LM317 current regulator by testing with different supplies such as batteries.

  • I'm looking into your schematic and will advise once I complete my review

  • Hi Carl,

    Could you please identify which signal on the schematic is represented by the scopeshot trace you attached above? I feel that this is not an issue with LM317. You have a circuit that loads the current regulated output of the LDO and depending on the state of Q5, sets output to HIGH or LOW for comms. This happens by changing the impedance at the VOUT node of the LDO. We can see from the scopeshot showing the abnormal behavior that the output remains high impedance instead of going low impedance. The LDO can handle much smaller time intervals of switching ON/OFF in the same signal profile. So, I feel this is an issue related to the modulation of controls for the Comms circuit. I will continue debugging once above question is answered.

    Best regards

    Ishaan

  • Hi Ishaan,

    Here's an overview of the test setup.

    Basically the LM317 supplies current into the circuit. The comms interface in the master (and similarly in slaves), use transistor Q6 to either pass current (i.e. the idle-line "mark" state) or turn the flow of current off (the comms "space" state).

    Hence the LM317 sees either a striing of 68R resistors (depending on the number of devices in the loop) in the "mark" state or an open circuit in the "space"  state..

    We're reasonably sure the issue is caused by the LM317 as:

    • We have tried a number of different Voltage supply sources including a 24V battery.
    • We have an older circuit that uses a BJT-based circuit as the current source and don't have the same issue
    • The issue is mitigated by adding 100nF in the C24 position across the 2WIRE+ supply.
      We do see current overshoot with the capacitor in-circuit but our inference here is that the energy from the capacitor is allowing the LM317 to reduce it's instantaneous current demand.

    We have pretty much settled on the addiution of a 100nF capacitor as the fix for the issue, however my goal of this investigation is simply to understand what is going on. I'm happy to concede that we are abusing the LM317 but would like to better understand how.

    Cheers,

    Carl Kamper.

  • Hi Carl,

    Thank you for a more detailed description of the issue. If it is possible, could you please make a similar measurement across R4? I want to compare the signal against your healthy and anomalous signal at DATA+ node.