Page 6 of 6 FirstFirst ... 23456
Results 51 to 59 of 59

Thread: Malum IK

  1. #51
    Join Date
    Jun 2013
    Location
    Tucson, AZ
    Posts
    248
    Rep Power
    25

    Re: Malum IK

    Click image for larger version. 

Name:	malum-hopper-v3.jpg 
Views:	31 
Size:	53.6 KB 
ID:	7535

    When designing my hopper I went through many many attempts/designs till I got it to feed properly without jamming. I would set them up with a tube that would self feed back into the hopper and run thousands of BBs through the loader till it would jam. I would then clear all the loose BBs away (I would leave it open as your showing above) and see where/how it jammed. Then I would redesign and repeat.

    Here's my critique of your loader:
    1) You've covered up over half the top surface minimizing the amount of time the BBs have to drop & settle into place
    2) The "open" slope I assume you have so the BBs can slide into the wheel. Its potentially doing the opposite, as the wheel turns it could be throwing the BBs out of the slots.
    3) You have nice flat hard surfaces for the BBs to jam against if the are not fully in the wheel slot (still falling into place). What I do is put a soft brush (pipe cleaner) along these types of edges to "softly" brush the not fully seated BBs away.

    While developing my loader, I would also put a little "shake" movement (quickly reverse direction a couple of times) while loading to free any jamming BBs. With my latest design its not needed and it only slows down the rate of fire.
    I'll try to post some pics this evening of my load set up.

  2. #52

    Re: Malum IK

    Thanks for the feedback! Truly helpful.

    Some more context: This is the result of several iterations of doing the same kinds of tests! It's good to know I'm following an established process. There are some features that avoid jams here that may not be obvious. Where I find jams in testing, I create surfaces that are angled, such that the wheel will push the BBs out rather than block. This is intended to work on area 3. (There's a triangular wedge thingie inside that edge, not visible from this perspective.) The main thought here is to have "one spot" where the BB/feeder interface happens, and work on the edges around that spot to make BBs either feed out, or feed in, but not jam. Having a larger open area from above (as my previous design did) led to more pinch points between feeder and mounting brackets, causing more jams. I imagine if I lift up the motor bracket far enough, that will be less of a danger, though?

    The steep trough into the wheel (area 2) is because, with a shallower trough, the BBs wouldn't "channel" as nicely; I'd actually get a larger potential area of jams. The draw-back is that, if the channel is too narrow, I end up with jamming/bunching on top of the channel. Having the open arc be big enough for four BBs has so far seemed to help. I may still go back and flatten this out, depending on how this goes.

    Area 1 is opportunistic storage space -- the previous iterations had a flat wall above the agitator wheel entrance. The storage may not be worth it -- it's very likely to bunch up in the areas that constrict. (This is the main problem with my existing design -- areas that narrow to 7-10 millimeters going around the micro metal motor, leading to bunching/jamming.) I'm currently at the point where I've fed, let's say, fifty BBs through the bottom construct without jamming, and I'm working on how to extend it up without creating new jams that didn't exist in the previous test. The edge of the sloping floor of the "second story" would be one of those possible jam surfaces.

    I really like the idea of soft "brush" walls. The West Point gentleman also used this in his design (he claims he got it from you, so props for spreading the knowledge!)

    Finally, prototyping in aluminum isn't really slower than prototyping in 3D printing. A mill can cut away metal at a much faster rate than a 3D printer can add plastic :-D The main drawback is that I sometimes am limited to the shape I can construct based on a plain 3D mill, no additional axes, and my ability to hold it.
    Last edited by jwatte; 05-29-2019 at 02:53 PM.

  3. #53

    Re: Malum IK

    In other news, I m in love with thread milling.
    No special tapping heads. No compression holders. No need to synchronize the spindle.
    Just friggin' cut the thread!
    The thread mills don't get stuck like regular taps, either. If I were to break one (hasn't happened yet,) it would just fall right out of the hole.

