Shop Learn P1 Docs P2 Docs Events
Halloween Hex - Page 2 — Parallax Forums

Halloween Hex

2

Comments

  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-09-25 15:45
    I know it takes more space, but perhaps consider:

    1. Creating your board so it handles servos in groups of 12, instead of 10.
    2. Provide separate Vdd taps for every six servos, using jumpers to minimize wiring if all 12 will use the same Vdd.
    3. I think it's okay for up to two sets of 8- or 10-pin ribbon cable to carry signal plus ground, and leave the power to a wholly separate input that, as an option, provides enough room for hardy screw terminals. If it were me, I'd put the power on either short side, and the signal cables on one long side.
    4. Any extra pins on the cable sets can be used for LEDs or sensors or something else you can decide later on. E.g., if using 10-pin, that gives you three lines for something like Ping signal connections, or whatever. Less wiring this way.
    5. Mounting holes are pretty important, as on these things the servo wiring get tricky and can pull and yank as the legs move (even with liberal tie-wrapping). So definitely make room for them next time around. Ideally I'd do four, and allow folks to mount with only two on opposing corners if they want. This is the Parallax method, and it seems to work well.

    Thanks for the ideas. I'll probably use your ideas and make another board but with mounting holes (which I agree are a good idea) the board will end up larger than I'd want to make using OSH Park. I'll use one of the other less expensive board houses and have a set of 10 made.

    It would be nice if there were an easy way to add servos to a Propeller (or other microcontroller) board.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2014-09-25 16:12
    Parallax had a Prop-based servo controller, but I think it's discontinued now. It had some of the same ideas I presented, though it only provided onboard for 16 servos. That sounds like a lot, but not when you and Bill are building 3DOF hexbots. I think it was too "in the middle" to be of much interest to builders.

    BTW, I had assumed you'd run ground for the servos through the IDC, and tie on this board, rather than tie somewhere else. I'm not sure if that's good or bad, though it means you don't need a separate junction box for all your battery connections. It also means you don't have a convenient place for power switches. (I tend to just disconnect batteries, rather than include switches. That's really cheap, I know!)

    One board has it mostly right, but it's not Prop-based. It's Lynxmotion's SSC-32 board. It's huge, but it's designed for hex's. They separate the power into two banks. It also has firmware onboard to do all the usual hex walking gaits. But I know you, Duane -- you'll want to code all that yourself!
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-02 14:52
    Back when I thought I had five QuickStart boards coming from RadioShack, I thought I'd just be using my expansion boards shown in post #29. However RadioShack send me an email telling my order for 5 QuickStart boards was cancelled since they didn't have any more in stock.

    I've started designing a board based on many of the suggestion Gordon made in post #30.

    Here's a screen capture of the DipTrace layout.

    attachment.php?attachmentid=111257&d=1412284223

    There are five holes for each servo channel. Two of the holes are used to connect the Prop's I/O pin to the signal position of the servo header. This connection can be made with either a resistor or a jumper. This is pretty much the way Bill has his RoboPi board laid out. I just added the ability to use smt resistors.

    It's possible to break the servo headers into five different groups. There are four groups of six and a single group of four.

    All the groups have the option of being powered from a separate source, from the group on either side of them or from the boards, 5V regulator. The last two groups also have the option of using 3.3V on the power pin. The final group also has the option of using Vin as the power source.

    I'm planning on using these connectors for the main power.

    2948.JPG

    The plug portion uses screw connectors but then the socket portion is soldered to the PCB. I'm not sure about the current rating of these connectors but they seem relatively heavy duty and they're really convenient to use.

    I just remembered I wanted to add an I2C connector. I/O pins P0 - P27 are brought out to the servo connectors. P30 & P31 are on the PropPlug connector but I still need to bring the I2C bus P28 and P29 out to a header so I can use then with other I2C devices.

    In case it's not obvious, there's a socket for a XBee in the lower center of the board. Only the 3.3V and ground connections are connected so far. I'll probably leave the XBee unconnected on the PCB and any connections made will need to be done with wire.

