LMZ34002 Spikes on Output

Hi J.

Changing to close coupled GND (coax to 50 ohm scope input) does improve the situation somewhat.

Now the spikes are down to +/- 100 mV - exactly in sync with the waveform at PH.

Do you think this is real or also attributable to measurement?

David

Hi David,

Please refer to our measurements as below, where ch3 was Vout in AC.  The output ripple (rms) should be less than 3.5mV, and the peak-to-peak was about 40mV (the BW was set at 20MHz as Jason stated).

Thanks,

Hung 

Hi Hung and Jason,

When I turn on band limiting then the picture becomes much more like yours (not surprising).

Could you please explain to me this procedure though. It does not seem technically correct to specify a maximum

level of ripple under a 20MHZ limit when much of the ripple energy lies above this limit. In a real world application

my down stream devices will still see this ripple. Unless of course I insert ferrites or similar along the way.

Thanks again

David

Hi David,

The ripple measurement at 20MHz BW has been a most commonly used standard in the power supply industry.  This would give readings consistant with what the end user system really saw.  On the other end, several factors can contribute to false ripple measurements (especially at much higher frequency) if they are not BW limited.  These factors include probe leads and scope setup which are not parts of actual system.

If you have further question, please send me an email at chung@ti.com

Thanks,

Hung