Post Go back to editing

LTC4020 simulation in LTSpice not doing what I expected.

Hi,

I’m trying to use the LTC4020 as a 12V to single-cell LiPo charger, at present only on LTSpice. The goal is for a 4A charging current and a simultaneous capability to supply 4A to the load.

I’ve adapted the test fixture in what I think is a reasonable way, though I have not yet closely examined stability.

The first oddity I observe is with the battery disconnected and a nominal 1A load, I do not see Vout at the 4.2V I expected, I see an output hunting between 2.4V and 3.6V. (trace-no-battery.pdf)

If I connect the battery simulation, again adapted from the test fixture, I see Vbgate possibly trying to trickle charge whilst the battery is too low, though I see very little current … I rather expected 4A/15 => ~250mA. As the battery voltage rises above about 3V, I expect to see charging start at 4A, but it does not and Vout soon falls below Vbat.

Only as Vbat approaches 4.2V does Vbgate finaly turn on the powerpath FET. But Vout remains below Vbat, so the circuit is now discharging the battery.

PDFPDFPDFLTC4020.asc

Parents
  • Unfortunately, the LTC4020 LTspice model is not working properly.  The problem is being addressed along with other issues, some which have higher priority.

  • Hi,

    I had rather suspected that may be the case.  There were a few aspects of the behaviour that seemed odd.

    Is there any guide to in what I can have reasonable faith and what is definitely broken?

    The next question then, of course, is does it look like my circuit _should_ do what I expect, subject to verifying currents and stability?

    I'm on a scary tight schedule with this design, so proceeding with many question marks over performance may well be unwise.  I do have a fall-back to an adapted existing design, but we'd hoped to execute a significant upgrade.

    Alternatively, perhaps someone could guide a reinforcement of our present design, which has a 12V->5V regulaulator, a 5V to LiPo charger (LTC4001) and a LiPo to 5V boost (LTC3872).   12V to LiPo charger would liberate all the 12V->5V suply for the load, and a LiPo to 5V/4A boost supply would support the LiPo chain in parallel.  Cheeky to ask, but I've done several days work on the LTC4020, that I can ill aford to lose.

    Thanks.

                         Gordon.

Reply
  • Hi,

    I had rather suspected that may be the case.  There were a few aspects of the behaviour that seemed odd.

    Is there any guide to in what I can have reasonable faith and what is definitely broken?

    The next question then, of course, is does it look like my circuit _should_ do what I expect, subject to verifying currents and stability?

    I'm on a scary tight schedule with this design, so proceeding with many question marks over performance may well be unwise.  I do have a fall-back to an adapted existing design, but we'd hoped to execute a significant upgrade.

    Alternatively, perhaps someone could guide a reinforcement of our present design, which has a 12V->5V regulaulator, a 5V to LiPo charger (LTC4001) and a LiPo to 5V boost (LTC3872).   12V to LiPo charger would liberate all the 12V->5V suply for the load, and a LiPo to 5V/4A boost supply would support the LiPo chain in parallel.  Cheeky to ask, but I've done several days work on the LTC4020, that I can ill aford to lose.

    Thanks.

                         Gordon.

Children
  • LTC3872 is already the new 4A boot supply.

  • LTC4009 seems apropriate and has a spice mdel.

  • I can't help you with what parts of the LTspice may or may not work.

    You could use the LTC4020 to perform the tasks you describe with one caveat.  The output voltage of the converter power stage will follow the cell voltage as long as it is charging (say 3.6V to 4.2V plus a 100mV drop across the PMOS device and current-sense resistor).  If this would not be an acceptable voltage range for your 5V application you would have to add a boost converter from the converter-power-stage output to generate the 5V.  The LTC4020 will automatically source cell power to the boost converter when power is removed from the input of the LTC4020.

    An option that may be suitable is to use the instant-on feature to establish a minimum output voltage (say 5V) from the power stage.  The down side would be that the PMOS would operate as a linear regulator with associated dissipation.  If your charging current is not high this option could be considered as a way to obviate the need for the 5V boost converter.  See datasheet, p. 27, 28 for a discussion of instant-on mode.

    I'm not the expert on the LTC4009, but it looks like a good choice if only buck conversion is necessary.  You would still require the boost converter on the output for well-regulated 5V, obviously.