    There are six mounting holes. The four around the far corners line up with the inner set of holes on my small hexapod's frame. The two holes toward the center of the board, allow the board to be used with the large hexapod frame.

    I wanted to post what I had so far in hopes of receiving additional feedback.

    I also plan to add a reset button.

    As I mentioned, the mounting holes are spaced to match up with the small hexapod's inner mounting holes.

    Here's one of Paul's photos showing his board mounted using these inner holes.

    attachment.php?attachmentid=90477&d=1331405935

    The outer set of holes can be used to mount a Propeller Proto Board sized board. I think the robot would look better with smaller PCB like the one shown on Paul's hexapod above. IMO, the larger board shown below doesn't look as good.

    attachment.php?attachmentid=98092&d=1356027992

    I think I like having the servo connected from below as I did above and I'll probably try to do this with my next board.

    As usual, I'm open to suggestions.
    820 x 835 - 120K
  • Paul K.Paul K. Posts: 150
    edited 2014-10-02 22:02
    Email me your address. I have a spare SSC-32 board which I can send out.

    ***Sorry Duane I meant my board. Have no clue where the SSC-32 came from.
  • Paul K.Paul K. Posts: 150
    edited 2014-10-06 21:14
    I've built a couple boards for controlling my bots based on the Prop. These are some of the things I found to be useful. I made these changes in my revisions as I moved from board to board.

    1. Power your servos from a Ubec. I use this one on full sized servos: http://hobbyking.com/hobbyking/store/__35022__TURNIGY_8_15A_UBEC_for_Lipoly_USA_Warehouse_.html . The micro hex can use the 7.5 version. The Ubec is a simple solution and eliminates components that you might need to put on the board. You can select 5V or 6V while you power the bot from a 2 cell Lipo -- Power in to Ubec which splits, one leg to the servos one to your board reg. I use them in all my hex's.

    2. If you use a Dpak package you can hide the reg's both of them under the Xbee. I did this in the Q4/Q2's main controller boards. Space saver.

    3. Bank your servos in 3's with 3 banks on each side. Its easier to plug in servo headers with some room. Plus 2 for a pan tilt of some type for a head. Which might have a sonar sensor or camera.

    4. Make one connection for AX-12 servos. If you move on to them, since they run on a network you only need one connection. The board will still work and you can use the extra pins for sensors.

    5. Add a tantalum capacitor on the 3.3V and 5V line. Will reduce noise if you add A/D in the future. It might not be necessary and can always be added by soldering on the bottom of board later. Not needed now but could be a issue fixer later.

    6. Place extra headers for 5V & 3V access. Mainly for future device hook-ups.

    7. Try to place the xbee towards the edge of the board. It easier to hook up antenna that way.

    Just things I wish I made right on boards in the past to same time and money.

    Here's a pick of my new servo board/rx I'm working on.

    Just a prototype.

    IMG_2201.jpg
    1024 x 768 - 153K
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-06 23:57
    Paul K. wrote: »
    1. Power your servos from a Ubec. I use this one on full sized servos: http://hobbyking.com/hobbyking/store/__35022__TURNIGY_8_15A_UBEC_for_Lipoly_USA_Warehouse_.html . The micro hex can use the 7.5 version. The Ubec is a simple solution and eliminates components that you might need to put on the board. You can select 5V or 6V while you power the bot from a 2 cell Lipo -- Power in to Ubec which splits, one leg to the servos one to your board reg. I use them in all my hex's.

    I recently purchased three of those UBECs. I'm using one with the mini hexapod. I had thought I'd use all three on the large hexapod until Gordon told me the servos I'm using can be powered straight from a 2S LiPo.
    Paul K. wrote: »
    2. If you use a Dpak package you can hide the reg's both of them under the Xbee. I did this in the Q4/Q2's main controller boards. Space saver.

    You're not concerned about heat?
    Paul K. wrote: »
    3. Bank your servos in 3's with 3 banks on each side. Its easier to plug in servo headers with some room. Plus 2 for a pan tilt of some type for a head. Which might have a sonar sensor or camera.

