Dome Nodes - v2

EMF was announced for next year :slight_smile:

I plan to improve on the DomeNode hardware after what we learnt previously.
Here’s what I have so far:

  • 24v-5v regulator is not precise enough. We should add a separate 5v LDO regulator alongside the 3.3v one to power sensitive devices (e.g. the neopixels).
  • Neopixels should be put on a separate ‘shield’. This would simplify construction dramatically, as these basic shields could easily be stenciled/reflowed. It also means testing will be easier as we separate the various functions.
  • Vertical RJ45s :slight_smile:
  • Potentially additional functionality testing. I populated the original board with microphone and speaker options, but it’s feasible we could shield these as well and do something more fun with them.
  • Also possibly allow fitting of a dev board instead of ESP module for easier soldering/development.
  • Easier programming, specifically using the special ESP programmer with built in reset/program ability.

I should also mention I was very pleased with some of the specific successes we had:

  • Speed of setup was quite impressive
  • The touch plate mounting holes on the PCB for the capacitive sensing worked really well (and I am quite pleased with their design)
  • We turned the design/manfacture of the boards around super quickly.

It also occurs to me that we could simply use some off-the-shelf neopixel ‘rings’ instead, or, the power RGB LED :slight_smile:

Whether you could use an additional 78(S)12 regulator in cascade to help step down the 24v supply. Similar to using multiple zener diodes to improve on regulation back in 60’s- early 70’s before the 78/79 series regulators were introduced. With the extra regulator, the available current drawn should go up with less voltage across input to output ( keeping within permissible dropout limits) while keeping with much the same power dissipation.

You will still need to use a discrete component dropper circuit if interfacing directly from the 24v power lines( see Don Lancaster’s CMOS Cookbook) consisting of a npn/ pnp transistor pair with appropriate bias resistors. Other than op- amps I don’t know what other ic logic circuit would run off 24 volts. I think the commercial Led power supply should cater for this in itself if it has a low voltage 1- 10 v control input that could be implemented in 3 ways according to psu’s type code and supplied instructions. For the external microcontroller/logic psu the 3v3 regulator has an absolute max. voltage input of 15 volts and whether this can be supplied/supported by the Led psu.
Following 9/7/2019 Boiler room meet where I examined the node display module I finally decided the 3v3 rail supply is not really necessary as there is a on board voltage regulator(in this case non decoupled!). In this case an extra 5v lo-dropout regulator could be used connected to the USB header for the display module chipset supply( breadboarded circuit has orange/amber LED ( following pc ATX supply colour coding) which its series resistor needs raising in value to 470R if a dual rail 5v supply option is implemented by replacing the appropriate regulator). The 3v3 volt regulators I had did not work. Noted extra header pins on display module is needed for 24 volt supply. Display module chipset uses an open collector/drain driver for which the logic power supply seems to be via a 4 way header very similar to USB headers found on pc motherboards.There are also 6 way 3 by 2 headers.
See ESR Components website for a commercial Constant Voltage 24v supply priced at over 55 pounds ex vat and delivery(for dimmable 24v 150W). Recently viewed more 24v LED PSU’s on Rapid Electronics website. Could split output via IP rated junction box with cable gland grommets (available at Toolstation/Screwfix/CPC Farnell or Rapid).
Communication with display module carried out by 2 RJ45 sockets or USB header? ( Have 2 Buccaneer style RJ 45 IP rated ones there at Boiler room).
Reading up on Audino module supply voltage says 7 to 12 volt recommended so could use an 7808/7809 regulator connected to Arduino’s dc in jack unless a 5v external source via Arduino’s 5v or Vin header pins has been derived. What has been built already other than my breadboarded logic psu with heavy duty heatsink?(All at Boiler Room).
Also have self built LM317 K supply with a 2x12volt 30watt ac input and an old zener/transistor -ve regulated 12/24volt output supply for burn in testing.Would like to see if these can support node display load testing with power resistor using I = V/R or R = V/I equations.

