TPS65251 noisy output and PGOOD goes low after several seconds

Moving to the correct forum!

Not much output capacitance...How did you come up with this design?

The output capacitance was calculated using Eqn. 12 and 13 from the data sheet.

For the 1.0v line I've got

dI_out = 0.25 A

Lo = 15 uH

Vout = 1.0 v

dV_out = 0.050 v

Iripple = 0.14

fsw = 570 Hz

Vripple = 0.020 v

This gives me 1.47 uF with Eqn 12 and 1.53 uF with Eqn 13. There's an additional 15.9 uF worth of caps on the line near the ic it is powering.

The 1.8v portion of the design is identical to another design that we have working so I've been primarily looking at the 1.0v output.

I usually design with a minimum of 1x22uF.  I would suggest you try that value and re-compensate accordingly.

John,

I made the modifications you suggested and I am still seeing the same powering on and off behavior. Do you have any other suggestions?

Thanks,

Dan

Is it possible you are seeing OCP hiccup?  Your Ilim resistors are set for the minimum output current.

The maximum current draw on the 1.8V line is 84mA and the max draw on the 1.0V line is 73 mA so I should be well below the OCP threshold.

Can you provide a waveform of the inductor current?

I don't have a current probe available to me so I can't get the inductor current waveform.

I've done some additional testing I am able to get the 1.8v output to come up and stay on if I disable the 1.0v output. When I do the opposite modification (disable 1.8v and leave 1.0v enabled) I still see the powering on and off of the 1.0v line and the 1.8v line outputs 300mV even though it is disabled.

How much input capacitance do you have?  Is it tightly coupled to the Vin and GND pins and located close to the IC?  It s always difficult to troubleshoot these types of problems over the forum.  Who is your local FAE?

The input capacitance is a single 10uF capacitor that is close to VIN1 and VIN2. I put a second 10uF capacitor in parallel with that one but there was no change.

We have also been getting some support via email from Joey Yuen. He recommended changing Rc1 to 10k and Rc2 to 3k along with changing the output capacitance on both bucks to 22uF. This modification fixed the 1.8v output but the 1.0v line is still not stable.

How are the Rc values being calculated? By my calculation for the 1.8v line Rc = 15.5k and for the 1.0v line Rc = 8.6k using the equation:

Rc = (2pi*fc*Vo*Co)/(gM*Vref*gm_ps)

where fc = 65000, Co = 0.000022, gM = 0.000130, gm_ps = 10, Vref = 0.8

That equation is a first order approximation for calculating Rc.  It should get you close. In any case, this does not look like a stability issue to me.  It appears that converter is shutting down.  During this operation, the other two converters are operating normally correct?

When hiccup over current is active for more than 10 msec, all the bucks will shut down and restart.  I suspect you have an OCP condition on one of them.  It would be good if we could get an example board in our lab or the local FAE in your lab to look at it.  It is difficult to do this type of troubleshooting over the internet.

Do below test:

Pull EN3 low, monitor VOUT1 and VOUT2. If VOUT/2 is stable, we can fix that the reson is FB3's high volgate.(3.3V)

Jason Wang

jason.wang@ti.com

system engineer