First Custom PCB

[set321go]

Hi Folks

I'm trying to make a stroke computer for rowing, there are some products out there but I thought it would be an easyish first real project. I've prototyped the idea on a mkr1000 and a lucky shield and it works fairly well, I did have some problems with the SPI and breadboard connections to an SD board but I think that was mainly due to poor connections and wires moving around (i had similar issues with i2c until i taped them all up).

I've been using kicad to put together a design and i'm at the point where I think i'm ready to place components on a pcb but it would be great if I could have some input from those more knowledgable than I (i'm a software guy no background in EE) on the schematic. I really struggled with were inductor sizing and I feel i've only really scratched the surface in that area. Some feedback on my power & charging circuitry especially would be much appreciated.

For a design like mine that is primarily going to be low current + low voltage how important is the inductor & capacitor sizing?

(I don't seem to be able to generate a jpg from kicad so i've attached a pdf)

[nAyPDJ]

Power flows from up to down, and signal flows from left to right. You mostly follow this convention, but you could improve things in a few spots.

I generally also prefer to use global labels in KiCad, since they allow me to indicate in which direction the signal is flowing (see attachment). This is just personal preference though.

What is the "-BATT" and why have it if it is equivalent to GND?

At first glance, 33u seems huge for the LM2731 boost, especially with 1.6MHz switching. Did you follow the "Detailed Design Procedure" in the LM2731 datasheet? Is it the right value? You display claims to have a typical power consumption of 18mA.

> It might be more efficient to drive directly from VBAT

Nope. The SYS pin on the bq25895 says it's connected directly to the battery unless Vbat is under 3.5V, in which case Vbat is boosted up to 3.5V.

For a design like mine that is primarily going to be low current + low voltage how important is the inductor & capacitor sizing?

Too big and it's going to increase the board size and cost. Too small and it won't work.

[set321go]

Thanks for your feedback, much appreciated, I will go and work on making things a bit more 'standard'

What is the "-BATT" and why have it if it is equivalent to GND?

I got rid of it already, I labelled it to remind me what was going on.

At first glance, 33u seems huge for the LM2731 boost, especially with 1.6MHz switching. Did you follow the "Detailed Design Procedure" in the LM2731 datasheet? Is it the right value? You display claims to have a typical power consumption of 18mA.

Yes but i didn't find enough info there for my limited knowledge I used another document from TI  "How to Select a Proper Inductor for Low Power Boost
Converter" http://www.ti.com/lit/an/slva797/slva797.pdf

I found this much more useful for my limited knowledge but looking back at my math

Code: [Select]

IL = (13 x 0.05) / (3.8 x 0.8) = 0.214
L = 3.8 x (13 + 0.5 -3.8) / (0.3 x 0.214 x 1.6e6 x 13.5)  = 32uH
I think I used 0.05 (50mA) for my output current. I selected 33uH because it was the closest standard size. I think maybe I did something wrong there because now that comes out as 26.5uH. Using 20mA as a starting point i ended up with 66uH.

[ebastler]

I am a bit confused about your supply voltages: 3.3.V, 3.8V, and 5V. Do you need them all?

[set321go]

I am a bit confused about your supply voltages: 3.3.V, 3.8V, and 5V. Do you need them all?

3.3v is the logic voltage (ideally i could just use this everywhere)
5v is the external input voltage (usb)
3.8v is the output voltage (i think) from the bq25895
13v is the output from LM2731 for the display

I took the output of bq25895 (3.8v) for the input of LM2731 (used for the display) as it was already there instead of using the 3.3v

[nAyPDJ]

Yes but i didn't find enough info there for my limited knowledge I used another document from TI  "How to Select a Proper Inductor for Low Power Boost
Converter" http://www.ti.com/lit/an/slva797/slva797.pdf

I found this much more useful for my limited knowledge but looking back at my math

Code: [Select]

IL = (13 x 0.05) / (3.8 x 0.8) = 0.214
L = 3.8 x (13 + 0.5 -3.8) / (0.3 x 0.214 x 1.6e6 x 13.5)  = 32uH
I think I used 0.05 (50mA) for my output current. I selected 33uH because it was the closest standard size. I think maybe I did something wrong there because now that comes out as 26.5uH. Using 20mA as a starting point i ended up with 66uH.

Huh, I guess that's right. My math works out to 37uH, as I was a little more conservative with the I_max=50mA, V_in(min)=3.5, DI_pk-pk=.4:

Code: [Select]

I_L(DC_MAX) = (13V×50mA)/(3.5V×.85) = 218mA
L = (3.5V × (13V+.38V−3.5V)) / (.2×218mA * 1.6MHz * (13V+.38V)) = 37uH

but 33uH would also work fine.

[set321go]

My first pcb's arrived (no assembly) there are a few issues as I expected (me not knowing what i'm doing mostly) there to be but the quality seems pretty reasonable. I got them from JLC. I have a couple of questions about about how to make some changes using kicad.

In the closeup example where the ground pins are connected to the fill zone they haven't been filled all the way up to the pad all around but seem to have branches, can I fix this in Kicad or do i just have to live with it?

On the microSD footprint there is no solder mask between the control pins which seems odd as there are components on the board with what looks like the same or smaller pitch pins, is this a problem with the default footprint can i just adjust the size of the pads

[obiwanjacobi]

In the closeup example where the ground pins are connected to the fill zone they haven't been filled all the way up to the pad all around but seem to have branches, can I fix this in Kicad or do i just have to live with it?

That is normal. If you would fill it all with copper, you would need a lot of heat to solder to that pad, because the ground plane would have to be heated too. I would not recommend changing that.

[set321go]

Thanks, i presumed there was a reason it was that way but didn't know why.

I adjusted the width of the footprint pads that were missing solder mask to make them narrower after looking at the datasheet. I think I would prefer having solder mask between the pads rather than having a bigger pad to solder to.

onto v2