Hi, Im using LT8631 for a high voltage stepdown boost converter.
When testing the proto type, the out put ripple is far more higher than the simulation and datasheet show.
Wheb testing the circuit on the board the layout is fine: Stable signals, less GND ripple.
The circuit is included in the picture.
We tested the application with:
Load = 30 mA, The LT8631 is running in Burst mode.
Vin from 25 to 85 volt.
Vout ripple is:
30 mVpp at Vin= 25V, in creasing to 80 mVpp at Vin= 75 V.
When checking the switchin signals the inductor peak current goes from:
0.28A at Vin= 25V increasing to 0.55A at 75V
The On time is :
350 ns at Vin= 25V decreasing to 180ns at 75V
When recalculating these values against the inductor and the capacitors; the output voltage
ripple is rather the same as measured. (far to high against datasheet and simulation)
The simulation regulates the ON time much better, to keep the peak current on a better value.
I can imagine that the chip is doing this not so idealy in the real world, the timing is rather fast.
So, we have a problem with this extensive ripple.
Also the Datasheet does not seem correctly.
Burst Waveforms at page 7, (lower right picture) shows a correct waveform.
While iL is increasing to 280mA (On time) the output voltage increases, And while iL is decreasing to 0 mA
the output voltage continues increasing. Logical: the inductor current is flowing all this time and is charging
the capacitor. The stored energy in the inductor is transfered to the capacitor.
Then:
Burst Waveforms at page 8, (high left picture) shows an INcorrect waveform. This is when Vin is 100V
Look at the iL is decreasing. This picture shows a NOT increasing Vout.
While iL is increasing the output voltage increases, BUT while iL is decreasing to 0 mA
the output voltage does NOT increasing.
Note that the peak current in this picture is 500 mA, just like we measured on our proto board.
Where is the inductor current flowing to, no where?
I expect that the capacitor will be charched when the inductor current decreases to 0, like the previous picture showed.
The capacitor will be charged, which will result in in a much higher ripple.
Just like we measured at our proto board.
Our question:
is this analysis correct?
Or should the ripple current be better?
page 7:
Page 8:
