WiFive55

Just successfully tested with my AirCannon & a "rapid paintball distribution" adaptor which shoots 10 paintballs in all directions, using a 100mS pulse of about 40psi.

The important bit is that the valve closes fast enough to keep enough air in the 24 litre tank to do 2 more shots.  Maybe 4 shots in total when I fully charged to 90psi next time.

As an interesting side-effect, when it goes off, it's *much* quieter than an uncontrolled release. 

Also tried a bit of spot-welding recently, using a repurposed microwave transformer.  Unfortunately the FOTEK SSR I had was only going ON with quite a high resistance - presumably a flakey fake.  So while not thinking properly, I tried it on the WiFive55 board with the inline SSR which is only rated at 3A.  Handy note: that SSR fails-ON, so beware.  Fortunately I bought 2 inline SSRs so I replaced it for the above AirCannon test.  Still waiting on another FOTEK, hopefully genuine!

When I designed the alpha board, I thought that most SIL-PCB relays and SSRs had a very similar footprint. But when I went to buy an SSR I found that by far the most common cheap SSR has the same layout but in a larger footprint. So for the alpha board test, I've put it on wires and tucked it underneath the PCB. 

Having killed a couple of the small BS170 MOSFETs by running without a gate resistor (which the original design had) I bought some 2N7000 MOSFETs which are cheaper and lower rated but should be good enough.  Annoyingly their TO-92 pinout is not the same but at least it's just rotated 180 degrees, so still fits fine.

I also reinstated the MOSFET's gate resistor with a 470 ohm to match the LED resistor, thus reducing variety & assembly costs.

A handy aspect of the flexible configuration by solder-jumper & easy-access pin headers means that with a single jumper wire from GPIO0 to GPIO2, I can test the transistor, MOSFET & output device using just the on-board trigger button.  No need to have an ESP-01 installed!

I then put this board into the original project box, replacing the separate 555 monostable & SSR boards.  All works fine.

The latest PCB design includes these and other changes and now that this last test is complete I'll be sending off for the new PCBs soon.  I'll probably get them assembled with the essential/cheap components as SMD, leaving regulator, header pins, sockets & screw terminals to the user.  And of course leaving the final output device free for the users choice.

Identify the need, list of goals & hacked together a proof-of-concept as per the details section.

Wrote (first generation) of software using WebSockets for good a selection of pulse patterns. All OK.

Read lots about submitting to PCB prototyping companies.  Settled on KiCAD, with the simplest viable THT design to improve chances of success.  First 3 boards were perfect first time.  Assembled in 3 variations: mechanical relay, opto-isolator & off-board SSR (will add onboard inline PCB SSR when I find it).  Only have a few little SMT regulators to hand to soldered on legs & a heat-sink.  All pics are of these first boards.

Test in every configuration with both "ESP-01" & "ESP-1 S".  Tweak a couple of component values.  All OK but identified a few opportunities to improve layout - particularly for mains traces.  Start the feature-creep!  Add MOSFET option.  Add edge castellations for direct Fotek SSD mount.  Test with 1-Wire DS18B20 - no additional components needed.  Update PCB design for next version including looking at how to mix THT & SMT for assembled or kit options.  Look at BoM management & component sourcing,

Start planning new software architecture to support multiple concurrent control streams.

Outline software for special cases using teamed boards.

ToDo:

Finish combined THT/SMT PCB design & get new boards made & assembled.  Q: How many?

Finish second generation software - no hurry as gen1 does what I need & still better than any other.