So the V0.2 boards came in today. Less than about a week and a half total turnaround time... Not bad!! That's about the last time we can use "not bad" though because everything after this is bad...

Okay maybe not everything. Lets break down each system and see how I messed it up this time:

5V in to 12.6V boost

Actually, this circuit entirely tests and works! So a great start.

TPS5100 Buck charging circuit

Well this would work, but in likely a sleep deprived state of changing the sense resistor from a 2510 to a smaller package, and likely trying to find one already in my PCB assemblers' loaders, I went from a .068 ohm resistor to a 68 ohm resistor. Those are not the same numbers. So When a battery was plugged in, it was charging at a current I couldn't even pick up with my shop power supply.

Also, in likely the same sleep deprived state, and trying to find the loaded version of the LED i had for the STBY indicator I somehow replaced this with an 0603 resistor. Then I just decided to do this 2 more times in other places on the board. But once putting the smallest SMD resistor I had around in (a 1 ohm), I got a good charging state out of it. Not that I could see that as the LED was a resistor :D

8.4V to 5V buck circuit

Apart from it's on LED being replaced by the resistor, this worked great. The only real change to this was putting a larger EL capacitor in, which shouldn't have done much anyway.

5V to 3.3V LDO

This actually worked great! Though thinking of the power budget, this IC was putting out 500ma at most. I wanted a little more overhead given the chip and that I'm using the amplified host board for the NRF24L01+, so I'll want a little more overhead here.

Integrated ESP

The good news here is that I was actually able to boot/reset it and have it show up via. a USB connection as a serial device. This means the supporting circuitry around the crystal and other strapping pins was good... I had to drop the upload BAUD down to 115200, likely because of my long and likely interfered-with USB trace. Near term solution here is to remove the data connection to the USB connector - reason being im only really using that to initially flash the chips. Otherwise, it's unused. So ill just break out pads to use close to the chip and make something i can clamp it into to program it...

But there was a bigger fundamental error here. Here I am assuming that an ESP32-S3/S2 just had integrated flash as every dev board I see hasn't had an external flash chip. You guessed it - I was wrong. In fact just a couple variants do, and most don't. So after 15 or so minutes of being connected to the chip trying to discern what flash it had it dawned on me... what if it doesnt have flash? Sure as shit, looking at the data sheet revealed that I was using the ESP32-S3 ... not one of the flash-equipped ESP32-S2FH2, ESP32-S2FH4 or ESP32-S2FN4R2. You see that F? That F stands for Flash. It also stands for the word most adequate for realizing these boards will take a lot of work to even make into frankenboards for testing.

Looking at the PCBA manufacturer's inventory though, those ESP32-S3 variants were either fairly sparse, or generally expensive... So I asked myself - why am I even using the S3? It's just the receiver so I definitely don't need the dual core, and I'm tested on the S2 anyway... So I completely redesigned that sector around the ESP32-S2FN4R2. Fun.

Oh, V0.3 so soon?

So here's the V0.3. Effectively addressing all of the above issues, likely still with some left undiscovered, but regardless will likely be flying over the atlantic sometime in the next couple weeks.

A Happy Ending - The Cue Boards

While sorta-scrapping the V0.2 receivers makes me a little sad, it's okay. I learned a lot for it and it was actually the first time I tried to integrate the MCU directly on board vs just socketing in a dev board. So failure here as marketed as learning and a lesson that still vastly dwarfs the cost of an actual Electrical Engineering education.

But in the box with those 5 failed receivers were 20 of these beauties of cue modules -

The fit inside the printed enclosure is perfect, and at least with the test bed made from a v0.1 receiver, worked perfectly, addressing the issues from the first revision. I have a suspicion that some of the ground path might not be super ideal, which might be limiting when I start really pressure testing simultaneous firing, but otherwise im super happy with the current state of them.

What's next

While im waiting on the next revision of the receiver boards, I'll see if just for fun, I can solder in some external flash to the v0.2 receivers and get them going... But apart from that, I'll do some heavier testing on the v0.2 cue boards and probably iterate a bit on their cases. I'll chain a few cue boards and test that all out, in addition to some end-to-end testing for firing, as well as continuity testing and battery voltage sensing.