Bare Bones Propeller

HumanoidoHumanoido Posts: 5,770
edited July 2013 in Propeller 1 Vote Up0Vote Down
Bare Bones Propeller

This updated Bare Bones Schematic is currently at version 1.4
and is subject to change without notice. It replaces versions 1.0,
1.1, 1.2, and 1.3. Use the blog's search feature to find any future
changes.

The suggestions from this thread, and some from another thread
are now incorporated into the circuit to maximize Propeller chip
performance, reliability, and facilitate ease of assembly for
beginners.

The circuit allows lower power operations, mainly 3-volts with
two (1.5 volts each) AA batteries, and experiments at lower
voltages. Bare Bones is a good reference circuit and starting
point for many projects.

By itself, Bare Bones is considered to be a simple machine
project with provided software on the schematic and the ability
to communicate with a serial terminal, providing virtual
representations.

The program runs in rcfast (~12 mHz) and at 1200 baud
to blink a serial terminal LED on and off.
«1

Comments

  • 35 Comments sorted by Date Added Votes
  • Mike GreenMike Green Posts: 22,601
    edited July 2013 Vote Up0Vote Down
    I strongly suggest having a 0.1uF ceramic capacitor connected between each pair of Vdd/Vss pins on either side of the package. The fact that your setup works doesn't mean that it will work reliably. The Propeller is made up of circuitry that switches very rapidly and even a few inches of wire between the package power supply pins and the power source can introduce enough inductance plus the battery's internal resistance so the supply voltage can drop very briefly very significantly. The 0.1uF capacitors mounted as close as possible to the package pins act as a high speed power supply buffer for these transient power demands.
  • PublisonPublison Posts: 9,942
    edited July 2013 Vote Up0Vote Down
    You are missing the bypass caps, as has been discussed before in your previous post. It's not a good idea to tell a noobie that this is all it takes to run a Prop. They will be back to ask why the PLL failed.

    EDIT: Mike types faster than me. :)
    Infernal Machine
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    Mike Green wrote: »
    I strongly suggest having a 0.1uF ceramic capacitor connected between each pair of Vdd/Vss pins on either side of the package. The fact that your setup works doesn't mean that it will work reliably. The Propeller is made up of circuitry that switches very rapidly and even a few inches of wire between the package power supply pins and the power source can introduce enough inductance plus the battery's internal resistance so the supply voltage can drop very briefly very significantly. The 0.1uF capacitors mounted as close as possible to the package pins act as a high speed power supply buffer for these transient power demands.

    Thanks Mike for this important reminder. The circuit will use a power supply with leads that end with .1uf decoupling capacitors. These will connect immediately near and closest to the Propeller chip pins, ground and +3V, on both sides of the DIP as the breadboard will allow.
  • mindrobotsmindrobots Posts: 6,489
    edited July 2013 Vote Up0Vote Down
    Humanoido wrote: »
    Thanks Mike for this important reminder. The circuit will use a power supply with leads that end with .1uf decoupling capacitors. These will connect immediately near and closest to the Propeller chip pins, ground and +3V, on both sides of the DIP as the breadboard will allow.

    You are making that assumption and NOT stating it in your design. Therefore, the circuit design is not following best practices and may not be reliable. Just add the two capacitors and the world is a better place and PLL's live to run another day! :smile:
    MOV OUTA, PEACE <div>Rick </div><div>"I've stopped using programming languages with Garbage Collection, they keep deleting my source code!!"</div>
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    Publison wrote: »
    You are missing the bypass caps, as has been discussed before in your previous post. It's not a good idea to tell a noobie that this is all it takes to run a Prop. They will be back to ask why the PLL failed.
    EDIT: Mike types faster than me. :)

    The bypass caps are not missing because two are included across power supply leads, where it connects to the circuit (at VSS/VDD on both sides of the prop). I consider the filtering they do to be part of the power supply to provide a clean signal and source.

    It looks like you type faster than me. :smile:
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    mindrobots wrote: »
    You are making that assumption and NOT stating it in your design. Therefore, the circuit design is not following best practices and may not be reliable. Just add the two capacitors and the world is a better place and PLL's live to run another day! :smile:

    You're right! I made the assumption the power supply design would have the required filtering and forgot to mention it. The next drawing version will have a power supply note to avoid any confusion.
  • tonyp12tonyp12 Posts: 1,873
    edited July 2013 Vote Up0Vote Down
    How are they getting the program in to the Prop?, Oh a $15 PropPlug
    Just go with my soon to be released ~$10 Prop_Traverser
    PropBreadBoard2s.jpg?psid=1
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 21,772
    edited July 2013 Vote Up0Vote Down
    Humanoido wrote:
    The bypass caps are not missing because two are included across power supply leads, where it connects to the circuit (at VSS/VDD on both sides of the prop). I consider the filtering they do to be part of the power supply to provide a clean signal and source.
    As was already pointed out here and elsewhere, you need to include the caps in your schematic! No one lookiing at your circuit will make any assumptions about components that are not explicitly shown. In any event, bypass caps are not the same as filter caps and are not part of the power supply: they're part of the working circuit used to reduce noise caused by the circuit itself, not noise from the power supply.

    -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
  • Mike GreenMike Green Posts: 22,601
    edited July 2013 Vote Up0Vote Down
    They are missing. They're not shown in the schematic you posted nor are they described in the associated text. You may have them in your setup, but you're communicating something significantly different. Although you consider them as part of the power supply, they're really part of the environment of the chip itself. With the switching speed of the Propeller, a few inches of wire make a huge difference.
  • Dave HeinDave Hein Posts: 5,370
    edited July 2013 Vote Up0Vote Down
    To be fair to Humanoido, he is just using the circuit that Parallax published in the Prop Manual, Prop Datasheet and even the manual for the Propeller Education Kit. These manuals don't show any bypass caps in their circuit diagrams, and if you search the manuals for the word "bypass" you won't find any occurrences of it. Of course, it is a good idea to use bypass caps if you want to avoid problems, but Parallax doesn't state this explicitly in their manuals.
  • Dave HeinDave Hein Posts: 5,370
    edited July 2013 Vote Up0Vote Down
    @tonyp12, the Prop_Traverser looks very interesting. When will it be available. I wonder if it would be possible to make a board that would essentially be a 40-pin socket with all of the support circuitry sitting under the Prop between the two rows of connectors. Has anybody done a board like that?
  • Mike GreenMike Green Posts: 22,601
    edited July 2013 Vote Up0Vote Down
    "to be fair" - Yep. I'm not blaming Humanoid for posting something that's now known to be poor design practice. I'm just insisting (with justification given) that it be posted correctly now. If Parallax documentation is also showing the same mistake, it should be corrected as well. The Propeller was felt to be a bit more robust in the past. The "PLL multiplexor" failure mode demonstrated the importance of connecting both sets of power supply pins using short PCB traces and bypass capacitors, that the absence of these measures had more consequences than just more noise.
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    Mike Green wrote: »
    The Propeller is pretty robust and doesn't fail easily. The most common damage to the Propeller is the blowing of the PLL multiplexor which disables the use of the PLL. Essentially the Propeller identifies and downloads programs properly, but the only clocks that work are RCFAST and RCSLOW. This failure commonly occurs with large voltage spikes on the power or ground supplies or problems with ground loops on the board.

    They are missing. They're not shown in the schematic you posted nor are they described in the associated text. You may have them in your setup, but you're communicating something significantly different. Although you consider them as part of the power supply, they're really part of the environment of the chip itself. With the switching speed of the Propeller, a few inches of wire make a huge difference.

    "to be fair" - Yep. I'm not blaming Humanoid for posting something that's now known to be poor design practice. I'm just insisting (with justification given) that it be posted correctly now. If Parallax documentation is also showing the same mistake, it should be corrected as well. The Propeller was felt to be a bit more robust in the past. The "PLL multiplexor" failure mode demonstrated the importance of connecting both sets of power supply pins using short PCB traces and bypass capacitors, that the absence of these measures had more consequences than just more noise.

