LM27762: Load on OUT- causes VOUT+ to drop.

... a bit more information: this appears to happen on 3 out of 3 boards. On the board I'm investigating, I have removed the ICs that are powered by +/-5V. All that is left hanging  on the +5V rail is about 3u3 of decoupling, and an LED drawing 1mA. Applying the load to -5V, the +5V drops in an RC style over about 1 second. Removing the load, it rises in about 3 seconds:

... I increased the flying cap C12 to 1u: The step load drop on -5V halved but the drop on +5V doubled.

Hi Brian,

Thank you for all the details. I am taking care of this request.

Please give me sometime to analyse and I will get back to you as soon as I can.

Regards,

Febin

Hi Brian,

My suggestion would be to use E96 -1% tolerance resistors for R1, R2, R3, R4 (Feedback resistors: please check the datasheet if necessary). 

Hence please use R1=316k, R3=309k, R2 & R4=100k and try to take the measurements again.

Please let me know if it works or if you need any further help.

Regards,

Febin

Hi Brian,

Additionally, I also guess that the device is running close to 100% mode. (VIN- 5.35V, VOUT-5V)

So, I would suggest to measure VIN with and without load with the same BOM. It should be lower with a load.

What is the maximum load applied at both outputs?

Please increase the VIN slightly (with a load) and take the measurements again. I hope this would solve the problem.

Hi Febin, this did not fix the problem.

Hi Febin,

This did not fix the problem. Increasing Vin by 0.1V, Vout+ increased by 0.03V under load.

I was testing with 200mA loads. Typically will be using 100-150mA.

Hi Brian,

1. Please confirm you are using 1% tolerance resistors. Have you used  R1=316k, R3=309k, R2 & R4=100k or resistive load? This is very important.

2. Please use a stable DC power supply of 5.5V.

3. Can you please provide oscilloscope plots for Vin, +/-Vout, +/-Iout, Cfly with Vin=5.5V, +/-Iout=200mA, Vout=+/-5V, Cin=10uF and Cfly=1uF. The other external capacitors as per your schematic. Please provide plots for both with and without load.

4. Do you have a chance to work with the EVM?

Regards,

Febin

Hi Fabin,

I am using 1% 100k & 316k for both. I could not get 309k quickly. The absolute voltage does not have to be exaclty 5.0V.

Vin is stable 5.45V. I can try to increase it a bit, but the stripped down board I was working on has died and now I have to use a board that I have to ship, so I'm a bit more limitied in what I can do.  I did try replacing the caps with recommendations from the power desiger, and the -5V output only made it to -1V.

I will try to get the measurements. Could take a while...

I don't have the EVM.

Hi Brian,

In the meantime, I would like to see the layout if that is possible.

Regards,

Febin