Can, except the apex is their own proto on top of a CAN physical layer. Mostly pretty annoying
Ah yes. It certainly is. It's all on github though by now.
Your power bar even has the same number of outlets as an EB8.
Reef-Pi Controllers: New all-in-one with power bar, and the original Pico
- Thread starter theatrus
- Start date
- Tagged users None
Ah yes. It certainly is. It's all on github though by now.
Your power bar even has the same number of outlets as an EB8.
OP
OP
Yeah - not all the modules were reversed but at least the key ones. Right now my design has everything actually wired to a physical CAN controller - I considered at least making sure it was dual wired to a SERCOM peripheral, but kept it simple this time around. The actual bus physical side is identical (12V, CANH/L, GND)
OP
OP
Yeah, at least for the inlet power. There is a CT around the line/neutral, which is run down to the bottom board. I may need to do a second pass on the circuit for this feature to add a much beefier low pass filter as many modern SMPSs are very very noisy
OP
OP
Logic board first articles, no all connectors mounted. Coming along!
There are a few fixes I identified from the load controller board I made (such as needing a pullup up RESET), but its a quick fix to do on the board are everything is pretty well spaced out. More soon, such as in-case shots.
There are a few fixes I identified from the load controller board I made (such as needing a pullup up RESET
), but its a quick fix to do on the board are everything is pretty well spaced out. More soon, such as in-case shots.
AbjectMaelstroM
Valuable Member
Logic board first articles, no all connectors mounted. Coming along!
There are a few fixes I identified from the load controller board I made (such as needing a pullup up RESET
Coming along nicely!
OP
OP
Some recaps from the weekend:
- The load tester / AC load has been fully built out. Some pictures.
Only a few mistakes which were fixed with some clever wire harness work (series resistor the fan PWM signal):
Its sharing the same MCU as the actual logic board, which made setup and testing much easier and carries over to the actual controller. Its a new setup for me, so getting used to the library stack, USB driver, etc all can be re-used.
The actual controller boardfg, which, minus forgetting to buy a connector, is looking like its a good fit. Note that I don't have the cosmetic face-plates designed yet, nor the top milled, but its starting to look pretty good!
The aluminum plate needs some light adjustment to let the locking tab of the XA connectors fully engage and disengage - I didn't account for its ramp like shape when cutting the clearance, but this doesn't prevent any further testing.
The board powers up nicely, all the voltage rails are where they should be, so now its a matter of lots of firmware.
Its also impressive on how small this ended up being - on your average keyboard its from the spacebar to the F-keys, and left to right from the far left out to the J key.
- The load tester / AC load has been fully built out. Some pictures.
Only a few mistakes which were fixed with some clever wire harness work (series resistor the fan PWM signal):
Its sharing the same MCU as the actual logic board, which made setup and testing much easier and carries over to the actual controller. Its a new setup for me, so getting used to the library stack, USB driver, etc all can be re-used.
The actual controller boardfg, which, minus forgetting to buy a connector, is looking like its a good fit. Note that I don't have the cosmetic face-plates designed yet, nor the top milled, but its starting to look pretty good!
The aluminum plate needs some light adjustment to let the locking tab of the XA connectors fully engage and disengage - I didn't account for its ramp like shape when cutting the clearance, but this doesn't prevent any further testing.
The board powers up nicely, all the voltage rails are where they should be, so now its a matter of lots of firmware.
Its also impressive on how small this ended up being - on your average keyboard its from the spacebar to the F-keys, and left to right from the far left out to the J key.
OP
OP
Power bricks can be huge and various orientations - its a pretty standard outlet spacing, but larger bricks can always use extensions:
OP
OP
Around there.
Project is still on-going, though I took an extended holiday break for much of November/December (much needed).
Currently, I'm working through the firmware. There are a few small fixes needed on the main board, but no major redesigns. Doing the aluminum cutouts is still pending - I've prototyped one in house but trying to outsource this work if I can. I like the build the first versions in house, but then try to not bottleneck things on how fast I can build things after something is proven
Currently, I have the following working:
- UF2 bootloader, shared with the Adafruit line of Feather boards, which means you can update it by just dropping code on a USB "drive" on your computer. I'm making some slight modifications to run with two banks of flash which the original doesn't.
- Basic GPIO and control of outlets and status indicators. I wrote a hacky driver for ReefPi control by using the file module, and will try to clean that up
- OLED display support is still not started (as promised)
- I'm working through the sensor ports. I made them "complicated" so its taking a little bit of time to work through the various modes. Specifically, there is a current compare function as well as a few functions to detect dumb sensors (floats) as well as 1-wire smart sensors (temp, pH), and reset stuck sensors, detect new sensors, etc. Each port runs independently.
- I prototyped a number of things using some throw-away C based on the MPLAB Harmony framework, which is useful. I however (as one does with no external pressures) started porting a lot of the logic and runtime to Rust. I'm planning on making use of the TinyUSB library with Rust bindings to support the USB front and not the Microchip libraries. The Microchip libraries have a crappy not-really-OSS license which I'd like to avoid wherever possible.
More updates soon!
OP
OP
Some new visuals: I'm finally on working out the tooling for machining the case top lids for the power outlets and mounting screws. All I have for this is a basic CNC6040 (Chinese metric) router, but with the upgraded spindles so it can run small tools quickly (24000RPM) and has a decent shot at shallow cuts in aluminum. This is not a precision, rigid, or even chip containing setup but will work fine for the first batch or so - plus everything is through holes so no controlled depth here.
The plan here is to use this stock plate as a fixture - I'll mill out a pocket in the plate so the actual case cap sits in the plate, giving me a chance of clamping it down. Not sure if I'll clamp it from the sides or the top and bottom, I may try to do sides and use small plate sections along the sides screwed into the plate. I also need to sort out the tool zeroing/centering setup for this fixture - I haven't gotten to a Haas MiniMill in the garage yet so no auto-probing setup for me
Since I don't have any tool changes, I want to try running everything on the same bit, and these cutters from Mari-Tool are a really good fit - bore out the #6 screw holes and cut the outlets with small corner radiuses.
The plan here is to use this stock plate as a fixture - I'll mill out a pocket in the plate so the actual case cap sits in the plate, giving me a chance of clamping it down. Not sure if I'll clamp it from the sides or the top and bottom, I may try to do sides and use small plate sections along the sides screwed into the plate. I also need to sort out the tool zeroing/centering setup for this fixture - I haven't gotten to a Haas MiniMill in the garage yet so no auto-probing setup for me
Since I don't have any tool changes, I want to try running everything on the same bit, and these cutters from Mari-Tool are a really good fit - bore out the #6 screw holes and cut the outlets with small corner radiuses.
speedstar
Well-Known Member
Nice way of making what you have work, seems like a really cool project, much cleaner that my DIY box with whips. I'll have to read back some to get the whole scope.
OP
OP
Working on the CAM files still for milling, and ordered a few more stock pieces of aluminum to work with (mostly 12"x12" 1/8" and 1/4" plates).
Last week I put together a quick revamped design based on the same exact CPU and I/O layout as the power bar to give my poor little freshwater cube light a proper LED controller and driver - it uses a long 100V string (x2) of Yuji 150mA LEDs. I'll put up design files, but the process gave me an inspiration to whip up another small toy, a top-off, dosing, or other level controller. Its small: fits in a small 50x80mm case, and gives 4 I/O ports which can either support a float sensor, or power (with PWM) a 12V pump. Also features a MXPV7002 pressure sensor which can be used a level sensor via pressure tubing.
As I said, its the same exact MCU and I/O structure, which means it effectively shortcuts the slowest thing for me to do, which is write firmware (I do software for a day job... so I try to keep things to a minimum). The goal would be for it to run stand-alone, or be directly controlled by Reef-Pi as is the 8 port power bar. And since that firmware is now in a useful shape, it means this project has some hope.
Again:
- Reef-Pi USB control
- CAN bus to extend USB (same protocol, same messages, same bus broadcast architecture)
- 4 ports which can act as either an input (12V->Signal) or a 12V output, with PWM (speed control). No reversal.
- Pressure sensor acting as a level sensor.
- Small aluminum case, pluggable terminal block.
The donor design was this AL3066 boost mode LED controller:
Which runs this revamped LED array, which is basically cobbled together from custom board with 28 5038 Yuji LEDs, a little heatsink, two sections of MakerBeam, and a 3d-printed tank glass bracket which pockets the MakerBeam:
Last week I put together a quick revamped design based on the same exact CPU and I/O layout as the power bar to give my poor little freshwater cube light a proper LED controller and driver - it uses a long 100V string (x2) of Yuji 150mA LEDs. I'll put up design files, but the process gave me an inspiration to whip up another small toy, a top-off, dosing, or other level controller. Its small: fits in a small 50x80mm case, and gives 4 I/O ports which can either support a float sensor, or power (with PWM) a 12V pump. Also features a MXPV7002 pressure sensor which can be used a level sensor via pressure tubing.
As I said, its the same exact MCU and I/O structure, which means it effectively shortcuts the slowest thing for me to do, which is write firmware (I do software for a day job... so I try to keep things to a minimum
). The goal would be for it to run stand-alone, or be directly controlled by Reef-Pi as is the 8 port power bar. And since that firmware is now in a useful shape, it means this project has some hope.Again:
- Reef-Pi USB control
- CAN bus to extend USB (same protocol, same messages, same bus broadcast architecture)
- 4 ports which can act as either an input (12V->Signal) or a 12V output, with PWM (speed control). No reversal.
- Pressure sensor acting as a level sensor.
- Small aluminum case, pluggable terminal block.
The donor design was this AL3066 boost mode LED controller:
Which runs this revamped LED array, which is basically cobbled together from custom board with 28 5038 Yuji LEDs, a little heatsink, two sections of MakerBeam, and a 3d-printed tank glass bracket which pockets the MakerBeam:
Last edited:
OP
OP
First prototype of the work holding tool for the power strip case, in this case the base plate. The idea is to have a repeatable fixture to hold the lid, meaning I can "crank" some lids and holes in a small amount of time without spending forever on setup.
This prototype is done on a scrap of 1/4" 6061, about 8x6" (its an off cut scrap). The intent is to use a piece of MIC6 tooling cast aluminum.
The bulk of the work was done with a 1/4" 3-flute 38 degree aluminum end mill. The detail work required a smaller tool - I had actually made a mistake and undersized the extra well, which meant I had to use a 2mm tool - however, it would be quite possible to use a 1/8" in the final version.
I cut the cavity and did a quick fit check - its in a useful tolerance. I still need to design a work holding setup for the plate - probably a plate I can screw on top clamping the piece to cut as though it were a clamshell.
This prototype is done on a scrap of 1/4" 6061, about 8x6" (its an off cut scrap). The intent is to use a piece of MIC6 tooling cast aluminum.
The bulk of the work was done with a 1/4" 3-flute 38 degree aluminum end mill. The detail work required a smaller tool - I had actually made a mistake and undersized the extra well, which meant I had to use a 2mm tool - however, it would be quite possible to use a 1/8" in the final version.
I cut the cavity and did a quick fit check - its in a useful tolerance. I still need to design a work holding setup for the plate - probably a plate I can screw on top clamping the piece to cut as though it were a clamshell.
Last edited:
OP
OP
Firmware work is on-going. Not much to share, except I can now share a common firmware build between the 8-outlet power bar and the high-voltage LED driver:
Tubifex worms for the planted tank are not included.
Tubifex worms for the planted tank are not included.
DirtDiggler2823
If I can't break it, it doesn't exist
New board in the works?Firmware work is on-going. Not much to share, except I can now share a common firmware build between the 8-outlet power bar and the high-voltage LED driver:
Tubifex worms for the planted tank are not included.
OP
OP
OP
OP
Where have the months gone! Some life events made a few periods of not much attention (all for the better in the end!), but back to the projects.
First off, some bad news. I was doing a bit of prep for getting the power bar certified (minimum FCC, looking at the scope and liability for electrical safety), and figuring out a few odds and ends.
Part of that process was scoping production, which hit some key snags: the processor on the board is currently expected to be available in inventory... in July... 2022. A whole year lead time, with no inventory to speak of anywhere except for the ~20 or so I have as prototypes. I'd just swap, except, well, its everything and every similar processor out there. STM32, SAMD/E, you name it, unless you have a bunch sitting in stock, you're not getting any. The trickle down effect of the chip shortage is not just high end processors and GPUs, its the little stuff getting pushed off.
I'll be continuing some interesting builds in the interim as I love exploring, but right now things remain in stasis for the power bar controller pending supply chain updates. When supply chain clarity happens, I'll work on finding the right purchase/group buy model for this and outsource the bulk of the fabrication.
First off, some bad news. I was doing a bit of prep for getting the power bar certified (minimum FCC, looking at the scope and liability for electrical safety), and figuring out a few odds and ends.
Part of that process was scoping production, which hit some key snags: the processor on the board is currently expected to be available in inventory... in July... 2022. A whole year lead time, with no inventory to speak of anywhere except for the ~20 or so I have as prototypes. I'd just swap, except, well, its everything and every similar processor out there. STM32, SAMD/E, you name it, unless you have a bunch sitting in stock, you're not getting any. The trickle down effect of the chip shortage is not just high end processors and GPUs, its the little stuff getting pushed off.
I'll be continuing some interesting builds in the interim as I love exploring, but right now things remain in stasis for the power bar controller pending supply chain updates. When supply chain clarity happens, I'll work on finding the right purchase/group buy model for this and outsource the bulk of the fabrication.
Nice to hear the wheels are still in motion but sad to hear about delays you can't control. This chip shortage is definitely real and I agree the little stuff is getting pushed off. When I first heard this 6+ months ago I never expected little old me would ever be affected, boy was I wrong. I can only imagine how many have been hurt much worse. The PCA9685 that you could get by the wheel barrel is out of stock everywhere, lately I've been getting a few from here and there but no stock until 2nd quarter 2022 and I have a feeling dates will be pushed as we get closer and as you know there's no alternatives. Same with the ATtiny85, only MLF package available as they are a bugger to program, it's getting scary when these chips feed you.
Looking forward to seeing some builds in the meantime.
Looking forward to seeing some builds in the meantime.
Similar threads