    If you knew this all along, you had every opportunity to let Parallax know this information (what you call poor design practice) so their material could be promptly updated. Perhaps you did inform Parallax and they did not want to change the schematic. One could say they were too busy, but this raises the question as to why they had time to issue two new updated manuals without the schematic change.

    Not everyone has time to read posts on the Forum. Many people get information from the source - the updated Parallax PEK kit manual and the updated Parallax Propeller Manual, for proper propeller wiring information. It has no information about decoupling capacitors.

    With the Bare Bones Prop, there are decoupling capacitors at the end of the power supply leads, by choice. The power supply is not on the schematic. An updated text about this config is added to the schematic page which I just uploaded to post #1. For clarity, the caps sit on both sides of the Propeller chip, on the breadboard pins closest to the prop's VSS & VDD lines, even though they are physically attached to the end feed lines of the power supply (and also the prop through the closest breadboard holes).

    A few inches of wiring? Absolutely not. The distance from the power supply end connector with decoupling caps to the Props VDD & VSS pins is very close to 1/16th inch. This is the best we can do with a breadboard plug-in design even though there are other effectual considerations inherent in the breadboard pad.

    It's news to me that the Propeller is now less robust as you say. Have you talked to Ken Gracey about this? Someone ran a hundred propeller chips for two years and they remained robust with no PLL failures whatsoever. "Large voltage spikes on the power or ground supplies or problems with ground loops" sound more like board problems.

    Maybe some effects are less or absent when running RCSLOW or RCFAST at a lower frequency without a clock, as in this circuit. As I mentioned in another thread, I have props running years, without decoupling capacitors, that have never failed. I don't think my props are handpicked over others.
  • jmgjmg Posts: 11,065
    edited July 2013 Vote Up0Vote Down
    Dave Hein wrote: »
    @tonyp12, the Prop_Traverser looks very interesting. When will it be available. I wonder if it would be possible to make a board that would essentially be a 40-pin socket with all of the support circuitry sitting under the Prop between the two rows of connectors. Has anybody done a board like that?

    This over/under is an interesting question.
    We have done DIP boards in the past, and placed quite a few parts in the DIP socket cavity.

    The problem with this, becomes if you then want to stack it as plug-able, you ideally need fine male pins out the bottom.

    Machine pins are costly, and fragile.

    The picture in #8 nicely sidesteps that, by attaching on top of a DIP package, and you can plug the DIP-Prop into a std DIP socket, which can plug into any target (if you want an element of cheap damage protection).

    PCB suggestions:
    To make it stronger, I'd probably add connections down to P0, P1, and I'd bring the handshake lines to solder-pads.

    - and you might be able to fit pads for a Si504 option- they come down to 2 x 2.5mm, and use a C1D interface to set Freq using a 4 byte real.
  • jmgjmg Posts: 11,065
    edited July 2013 Vote Up0Vote Down
    Humanoido wrote: »
    As I mentioned in another thread, I have props running years, without decoupling capacitors, that have never failed. I don't think my props are handpicked over others.

    Do you really mean no caps whatsoever in the Vcc-Gnd, anywhere ? - or did you mean 'without local decoupling capacitors"
  • PublisonPublison Posts: 9,942
    edited July 2013 Vote Up0Vote Down
    Humanoido wrote: »

    Maybe some effects are less or absent when running RCSLOW or RCFAST at a lower frequency without a clock, as in this circuit. As I mentioned in another thread, I have props running years, without decoupling capacitors, that have never failed. I don't think my props are handpicked over others.

    Over the years, it has been found that lack of proper decoupling capacitors lead the failure of the PLL circuit.

    In your case, you may never see the failure because you are only relying on the RC clock circuit . The PLL might have been blow but you would never know it until you tried to run a standard 5Mhz crystal program.
    Infernal Machine
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 21,772
    edited July 2013 Vote Up0Vote Down
    Humanoido wrote:
    With the Bare Bones Prop, there are decoupling capacitors at the end of the power supply leads, by choice. The power supply is not on the schematic
    As Mike and I have both pointed out, the decoupling caps are NOT part of the power supply; they are an essential part of the Propeller circuit itself. So even though the power supply is not on the schematic, the decoupling caps need to be. This is the case with every schematic I've ever worked with when the power supply is external to the board. If a component is on the board, it needs to be on the schematic. Period.

    Making a simple circuit look simpler by omitting parts is not helping anyone.

    -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
  • Cluso99Cluso99 Posts: 13,370
    edited July 2013 Vote Up0Vote Down
    Just to enforce that "the decoupling caps are NOT part of the power supply; they are an essential part of the Propeller circuit itself"
    If you plug in your power supply circuit (with the pair of 0.1uF caps on the leads as I presume you mean) and your power supply is powered on - presume by a 3V battery), it is possible to likely that the voltage spike at the prop pins could damage the props PLL or worse.

    And, just a pair of 0.1uF is still IMHO not acceptable. You also need at least a 10uF cap there too, presuming you are not drawing considerable power from the supply. Overclocking shows up the problems of unsatisfactory power supplies because it exaggerates the problem.
    My Prop boards: P8XBlade2, RamBlade, CpuBlade, TriBlade
    Prop OS (also see Sphinx, PropDos, PropCmd, Spinix)
    Website: www.clusos.com
    Prop Tools (Index) , Emulators (Index) , ZiCog (Z80)
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 21,772
    edited July 2013 Vote Up0Vote Down
    Cluso99 wrote:
    You also need at least a 10uF cap there too.
    True enough. That would be the filter cap, as opposed to the bypass caps we've been discussing. It's particularly vital with long external power leads or noisy supplies. It's also load-dependent. If your board drives high current loads, even 1000uF would not be too much for such a local charge reservoir. The rule I follow is this: 10uF or more where external power enters the board. More as load conditions dictate.

    But one thing at a time. I didn't mention filter caps here -- yet -- since I did not want to cause cognitive overload. :)

    -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
  • Heater.Heater. Posts: 20,554
    edited July 2013 Vote Up0Vote Down
    I don't think I would fault Parallax for not showing decoupling and or filter caps in a minimal Prop schematic.
    One often sees minimal circuits like that that are intended to show a concept or a use case of a device, the clutter of "housekeeping" components being left out for clarity. It is often expected that a designer will be aware of these details and take care of them.
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    Publison wrote: »
    Over the years, it has been found that lack of proper decoupling capacitors lead the failure of the PLL circuit. In your case, you may never see the failure because you are only relying on the RC clock circuit . The PLL might have been blow but you would never know it until you tried to run a standard 5Mhz crystal program.

    Yes, exactly. You just proved the complete reliability of this circuit, even without decoupling capacitors.
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    jmg wrote: »
    Do you really mean no caps whatsoever in the Vcc-Gnd, anywhere ? - or did you mean 'without local decoupling capacitors"
    For this reference, the power supply is very stable and well regulated and has its own filtering capacitors. The reference Propeller circuit tests were without decoupling capacitors connected locally to the chip at VSS/VDD.
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    Heater. wrote: »
    I don't think I would fault Parallax for not showing decoupling and or filter caps in a minimal Prop schematic.
    One often sees minimal circuits like that that are intended to show a concept or a use case of a device, the clutter of "housekeeping" components being left out for clarity. It is often expected that a designer will be aware of these details and take care of them.

    I agree. There is absolutely no fault of Parallax. The bare bones circuit is intended to show the concept without the clutter. It's a hobby project and others wishing to work with it may do so and can take care of any details they wish.
  • Heater.Heater. Posts: 20,554
    edited July 2013 Vote Up0Vote Down
    Humanoido,
    Yes, exactly. You just proved the complete reliability of this circuit, even without decoupling capacitors.