    I have banks of 6 right now. I wondered about adding some space between some of the headers but by the time I thought of it, I wasn't in the mode to nudge a bunch of traces over a fraction of an inch.

    I've been trying to keep the headers on a 0.1" grid but I think adding some space between groups of servos is probably more important than keeping the headers on a 0.1" grid.
    Paul K. wrote: »

    4. Make one connection for AX-12 servos. If you move on to them, since they run on a network you only need one connection. The board will still work and you can use the extra pins for sensors.

    After you mentioned this in a PM, I added an extra 2.5mm pitched header to the board. I think it would be fun to make a hexapod with AX-12 servos. I'm not sure if I have enough right now, but if I removed all the AX-12 servos I have from other projects, I might have a total close to 18.

    I just realized I forgot to add one other connector. I want to add a connector for Spektrum satellite receivers. Apparently it's possible to communicate with Spektrum RC gear using a serial connection. I've done this on the transmitter side, but I haven't done this with the receiver. I want to add a connector to make it easy to add one (or more) receivers.
    Paul K. wrote: »
    5. Add a tantalum capacitor on the 3.3V and 5V line. Will reduce noise if you add A/D in the future. It might not be necessary and can always be added by soldering on the bottom of board later. Not needed now but could be a issue fixer later.

    I have pads for tantalim caps near the voltage regulator, but I don't have provision for extra caps near the servo connectors. As you say, I can still add these to the solder side of the servo connections.
    Paul K. wrote: »
    6. Place extra headers for 5V & 3V access. Mainly for future device hook-ups.

    This is another thing I wondered about. All of the servo banks have a jumper to allow them to be connected to the regulated 5V line and many also have the option of being powered by 3.3V but I don't have extra power only headers. I'll probably want to do this on a future revision.
    Paul K. wrote: »
    7. Try to place the xbee towards the edge of the board. It easier to hook up antenna that way.

    This is one reason why I turned the XBee around from the previous version of the board. If I had planned better, I could have gotten the XBee closer to the edge. Again, something to think about for another revision.

    Thanks for the many suggestion. Let me know if there's something with this latest version you see which could be improved.

    Here's the front of the board.

    attachment.php?attachmentid=111314&d=1412662043

    And the back.

    attachment.php?attachmentid=111315&d=1412662044

    I only connected the rx and tx lines on the XBee to Propeller I/O pins. Three other pins have solder jumpers which could be used to connect the pins if I decide it would be useful to do so.

    I also included pads for a Nordic nRF24L01+ module so these could be used instead of the XBee if one didn't mind using the extra I/O pins.

    I found some "long distance" versions of the Nordic transceivers and I wanted the board to be able to use these if I chose. This meant adding two extra ground pins (not on a 0.1" grid).

    This is a relatively simple board but it took me a ridiculous amount of time to get it laid out. I have laid out Propeller based PCBs before, but this is the first one I'll likely have made.

    I thought I was ready to submit this board but I want to add the connection for the Spectrum receivers so I have a bit more work still.

    I tried to keep power traces wide. The servo power traces are 40mil and the 5V and 3.3V traces are 20mil. Most of the data traces are 14mil which is likely wider than needed but I've read it's a good idea to to use wide traces if there's room on the board.

    Both the top and bottom of the board have copper pours connected to ground.

    The board could be run from up to three separate power sources. Two of the power sources have large screw terminal connectors. Any single power connection could be routed to the rest of the board by using the appropriate jumpers.

    Again, I welcome suggestions.
    588 x 759 - 129K
    581 x 754 - 114K
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-07 11:29
    Duane Degn wrote: »
    I thought I was ready to submit this board but I want to add the connection for the Spectrum receivers so I have a bit more work still.

    I changed my mind about adding a satellite receiver connector. I decided my sanity couldn't tolerate any more changes in this version of the board.

