Project: Melissa Hands - Perhaps a little easier..

Discussions regarding building a walking robot at home. Most of the robots participating at Robo-One competitions are custom fabricated.
135 postsPage 4 of 91, 2, 3, 4, 5, 6, 7 ... 9
135 postsPage 4 of 91, 2, 3, 4, 5, 6, 7 ... 9

Post by PaulL » Mon Jun 10, 2013 6:04 am

Post by PaulL
Mon Jun 10, 2013 6:04 am

New version, 2nd run. The first run was scrap, the carpet tape didn't hold properly. When I do this alu sheet, I use particle board as sacrificial board, then carpet tape, then the 6061 alu sheet.

With the small brackets and the separation problem, I decided to try a new method: mechanically securing each part with screws. You can see in the photo, all parts have at least 2 screws. These are M2 4mm screws, same as the machine screws for the RN-1. This worked out great. First off, the parts don't separate. Secondly, the quality of the cut is much better.

I'm almost kicking myself I didn't do this a long time ago. I've had this "separation" problem before, and I typically scrap and re-do. Using screws in the part before running the end mill almost guarantees the parts will come out right.

In the 2nd pic, the parts look terrible, but that's the result of chips and glue from the tape. I use lighter fluid to clean the glue off the parts, and after clean up, they look great. It's late, but I pulled one part, a small bracket (on right in the pics) to see how it looks cleaned up. I was impressed. In examining the edges of the other parts (where not covered in glue and chips), the edges are much nicer.

Image
Image
New version, 2nd run. The first run was scrap, the carpet tape didn't hold properly. When I do this alu sheet, I use particle board as sacrificial board, then carpet tape, then the 6061 alu sheet.

With the small brackets and the separation problem, I decided to try a new method: mechanically securing each part with screws. You can see in the photo, all parts have at least 2 screws. These are M2 4mm screws, same as the machine screws for the RN-1. This worked out great. First off, the parts don't separate. Secondly, the quality of the cut is much better.

I'm almost kicking myself I didn't do this a long time ago. I've had this "separation" problem before, and I typically scrap and re-do. Using screws in the part before running the end mill almost guarantees the parts will come out right.

In the 2nd pic, the parts look terrible, but that's the result of chips and glue from the tape. I use lighter fluid to clean the glue off the parts, and after clean up, they look great. It's late, but I pulled one part, a small bracket (on right in the pics) to see how it looks cleaned up. I was impressed. In examining the edges of the other parts (where not covered in glue and chips), the edges are much nicer.

Image
Image
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Mon Jun 17, 2013 11:41 am

Post by PaulL
Mon Jun 17, 2013 11:41 am

I tapped and assembled rev 4 with separate small angled brackets. I'm not happy with these small brackets. Too many screws, and the look isn't quite what I was hoping for. I'll end up going back to something similar to rev 3.

However, the dimensions are pretty much set, so I'm going to continue and attempt to build this into a working hand.

Next up is to cut some teflon hubs for the servos. I have some thin teflon sheet on order (arrives today), this should work better than trying to cut them from thicker stock.

From here, I need to make sure the hub size is right, then mount the servos and strips and see how things go.
I tapped and assembled rev 4 with separate small angled brackets. I'm not happy with these small brackets. Too many screws, and the look isn't quite what I was hoping for. I'll end up going back to something similar to rev 3.

However, the dimensions are pretty much set, so I'm going to continue and attempt to build this into a working hand.

Next up is to cut some teflon hubs for the servos. I have some thin teflon sheet on order (arrives today), this should work better than trying to cut them from thicker stock.

From here, I need to make sure the hub size is right, then mount the servos and strips and see how things go.
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Tue Jun 18, 2013 9:29 am

Post by PaulL
Tue Jun 18, 2013 9:29 am

A friend at work likes the latest one best. I'm still leaning towards rev 3.

Teflon sheet of .010" and .015" showed up today. It looks like .015 is going to be the best bet. Still need to cut servo hubs, cut some strips, and mount some servos and test.

Pics of latest:

Image

Image

Image

Image

Image

Image
A friend at work likes the latest one best. I'm still leaning towards rev 3.