    There are a nice series of single-tooth cutters from Micro 100, available on Amazon, and other tool suppliers, in a variety of sizes. They are imprial, but they have ample additional spacing,and they are single-tooth, so the metric threads I cut are simple with those tools. I have a 0.100 for cuttin M3, a .180 for cutting M6, and a .250 for cutting M10, but they can really cut a variety. (There are also .060 and .080 for the M2/M2.5 needs; I don't have them yet.)

    Separately, the nice Sandvik Coromant M2.5 cutter I use is a three-tooth cutter -- it can only cut 0.45 pitch. There appears to be two sellers on Amazon; one that charges list price at $230, and another that charges presumably cost-plus-10% at $38. Needless to say, I prefer to buy when the second one has stock :-)
    (Or these are Chinese impostors that aren't actually made in Sweden -- seems to work fine for me, though, so I'll take the savings! Is impostor tool brand names a thing?)

    Also: Threaded hole brackets beats those darn nuts in the servo flanges by a mile. That part worked out great!

  4. #54

    Re: Malum IK

    SO I haven't been very active on the gun thread, because I realized something else was broken, too.
    The new video transmitter, with "SmartAudio(tm)" 4800 bps half duplex interface, needs a serial port to talk to it.
    The MinimOSD I'd like to use on top of video to send telemetry back also needs a serial port.
    But I'm already using the three ports on the Teensy 3.2 on the current board (that I re-purposed from another project from last year.)
    So I need a bigger MCU.
    So I need a new carrier board.
    So I might as well integrate the bits that were floating around as separate little Chinese boards in the current 'bot: separate video regulator, motor driver for the agitator, and so forth.
    Might as well add a video switch for a possible two-camera system on one video channel.
    After picking parts and making schematics, and finally an intense week-end of placing and routing, this is what my loyal robotic minions (a k a MacroFab) will provide me:

    Click image for larger version. 

Name:	malum-board-v2.jpg 
Views:	22 
Size:	93.5 KB 
ID:	7562

    I'll drive it with a Teensy 3.6.

    I also added some other goodies, like dedicated circuitry to condition the TCRT-1100 IR sensor I used to use for gun cycle detection; a PWM driver output for an UV LED should I manage to put tracer BBs in there; and a switchable power outlet for some distance sensing apparatus TBD so I can tell the OSD video overlay to tell me when I'm about to back into a building.

    I also switched out the standard "P channel MOSFET" power switch I usually use for an automotive load switch from ST, mainly to get some experience, but also because it has a built-in current consumption sensor.

    We'll see how well it works in two weeks when it shows up on my doorstep. Maybe I built a solid piece of robot heart. Or maybe I built a 5V 1.2 MHz oscillator; we'll see!

    Also, the video switch was an interesting challenge. All "analog switch" parts I could find have 0.3V protection diodes from the input channel to the ground and VCC rails. The problem with that is that a video signal is 1.2V P-P centered on ground, and thus would get clipped by those kinds of switches.
    I ended up finding an application note by Analog Devices about how to use their high-bandwidth opamps with output disable as a video switch, and that looked alright so I cribbed that design, and their chip comes with three opamps in it; I only needed two for the switch, so the third is what now conditions the IR sensor feedback.
    Of course, THAT in turn needed an 1.8V reference to re-center the video signal between the rails, AND an inverter to enable one input when disabling the other, so the part cound kind-of grew a bit there ... (Why not use two pins from the seemingly inexhaustible Teensy 3.6 pin cornucopia? Because I already allocated functions to all of the pins, that's why! Can't add a current sensor and then not wire it up, and so on and so forth ...)

    Click image for larger version. 

Name:	video-switch-circuit.jpg 
Views:	21 
Size:	32.6 KB 
ID:	7563

    But, with all that done, I'm back to thinking about gun magazines, and how great it would be if I would just take Dhon's 3D model and cut it out on the mill and it then worked! (There are some nontrivial areas for a 3 axis mill on that part; I can't perfectly replicate it, so we'll see how that goes ...)
    Last edited by jwatte; 4 Weeks Ago at 10:41 PM.