    I was impatient enough to submit the board to OSH Park. I usually don't use OSH Park on boards this large. Three boards cost me $34.65.

    The boards I mentioned in post #29 arrived from OSH Park a few days ago. I'm presently soldering headers on to the boards. Hopefully these little boards will allow me to take the large hex on a test walk.
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-08 11:39
    I populated a couple of the servo extension boards.

    Here's a photo showing the two populated boards and the remaining bare PCB.

    attachment.php?attachmentid=111330&d=1412792872

    Here's a photo showing the board connected to the power supply and the QuickStart board. The servos have obviously not been connected yet.

    attachment.php?attachmentid=111329&d=1412792870

    The ribbon cable connects the ten signal pins to the QuickStart I/O pins. Each board has a ground wire which connects with the QuickStart. I'm using four of the power connection positions to connect the board with the battery. There are two positive and two negative wires.

    I've linked the two power buses on the boards together with some copper wire.

    I have a bit more work before I'm ready to connect the servos and power up the hexapod but I'm hoping to get the large hexapod walking sometime today.
    780 x 560 - 160K
    490 x 537 - 105K
  • Bill HenningBill Henning Posts: 6,445
    edited 2014-10-09 09:14
    Sheesh I get busy for a while, and a lot happens on the forum.

    Nice work Duane! I'll be keeping an eye on your hexapod!
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-09 09:23
    They might also be different from the ones sold by Fingertech, in which they have specs going up to 9.0 volts. You can see if they match the pictures.

    My servos don't like the higher voltage.

    I have a setup program which only activates one servo and a time and oscillates the servo to identify it. This lets me know which servo is connected to which pin so I can plug the servos in without worrying about the order. IMO, it's easy to change the pin constants than to try to keep all those wires straight (as in identified).

    The programmed oscillation motion is pretty slow, just enough to let me know it's moving. When I had my bench supply set to 7.8V some of the servos would oscillate very quickly and were obviously not following the pulses to set their position. The oscillation returned to normal once I turned the voltage down.

    It looks like I'm going to need those voltage regulators after all.
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-09 09:32
    Nice work Duane! I'll be keeping an eye on your hexapod!

    Thanks Bill.

    The PCB took me longer to layout than I had thought it would. I had used the RoboPi for a while, but I'd have to rig up some other way to power two of the servos (if not using 5V) so I'm saving the RoboPi to use with a RPi on a different project.

    I need to get back to work on the software. My initial attempts to get the hexapod to rotate didn't work correctly.
  • Bill HenningBill Henning Posts: 6,445
    edited 2014-10-09 12:45
    You are welcome :)

    What I ended up doing for HexPi is using two short servo extension cables, cutting the 5V from RoboPi, and inserting power for two more servos that way.

    "Real Work (tm)" has been keeping me from HexPi, P1V, and a lot of other fun projects :-(
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-20 14:47
    The PCBs arrived a few minutes ago!

    attachment.php?attachmentid=111551&d=1413841514

    I think I have the parts to populate at least one of the boards.

    I get a flu shot at my wife's work on Thursday. She is hoping the hexapod is working well enough by then to show her coworkers. I think it will be ready by then.
    793 x 703 - 304K
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2014-10-20 18:25
    Duane Degn wrote: »
    My servos don't like the higher voltage.

    Hmmm. Wonder what's different between yours and the ones Fingertech sells? While it's not unheard of for the Chinese factories to change a product and call it the same thing (sometimes by using a whole different factory), I wonder if there's something else going on here.

    Out of curiosity, what is the refresh rate of your servos? And have you tried a different frame period? I'd be interested in what happens if you go slightly longer between refreshes, like from 50 Hz to maybe 40 or even 30. But of course, I wouldn't want you to damage your servos.

    The boards look excellent, BTW. Nicely done, as usual.
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-26 21:02
    Hmmm. Wonder what's different between yours and the ones Fingertech sells? While it's not unheard of for the Chinese factories to change a product and call it the same thing (sometimes by using a whole different factory), I wonder if there's something else going on here.

    Out of curiosity, what is the refresh rate of your servos? And have you tried a different frame period? I'd be interested in what happens if you go slightly longer between refreshes, like from 50 Hz to maybe 40 or even 30. But of course, I wouldn't want you to damage your servos.

    I'm using the standard servo object by Beau so they're being refreshed at 50Hz.

    I haven't tried other rates yet but I'll give it a try.

    I'm not sure if I can modify Beau's object to a different rate but I'm sure I could hack together some code to drive a couple of servos at arbitrary rates.

    Not all the servos behaved poorly at 8V but several did. I didn't pin down the exact voltage they started to act up on.

    One thing which could have contributed to the problem was my bench top supply. I wasn't powering the servo off a battery but I used my bench supply. Since I was only enabling a single servo at a time to identify which servo was connected to which I/O pin, I don't think the current draw was a problem but I've had several motorized projects not work well when powered from my bench top supply so I'll probably try the servos again using a battery.
    The boards look excellent, BTW. Nicely done, as usual.

    Thanks but unfortunately you're wrong with the "Nicely done" comment. These boards are Prop killers. I've killed four PLL circuits with two of these boards.

    I even used copper wire to connect all four grounds across the top of the chip before powering it up and the PLL still blew.

    While I've had a bunch of different custom PCBs made, this is my first custom Propeller board and I obviously got it wrong. Before I attempt to redesign this board, I'll submit a small Prop PCB to make sure I can properly route the power and ground traces.

    For now, I plan to use a QuickStart board to control the large hexapod.

    I keep trying to tell myself "it wouldn't be as rewarding if it were easy".

    I'll give more detail of what I think I did wrong when I have more time (probably after Halloween). I appreciate your feedback and I do plan to redesign this board in the very near future.
  • ercoerco Posts: 20,216
    edited 2014-10-26 21:53
    Duane Degn wrote: »
    I keep trying to tell myself "it wouldn't be as rewarding if it were easy".

    If it was easy, anyone could do it.

    If it was perfect, you couldn't tell it was homemade.


    Keep going, Duane!
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-10-30 20:58
    erco wrote: »
    Keep going, Duane!

    Sir! Yes, sir!

    I do need to admit to procrastinating a bit, but I've been back to work on the Halloween Hex lately.

    I cut of piece of the fantastic robot building material expanded PVC (which I learned about from the guy who wrote Robot Builder's Bonanza), to use as a "lid" of sorts. The QuickStart board and voltage regulators are mounted on the bottom of this platform and a pair of pan tilt gizmos are mounted to the top.

    attachment.php?attachmentid=111704&d=1414727617

    attachment.php?attachmentid=111703&d=1414727614

    I plan to add some sort of eyes to the pan and tilt brackets.

    I just wanted to let you all know I'm actively working on this. Hopefully it will be walking around with a bowl of candy on this platform tomorrow evening.
    752 x 637 - 144K
    843 x 630 - 206K
  • ercoerco Posts: 20,216
    edited 2014-10-30 21:33
    I have to build a new servo control board for my pumpkin flamethrowers tonite. I think my old one got built into a project. Here's to staying up late tonite to scare the kiddies tomorrow!
  • ercoerco Posts: 20,216
    edited 2014-11-01 02:40
    Awaiting your Halloween Hex report, good sir!
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-11-04 14:40
    erco wrote: »
    Awaiting your Halloween Hex report, good sir!

    It's getting close to being ready. How many more days do I have?
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-11-04 15:13
    As many may have guessed, I didn't get this large hexapod up and running in time for Halloween this year. I did have my smaller hexapod running and let kids control it with the wireless Wii Nunchuck I'm currently using as a controller.

    I finally got around to opening up one of the TowerPro MG996R servos I purchased off ebay.
    Duane Degn wrote: »
    I did order 4 996R servos from ebay today.

    The TowerPro MG996R servos were less expensive than the HobbyKing HXT12K servos so I shouldn't have been too surprised to find the TowerPro MG996R servos didn't have bearings inside. The final gear uses a brass bushing on top and a plastic bushing on the bottom. The HXT12K have two sets of ball bearings on the final gear.

    The TowerPro MG996R are a little slower than HXT12K servos but I didn't notice a perceptible difference in torque as I tried to hold the servo horn stationary (not a very good torque test).

    The TowerPro MG996R servos were significantly louder than HXT12K servos.

    I ordered a couple more of the HXT12K servos from HobbyKing and I was surprised to see how different the new servos were compared with the older version. The old version looks like all of the gears are brass while the new version have several silver colored gears. Are these silver colored gears aluminum?

    The PCB and pot on the new HXT12K servos are different from the older version of the HXT12K. One of the stranger differences to the new H is its case. The new case has the word "TowerPro" molded into the plastic. I think this is strange since the TowerPro servo didn't have this feature.

    I took some photos of the insides of the servos. I'll try to make some time to upload some of these pictures.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2014-11-05 16:16
    Duane Degn wrote: »
    I finally got around to opening up one of the TowerPro MG996R servos I purchased off ebay.

    The TowerPro MG996R servos were less expensive than the HobbyKing HXT12K servos so I shouldn't have been too surprised to find the TowerPro MG996R servos didn't have bearings inside. The final gear uses a brass bushing on top and a plastic bushing on the bottom. The HXT12K have two sets of ball bearings on the final gear.

    Got you, too, eh? Bill bought these same servos, and had a number of DOAs and other issues, which he posted here and on his Web site. I'll repeat the link to a good (honest) review of the new crop of these:

    http://www.youtube.com/watch?v=0d4qif7m6WQ

    From the video: "Quite literally, this is THE worst servo ever made."
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-11-05 18:55
    Got you, too, eh? Bill bought these same servos, and had a number of DOAs and other issues, which he posted here and on his Web site. I'll repeat the link to a good (honest) review of the new crop of these:

    http://www.youtube.com/watch?v=0d4qif7m6WQ

    From the video: "Quite literally, this is THE worst servo ever made."

    Gordon, I don't have any MG995 servos to compare against my MG996R servos but I was under the impression the MG996R servos were not the same as the MG955 servos.

    One reason I purchased some of the MG996R servos was because of a "mini review" I saw on YouTube. The mini review about the MG996R servos was on the same YouTube channel as the MG995 review you linked to.

    I just found the T Pro MG996R mini review video again and I see the servos in the review were labelled "T Pro 996R". My servos were labelled "TowerPro 996R". I which the guy on UltimateRCnetwork had opened one of the servos so I could see it he had the same version I have.
  • ercoerco Posts: 20,216
    edited 2014-11-05 19:37
    Say what you like, Gordon. The MG995 may suck, but it sucks big time and consistently across all reviews! :)


    http://www.rcmodelreviews.com/mg995review.shtml "world's worst servo"

    http://www.amazon.com/ss/customer-reviews/B004OQU1UU "its a piece of junk don't buy this ever"
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2014-11-05 22:45
    Duane Degn wrote: »
    The mini review about the MG996R servos was on the same YouTube channel as the MG995 review you linked to.

    You mentioned you bought Tower Pro and the guy's review is for T-Pro. These Chinese factories play fast and loose with brand names. Unless this guy (who seems knowledgeable) on YouTube says T-Pro and Tower Pro are the same brand, I would not assume that. I would have thought these were in fact different servos, and not just because they are one digit apart. Tower Pro servos that you buy today are not the same ones that you could get years back. Whole new kettle of corn.

    Or, you might get lucky and the "R" means Revised ... for the better, let's hope.
  • ercoerco Posts: 20,216
    edited 2014-11-06 09:22
    Q: What do you call a master carton of MG995s on the ocean floor?
    A. A good start!

    Q: What's the definition of a tragedy?
    A. A box of MG995's falling over a cliff. The box wasn't full.

    Take any lawyer joke and insert MG995.
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2014-11-06 10:56
    Well, in any case, it was the Tower Pro MG995 that Bill had so much trouble with, and not the "Tower Pro MG996R."

    There are videos on YouTube with both the Tower Pro and T-Pro brands, and while they might be the exact same servo, they might not. On the Tower Pro Web site they only seem to show Tower Pro and TorqPro as their brands. T-Pro may be a copy from another factory, or a custom label for certain sellers, or something else.
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-11-06 12:35
    There are videos on YouTube with both the Tower Pro and T-Pro brands, and while they might be the exact same servo, they might not.

    To add to the confusion, there's also "Towardpro" servos. This video opens up a Towardpro MG996R servo. While the gears of the "Towardpro" servo look like the gears inside the "Tower Pro" servos I have. The electronics don't look the same.

    Here's a section from a screen of from the Towardpro MG966R video.

    attachment.php?attachmentid=111806&d=1415305058

    Here's the photo of the PCB on my TowerPro MG996R.

    attachment.php?attachmentid=111810&d=1415305065

    Obviously the two PCBs shown above are different. In case anyone is interested, I've attached a closeup of the two chips below (at the bottom of this post with the thumbnails).

    Here's an overall view of the TowerPro MG966R parts.

    attachment.php?attachmentid=111807&d=1415305060

    Here's a closeup of the bushings.

    attachment.php?attachmentid=111808&d=1415305061

    Another problem with the TowerPro MG996R servos is the way the shafts supporting the gears terminate directly into the plastic case. Nicer servos use little brass inserts to protect the plastic and to keep the shafts properly positioned. You can see these inserts in the HXT12K photos below.

    As if things concerning the name "Tower Pro" weren't confusing enough, the new HXT12K servos I purchased from HobbyKing have the words "Tower Pro" molded into the servo's case.

    attachment.php?attachmentid=111813&d=1415305071

    The newer HXT12K servos had several "silver" colored gears. One of these gears had five spokes.

    attachment.php?attachmentid=111812&d=1415305069

    The new servos also use square pots rather than round pots which are found in the older HXT12K servos.

    Here's the insides of an older HXT12K servo.

    attachment.php?attachmentid=111811&d=1415305067

    Both the 2011 and 2014 HXT12K servos use two sets of ball bearing on the final gear. It looks like the newer HXT12K servos are still pretty good quality, IMO the older version work better than the newer version since the new version tends to get hotter than the older version.
    331 x 474 - 385K
    786 x 530 - 138K
    515 x 446 - 79K
    402 x 588 - 108K
    859 x 513 - 144K
    725 x 577 - 159K
    772 x 565 - 170K
    606 x 766 - 178K
  • GordonMcCombGordonMcComb Posts: 3,366
    edited 2014-11-06 13:53
    Just by quick look, the Hobby King 12K's appear to have higher quality. The inserts for the axles are metal, rather than just plastic, like in the Tower Pro (or TowerdPro) model. The 996R is also obviously interior due to the lack of ball bearings, using just bushings. It really doesn't make much sense to have metal gears and simple metal and/or plastic bushings.

    I wouldn't worry about the "d" in the the TowerdPro name. It stands for Duane, to show these were specially made for you! Seriously, for the Chinese and Taiwanese, they probably wouldn't noticed the typo caused by fat fimgets.

    The spoked gear, while not ideal, is probably okay if it's the one right off the motor. Fairly little torque exerted on that gear. In all servos it's the most lightweight.
  • Duane DegnDuane Degn Posts: 10,574
    edited 2014-11-09 18:32
    It's alive!

    With 356 days to go, I have the hex up and walking.

    The video is uploading now but to give you an idea of the awesomeness of this robot, here are a few pictures.

    attachment.php?attachmentid=111914&d=1415586293

    attachment.php?attachmentid=111915&d=1415586382

    One of the many parts of the program I still need to fix is the eye display code.

    YouTube says the video just has another two minutes so come back soon for more hexapod goodness (but keep your expectations kind of low (it was a first test after all)).
    715 x 590 - 215K
    766 x 591 - 257K
Sign In or Register to comment.