What makes the perfect FLiP project board? Leave your ideas here please!

Ken GraceyKen Gracey Posts: 6,285
edited April 14 in Propeller 1 Vote Up0Vote Down
Hey there,

Like many of you, I'm coming across projects where I'd like to dedicate the FLiP Multicore Microcontroller but I've never designed a PCB (it's been the ONE SINGLE goal at Parallax I have yet to achieve) so I'm usually plopping it on a breadboard or using some perf-board with point-to-point soldering. It's time that we make a single "FLiP Project Board". I'm looking for your ideas to make this board exactly what it needs to be for most of us. I'd like you to deposit your ideas here. To get this discussion going, I'm going to pose some basic ideas which are all mine, but you can comment on them.

FLiP Project Board Ideas

3"x4" standard Parallax size
No manufacturing done at Parallax - PCB + all solderable parts provided in a bag
Less that $15-20 cost to customer - maybe even $10 - cheap enough to hand out now and then with a FLiP purchase (low-cost to allocate to a project, and use another instead of rework a current project)
No power supply or SMT parts
2.1 mm power jack location
Sockets for FLiP Module
4-6 3-pin header locations for servos or sensors

That list wasn't intended to steal the fun or identify the requirements for this project.

I've asked this same question before in another thread and I didn't follow through. I couldn't find that thread but somebody will locate it and post it here for x-ref. This time we will follow through and have this board ready by June since it's such a simple effort.

If you're interested, share your ideas and let's just get this one done!

Ken Gracey
Parallax Inc.
«1

Comments

  • 55 Comments sorted by Date Added Votes
  • Ken Gracey wrote:
    3"x4" standard Parallax size
    Actually, that's 3.05" x 4", with 1/8" corner holes on 2.75" and 3.7" centers.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 22,027
    edited April 15 Vote Up0Vote Down
    How about a 4-digit 7-segment LED display? They're cheap and multiple-sourced. Without a shift register, it will use up 12 pins, but so what? If it's multiplexed 1:8 1:32, you don't need a current driver, just four resistors for the common anodes or cathodes.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • jmgjmg Posts: 12,087
    How about a 4-digit 7-segment LED display? They're cheap and multiple-sourced. Without a shift register, it will use up 12 pins, but so what? If it's multiplexed 1:8, you don't need a current driver, just four resistors for the common anodes or cathodes.

    -Phil
    For more info on 4 digit LEDs and some uses, see this thread ....
    https://forums.parallax.com/discussion/168215/2-clock-kit/p1

    You would need a high brightness LED model, in order to drive from MCU pins.
    The 4 digit examples in the clock thread use shunt-drive which is simpler circuit wise, but a little more power wasteful.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 22,027
    edited April 15 Vote Up0Vote Down
    jmg wrote:
    You would need a high brightness LED model, in order to drive from MCU pins.
    True. I did a circuit like that with a regular-brightness display, and it's pretty dim. A step up would be to use a ULN200x ($.201 @ 1000 pcs. from Mouser) to drive the common cathodes, and multiplex 1:4. You'd need eight resistors instead of four. I've done that, too, and have gotten more-than-adequate brightness.

    As a plus, the three remaining ULN200x drivers could be used for something else.

    -Phil

    Correction to post above: With four resistors and without a current driver, you have to multiplex 1:32. That's why the display I did that way is so dim.
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • Publison wrote: »

    That's it! Nice find Jim (missed you last weekend, too!).

    Ken Gracey
  • jmgjmg Posts: 12,087
    jmg wrote:
    You would need a high brightness LED model, in order to drive from MCU pins.
    True. I did a circuit like that with a regular-brightness display, and it's pretty dim. A step up would be to use a ULN200x ($.201 @ 1000 pcs. from Mouser) to drive the common cathodes, and multiplex 1:4. You'd need eight resistors instead of four. I've done that, too, and have gotten more-than-adequate brightness.
    As a plus, the three remaining ULN200x drivers could be used for something else.

    This one
    https://lcsc.com/product-detail/Led-Segment-Display_0-56-Digitron-red-RED_C109200.html
    claims 60mcd at 20mA, and is 40c/100 - could be worth trying ?

  • Would it be something that still has a prototyping area on it but just has a few onboard features - kind of like a Demoboard/Activity Board/BOE (sorry, it sounds like a bit of a stupid question as it rolls off my fingers, but I wasn't sure)?

    Maybe something to break out P28 and 29 from the FLiP, for I2C. I know any pins could technically be used for it, but I think the FLiP already has pullup resistors on these pins? (unless all of the I/O's were already going to be exposed - disregard then).
    Location for a WX module socket (and wired up to the the FLiP such that it could be programmed wirelessly)
    Socket to fit the WS2812 modules Parallax sells #28085
    Sockets for a few other common I/O, like the PS/2 Mouse/Keyboard modules...although not sure if these are sold anymore, so maybe not. Some kind of buttons or joystick, like maybe the simple 5-position switch #27801
    Socket(s) for whatever other I/O devices seem to be the best sellers
  • jmgjmg Posts: 12,087
    Ken Gracey wrote: »
    No manufacturing done at Parallax - PCB + all solderable parts provided in a bag
    No power supply or SMT parts

    Hmm, I think all the 3 colour digital LEDs are surface mount, seems a shame to exclude them ?
    Perhaps a 'no fine pitch SMT' is a better rule ?

  • Take the QuickStart pcb and replace the meat of it with a pair of SIP sockets for the Flip, and add a few prototyping holes between the SIP sockets (for sigma-delta experiments). Leave on the non-interfering LEDs and buttons, adapted to kit form. Use the same dual role header on the side as on the QuickStart, but thru-hole for the kit, and keep the pinout of the leds, buttons and headers, so that it can leverage the QuickStart code and tutorials.

    QuickStart is (was) 2" x 3". Extend to 3" x 4" with room for a white plugboard, but have more prototyping holes in the area where the white plugboard would sit. The plugboard can be an add-on. There should be room for a 2.1mm power plug and a pair of holes at 0.1" for a two-pin polarized header as an alt power input. There could be dedicated servo or i2c or spi headers with positions for pullup resistors where required.

    Bom: pcb, 2*sip header, 1*2row header, 74HC541 dip, 8 leds. sip resistor packs, 2.1mm socket, long header strip and optional pullup resistors.
  • frank freedmanfrank freedman Posts: 1,411
    edited April 14 Vote Up0Vote Down
    Personally, when I work with the prop, I use a basic four component row breadboard, more than enough space and max flexibility. Sometimes cheat with a recently acquired PPDB. My wish list would be a project board with a supply I could power from a wall-wart (5,9,12,18V DC, AC may cost to much given Ken's sell range) plug that would provide 1.8, 3.3, 5V for logic, and +/- 3V or greater analog supply for real world to digital interfacing. I hate to suggest using a particular layout because that may work for me and not others. Maybe a part area like the dev board with the flip socket on the edge and the prop pins and power supply pins along the same/an edge.

    @Ken, noticed the req list said no power supply parts, but that would be on my list for doing anything beyond straight logic. Yes, some op amp / analog can be done single supply, but sometimes +/- supply is just easier. Still want only one external source going to the project board. Easier, lower cost, and just less stuff in the box.
    Ordnung ist das halbe Leben
    I gave up on that half long ago.........
  • Take the QuickStart pcb and replace the meat of it with a pair of SIP sockets for the Flip, ...
    Perhaps most importantly of all, with the expanded board size, there would be room to silkscreen the dual-row receptacle pinout. :)

    The three-pin servo-style headers are important -- the more the better -- with voltage selection shunts for every two positions.

    I don't know where Parallax buys their connectors, but I've always found them to be the tail that wags the dog price-wise.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • msrobotsmsrobots Posts: 2,045
    edited April 15 Vote Up0Vote Down
    +1 for the WX socket, wireless programming is important.

    I haven't paid much attention to later project Boards, I still use the old 5-pack serial version most of the time, I had some project in mind witch never realized, and bought some couple of those 5-packs.

    What I really like is the processor in the center to build around it with very short wires. Besides not having build in USB. What I do not like is that ground strip placed somehow always in the way, not needed at all.

    For me the PE-KIT is still the best way to start with a propeller, It shows that you just need a couple of parts to get a P1 running.

    Since it should be for prototyping and one-offs, things like VGA sockets or PS2 sockets seem to be to restrictive to me, I have way more propeller then VGA monitors.

    I would say 5V and 3.3V regulators, power jack, 2 power and 1 ground rail along one side, WX socket, FLIP socket, 4-6 3pin connectors with positions for jumpers to connect power and ground to the pins or not, to be able to use the header for different purposes.

    I never liked the SMD LEDs, too small. Regular thru hole parts would be better for soldering disabled people like me.

    Please do not waste any space for the quick-start buttons. IMHO they are just terrible and useless. You need to run a driver in a COG to use Buttons? Simply insane. Just throw some pushbuttons into the bag.

    my two cents,

    Mike

    I am just another Code Monkey.
    A determined coder can write COBOL programs in any language. -- Author unknown.
    Press any key to continue, any other key to quit

    The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this post are to be interpreted as described in RFC 2119.
  • Tracy AllenTracy Allen Posts: 6,276
    edited April 15 Vote Up0Vote Down
    Agreed; labels all over, a good silk screen. There was once aftermarket labeling strip, very effective--Is one of those shipping with QuickStart now? I don't know. I used a Brother labelwriter tape.
    IMG_3953.JPG
    ...connectors, but I've always found them to be the tail that wags the dog price-wise.
    True enough, but we want good quality, so that the springs don't wear out too soon even with the range of wire sizes people are going to use. The worst thing for experimenting is when those contacts are flaky.
    576 x 432 - 112K
  • msrobots wrote:
    Please do not waste any space for the quick-start buttons.
    I'd have to agree. Their utility and reliability were pretty sketchy.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • msrobots wrote:
    Please do not waste any space for the quick-start buttons.
    I'd have to agree. Their utility and reliability were pretty sketchy.

    -Phil

    Yeah, yeah, these are definitely OUT.

  • K2K2 Posts: 607
    edited April 16 Vote Up0Vote Down
    Ken Gracey wrote: »
    PCB + all solderable parts provided in a bag
    Less that $15-20 cost to customer - maybe even $10...
    @Ken, as far as I'm concerned, you've already made the single most important design decision. The idea of an inexpensive assemble-it-yourself Propeller board is nothing short of inspired. I'd rather use the Propeller than any other uC in existence. The only reason I dabble in other chips and boards is price.

    What I'd like to see on such a board:
    * The few pads it would take to implement one or two Sigma-Delta ADCs.
    * The few pads it would take to add resistors for a video DAC for composite video output.
    * Pads for a first-order (cap and resistor) low-pass filter for PWM signal generation.
  • K2 wrote: »
    Ken Gracey wrote: »
    PCB + all solderable parts provided in a bag
    Less that $15-20 cost to customer - maybe even $10...
    @Ken, as far as I'm concerned, you've already made the single most important design decision. The idea of an inexpensive assemble-it-yourself Propeller board is nothing short of inspired. I'd rather use the Propeller than any other uC in existence. The only reason I dabble in other chips and boards is price.

    What I'd like to see on such a board:
    * The few pads it would take to implement one or two Sigma-Delta ADCs.
    * The few pads it would take to add resistors for a video DAC for composite video output.
    * Pads for a first-order (cap and resistor) low-pass filter for PWM signal generation.

    By the way, if I could get bags @ 5/$50, I'd forego everything but a crystal, an EEPROM, a 10K resistor for SDA, the Propeller, a four-pin terminal strip for programming, and the board itself. In other words, a solder-it-yourself breakout board.

    at least throw in some caps...

    a PE style break out board, another nice idea.

    But @Ken wants to build a board to plug in the FLIP module, a high volume ordered, quite affordable complete running P1 without need for much soldering at all.

    Mike
    I am just another Code Monkey.
    A determined coder can write COBOL programs in any language. -- Author unknown.
    Press any key to continue, any other key to quit

    The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this post are to be interpreted as described in RFC 2119.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 22,027
    edited April 15 Vote Up0Vote Down
    A couple points:

    1. I'm not optimistic that a sigma-delta ADC will work with the FLiP. The reason is the long lead lengths from the pins to the chip. Moreover, Ken said no SMD parts. Through-hole parts only add further inductance to an already lengthy feedback path.

    2. Ken's budgetary parameters are quite restrictive. Even a $20 MSRP dictates no more than a $6.50 BOM for parts -- preferably $5 or less -- and that's the high end of his price range. He's hoping for an MSRP around half that. Don't forget that somebody has to order and kit the parts, and kitting is not a trivial exercise. The BOM will likely be dominated by the PCB, since 3.05" x 4" isn't exactly small, and there won't be any advantages to be had by panelization. As a consequence, I think we're looking at a pretty minimal carrier board design.

    So what all can be crammed onto that board and still bring the total cost below $6.50 -- well below if possible? Remember, everything needs to be multiple-sourced, so an out-of-stock situation won't create an inventory issue. Aside from the PCB itself, we're talking strictly jelly-bean components here. That's the challenge.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • jmgjmg Posts: 12,087
    K2 wrote: »
    By the way, if I could get bags @ 5/$50, I'd forego everything but a crystal, an EEPROM, a 10K resistor for SDA, the Propeller, a four-pin terminal strip for programming, and the board itself. In other words, a solder-it-yourself breakout board.
    Those are all already mounted on FLiP, and this is a board for FLiP.
    1. I'm not optimistic that a sigma-delta ADC will work with the FLiP. The reason is the long lead lengths from the pins to the chip. Moreover, Ken said no SMD parts. Through-hole parts only add further inductance to an already lengthy feedback path.
    It's worth checking, as some form of ADC is a natural next step.

    My guess is it would be 'good enough', but if performance is too poor, you could look at including a cheap MCU for ADC.
    So what all can be crammed onto that board and still bring the total cost below $6.50 -- well below if possible? Remember, everything needs to be multiple-sourced, so an out-of-stock situation won't create an inventory situation. Aside from the PCB itself, we're talking strictly jelly-bean components here. That's the challenge.

    Your suggested 4 Digit LED fits inside that easily at ~ 40c

    Footprints cost nothing to include, just the time to think about which ones to allow for, the point being not all parts allowed for, need to be always included.

    eg a VGA HD15 is a modest footprint, in the compact versions, but not something you would drop into every bag - just give the part code.
    Likewise a Character LCD footprint, reasonably standard, but a Char LCD is also not something you would drop into every bag.

    I'd include SMD 3-4 Colour LED footprints too, as those demo well.

    Other footprints :
    Photo-interrupters are useful low cost sensors
    https://lcsc.com/product-detail/Photo-Interrupter_ITR9606-F_C80816.html large, and sub 20c
    or smaller
    https://lcsc.com/product-detail/Photo-Interrupter_ITR9907_C135482.html
    or common
    https://lcsc.com/product-detail/Photo-Interrupter_ROHM_RPI-352_RPI-352_C79953.html


    Tact Switch ? - always popular, these look to be sub 2c
    https://lcsc.com/product-detail/Tactile-Switches_6x6x14-Plastic-head_C10890.html 1.6c/600


  • frank freedmanfrank freedman Posts: 1,411
    edited April 15 Vote Up0Vote Down
    I would still argue for regulated power with at least 12V input. Even if you don't put in a negative rail, 3.3 and 5 will always be usefull. And lots of 0.1 centered pads. As to center or edge placement for flip, that ranks right up there with which components you will include; wanted by some, in the way of others or ignored entirely. If you are going after learners, add the goodies, multichannel ADC, servo jumpers, led etc, but looking at some other posts, someone with a plan will neither appreciate nor want "stuff in the way". Just need power and pads to wire wrap, point to point or dead bug it. Have not ordered a flip yet, but if it weren't for a PPDB, a GG board, three propdev boards (gave away a quick start a few years ago) I would absolutely get one or more. But only if it came without pins soldered in or had the round machined pins in place The big square pins most of these things use always feel like they are tearing up my breadboard when inserted. So I chuck the supplied pins and put in pins ordered for that purpose.

    Thinking about it, if it did not deprive one of seeing just how little is required to make a prop work/useful, they could upgrade the Prop Ed kit with a flip instead of a bare prop. May even save a newcomer from frying their first and only prop chip.
    Ordnung ist das halbe Leben
    I gave up on that half long ago.........
  • K2K2 Posts: 607
    edited April 18 Vote Up0Vote Down
    msrobots wrote: »
    But @Ken wants to build a board to plug in the FLIP module...
    You're right, of course. Wishful thinking led to momentary delusion.
    ...a high volume ordered, quite affordable...P1.
    If you say so. ;) I put 100 FLiP modules in my cart and the total is $2770. I put 100 STM32F103 "blue pill" boards in my cart and the total is $218. It really doesn't matter, though. Anyone can afford a Propeller for those uses that demand a Propeller.
  • I would still argue for regulated power with at least 12V input.
    Agreed. I seem to buy these things every time I pop into All Electronics.

    -- https://www.allelectronics.com/item/mb-ps/breadboard-power-supply/1.html

    There are a lot of interesting breadboard-friendly modules to play with, and these ensures plenty of power for the kinds of things I do.

    Jon McPhalen
    Hollywood, CA
    It's Jon or JonnyMac -- please do not call me Jonny.
  • Tracy AllenTracy Allen Posts: 6,276
    edited April 15 Vote Up0Vote Down
    About the power supply, do recall that the FLIP has a rather sophisticated power supply scheme. There is a switcher that can provide up to 1.8A at the 3.3V output from either USB or the external power. Also it can supply 1.5A current limited at the USB_5V output, but only when the FLIP is powered through USB by an external supply/brick. (The usual 0.4A limit applies on computer USB).

    The secondary supply input is limited to 9V maximum input though, and when the FLIP stands alone without external power, the USB_5V supply is not available. That is too bad, and I agree that an on-board 5V regulator that could run from 12V or more and could supply the regulated 5V power to both the FLIP and to the projects in standalone mode would be very useful, essential.
  • Tracy AllenTracy Allen Posts: 6,276
    edited April 15 Vote Up0Vote Down
    About the sigma-delta. Only SMT pads. The QuickStart has it on p8/p9. The path on FLIP from those pads to the edge of the pcb is not too bad. It would be simply a demo for people who want to take a stab at it.

    IMHO, pads or holes for other demos and devices (special ADC, video etc.) dedicated to certain pins can be included, but strictly as a user option. All pins should be available, with the usual exceptions for p28 to p31 as on the FLIP.

    About the QuickStart touch pads. I agree they are flaky and a bad option for either a beginner or a pro. However some kind of switch input is needed for instant gratification with the LEDs. The kit could have locations for through-hole SPST switches to ground, with pull-up resistors, but to keep with the philosophy of not dedicating any pins, the switches/pull-ups would have to be isolated with diodes. Probably dirt cheap 1N4148s, since diode arrays tend to be pricey. The kit would not need to include all 8 switches.




  • About the sigma-delta. Only SMT pads. The QuickStart has it on p8/p9. The path on FLIP from those pads to the edge of the pcb is not too bad. It would be simply a demo for people who want to take a stab at it.
    I think I'd rather see through-hole pads for a two-channel, 8-pin ADC chip. I2C would be optimal.

    -Phil
    “Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away. -Antoine de Saint-Exupery
  • Absolutely has to have 12 to 16 volt input jack with filtered & regulated 12, 5 and 3.3 voltages, in what ever configuration is inexpensive and reliable, ie. short circuit proof. FLiP already has a great 3.3 power supply. Standard PCB layouts (outline and holes) for Servo style headers, VGA connector with resistors, Mouse & keyboard, ADC & DAC layouts for common IC chips, with voltage divider resistors. In short, if I have the layout and holes, I can order the discrete parts on-line pretty easily. Likewise, any kind of specialty bread-board-friendly breakout boards I can usually find at All Electronics, Adafruit, and other on-line suppliers. Whatever the final design it must & should highlight the capabilities of the Propeller chip and FLiP module.
  • Unless Ken wants to position the board as a educational tool, building a board for flip seems a non-starter given all the flavors that wold be needed to make most people happy. The greatest value of the flip stems from the fact of drop in and build the project. The fact that I can just drop a fully operational prop into a dip position anywhere on the (bread)board and concentrate on the project at hand provides huge value in time saved. Maybe a good kit would be the Flip, a breadboard, and some of the same components as the Prop Ed kit. Maybe a CD or (think show promo gimme priced) USB stick copy of the Prop Ed manual, Prop manual, App notes, and prop tools with Ed examples propellant etc......
    Ordnung ist das halbe Leben
    I gave up on that half long ago.........
  • jmgjmg Posts: 12,087
    About the sigma-delta. Only SMT pads. The QuickStart has it on p8/p9. The path on FLIP from those pads to the edge of the pcb is not too bad. It would be simply a demo for people who want to take a stab at it.
    I think I'd rather see through-hole pads for a two-channel, 8-pin ADC chip. I2C would be optimal.
    That could be allowed for, but a quick search at Digikey reveals the trends of DIP becoming scarce, and all parts moving to SMD.
    Outcome is cheapest DIP-ADC is $1.30+/3k, which is way above the cheapest MCU. (ouch)

    The Prop SDM needs only manage 8~10 bit, to compete with the lowest end ADCs. Possibly a 74AHCU04 could boost the ADC specs a tad, with Vcc buffering, but those are 17c/3k and tagged non-stock in DIP or (~10c in SO14).

    Digikey lists SMD MCUs with ADC's from 14c/3k, cheapest with +HW i2c is STM8, in SO8 29c/3k, Cheapest in DIP is 45c, or with DIP+HW i2c, 63c/3k (climbing, but still well under DIP-ADC)
    - you'd probably want a MCU with a bootloader.

    DIP MCUs from Asia are more common, but even there SMD is new designs. The new SO8 STC8H04A10 claims 0.7 yuan (~11c), or STC8H1K08S2A10 SO16 ~17c and those have a USB loader..

    the 150mil SO packages are 1.27mm, which I think is not fine pitch smd, so can be soldered. A footprint with long pad fingers makes hand soldering even more forgiving.
  • jmgjmg Posts: 12,087
    About the QuickStart touch pads. I agree they are flaky and a bad option for either a beginner or a pro. However some kind of switch input is needed for instant gratification with the LEDs. The kit could have locations for through-hole SPST switches to ground, with pull-up resistors, but to keep with the philosophy of not dedicating any pins, the switches/pull-ups would have to be isolated with diodes. Probably dirt cheap 1N4148s, since diode arrays tend to be pricey. The kit would not need to include all 8 switches.
    The link I gave above has Tact buttons for ~1.6c/each in modest volumes. The clock kits have 2 or 3 buttons, so agree 8 is not needed.
Sign In or Register to comment.