ADP2120 operating in a discontinuous mode when synchronized.

Kevin,

A schematics, from your schematics,  some scope plots (VIN,  SW, Vout) would be great to have to be able to help you.

LucaV

If the ADP2120 is not synchronizing maybe your clock source is not compliant with the specs.

100ns min pulse and VH=1.3V, VL=0.4V

LucaV

The circuit is synchronizing correctly. The schematic is essentially the same as the application circuit in the data sheet (figure 50) with L= 2.2uH inductor, Cin = 2x 1u, Cout = 2x 10u. The circuit appears to be operating in the Discontinuous conduction mode, not the PFM mode. I've inserted a couple scope shots of the switch waveforms.

As is, the circuit is generating a lot of low frequency noise. I believe this is due to the varying pulse width modulation. I want to get it to operate in the Continuous conduction mode where there should be less modulation of the inductor current. I *think* that increasing the inductor value should accomplish this. I would appreciate any info or links to any ADI documents that can offer insite into this.

Kevin,

thanks for the pictures, this is indeed DCM and if the part was synchronizing correctly then it would be operating in CCM.

I would be interested to see if you force MODE pin high always if the part goes to CCM or not. Also when synchronized I would be interested in seeying what is the delay between the synch and the SW transition.

I do not have an explanation for what you are seeying, how many boards have you built? do they all behave the same?

Do you have R1 between PVIN and VIN? have you probed pin 1?

LucaV

I have the normal 10 ohm resistor between PVIN & VIN. Pin 1 looks pretty

stable; some hash but no major ripple.

What I believe is happening is that the stored energy in the inductors

magnetic field has completely discharged into the output capacitors

before the PWM cycle ends. At that point the the internal diode stops

conducting and the switch voltage moves slowly to the output voltage.

Increasing the output load makes the switching waveform look pretty much

ideal (CCM). Doubling the inductance also does, I guess by making it

take longer to energize/de-energize the inductors magnetic field.

Is there a limit to the inductor size? The data sheet lists an extremely

limited range (table 5 & 6) of 1 - 2.2 uH. Is there any problem using a

4.7 - 6.8 uH inductor?

Kevin McKee

Project Manager

Stanford Research Systems

1290-D Reamwood Ave

Sunnyvale, CA 94089

408-744-9047, x237

Kevin,

your description would be accurate if the part did not support CCM. I have the feeling your IC is not soldered well and the Sync pin is not getting the signal. How many boards have you built? As you described above the 1.4V part works as expected so there is something going on with this 3.3V design.

The part has internal compensation and it designed to be stable with that range of inductors.

LucaV

Luca,

The part is perfectly synchronized, with the falling edge of the switch

waveform perfectly stable. But there may be an issue with the part or

soldering since 2 other boards do not exhibit this behavior.

Of course the first board you check always behaves weirdly.

Kevin McKee

Project Manager

Stanford Research Systems

1290-D Reamwood Ave

Sunnyvale, CA 94089

408-744-9047, x237

Thanks for confirming my suspicion and good luck with the project.

LucaV