    There is no such proof in there. Perhaps a Prop on a breadboard using it's RC clock works for years.

    Fine but what if it's PLL has been blown by a dodgy power setup?

    You would never know. Until one day you decide to add an XTAL or pull the Prop off the bread board for some other circuit.

    A circuit that damages it's components is not good. Especially among hobbyists who tend to recycle components a lot.
  • jmgjmg Posts: 11,065
    edited July 2013 Vote Up0Vote Down
    Humanoido wrote: »
    For this reference, the power supply is very stable and well regulated and has its own filtering capacitors. The reference Propeller circuit tests were without decoupling capacitors connected locally to the chip at VSS/VDD.

    Ah, so really, you are merely using the power supply capacitors, and the variable is then the lead length.
    Every decoupling circuit has some inductance between the die, and the capacitors.

    Another reason designers use local caps, is passing EMC - did you EMC test your 'no caps' design ?
    Checked it for false resets on radiated energy ?
  • HumanoidoHumanoido Posts: 5,770
    edited July 2013 Vote Up0Vote Down
    jmg wrote: »
    Ah, so really, you are merely using the power supply capacitors, and the variable is then the lead length.
    Every decoupling circuit has some inductance between the die, and the capacitors.
    Another reason designers use local caps, is passing EMC - did you EMC test your 'no caps' design ?
    Checked it for false resets on radiated energy ?

    Yes the lead length from the power supply routes to the first master loader chip and I just ran it and it worked fine so there was no issue. As I recall, with more and more props added, it became a little noisy at around 50-60 props and the circs were redesigned into three partitions as explained in the Big Brain thread. I scoped it out at one point and settled on the acid test. In using breadboards there's a way to create grounding islands and I experimented with this, and will likely do more work in this area to improve circuits with large numbers of props. In the BSS, I built shielding, but this didn't seem necessary with the props. As you know I'm very interested in the signals created by the chip and at one time developed a kind of radiative non-invasive "brain wave" monitor for chip study. A lot of things change inside the chip depending on the program running and I'm interested in sensing these changes for knowledge of the chip without looking at the code.
  • ctwardellctwardell Posts: 1,639
    edited July 2013 Vote Up0Vote Down
    So is this an ego issue now?

    Multiple people have pointed out an known issue and suggested the simple addition of two very inexpensive components to the basic circuit to help with the care and feeding of the PLL.

    Like others have said, it's a known issue, so regardless of the initial Parallax documentation it should be properly documented going forward.

    I don't understand the resistance in making the change.

    You obviously have a passion for the Propeller and other Parallax products and you like documenting your projects and teaching others, so why risk setting them up for possible failure by not using the known best practices?

    C.W.
  • Duane C. JohnsonDuane C. Johnson Posts: 955
    edited July 2013 Vote Up0Vote Down
    What is it, is the weather to hot?

    Boys, boys!!

    1. It's always bad practice to not use good engineering common sense, even if it works and you got away with it.

    2. Engineering spec sheets rarely show everything. Their purpose is to show the important circuit concepts.

    Get over it!!!

    Duane J
  • Dave HeinDave Hein Posts: 5,370
    edited July 2013 Vote Up0Vote Down
    One issue is that the PEK still does not specify using bypass caps as far as I know. Maybe it's been corrected, but I don't think so. Parallax is a great company, but one thing they should work on is updating their documents and software.

    Humanoido, add bypass caps to your schematic. It's the right thing to do. Claiming that they're hidden in the power supply doesn't make sense.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 21,772
    edited July 2013 Vote Up0Vote Down
    2. Engineering spec sheets rarely show everything. Their purpose is to show the important circuit concepts.

    'Possibly so when your audience is engineers, who can fill in the gaps from experience. When your audience includes newcomers to the art, a somewhat higher standard is called for. When achieving even a modest standard is blocked by intransigence, yes, people will complain.

    -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
Sign In or Register to comment.