adatadate / etta_and_gella Goto Github PK

Uploaded a few test exports to Graphics > Silk Exports

@ADataDate let me know if these are importing okay and anything I need to change.

Labeled silk needs a cross check and I need direction on what the two side pads are on the new trapezoid (connected to D7 and A3).

General thoughts - made all GNDs pop silk with mask behind it.
I like labels following edges on controller vs original all aligned in one direction.
I will be unavailable tomorrow 5am-5pm eastern time, but can pop back online to adjust artwork as needed.

left and right sides of all triangles; pins could be centered on each side, pins are currently a little closer to the top than the bottom

I also think it would be aesthetically pleasing to have all the LED pins be centered on the side that they lay on. It also doesn't seem like these pins have a consistent pitch. maybe group them with the proper pitch, group select and then rotate to avoid having to do trig?

Will Cap C10 allow for space for power silk?

Maybe move C21/C25 to the horizontal centerline of the LED wings, little symmetry thing but hey it's jewelry right? (You could also probably replace the 3 power caps with a single capacitor as the cap is just providing electrons when we go from 0-full white)

Toss a 470 Ohm resistor between Din and your first LED. Do this pretty close to the first LED. (Not completely necessary as this is usually for setups with way more pixels)

maybe add some LED's for power, maybe a simple user indicator one like the pin 13 LED on a RedBoard. I understand this may cause unwanted light in the final product but it's nice for development and you could always have some trace jumpers to deselect them

That's all I've got for ya, it looks great, the workshop will have a blast this board is unique and robust as heck. I wonder if people will trade modules with each other?

On eBay

I think the others mentioned most of the important issues. The only other thing that stood out to me:

• It looks like you only have 5x 1.0uF caps decoupling for the APDS-9960. I'm worried this won't be enough. I could not get the APDS-9960 to function with a single 10uF cap, which is why I used a 100uF tantalum on the SparkFun breakout. If you can't fit 10x 10uF caps in there, I definitely recommend going back to the 100uF tantalum.

Source as many parts from SparkFun storefront as possible with discount - dark blue lines in the spreadsheet

Obtain information to source parts that can't currently be found - light blue lines on the spreadsheet.

Source findings, chain, clasps, leather scraps, etc. (ideas?)

Consider 12C/radio only comm approach for next round.

Initial review:

1. Make cutouts uniform
2. Add descriptive top and bottom silk
3. Add silk for the flair -- -> need a combined image for this collab for bPlace
4. Beautify those fugly traces
5. Put decoupling caps in their place

Hardware will be back this week sometime.

• Looks really good!

• For the punch-out at the top of the board, you probably (?) don't need the line of drill hits on the bottom edge (that are effectively in the route out below it).

• Also, I'd suggest either pre-removing the above section (I'm guessing a user will always need to do this to connect USB), or dividing it with a route-out vertically down the middle to make it easier to snap off each half. BUT this will only work if you ensure that ground is available on both sides, which it may not currently be.

• For the traces that go through the mouse bites, spend a bit of time to ensure they're right down the center. (When I was prototyping LilyMini with a very similar design, someone from Advanced Circuits called me concerned about the drill tolerances in this situation.)

• Unless I missed them, consider adding 4.7k pull-up resistors for the I2C bus at the processor, preferably with a 3-way jumper on the VCC side so they can be disabled if necessary. The weak PUs on the I/O lines can be used instead, but in general they're not as reliable (I don't recall the SAMD specifics in this regard).

• (Minor) J29 and J19, and J17 and J28 overlay each other on the schematic.

Those are the things I noticed at first glance, I'll keep looking and add to this if I have other suggestions.
Again super cool board, can't wait to see it up and running.

Hey Mary, I hope you're feeling better!

This is a REALLY cool concept, I love the geometry and the layout! I think this will be totally successful when you're done with it.

I don't see any obvious problems in the schematic, so I'll talk about a few design paths you might consider depending on your near-term (get through this workshop!) and long-term goals.

If you're looking to prune the BOM for cost savings, I just brought up a SAMD21 with surprisingly few external parts (e.g. no USB HOST stuff), see the attached 4x4 button pad schematic for details. I'll also attach a schematic for the LilyMini, I was pretty proud of the power circuit that uses a MOSFET as a "perfect diode" to switch seamlessly between USB and a coin cell. I do think that if you're going to include a rechargeable coin cell, you should add a charger chip to charge it via USB (if not for the workshop version, then definitely for the production version).

Since the MSGEQ7 seems to be EOL, maybe just run the audio out of your mic/amp to an analog input on the SAMD21? It should be fast enough to sample audio frequencies. I think the SAMD21 has a true analog output as well if you want to run it through an amp to a small speaker in the future.

In general, the SAMD21 has quite a bit of horsepower and may be able to take over the functionality of some of the external parts like the haptic driver (replace it with a simple transistor driver and let the SAMD PWM it as needed). This would lower the cost, but shifts the hardware problem to software...

Since time is so short, I'd consider simplifying the boards as much as you can for this workshop, and use the workshop experience as beta-testing to help direct the next version.

I'm really excited to see this brought to life! Good luck and let me know what I can do to help.

@GellaCraft @mgrusin @AndyEngland521 @ShawnHymel @NPoole

Just pushed final board layout. It is still missing silk but Angela and I will figure that out later. Would love another pass and final thoughts.

Thanks Everyone!!!

I've got some notes!

D+ and D-, have you tested these to make sure theyve got the same impedance? Might be better to rotate the SAMD21 180 degrees but thats a lot of rerouting.

DI is mislabelled on the left neopixel wing

Speaking of labelling the neopixel wings, I think it might be a better idea to label them with the corresponding data pin instead of DI, as otherwise people will have to trace the traces back to find out which data pin they need to write to.

Net N$8 is unlabelled and also doesn't connect tot anything? perhaps pointing out that this is a thru connection or something, or even just getting rid of it, idk what i'd use it for after popping the board apart, but if you have a reasoning behind it, go ahead and keep it, I think it should definitely be labelled.

via between C21 and D33 seems to be misplaced

D9 signal gets a little wonky while exiting the right LED wing

Labels could be more consistent (Centered over the pad with the bottom of the label facing the pad)

label reset button?

many traces have extra little kinks, might be worth cleaning up some of the angle errors (Also the angle errors might be what's causing your crashes, I've had similar issues with angle errors)

Label VIN switch?

Relabel VCC to 3V3? Idk I just like it better