Dyne's LD-F1 (D-O concept) build

[kresty]

Looks pretty cool! Access for changing batteries and stuff is always important in a build. I like how you took the idea of a build from a concept and ran with it.

[Dyne]

Mid-dragoncon report, so I'll be brief:

LD-F1 is here, but is suffering a few mechanical issues (slipping belts, a nonresponsive servo, the lights seem to not be working -- one of them due to a wire coming off the LED Sequin).

I'm not sure what the issue is on the first problem, as they were fine before the con. The servo is probably down to a loose connection somewhere. The lights could be another loose connection or just the ESP32 not booting properly. I don't have my iron here and I'd have to extract the LED, so a fix for the sequin will have to wait.

In the meantime I have my new MSE droid, which is very popular, so I'm not lacking for something to drive.

[kresty]

Well, in the movies they're always repairing droids.... there's a reason for that!

Sent from my SM-S901U using Tapatalk

[Dyne]

Well, in the movies they're always repairing droids.... there's a reason for that!

Sent from my SM-S901U using Tapatalk

True. Even my mouse required multiple repairs and screw replacements.

I decided to bring Eldee out again yesterday (last chance), even if I was only going to keep his head level. While fiddling around to reverse a prior radio swap I found the magic position where the jumpers from the receiver to the head servos worked. The lights (aside from the sequin) also randomly started working again.

That left just the belt skipping issue, which really mostly just happened when turning on the carpet. At least he didn't throw or break any belts or damage the chassis, as far as i know.

Sent from my SM-A326U1 using Tapatalk

[Lichtbringer]

Maybe it's due to a bad motivator.

I've seen that before.

[Dyne]

Definitely.

To that end, I really need to solve the belt issues on this droid.  (I'm perversely annoyed that he didn't break a single belt after I went to the trouble of bringing five spares for each type.)

I should probably look into commercial belts, which have the advantage of being fiberglass reinforced and a bit more flexible. LD-F1's belts are all made to 9 mm wide HTD-5 (5mm pitch) specs.  He uses a 40 tooth (200 mm) primary stage belt, and a 58 tooth (290 mm) secondary stage belt on each side.

Finding belts with these exact tooth counts is tricky, but not impossible. McMaster-Carr doesn't even carry belts in the right spec with less than 75 teeth (375 mm).

It's easier to find them on a specialty shop. For example, I found both on VBeltOutlet right away. The 40Tx9mm belt is currently $1.10 and the 58T x 9mm is $1.20. The drawback is that shipping alone for 4 each of those costs almost $37 which is ... excessive.

On Amazon, I can find the 40T belts for $3 to $5 each.

I'll keep looking around, I think.

[Dyne]

Recently, I remembered that there was another option available for at least partially improving the belt situation: replace the primary belt and its pulleys with a pair of gears.

Just to explore the idea, I designed a pair of herringbone gears -- the green one for the brushless motor and the yellow one as a replacement for the larger part of the compound intermediate pulley.  Assuming that these gears work well, that would leave only the 58T belt on each side, which means less things able to slip.


The gears have the same gear ratio that the original pulleys had -- the green motor gear has 15 teeth, and the larger gear has 45, so 3:1.  In the current droid, the motor shaft is 51.747 mm on center from the axle of the intermediate pulley.  However, the geared version will need to sit closer -- meshing, in fact, which makes them 35mm on center.  

That means I'll have to redesign the front side walls again to move the motor mount.  To keep the gear diameters roughly similar to the pulleys they replace, I kept the module of these gears fairly low (1 mm).  I might have to increase it if the gears aren't strong enough or they don't mesh precisely enough, and I'd have to change the side wall yet again.

As you can see in the CAD, there aren't any big issues with the parts interfering, though I might need to put a washer beneath the motor gear and lower its position accordingly (so that the black shaft part is less likely to contact the yellow gear at the edge nearest the motor).  It's possible the secondary belt may rub the motor if it doesn't seat well, so I'll have to check that.

Currently, each intermediate pulley uses a pair of 623ZZ bearings, which are 10 mm OD, x 3 mm ID x 4 mm thick.  As it happens, I've been working to upgrade the front axles and steering on the Baddeley Mouse droid to use M4 bolts, and as part of that design, I obtained a pack of MR104ZZ bearings.  These bearings are 10 mm OD x 4 mm ID x 4 mm thick -- a drop-in replacement if I want to upgrade these pulley axles to M4 as well.  (Another good reason to redesign the side plates.)

That sounds like a good idea to me.  The belts put a lot of tension on the M3 hardware I currently use as pulley axles, but since there are two belts pulling in opposite directions, the net result is kind of a wash.  With only one belt, the force would instead act unopposed in a single direction, which might be more likely to bend the axle.  (M3 axles getting bent is one of the two big reasons I'm upgrading the Mouse droid to M4.)

Edit: And here is the new arrangement with the revised front outer side wall. I've put a recess for an M4 lock nut in the back of the wall, but the distance from the underside of the hex head to the outside of the yellow part is just under 50 mm, so 55 mm is likely the shortest bolt that will work.


One drawback of this arrangement is that I'll once again lose my new easier belt installation option, and will instead have to take bits off.


As for the secondary belt, I can't really directly replace that one with gears, because the replacement for the integrated 42T pulley on the track's drive sprocket would need to be larger in diameter than the sprocket (and the track) in order to reach the smaller bit of the yellow part above.  It might be doable with a gear train, but then the challenge would become finding a way to mount axles for the additional gear(s).


Oh, while I'm here, I decided to get an early start on a possible build for next Dragon Con... the OTHER D-O concept from Luke Fisher (third image here).  After Dragon Con 2022, I considered doing this version for 2023, but there was a lot of doubt that I'd be able to build anything at all, and by the time I decided to, it was too late to start something that involved that much design work.


This one is being done as a mod for the Baddeley v2 D-O (or at least what's in the Fusion file). No idea if or when I'll get further than just cosmetic CAD, but I will likely start a build thread if I do.

[Dyne]

It's been awhile, but I finally did some more work on LD-F1.  Specifically, I just finished implementing the gear drive system that I mentioned in my previous post.

As a reminder, the old system had a two stage belt drive with 3D printed TPU belts using an HTD5 x 9mm wide profile.  The primary belt had 40 teeth - which determines length of the belt - and had a 3:1 reduction (10T pulley on motor to the 30T pulley on the compound intermediate stage).  The secondary belt has 58 teeth and has a 4.2:1 reduction (10T pulley of the intermediate stage to the 42T drive pulley).

The clearest view of these parts (except the secondary belt) is probably back in post #36.

The new setup would replace the primary belt on each side and the relevant pulleys with herringbone gears (with the same gear reduction, as described in my last post), halving the number of possible sources of slippage.

I printed the new parts some time ago, but I couldn't actually install them until I ordered new M4 hex bolts for use as the axles for the new intermediate pulleys.  Also, I knew installation would require some significant disassembly of the droid's front end (it requires an entirely new front outer wall on each side), so it took awhile to work up the motivation to do it.

Here is the front section of the droid:


The right side has been upgraded, while the left side still has the original parts.  Notice that the right outer wall is a different color than the other black parts (because it's painted, not black plastic).  You can see how much closer the motor sits to the intermediate gear on the upgraded side.  

Here is the new setup.





This shows pretty clearly why order of operations matters for the installation. The remaining belt needs to go around the small pulley inboard of the larger gear. Since the two gears mesh, there's no way to put that belt on once they are installed.  I need a gap to allow the belt to go between them.

So long as the intermediate gear is in place, I can't actually take the motor gear off without removing the motor entirely, and I can't take the motor off (without further disassembly) because the mounting thumbscrews are hidden between the inner and outer walls. Which leaves only loosening/removing the intermediate gear.

Thankfully, I designed the new mounting post for the part so that I'd have the option to put in its axle bolts pointed toward the center of the droid (possibly with a lock nut in a recess between the walls), not just pointing outward as I originally considered doing when I was still thinking of using a long M3 bolt.

Doing it that way means I can loosen or entirely remove the intermediate gear without disassembling anything else.  I don't even need to use that hypothetical lock nut between the walls, since the mounting post can be tapped for the axle bolt and is thick enough to keep the bolt from backing out by itself.  

There's only one issue preventing that...

At the bottom of the last photo, directly below the intermediate gear, you can see where I've started slicing off the support arm for its old M3 axle. This arm is no longer necessary or useful (if it ever was).  In fact, it's actively in the way -- once the front end is actually on the droid, the arm hinders my ability to get a ratchet onto the gear's axle bolt. So it needs to go.


I haven't tested the new drive yet (I need to break out my radio first, and eventually put the tracks back on) but I'm pretty happy with the motion when turning the system by hand.  

With the belts it always required a small effort to start turning and felt sort of ... granular?  Bumpy?  Now it feels smoother and pretty low-effort.  With any luck, that also means it'll slip a lot less.  (My suspicion/hope is that this startup resistance, exacerbated by the full weight of the droid, plus the higher rpm on the motor side, meant that the 10T pulley that was previously on the motor was the most likely place for slippage to occur, not the 10T pulley on the intermediate stage that still exists. But i could be wrong.)

I could still stand to improve the remaining belt's tension, but I'd have to figure out where to cram in an idler/tensioner to do that.  I had some thoughts on that, but I haven't implemented them in this iteration and I'm not sure they are feasible.

[Dyne]

I did my "benchtop" powered test and it works.  





Still noisy, but that's no surprise.  I made some tweaks after this (a little lithium grease on the gears, tighten the right-hand motor spindle, center the front axle properly, etc.) and finally finished cutting off the motor arms.

[Dyne]

Compare that to the original test of the track mechanics back in 2022 (forum only allows link one video per post):