For a discrete Led string driver circuit from a pi microcontroller board’s ULN Darlington driver outputs, thought about using a compound complementary driver(npn driver operating a pnp power transistor connected similar to a Darlington pair) for the LED’S. Following the ‘Better Buffers’ article in Radio and Electronics World back in the late 1990’s. Also mentioned in Power Supply Projects and Epemag. Driver transistors try Zetex ZTX451 or 651 series. For pnp power try TIP32,42 series. Chief advantage is lower turn on voltage over normal Darlington transistors. Hfe current gain question is just that of the driver transistor in this connected arrangement. Npn power are TIP31,41 and alternative driver transistors are BC 337,327 npn,pnp respectively. After 21.6.2019 meeting, I’d left out an extra 470 to 1k ohm resistor between the TIP41 base and the 0v ground rail. This should rectify the floating driver transistor rail defect on the sketch output driver circuit diagram I came up with(now corrected). For hard to get components try Èsr Components for the complimentary TIP power series or Cricklewood Electronics for the Zetex driver series. The BC337npn and BC327pnp driver stage transistors are 45 volt 500mA(0.5 amp) although some suppliers quote 800mA. The power transistors finally settled on are TIP42C for 24 volt supply operation( following electronics author R.A Penfold’s rule of power semiconductors being rated at 4 times supply rail voltage). These are connected to the positive supply rail mainly to provide a pull down to 0v across load voltage when switched off therefore compliant with latest saftey regulations.

Just thinking about a stand-alone constant current driver using the LM317 T variable voltage regulator ic. A foreseeable problem is the voltage drop across itself. Wonder if it could be used in place of the passive ballast resistors on the TIP41/42C outputs on the discrete Led string circuit when running off 2 (3 with LM338 x2/3 in parallel derived 24v supply voltage regulator) SLA/car batteries in series. Has anyone thought about using 2-3 L296 switch- mode regulators? Thought about using power Mosfets for their claimed low voltage drop but as yet have no datasheets/books/schematic diagram about constant current operation which places the circuit into the analog category. Just thinking on what components already there and come up with a current sink circuit formed from a single transistor with 2 diodes in that transistor’s base circuit as shown in Babani Publishing Circuit Source Book 1 and Circuit Exchange International website in Power Circuits section.
New topic about input switches - use either 74HC14 with RC circuit or 74HC74 to debounce any mechanical switch.
23/07/19 Things are looking up for the microcontroller psu in that the 3v3 supply rail is now working. Turned out that the LD33V regulator has a different pin out to the 78 series convention( similar to the LM317 exchanging adjust for ground terminals). The only foreseeable disadvantage is that it will require an insulation kit if mounting on a common heatsink with the other 2 regulators. A further note if using a commercial Led psu is that some versions do have an auxiliary 15 volt max output for microcontrollers. Would suggest reducing the 1st input regulator to 9v 78S09 in this case, as currently has the 78S12. With thanks to Matt(administration) for getting the 3v3 LD33V regulators working. For the microcontroller ( other posts say a pi is going to be used) there is a Velleman VMA404 2.5A dc- dc converter module from esr.co.uk. It’s output is adjustable and will need to be set to 5v for the pi microcontroller and 3v3 for display node chipset. If the input voltage gives too much power dissipation use a zener- transistor pre regulator of about 12 - 18 volts. A LM338 could also be used as a pre regulator if 2 of the forementioned switched mode modules are used. Alternatively use a 12v open frame switched mode supply with the Velleman modules connected via diodes (UF5402/4) to the output.

Looks like the project is coming on a treat !

I’ve been having a good think about the firmware and overall architecture for DomeNode v2. Hopefully, I’ll be getting started on it soon.

I have a small side project of a music visualiser that I’ve started to work on - I’m hoping it will provide a slightly simpler entry into the complexity of the DomeNode solution as a whole.

My thoughts, so far:

  • We should have a router in the space as part of the setup that all of the nodes are connected to. I will provide one. This can then be taken to wherever the dome is set up.
  • This router should also be accompanied by some sort of central server, e.g. a raspberry pi. This pi will provide several features:
  • A central firmware update repository for automatic OTA updates
  • An MQTT server that the nodes can communicate to, for a central logging solution (for easier debugging and monitoring). I choose MQTT because I already have experience with it and embedded ESP32 devices. We could also use something like rsyslog.
  • Some sort of central control (someday!)

I have recently discovered ArtNet and will be experimenting with this. It’s essentially a way to configure a lighting ‘fixture’ and control multiple fixtures from a central point. The particulars of a light show are defined by a compatible lighting design software, of which several I’ve seen are open source.