Shop OBEX P1 Docs P2 Docs Learn Events
Cluso's P2SBC board(s) - Developer and OS and CPM board (a new question) — Parallax Forums

Cluso's P2SBC board(s) - Developer and OS and CPM board (a new question)

Cluso99Cluso99 Posts: 18,069
edited 2020-10-10 20:48 in Propeller 2
Obsolete
New thread started here...
forums.parallax.com/discussion/172262/clusos-p2retroblade-and-p2maxiblade-interest-please/p1?new=1

Update
I’ve renamed the thread from “Cluso’s P2 OS and CPM board (preview)
The specs have changed from those listed below so please see the latest posts. I’ll start a new thread once these boards are ready.


Here is a preview of my P2 PCB capable of running my P2 OS and Z80/CPM :sunglasses:

Features
* P2 (P2X8C4M64P) with 20MHz SMT xtal
* Transistor reset circuit (as per P1) - optional by pcb solder link
* LDO regulators for 3V3 and 1V8 (no switching power supplies)
* 1 TTL Serial port connector (can connect to cheap $1.50 CP2102 USB/TTL board) - for downloading code, PC connection, etc
* microSD - Boots from microSD (no Flash chip on pcb)
* VGA female connector
* 2 x microUSB connectors (direct to P2) - can connect to USB/PS2 keyboard and/or mouse, or provide 5V power to pcb
* 20 uncommitted I/O - 2 x 12x0.1" pitch connectors on 1" centres providing 3V3, GND and either P00-P09 or P38-P47 direct connections to P2.
* Size 1.3"x2.1" (VGA over hangs board by ~0.35") which minimises the footprint if the VGA connector is not fitted.

This board will not do everything for everyone! It will meet the low cost and simplicity of design though it does have enough general features for many uses.

Currently I am not certain whether I will produce this pcb because I am unsure of demand.
So I'm interested to hear your thoughts for the board as-is.

P2_CPM_PCB_20200915.jpg
Note: The 3D connectors shown for the microSD and microUSB connectors shown are not the ones used as I don't have the correct 3D part files.


P2_CPM_PCB_20200915.jpg
472 x 737 - 95K
«1

Comments

  • Peter JakackiPeter Jakacki Posts: 10,193
    Looks good Ray, even though it is not a "Flash" board :)
    Is it being powered via microUSB? How do you split the power and HID?
    I think the USB-A female to microUSB male cables that you would need for keyboards are OTG types which might make a difference maybe.
    Are you doing this on a 2-layer?
  • Cluso99Cluso99 Posts: 18,069
    edited 2020-09-15 01:59
    Peter Jakacki wrote: »
    Looks good Ray, even though it is not a "Flash" board :)
    Is it being powered via microUSB? How do you split the power and HID?
    I think the USB-A female to microUSB male cables that you would need for keyboards are OTG types which might make a difference maybe.
    Are you doing this on a 2-layer?
    Thanks Peter. No, it's not a "Flash" board ;)

    Can be powered by either microUSB or the TTL header which may be from a ~$1.50 CP2012 board or just the 5V & GND pins. All are in parallel.
    https://ebay.com/itm/CP2102-USB-2-0-to-TTL-UART-Module-6Pin-Serial-Converter-STC-Replace-FT232-Black/382658228394?epid=13013617291&hash=item59183530aa:g:AzIAAOSwgGZcB3PW
    Need to check the microUSB (OTG?) to USB-A male cable connections to see if they will work.

    Most people already have microUSB to USB-A male cables. Then just add a USB-A female to USB-A female gender bender ~$1.50
    https://ebay.com/itm/USB2-0-Type-A-Female-to-Female-Adapter-Coupler-Gender-Changer-Connector/402362005886?hash=item5daea4f97e:g:THIAAOSw3WZfNkL3

    Or, cut the USB-A end off the microUSB to USB-A male cable you already have, and replace it with the cable on your keyboard (provided you want to donate the keyboard to this project). A keyboard is only worth ~$10 these days. Same can be used for a mouse if desired but will require software to ustilise the mouse.
    I'm thinking a P1 board (like my P8XBlade2) could be used to replace the keyboard IC and cable, converting the keyboard to a 3V3 serial TTL connection.

    Yes, 2 layer. I'm contemplating adding provision for (but not fitted) underside bypass capacitors which might help the ADC. I did this on my original 1"x1" layout but it makes production more difficult so I shelved the design for now at least.

    My expectations are that 500mA will be the max so I don't see any problems with this design as far as heat or power problems. IMHO the heat and current issues are way overstated and may only be present in complex edge case designs.
  • dgatelydgately Posts: 1,630
    I don't see a power-on LED? And, possibly a single LED from an available I/O pin?

    Just might help diagnosing start-up for new users.

    dgately
  • TubularTubular Posts: 4,702
    Nice rendering Ray

    It'll be interesting how warm it gets using LDOs. Hard to predict, better to just try it
  • Cluso99Cluso99 Posts: 18,069
    Tubular wrote: »
    Nice rendering Ray

    It'll be interesting how warm it gets using LDOs. Hard to predict, better to just try it
    I have a pair of diodes in the 1V8 input path which will drop a bit - 0.7V-1.1V each depending on current.

    I've measured my P2EVAL 5V current and the max I read was 140mA. While there is certainly more work available, I'm just not seeing any outrageous power usage in the target for this board.
  • Peter JakackiPeter Jakacki Posts: 10,193
    The P2EVAL 5V current equates to 5Vx140ma = 700mw. But that is a switcher and most of the current is to the 1.8V supply so expect 700mw/1.8 = 389ma for the 1.8V LDO. Maybe you'd be better off cascading those LDOs if you haven't done so already.

    The other problem is with such small diodes that they could reflow themselves right off the pcb with the heat that they will generate. With Schottky diodes you can have high currents because they have a much lower voltage drop but here you want the voltage drop and you want the current. Maybe you can't have both.
  • Christof Eb.Christof Eb. Posts: 1,197
    Great!
    This board does make a lot of sense for me! Most of my projects do not need too many pins. Breadboard or experimenter pcb friendly! SD card is very useful!
    Those 2 Leds would make sense but can be added easily by the user.
    Perhaps you can add 5V to the right inline row?
    Spi Ram would be nice.
    But it is time, that a board like this becomes reality!
  • RossHRossH Posts: 5,462
    Cluso99 wrote: »
    Here is a preview of my P2 PCB capable of running my P2 OS and Z80/CPM :sunglasses:

    How much to add an 8 inch floppy disc drive? :)

  • Cluso99Cluso99 Posts: 18,069
    edited 2020-09-15 07:12
    RossH wrote: »
    Cluso99 wrote: »
    Here is a preview of my P2 PCB capable of running my P2 OS and Z80/CPM :sunglasses:

    How much to add an 8 inch floppy disc drive? :)
    My CPM machines only ever had 5-1/4" so I can add that. I do have an 8" floppy diskette, but not a drive ;)

    Peter suggested I could trim an 8" to 5-1/4" just like I did for the 16 I/O P1 ;)
  • Cluso99Cluso99 Posts: 18,069
    edited 2020-09-16 03:07
    Did a test of the regulator circuit with a 500mA load.
    The pair of diodes drop 2.08V at 500mA which is about what I expected.
    There is basically no copper for heatsinking on the diodes or regulators because each pin is soldered to the edge of a thru hole on a piece of protoboard.

    Yes, the regulators get quite warm and the diodes do get a bit hotter. This is subjective of course. Neither are as hot as I had the original glob top P2 running. I've been turning my garage upside down looking for the temp probe but alas I cannot find it :(

    While I doubt the diodes are going to unsolder themselves provided there is sufficient copper on the pcb (which I have space for), there isn't going to be enough difference to bother with the diodes. So I am just going to series the regulators as Peter suggested.

    For 5V 500mA, the 3V3 regulator will drop 1V7 @ 500mA = 850mW. Assuming no 3V3 usage (worst case) then the 1V8 regulator will drop 3v3 - 1v8 = 1.5V @ 500mA = 750mW. Assume the P2 with 1V8 @ 500mA = 900mW. Of course there will be some 3V3 current and hence power. So provided the total current draw is limited to under 500mA then there will be no problem with this design. Sure, there will be some things you might not be able to do with this design, but they will be minor IMHO.

    I found a microUSB cable to female USB-A and it meters correct so this is good alternative to using a gender bender.
  • msrobotsmsrobots Posts: 3,709
    Nice.

    I want at least one.

    Mike
  • hinvhinv Posts: 1,255
    Excellent! Mark me down for one. I really like the idea of a low cost solution of a P2 hanging off a USB port so I can program it via Linux PC, and this would fit the bill. Even better it will run your P2 OS and CPM with just a keyboard and VGA? Sound's like a winner to me, and would fill a currently missing low price point entry into the P2 space. I would love to see something like Sphynx on this.

    What happens to the uncommitted IO's. Is it possible to provide holes on the board even big enough for a 30guage wrap/bodge wire to get soldered into?
  • Cluso99Cluso99 Posts: 18,069
    hinv wrote: »
    ... Even better it will run your P2 OS and CPM with just a keyboard and VGA?
    Yes
    . ..
    What happens to the uncommitted IO's. Is it possible to provide holes on the board even big enough for a 30guage wrap/bodge wire to get soldered into?
    No. May as well use 0.050” connector, so P2D2 is the better fit for this.
  • hinvhinv Posts: 1,255
    Decisiveness is good. Please let me know where to order. Your previous boards have been great!

    Thanks,
    Doug
  • hinvhinv Posts: 1,255
    Oh, also, please accept Bitcoin Cash as it can make it a whole lot easier to accept payments.
  • Cluso99Cluso99 Posts: 18,069
    edited 2020-09-22 08:01
    Today I did some measurements on a prototype for the power supply...

    By daisy chaining the 3V3 regulator output to the 1V8 regulator input, at 545mA (load of 10 * 3R3 1/4W resistors) the 1V8 regulator gets too hot (I measure 80C) and shuts down. If I power the 1V8 regulator via two series diodes which gives an input of 3V, the regulator is fine at 545mA (I still measure 80C). The diodes get to more like 85C.

    Now remember, basically there is no copper heatsinking on these diodes or regulators in my test setup, so in practice these parts will run cooler.

    I light of these tests, I have decided to make two versions of the pcb.

    P2_MiniBlade
    This pcb (shown below) will now only have one miniUSB and the pair of I/O 0.1" pitch headers will now be 10 way instead of 12 way (retaining 1.0" pitch between connectors), so there will only be 8 I/Os (P0-7 and P32-39) brought out to the headers. A pair of LDOs will be used with the 3V3 feeding the 1V8 regulator (ie without diodes). The pcb now reduces in size to 1.1" x 2.075" 2.0” with the VGA connector overhanging the sides of the pcb slightly as well.
    Here is a 3D model
    P2_MiniBlade.jpg

    P2_MaxiBlade
    This pcb will use the dual 1A switcher IC MPM38111, will retain two microUSB connectors, and will be extended to bring out 12-16 I/Os per side (to be decided) by using a larger pcb.     See post below.
    P2_MiniBlade.jpg
    468 x 739 - 84K
  • PropGuy2PropGuy2 Posts: 360
    For my use I need something more utility -. the 1v8 and 3v3 LDO regulators, Flash memory, and USB port for programming, and of course the majority of the i/o pins brought out to some type of connector / termination, that's it. A modified "P2 Edge" type board would be great. The uSD, VGA and other devices I can add myself for what ever project I am working on.. Its the pick & place and soldering of the P2 chip I can not easily do. Several months ago, someone on this forum mentioned that they bought a tray of P2 chips for a bare bones PCB, not sure who that was. My 2 cents.
  • Cluso99Cluso99 Posts: 18,069
    Update...

    P2MiniBlade
    * pcb is 1.5"x2"
    * two 20x0.1" headers with GND, 5V, 3V3, GND and 16 I/Os
    * one 6x0.1" header with RST, SO, SI, 5V, GND, nc for connecting to CP2102 USB-TTL pcb
    * two microUSB connectors for keyboard and mouse and/or 5V supply and/or possible PC connection via USB and P2 software
    * microSD socket
    * VGA female socket
    * 20MHz xtal
    * 1V8 from two SOT223 1A LDO regulators in series (steps down via 3V3)
    * 3V3 from one SOT89 1A LDO regulator

    Notes:
    * No Flash chip
    * At 600mA for the 1V8 supply the 3V3 regulator will dissipate 1.02W, the 1V8 regulator will dissipate 0.9W, and the P2 will dissipate 1.08W excluding the 3V3 dissipation.
    * Probably will not do the P2MiniBlade now

    P2_MaxiBlade.jpg
    P2_MaxiBlade.jpg
    846 x 764 - 130K
  • Cluso99Cluso99 Posts: 18,069
    Checked the VGA footprint and discovered that the footprint I used is incorrect. This has extended the pcb 0.050" to 2.05" :(
    I cannot reduce the pcb at the top end without moving the 6pin header in the center of the pcb off the 0.1" pitch :(
  • TubularTubular Posts: 4,702
    If you need them, there are 'slimline' VGA connectors that typically compress the usual 3 rows down to 2. They're a bit less robust however. Some have reinforcing arms etc. I think Parallax used a good one (edac?) with the C3
  • Cluso99Cluso99 Posts: 18,069
    edited 2020-09-24 09:10
    Tubular wrote: »
    If you need them, there are 'slimline' VGA connectors that typically compress the usual 3 rows down to 2. They're a bit less robust however. Some have reinforcing arms etc. I think Parallax used a good one (edac?) with the C3
    4uconn have them but with an MOQ of 2,000 :(

    Anyway, I have compressed it back to 2"x1.5" and hoping I can move the 6pin hdr to keep it on the standard 0.1" grid.
  • jmgjmg Posts: 15,173
    Tubular wrote: »
    If you need them, there are 'slimline' VGA connectors that typically compress the usual 3 rows down to 2. They're a bit less robust however. Some have reinforcing arms etc. I think Parallax used a good one (edac?) with the C3

    Higher density ones, (like EDAC Inc. 634-015-274-992) are more expensive than the older, larger ones.

    These
    https://lcsc.com/product-detail/D-Sub-Connectors_D-SUB-15pin-three-rows-Female90degree-black-pitch2-29mm-Tin-plating-With-hex-screws4-8mm_C75754.html
    are ~20c/100

    So it's a trade off of what you want more : lower BOM, or lower PCB area.
  • TubularTubular Posts: 4,702
    I think the ones I used are Norcomp, around $1.50 to $3 ea
    https://au.mouser.com/Search/Refine?Keyword=200015213
  • TubularTubular Posts: 4,702
    I can post some of the norcomp vga to you if you like

    Also ebay (and everywhere) have the 15 pin M-M mini gender changers if you want to plug directly into monitor they are only a dollar or two
  • Peter JakackiPeter Jakacki Posts: 10,193
    Those VGA connectors were fine on big PC boxes but are horridly huge on tiny pcbs. Then you have to use those horrid VGA cables as well.

    Why don't you just put down a 16-pin header and use one of those common IDC to VGA adapters found on ebay? That way you can choose to use a socket or a direct plug into the monitor.

    wholesale-IDC-to-Mini-VGA-Cable_17414581094d4faa2ba97894web_up_file_398x398.jpg
    398 x 398 - 23K
  • AwesomeCronkAwesomeCronk Posts: 1,055
    This looks like a great way to get into the p2!

    Was this done in KiCAD?
  • Cluso99Cluso99 Posts: 18,069
    Yes. My first board with Kicad. I’ve used (owned) Protel 95 for many years and Protels Advanced PCB V3.1 and Autotrax before that. Done lots of commercial boards with it. see www.clusos.com for some pics. I tried Protel 98SE but not so impressed so decided to try Kicad. Altium got way too expensive for small companies.

    Kicad has been fairly easy to learn the basics. I haven’t exported the gerbers yet though.
  • jmgjmg Posts: 15,173
    Cluso99 wrote: »
    * microSD - Boots from microSD (no Flash chip on pcb)

    Flash chips now come down to 2x3mm and even 1.5 x 1.5mm DFN, so there may be space to allow for one of those small packages ?
    eg GD25D80CKIGR 8MBIT 3.3V SPI USON8 1.5x1.5x0.45mm
  • hinvhinv Posts: 1,255
    @Cluso99 Wouldn't it be easier and less expensive to just make the P2_MaxiBlade? How much smaller would it be?
  • hinvhinv Posts: 1,255
    Nevermind...seems you came to the same conclusion.
Sign In or Register to comment.