Teflon sheet of .010" and .015" showed up today. It looks like .015 is going to be the best bet. Still need to cut servo hubs, cut some strips, and mount some servos and test.

Pics of latest:

Image

Image

Image

Image

Image

Image
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Sun Jun 23, 2013 12:22 am

Post by PaulL
Sun Jun 23, 2013 12:22 am

I made two servo hubs today. The first was an attempt to use teflon strip with a screw and separate hub. The second was a one-piece hub and strip, as seen in the pics below. No pics of the first attempt, they're junk parts. The first hub didn't work out - not enough material for the threads, and the thread fit was loose once the teflon hub was stretched over the spline on the servo.

Below pics with one piece hub and strip before mounting stuff:

Image

Image

Image

Image

And then I went ahead and wired up a servo, mounted it w/ tape, drilled and mounted the teflon strip to the finger tip, and actually moved the finger by servo. Below pics are various stages of motion - mid point, then extended, then retracted, all by servo, no manual manipulation here. The point here is, the design is sound, and things work as expected.

Image

Image

Image

The motion is good, but the servo is glitchy, there's a problem with the pot inside - I can get it to act right if I put tension on the wire. I need to open it up and see what the problem is. The servo wires are pretty stiff, and mess with the pot through the board and wires to the pot. In a few I repaired, I used really fine and flexible wire from the pot to the board. Tedious, but it fixes the problem.

Also, the extended finger position isn't very flat because I need to leave more length on the strip next time.

I'll also make the hub a bit smaller in diameter to give more room for excess material in the inside corner where the strip meets the hub.

The problem and reason I didn't try this design (one piece hub / strip) first is that the 1/8 inch (.125", about 3 mm) end mill can't cut into the inside corner where the strip meets the hub - this area has to be cut out by hand, and it's easy to cut through .020" (half a mm) of teflon and ruin the part. On top of that, I wasn't sure how the mill would handle cutting a .020" thick strip out. I thought the end mill would catch the strip and tear it up, but a small miracle happened - it came out pretty well first try.

I ran the servo using a Pololu Micro Maestro servo controller, the same one I'm going to mount to the hand when it's done. I tested movement from the PC, and the result was pretty lifelike with some acceleration added.

These should be pretty neat when done. :)
I made two servo hubs today. The first was an attempt to use teflon strip with a screw and separate hub. The second was a one-piece hub and strip, as seen in the pics below. No pics of the first attempt, they're junk parts. The first hub didn't work out - not enough material for the threads, and the thread fit was loose once the teflon hub was stretched over the spline on the servo.

Below pics with one piece hub and strip before mounting stuff:

Image

Image

Image

Image

And then I went ahead and wired up a servo, mounted it w/ tape, drilled and mounted the teflon strip to the finger tip, and actually moved the finger by servo. Below pics are various stages of motion - mid point, then extended, then retracted, all by servo, no manual manipulation here. The point here is, the design is sound, and things work as expected.

Image

Image

Image

The motion is good, but the servo is glitchy, there's a problem with the pot inside - I can get it to act right if I put tension on the wire. I need to open it up and see what the problem is. The servo wires are pretty stiff, and mess with the pot through the board and wires to the pot. In a few I repaired, I used really fine and flexible wire from the pot to the board. Tedious, but it fixes the problem.

Also, the extended finger position isn't very flat because I need to leave more length on the strip next time.

I'll also make the hub a bit smaller in diameter to give more room for excess material in the inside corner where the strip meets the hub.

The problem and reason I didn't try this design (one piece hub / strip) first is that the 1/8 inch (.125", about 3 mm) end mill can't cut into the inside corner where the strip meets the hub - this area has to be cut out by hand, and it's easy to cut through .020" (half a mm) of teflon and ruin the part. On top of that, I wasn't sure how the mill would handle cutting a .020" thick strip out. I thought the end mill would catch the strip and tear it up, but a small miracle happened - it came out pretty well first try.

I ran the servo using a Pololu Micro Maestro servo controller, the same one I'm going to mount to the hand when it's done. I tested movement from the PC, and the result was pretty lifelike with some acceleration added.

These should be pretty neat when done. :)
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by limor » Mon Jun 24, 2013 3:49 pm

Post by limor
Mon Jun 24, 2013 3:49 pm

looks beautiful!

what's the pcb on the back of the hand?
fingers seem to be clamped only one screw, how are they prevented from moving?
i don't quite get how the finger actuation works with that strip attached to the servo.
can you please post a video?
looks beautiful!

what's the pcb on the back of the hand?
fingers seem to be clamped only one screw, how are they prevented from moving?
i don't quite get how the finger actuation works with that strip attached to the servo.
can you please post a video?
limor offline
Savvy Roboteer
Savvy Roboteer
User avatar
Posts: 1845
Joined: Mon Oct 11, 2004 1:00 am
Location: London, UK

Post by PaulL » Mon Jun 24, 2013 5:43 pm

Post by PaulL
Mon Jun 24, 2013 5:43 pm

Thanks, Limor!! :)

The PCB is a Pololu Micro Maestro servo controller. Having the controller on the hand keeps me from having to run 6 x 3 conductor servo wires (5 fingers + wrist) back to the torso. With this board, I can use tx only, so +v, gnd, and tx, 3 wires from the hand / wrist up. The wrist rotation is something I worked more on Sunday, it's getting close to finished.

The fingers do have only one screw, but they have a moulded pin as well. If you look in the pictures above, there's one showing the underside of where the fingers mount - you can see the screw protruding through, as well as the pin moulded into the "mount" link of the finger. They won't come out of alignment, and won't come loose once some locktite is added.

I'll post a video soon. ;) The strip is on the inside of the axis of the joint (held there by bluish gray UHMW blocks in the fingers), so it pulls on that side (the inside), curling the finger. If extended (strip pushed out into the finger), the finger flattens out. The strip is mounted to the finger tip. Teflon causes the joints to bend simultaneously, whereas for example a wire cable would not.

It's like a "Melissa Hand" with its nylon wire ties, I'm just using a wider, thinner strip of Teflon instead, with separate servos for each. The thumb servo will be located against the wrist plate (90 degrees rotated clockwise in relation to the finger servo in the pics), with the servo spline towards the thumb. I may have to put some pins / guides in there to keep the thumb strip from folding / buckling when extending the thumb - the thumb is the only area where this is likely to be needed.
Thanks, Limor!! :)

The PCB is a Pololu Micro Maestro servo controller. Having the controller on the hand keeps me from having to run 6 x 3 conductor servo wires (5 fingers + wrist) back to the torso. With this board, I can use tx only, so +v, gnd, and tx, 3 wires from the hand / wrist up. The wrist rotation is something I worked more on Sunday, it's getting close to finished.

The fingers do have only one screw, but they have a moulded pin as well. If you look in the pictures above, there's one showing the underside of where the fingers mount - you can see the screw protruding through, as well as the pin moulded into the "mount" link of the finger. They won't come out of alignment, and won't come loose once some locktite is added.

I'll post a video soon. ;) The strip is on the inside of the axis of the joint (held there by bluish gray UHMW blocks in the fingers), so it pulls on that side (the inside), curling the finger. If extended (strip pushed out into the finger), the finger flattens out. The strip is mounted to the finger tip. Teflon causes the joints to bend simultaneously, whereas for example a wire cable would not.

It's like a "Melissa Hand" with its nylon wire ties, I'm just using a wider, thinner strip of Teflon instead, with separate servos for each. The thumb servo will be located against the wrist plate (90 degrees rotated clockwise in relation to the finger servo in the pics), with the servo spline towards the thumb. I may have to put some pins / guides in there to keep the thumb strip from folding / buckling when extending the thumb - the thumb is the only area where this is likely to be needed.
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Tue Jun 25, 2013 10:21 am

Post by PaulL
Tue Jun 25, 2013 10:21 am

Video.. ;)

phpBB [media]
Video.. ;)

phpBB [media]
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Tue Jul 02, 2013 8:07 am

Post by PaulL
Tue Jul 02, 2013 8:07 am

A bit of an update. I'm working on hip / wrist joints on my RN-1. I want the next set of hand brackets to be the final design, and if I need to make any adjustments on the wrist mount plate, I want to know before I cut another bracket set. So!

I've been working up the CAD drawings for the wrist and hip mounts. I need to spend a little more time on them, but they're close to done.

As for the wrist / hip joint itself, I've tested a few things. First was an off-the-shelf thrust bearing. It was a total of 4mm thick, and all steel. This made for a nice rotation joint, but it was heavy and added 4mm to the leg length without counting a 2mm thick pulley that will be used to actuate the joint. I contemplated using teflon on teflon for the thrust bearing, but some testing revealed "sticking" in the joint, not something I want. I machined a new joint hub on my lathe and bought some peel-and-stick washers made with teflon and tried those the other day, but these also stick.

My latest attempt will involve two .020" (half mm) "shim washers", .040" (1mm) bearing balls, and a .031" (just under 1mm) thick UHMW ball cage. I'm waiting on the parts I ordered yesterday. The one part to make is the ball cage, which will be like a washer with holes for the balls. This will result in a .080" (approx 2mm) thick thrust washer stack. The pulley will be 2mm thick 6061 alu, the thickness is to hide the screw heads used to mount the pulley to the associated bracket. I'm also using a 10mm OD bearing for axial load.

In case anyone is curious, I order these kinds of parts from http://www.mcmaster.com. I also order tooling (end mills, taps, drills, etc) from them.

The only unknown in the thrust bearing design is the surface finish of the shim washers. I'm hoping it's smooth enough for 10 or so 1mm balls, but I won't know that until they show up.

The final stack height of the rotating joint (between the brackets) will be .160" (about 4mm): .080" pulley, .020" washer, .040" balls, .020" washer. This means the arms and legs will get lengthened by this amount. The wrist will get a little more than this from the last stock servo to the hand as a result of adding a bracket to mount the wrist. I'm hoping to keep this down to .080", maybe a little less.

The wrist will be actuated by a servo slung off the back of the arm or off the side of the arm (still debating). It'll be a little bulky, but there aren't many choices here.

In all, I doubt the stock RN-1 servos would be enough to move this bot - he's getting heavy. From the Z530 PC board to the full size servos added for hips and wrists to the 10 micro servos in the hands and 3 in the neck, I think the HSR-5498SG's will be near the edge of what's needed. I'm curious as to what the all-up weight is going to be. Heck, the HSR-5498SG's add weight.

One thing's certain - this is going to be a really cool bot when done, a great platform to demonstrate my software. :)
A bit of an update. I'm working on hip / wrist joints on my RN-1. I want the next set of hand brackets to be the final design, and if I need to make any adjustments on the wrist mount plate, I want to know before I cut another bracket set. So!

I've been working up the CAD drawings for the wrist and hip mounts. I need to spend a little more time on them, but they're close to done.

As for the wrist / hip joint itself, I've tested a few things. First was an off-the-shelf thrust bearing. It was a total of 4mm thick, and all steel. This made for a nice rotation joint, but it was heavy and added 4mm to the leg length without counting a 2mm thick pulley that will be used to actuate the joint. I contemplated using teflon on teflon for the thrust bearing, but some testing revealed "sticking" in the joint, not something I want. I machined a new joint hub on my lathe and bought some peel-and-stick washers made with teflon and tried those the other day, but these also stick.

My latest attempt will involve two .020" (half mm) "shim washers", .040" (1mm) bearing balls, and a .031" (just under 1mm) thick UHMW ball cage. I'm waiting on the parts I ordered yesterday. The one part to make is the ball cage, which will be like a washer with holes for the balls. This will result in a .080" (approx 2mm) thick thrust washer stack. The pulley will be 2mm thick 6061 alu, the thickness is to hide the screw heads used to mount the pulley to the associated bracket. I'm also using a 10mm OD bearing for axial load.

In case anyone is curious, I order these kinds of parts from http://www.mcmaster.com. I also order tooling (end mills, taps, drills, etc) from them.

The only unknown in the thrust bearing design is the surface finish of the shim washers. I'm hoping it's smooth enough for 10 or so 1mm balls, but I won't know that until they show up.

The final stack height of the rotating joint (between the brackets) will be .160" (about 4mm): .080" pulley, .020" washer, .040" balls, .020" washer. This means the arms and legs will get lengthened by this amount. The wrist will get a little more than this from the last stock servo to the hand as a result of adding a bracket to mount the wrist. I'm hoping to keep this down to .080", maybe a little less.

The wrist will be actuated by a servo slung off the back of the arm or off the side of the arm (still debating). It'll be a little bulky, but there aren't many choices here.

In all, I doubt the stock RN-1 servos would be enough to move this bot - he's getting heavy. From the Z530 PC board to the full size servos added for hips and wrists to the 10 micro servos in the hands and 3 in the neck, I think the HSR-5498SG's will be near the edge of what's needed. I'm curious as to what the all-up weight is going to be. Heck, the HSR-5498SG's add weight.

One thing's certain - this is going to be a really cool bot when done, a great platform to demonstrate my software. :)
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by mantrid » Tue Jul 02, 2013 12:19 pm

Post by mantrid
Tue Jul 02, 2013 12:19 pm

You mean you have time for software too !

On your servo issue - I dont know if you've seen this: http://www.automation.co.uk/components/smart-motors/
It was an ad in the june issue of Industrial Technology that attracted me to them - their "combitronic" stuff looks interesting.
It may be too expensive or may not help but I thought it worth noting.
You mean you have time for software too !

On your servo issue - I dont know if you've seen this: http://www.automation.co.uk/components/smart-motors/
It was an ad in the june issue of Industrial Technology that attracted me to them - their "combitronic" stuff looks interesting.
It may be too expensive or may not help but I thought it worth noting.
mantrid offline
Savvy Roboteer
Savvy Roboteer
Posts: 53
Joined: Mon Oct 22, 2012 2:56 pm

Post by PaulL » Wed Jul 03, 2013 11:41 am

Post by PaulL
Wed Jul 03, 2013 11:41 am

Absolutely, I have time for software! The software is really why I'm doing all of this. I've been working on the software for a few years now.

Thanks for the link!! Seems like good hardware, but it's too big for what I'm doing. The hand is less than 2.5 inches (63.5 mm) from closed knuckle to wrist plate, with the micro servos inside of the hand. It's being built for my Robonova-1, keeping as close to stock dimensions as possible.

My parts order from McMaster should show up today. I should have a nice rotating joint by the end of this weekend. :)
Absolutely, I have time for software! The software is really why I'm doing all of this. I've been working on the software for a few years now.

Thanks for the link!! Seems like good hardware, but it's too big for what I'm doing. The hand is less than 2.5 inches (63.5 mm) from closed knuckle to wrist plate, with the micro servos inside of the hand. It's being built for my Robonova-1, keeping as close to stock dimensions as possible.

My parts order from McMaster should show up today. I should have a nice rotating joint by the end of this weekend. :)
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Thu Jul 04, 2013 5:48 am

Post by PaulL
Thu Jul 04, 2013 5:48 am

My parts arrived. The surface finish on the shim washer is really smooth, these should work great for a bearing surface.

I've been finding lately, it's one thing to draw up a design in CAD, it's another to actually see what the physical parts look like.

The shim washers are 18-8 steel, and not as hard as some would recommend for a bearing surface. But, this isn't a constantly rotating surface, and not under any great load, so the washers should be OK. I accidentally ordered .2 mm instead of .5 mm, but the .2mm should work out OK, not as flimsy as I thought they would be. I'll still change the joint pulley to give more supporting surface for the washer. I was hoping to keep the two pulleys exactly the same, but I guess it really doesn't matter since it's all CNC'd.

The 1mm bearing balls are TINY, smaller than a pin head. I almost feel ripped off for what I paid for them, as together they would swim on a postage stamp, but they're supposed to be bearing grade - how could I possibly tell? :) The balls come in packs of 250, so I'm thinking I'll probably use a few more than I originally intended to distribute the working load. I need to build 4 thrust bearings out of 250 balls, so I'll still have some left over. I'm fairly certain I'll lose a few along the way.

I also received some .020" (half mm) piano wire for pushrods for the neck. These should work out well with the micro servos. I've said it before, but the neck will be 3 DOF - rotation with tilt in two directions, using a swashplate-like design with a universal joint. I've done a few CAD drawings for the neck, the design is almost finished for that as well.
My parts arrived. The surface finish on the shim washer is really smooth, these should work great for a bearing surface.

I've been finding lately, it's one thing to draw up a design in CAD, it's another to actually see what the physical parts look like.

The shim washers are 18-8 steel, and not as hard as some would recommend for a bearing surface. But, this isn't a constantly rotating surface, and not under any great load, so the washers should be OK. I accidentally ordered .2 mm instead of .5 mm, but the .2mm should work out OK, not as flimsy as I thought they would be. I'll still change the joint pulley to give more supporting surface for the washer. I was hoping to keep the two pulleys exactly the same, but I guess it really doesn't matter since it's all CNC'd.

The 1mm bearing balls are TINY, smaller than a pin head. I almost feel ripped off for what I paid for them, as together they would swim on a postage stamp, but they're supposed to be bearing grade - how could I possibly tell? :) The balls come in packs of 250, so I'm thinking I'll probably use a few more than I originally intended to distribute the working load. I need to build 4 thrust bearings out of 250 balls, so I'll still have some left over. I'm fairly certain I'll lose a few along the way.

I also received some .020" (half mm) piano wire for pushrods for the neck. These should work out well with the micro servos. I've said it before, but the neck will be 3 DOF - rotation with tilt in two directions, using a swashplate-like design with a universal joint. I've done a few CAD drawings for the neck, the design is almost finished for that as well.
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Fri Jul 05, 2013 3:53 am

Post by PaulL
Fri Jul 05, 2013 3:53 am

Today I tried to mill the ball cage for the thrust bearing for wrist rotation from thin UHMW. It didn't go well.

Since I had some thin teflon sheet lying around, I figured I would try that instead. Much better. See below, .635 mm thick, 16 mm ID, 10 mm OD, 20 x ~1.5 mm ID ball pockets. Good enough. :)

Image
Today I tried to mill the ball cage for the thrust bearing for wrist rotation from thin UHMW. It didn't go well.

Since I had some thin teflon sheet lying around, I figured I would try that instead. Much better. See below, .635 mm thick, 16 mm ID, 10 mm OD, 20 x ~1.5 mm ID ball pockets. Good enough. :)

Image
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Fri Jul 05, 2013 5:10 am

Post by PaulL
Fri Jul 05, 2013 5:10 am

I did a test assemble to check out the bearing, I have to cut a new alu hub and a 2mm pulley for the finished joint (I have 1 mm pulleys, but I'm not going to use those, need another mm to hide the screw head to mount the pulley).

The 20 x 1mm balls are very tedious to put in the ball cage. :)

It works great. It's certainly not ideal for continuous rotation, but it's great for what I built it for, a hip / wrist joint. No lateral or horizontal play. Much smoother than the other testing I've done, even smoother than the purchased roller thrust bearing. There's still some noise in this thrust bearing (you can feel it if turned quickly), I attribute that to the surface finish of the shim washers. Even still, it's good enough.

Shim washer under cage, balls in place:

Image

Upper washer in place, this is the thrust bearing "sandwich":

Image

Assembled for testing - spacing washers to the left, white is the teflon ball cage. The spacing washers will be replaced with a 2mm pulley in the final design:

Image

View of alu hub that mates with the axial bearing on the opposite bracket and keeps things in line. This is tapped, the nut is to lock the screw in place:

Image

Axial bearing side:

Image

This is exactly what I was hoping for. :D
I did a test assemble to check out the bearing, I have to cut a new alu hub and a 2mm pulley for the finished joint (I have 1 mm pulleys, but I'm not going to use those, need another mm to hide the screw head to mount the pulley).

The 20 x 1mm balls are very tedious to put in the ball cage. :)

It works great. It's certainly not ideal for continuous rotation, but it's great for what I built it for, a hip / wrist joint. No lateral or horizontal play. Much smoother than the other testing I've done, even smoother than the purchased roller thrust bearing. There's still some noise in this thrust bearing (you can feel it if turned quickly), I attribute that to the surface finish of the shim washers. Even still, it's good enough.

Shim washer under cage, balls in place:

Image

Upper washer in place, this is the thrust bearing "sandwich":

Image

Assembled for testing - spacing washers to the left, white is the teflon ball cage. The spacing washers will be replaced with a 2mm pulley in the final design:

Image

View of alu hub that mates with the axial bearing on the opposite bracket and keeps things in line. This is tapped, the nut is to lock the screw in place:

Image

Axial bearing side:

Image

This is exactly what I was hoping for. :D
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Sat Jul 13, 2013 9:13 am

Post by PaulL
Sat Jul 13, 2013 9:13 am

I finished up the bracketry drawings for the hips and wrists. These are modified from the Robonova stock design in that they're designed to mount the required servos that will rotate the joints and accomodate the bearing while keeping the servo hub and the pulley in line.

In finishing out the CAD work for the brackets, I came upon the same problem I've had before - accuracy required for bending. The new brackets need some 7 different spacing shims with the fixed backstop I'm currently using. That's absurd. The shims are tedious, and easy to misalign square, as they've been pretty narrow up to now. I'm not happy with the whole "where will the bend end up?" problem.

Frustrated with the limitations and problems with the fixed backstop, I came up with a design that incorporates a 4 inch (a bit over 100 mm) DRO, a piece of threaded rod and a nut, and a slide that will function as a backstop. I drew up the design in CAD last weekend. The DRO showed up yesterday. The only thing remaining before I start cutting the parts is time drawing up some mounts for the DRO. I hope to get this thing cut and built by the end of the weekend, maybe by next weekend, we'll see how it goes.

I looked around on the 'net, trying to find small bending brakes with precisely adjustable backstops - they're not easy to find. There are no DIY's or other efforts out there that I've been able to find, it seems no one has tried to do such a thing to a brake for hobby / home use.

It's actually not all that complicated AND pretty cheap to do - a couple of sliding rails, two sheets, one sheet mounted, one sheet to slide, a screw with a nut and follower mounted to the sliding sheet, DRO attached.

The DRO is pretty cheap, was around $50 USD shipped:

http://www.shars.com/products/view/7587/4quot_Horizontal_Digital_Machine_Scale

The unknown at this point is the clearance needed for the rail and sliding sheet. I've added a few thousandths of an inch of spacing, but I've also got some adjustment beyond that which should allow me to tune the fit.

I will post updates for the brake backstop mod once I've got it in place.
I finished up the bracketry drawings for the hips and wrists. These are modified from the Robonova stock design in that they're designed to mount the required servos that will rotate the joints and accomodate the bearing while keeping the servo hub and the pulley in line.

In finishing out the CAD work for the brackets, I came upon the same problem I've had before - accuracy required for bending. The new brackets need some 7 different spacing shims with the fixed backstop I'm currently using. That's absurd. The shims are tedious, and easy to misalign square, as they've been pretty narrow up to now. I'm not happy with the whole "where will the bend end up?" problem.

Frustrated with the limitations and problems with the fixed backstop, I came up with a design that incorporates a 4 inch (a bit over 100 mm) DRO, a piece of threaded rod and a nut, and a slide that will function as a backstop. I drew up the design in CAD last weekend. The DRO showed up yesterday. The only thing remaining before I start cutting the parts is time drawing up some mounts for the DRO. I hope to get this thing cut and built by the end of the weekend, maybe by next weekend, we'll see how it goes.

I looked around on the 'net, trying to find small bending brakes with precisely adjustable backstops - they're not easy to find. There are no DIY's or other efforts out there that I've been able to find, it seems no one has tried to do such a thing to a brake for hobby / home use.

It's actually not all that complicated AND pretty cheap to do - a couple of sliding rails, two sheets, one sheet mounted, one sheet to slide, a screw with a nut and follower mounted to the sliding sheet, DRO attached.

The DRO is pretty cheap, was around $50 USD shipped:

http://www.shars.com/products/view/7587/4quot_Horizontal_Digital_Machine_Scale

The unknown at this point is the clearance needed for the rail and sliding sheet. I've added a few thousandths of an inch of spacing, but I've also got some adjustment beyond that which should allow me to tune the fit.

I will post updates for the brake backstop mod once I've got it in place.
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

Post by PaulL » Wed Jul 24, 2013 9:40 am

Post by PaulL
Wed Jul 24, 2013 9:40 am

It's been a couple weeks. Have I stalled on this project? No way. I've been busy building the brake backstop to better bend brackets.

The hip and wrist brackets need a number of different bends, and that would've meant more shims. I have more bending to do in the future, so it only makes sense to find a better way now.

Pictures:

Image

Image

This is the end of the threaded rod where the knob goes. I still need to make a knob for it, but it can be turned by hand:
Image

The follower block that mounts to the scale, holds the nut, and mounts to the backstop plate is on the left side of the threaded rod:
Image

Here, the follower block protruding through the top plate to the scale bracket. Above the threaded rod, there is a horizontal screw, this is a stop for zero:
Image

There are four holes on the end where the backstop plate slides forward (on the left in the first two pictures). In these pictures, the backstop is fully forward. There's a flat plate on the brake to be located under the backstop plate when mounted.

This should work great. There is some backlash in the screw / nut, but this was expected - it's not a precision screw and there's no means of backlash control. This doesn't really matter, but the backstop should always be moved towards the front of the brake as a last move when adjusting. This eliminates the backlash affecting the bend results, and the backstop is very solid, no play against the backstop after adjustment.

Specs:

* 4 inches of travel.
* Adjustable zero stop.
* Perfectly square.
* Teflon rails and strips, should last a while.
* Adjustable rails for tight fit.
* Brass threaded rod, steel follower nut.
* All alu is 6061-T6.
* Reduced height to fit under fingers when bending 1mm alu sheet.

Left to do:

* Make a knob.
* Cut down some countersunk screws to fit (4 of these).
* Coutnersink mounting holes and mount to the brake.

Then, I'm back to bending bot brackets. :)
It's been a couple weeks. Have I stalled on this project? No way. I've been busy building the brake backstop to better bend brackets.

The hip and wrist brackets need a number of different bends, and that would've meant more shims. I have more bending to do in the future, so it only makes sense to find a better way now.

Pictures:

Image

Image

This is the end of the threaded rod where the knob goes. I still need to make a knob for it, but it can be turned by hand:
Image

The follower block that mounts to the scale, holds the nut, and mounts to the backstop plate is on the left side of the threaded rod:
Image

Here, the follower block protruding through the top plate to the scale bracket. Above the threaded rod, there is a horizontal screw, this is a stop for zero:
Image

There are four holes on the end where the backstop plate slides forward (on the left in the first two pictures). In these pictures, the backstop is fully forward. There's a flat plate on the brake to be located under the backstop plate when mounted.

This should work great. There is some backlash in the screw / nut, but this was expected - it's not a precision screw and there's no means of backlash control. This doesn't really matter, but the backstop should always be moved towards the front of the brake as a last move when adjusting. This eliminates the backlash affecting the bend results, and the backstop is very solid, no play against the backstop after adjustment.

Specs:

* 4 inches of travel.
* Adjustable zero stop.
* Perfectly square.
* Teflon rails and strips, should last a while.
* Adjustable rails for tight fit.
* Brass threaded rod, steel follower nut.
* All alu is 6061-T6.
* Reduced height to fit under fingers when bending 1mm alu sheet.

Left to do:

* Make a knob.
* Cut down some countersunk screws to fit (4 of these).
* Coutnersink mounting holes and mount to the brake.

Then, I'm back to bending bot brackets. :)
PaulL offline
Savvy Roboteer
Savvy Roboteer
Posts: 423
Joined: Sat Sep 15, 2007 12:52 am

PreviousNext
PreviousNext
135 postsPage 4 of 91, 2, 3, 4, 5, 6, 7 ... 9
135 postsPage 4 of 91, 2, 3, 4, 5, 6, 7 ... 9