Post Go back to editing

LT3471 Negative Supply

I have designed a ±20V power supply using this device and I have used the same design on a previous board which worked well. This time I am using an increased power supply (from 8V to 12V) and the only other change is the input capacitor was swapped for an electrolytic of the same value. When I tested the power supply the positive side works but the negative side was drawing a lot of current and the output was 0V. I tried a couple different capacitor values at C58 and after a couple tests put the value back to the original value to get images from probing of the switching node however at that stage it completely burnt itself out and I am now measuring 8Ω between SW2 and GND. I can't understand why this circuit worked in the previous design and not in this one. Any advice or indication what is causing this problem would be greatly appreciated.

Parents
  • The abs max voltage for SHDN and Vin is 16V. With the increased voltage from 8Vin to 12Vin, you may have enough ringing to violate this abs max. Adding a ceramic cap of at least 4.7uF in parallel with C54 will reduce the ringing. Recommend keeping the C54 electrolytic to reduce parasitic LC ringing (a larger electrolytic 47uF to 100uF will help further).

    Other observation:  You will get better negative supply performance if L9 and L10 inductors are the same type and value.

  • That's helpful, thanks, I think I will make those capacitor changes. However I don't think this is the cause of the issue as it's only the switch on the negative side that's damaged. Having played around with the simulation yesterday I think the problem could be D8, it has a voltage rating of 30V and I think the increase in input voltage was enough to increase the back EMF from L10 above this and damage the diode leaving SW2 unprotected. Do you think this would be a sensible explanation for what's happened here? I'm waiting on parts so I haven't actually been able to try this and verify it works.

Reply
  • That's helpful, thanks, I think I will make those capacitor changes. However I don't think this is the cause of the issue as it's only the switch on the negative side that's damaged. Having played around with the simulation yesterday I think the problem could be D8, it has a voltage rating of 30V and I think the increase in input voltage was enough to increase the back EMF from L10 above this and damage the diode leaving SW2 unprotected. Do you think this would be a sensible explanation for what's happened here? I'm waiting on parts so I haven't actually been able to try this and verify it works.

Children