Email or username:

Password:

Forgot your password?
3 posts total
Andrew Zonenberg

There are three main types of stroopwafels.

Igneous, the most common, are formed by melting a sticky filling and pouring it between two waffles.

Sedimentary stroopwafels are commonly encountered at the bottom of a large container of igneous stroopwafels.

The rarest form is the metamorphic stroopwafel, formed by subjecting igneous and/or sedimentary stroopwafels to extremes of heat and pressure until they fuse together.

Andrew Zonenberg

After some ngscopeclient updates and a busy week at work, back to playing with the LTC3374 test board.

Here's channel 1, with back side bulk cap and no high freq cap.. Input ripple ranges from 38 mV p-p / 1.4 mV RMS at no load to 57 mV p-p / 2.5 mV RMS at max design load of 1A.

Output ripple is *slightly* dependent on load, but much less of a shift. Pretty consistent around 21 mV P-P. / 5 mV RMS.

Now let's see what happens when we run the same filter graph on the other supplies.

Andrew Zonenberg

Channel 2, bulk cap on back side plus high freq cap on the front. No changes to the output side.

Input ripple is greatly improved, ranging from 21 mV p-p / 0.86 mV RMS to 33 mV p-p / 1.7 mV RMS.

Output ripple is only slightly improved, hovering around 19 mV p-p / 4.65 mV RMS regardless of load.

So the high freq cap had a massive impact on input ripple, but fairly small improvement on the output side.

Andrew Zonenberg

First ESP32-PICO-V3 cross section has a few artifacts, but I have more epoxy to use up and I want to get better at making really nice looking cross sections anyway. So I might do another. I have a few extra chips and tow more packs of epoxy to use up.

First off, I probably should have read the datasheet. It's not just the ESP32 and flash chip, there's a 40 MHz crystal and some matching components too. So it's a full SiP with embedded passives on a four layer substrate! This is a lot more interesting to look at than just a single die, which is good.

Less good: the die is way off center in the package. This is fine if you're expecting it, but I ended up sanding into the die using coarse sandpaper during the early stage of the cross section and creating some cracks which are difficult to get rid of. Lesson learned, I'll be prepared next time.

First ESP32-PICO-V3 cross section has a few artifacts, but I have more epoxy to use up and I want to get better at making really nice looking cross sections anyway. So I might do another. I have a few extra chips and tow more packs of epoxy to use up.

First off, I probably should have read the datasheet. It's not just the ESP32 and flash chip, there's a 40 MHz crystal and some matching components too. So it's a full SiP with embedded passives on a four layer substrate! This is a lot more interesting...

Andrew Zonenberg

Colored interference patterns on the plates of the ceramic capacitor seen at low mag. You can tell just how dense modern MLCCs are!

Go Up