  5. #55

    Re: Malum IK


  6. #56
    Join Date
    Sep 2010
    Location
    ಠ_ಠ
    Posts
    2,280
    Images
    27
    Rep Power
    274

    Re: Malum IK

    How has your experience with MacroFab been? Spent an hour or so assembling some PCBs* for small scale production (Seeedstudio PCBs and stencils plus low-temp solder paste), but hand producing these things does not scale well without building an actual stencil rig and a pick-and-place machine. If I finally had my BuildOne, I would be able to bash out both of those devices quite easily (could even add solder paste dispenser to the pick-and-place).


    *4-channel +/-5V Op-Amp feeding 4-channel 16-bit I2C ADC with level-translation that mounts to a RaspberryPi. Other board is barrel jack, P-FET power switch, fused >20W DC-DC converter for +5V, isolated 5V-to-5V DC-DC for -5V, and bunch of connectors and capacitors.
    Please pardon the pedantry... and the profanity... and the convoluted speech pattern...
    "You have failed me, Brain!"
    bleh

  7. #57

    Re: Malum IK

    I will be interested as well on how well the MacroFab works, especially for small runs... Like probably < 5...

    I think my first T4 breakout board, that I had Seeedstudio fabricate arrived yesterday at our mailbox in town, So hopefully will try to assemble one soon, using Beta T4 (probably older one without the updated SD pins...).. But sometimes it would be great to have someone else do the soldering...

    @tician - as for BuildOne, maybe we will see them this year, at least they were able to print a boaty... I have my other 3d printer, which is currently out of commission as the USB connector broke off the driver board... Will soon try to solder something on it to make it work again...

  8. #58
    Join Date
    Jun 2013
    Location
    Tucson, AZ
    Posts
    248
    Rep Power
    25

    Re: Malum IK

    Both artans and I from RTEAM have used Macrofab. I'd say our experience with them has been fine. You pay a premium (as opposed to some place like OshPark) but you don't have to go through the hassle of building the boards yourself.

    I mainly used them right after they launched. Their parts library was a bit small at the time, so it was a bit of a pain finding alternate parts. I found it much easier if you did stick to their parts library whenever possible as they generally had those parts on stock. I would hope their parts library/part stock has grown since I last used them a couple years ago.

  9. #59

    Re: Malum IK

    I've used Seeed before, and I've used MacroFab a few times (up to quantity 10.) My experience with Seeed was "acceptable" and my experience with MacroFab was "great!" MacroFab does charge American wages for the labor, and a small mark-up for sourcing the parts for you (and, I guess, receiving and setting them up ...) They do solder paste with a "jetter" which I gather is like an inkjet printer for solder paste :-)
    The trick with MacroFab is to use their "house parts" wherever possible, because you don't pay labor for that, including not paying labor for through-hole parts. They charge about $1.60 per through-hole solder point at the prototype volume, so a 2x3 100mil pin header that's a house part is $0.29 installed and ready; a 6-pin JST PH connector is $10.
    I have not used their volume services, but I know the OpenMV people do, and are happy with it (they had some scary BGA and image sensor alignment things that required careful collaboration for their production.)

    Also: OSH Park is great for boards, but a board does not a fully completed part make :-) And even when I use OSH Park, I can't successfully reflow double-sided boards in a toaster oven. (I've done single-sided before.) MacroFab pricing includes 4-layer boards (i e, no discount for 2-layer) and double-sided mounts.

    Their parts library has not really grown much since opening. In fact, there's nothing between 18 kOhm and 100 kOhm in their resistors ... But they integrate with DigitKey and Mouser and Arrow, so just pick a standard vendor part. The main reason to use their part library is not paying placement fees; turnkey/ordered parts add that cost. (Presumably they have their pick-and-place machines loaded with their house parts, and anything custom has to be manually arranged, either into the machine, or manually placed, depending on volume.)

    Also, their standard parts are 0402, 0603 and 0805. They will do down to 01005 without upcharge, but again, you then pay the "non-house part" placement cost.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •