Spinningmagnets' RC Friction-Drive Build-log

[spinningmagnets]

StemMount.jpg (85.36 KiB) Viewed 1061 times

http://endless-sphere.com/forums/viewtopic.php?f=28&t=22245
http://www.niagaracycle.com/index.php?cPath=129&sort=2a&page=2
http://www.jensonusa.com/store/sub/147-Stems.aspx

I have been very impressed with Hillhaters use of a bikes handlebar stem as a mount option, and I am trying to figure out how hard/easy it would be. AJ pointed out that stems are available with an adjustable angle, and that solves several problems for me. Niagra bicycle parts has 38 pages of stem selections, so I am collating the adjustable-angle stem options available to me.

The vertical post that inserts into the steer-tube of a bike is the quill (like the quill of a feather-pen being dipped in ink). The length from the center of the quill-clamp to the center of the handlebar-clamp is often referred to as the reach.

The quill-clamp inside-diameters (ID) are either not listed, or are listed as 1-1/8" (1.125", 28.57mm)

Bar-clamp ID's listed are in millimeters and are
22.2, 25.4, 26.0, 31.8

Reach are also in mm, and are
80, 90, 95, 100, 105, 110, 120, 125

If you wanted to use a fixed stem, just like Hillhater, they also list the "rise", which can be zero (90-degrees in relation to the quill) up to a steep angle. Fixed stems also present an intriguing option. The bar-clamp end often has a removable cap, which might allow the stem to be attached to a bike-frame seat-tube, rather than the removable seat-post (though the biggest bar-clamp ID listed is 31.8mm).

Some stems for downhill (DH), Dirt-Jumper (DJ), and BMX bikes are short and quite strong, and many of them have zero rise, such as the pic below.



There are various adapter-sleeves called seat-post shims, split-bushings, bore-reducers, etc. They may allow a certain stem to be used with a particular diameter of tubing. I have made these in the past, and my favorite materials are a section of brass plumbing pipe, aluminum crutch-tube, or electrical conduit.

The original diameter of the pipe is not so important as the wall-thickness. You can spin the pipe by some method and file/sand the OD until you achieve the proper wall-thickness, then chop to length, and slice one side to allow it to spread out to a useable diameter. I have also cut the pipe into two half-sections when appropriate.

http://problemsolversbike.com/index.php/products/seatpost_shims
http://www.niagaracycle.com/index.php?cPath=108_188

TwoPartBushing.jpg (7.05 KiB) Viewed 1056 times

[Kepler]

Good info Spin. Great being able to get high quality parts off the shelf. I really like this idea of using an adjustable handle bar stem however, I come across so many bikes with larger 31.8mm seatpost. Still, for those with 27.2mm seat posts, this is very tidy way of doing the job.

[spinningmagnets]

Thanks for the kind words! Are you as busy as I mentioned in several posts, that I'd thought you would be?

Good point on the seat-post diameters (thanks for reminding me!), the cheapest bikes often have a 1.0" diameter post, and two of my bikes have the 27.2mm variety. E-commuters seem to eventually end up with a mountain bike to affordably get a front disc brake, fat tires, and suspension for the potholes/curbs we are forced to pass over. Bigger diameter posts seem to be the trend for upscale MTBs...

If a stem-mount looks like a solution for a problematic application, a seat-post shim can be used to insert a smaller diameter seat-post into a bigger seat-tube. Here's one list of seat-post shims that are only $7, and should cover most adaptations.

http://endless-sphere.com/forums/viewtopic.php?f=28&t=21365&start=45#p318293

[Kepler]

Very busy at the moment. Re designing to make manufacture more afordable and basically lower the parts count. Also looking at some additions to the drive (avatar is a bit of a sneak preview ) Interface refinement is the biggest focus at the moment though as I firmly believe this the key to bringing RC friction drives to the people.

[spinningmagnets]

HA! Just noticed the new avatar before I read what you wrote! I should have known that there was no dust collecting on your R&D bench! I suppose you can sleep when you're dead!

Between the two-speed Kepler-drive, and the throttle interface, I believe you will have solved the remaining problems of adapting RC motors to E-bikes for the masses (since both of those solutions will reduce the peak amps the ESC sees). The interface in particular is something I anticipate to be a huge seller for you, as it sounds like it will be useful to just about any RC application. The other option for designers who crave a medium-power RC-motor is to add hall-sensors and use a larger 6-FET Infineon controller, so your gadget is a very easy decision.

Matt's recommended combination (Astro 3220 motor with Castle Creations HV160A ESC) is truly wonderful, but at $700 its beyond my budget right now.

Jeremy Harris has designed a compact and affordable throttle-signal-translator which will find its niche, but it doesn't have the wide range of solutions and features that you have described for your unit. I would say "best of luck" but I don't think you need any more luck, you are clearly on your way by virtue of your research, wise choices, and hard work!

[Hillhater]

The other option for designers who crave an RC-motor is to add hall-sensors and use a larger 6-FET Infineon controller,....!

SM, I have also considered this option, however,i am having trouble finding a sensored controller that will happily work on the 18-22v range and cope with the high current 60-70A, needed for these friction drives.
do you know of any ? i suspect it would be another "custom" built unit.

[spinningmagnets]

As far as controllers for the amp-load, I "think" it depends on what your build is drawing, so you may even need the 9-FET or even a 12-FET controller (even larger and more expensive).

If you contact Lyen, I'm sure he could set you up with an Infineon that will use the voltage range you need. I've been given the impression the voltage range is a minor change, but I don't know. Edit: Gwhy is using a Keywin-E-crazyman controller with $30 mods (= $55 total) on an RC motor using hall-sensors, and he sounds pleased with the performance.

I haven't done much homework on those options, once I determined that I wanted to explore the 85A ESC on a small diameter roller as my first RC project. There may be several unforseen roadblocks to using an Infineon/halls that I don't know anything about. He did mention that the lower kVs (170/130) were more doable for those units for some reason.

[spinningmagnets]

My first shipment from online metals arrived by UPS, took 4 days, prices are reasonable for small pieces, I'm happy with them so far.

http://www.onlinemetals.com/merchant.cfm?step=2&id=60

FD_013.JPG (118.24 KiB) Viewed 1202 times

[EVTodd]

Try to find a local supplier for aluminum. Any of the online places are 2 to 4 times higher than they should be. Seriously, their prices are nuts.

My local surplus warehouse says I'm their biggest aluminum customer and I know I'm even getting a bit screwed there but man, those online prices are WAY out there.

[D-Man]

Lowes , Home Depot, and OSH will happily screw you on metal prices.

One thing I don't see anyone talking about is that there must be something felt with a drive connected to the seatpost while your sitting on it. Maybe vibration or the seatpost fore and aft movement? Seems like that would be annoying.

[EVTodd]

Lowes , Home Depot, and OSH will happily screw you on metal prices.

I didn't really mean national retail chains. I'm talking metal suppliers.

I've personally never felt the motor even though I have it mounted via seat post rack. I don't think it's a big issue.

[spinningmagnets]

Yes Todd, I agree. I am in a remote region two hours from Kansas City, and one hour from Topeka, currently surrounded by beef ranches and farms (what the hell is Sorghum and Milo?). I did find a small local supplier called "Custom Cut Metal", and I was pleasantly surprised at how affordable the pieces in the first pic of this thread were. Plus not having to wait 4 days for UPS is nice. (CCM, 8989 Green Valley Dr, Manhattan, KS 66502-9005, Ph# 785-537-0441 ) ‎

The local supplier has half the stuff I wanted, and the shipping cost from Online-Metals was cheaper than the gas to go to Topeka/KC for hard-to-source bits. I'm not done fiddling around with the final design yet (now trying to incorporate the adjustable stem above), I only wanted one foot of the odd-dimension stuff.

If anyone wanted to copy what I've done so far, I think Online-Metals might be a viable option to build one drive. However, I'm sure anyone in a major city could find a better source with a little effort. Find a local fabrication and welding shop (especially if they make aluminum trailers, their recycle scrap-bin is wonderful and cheap), then just ask where they buy small orders of metal.

[daveo]

ESC is a bit light on. The 100 is a better choice but out of stock. (The 20 I recently bought might have contributed ) That being said, I am hammering an 85 at the moment and doing all I can to smoke it. Bloody thing wont break I have hit the high temp limit on it a few times though.

Noobie question: If the 85A is a bit light, couldn't you just use a 80 or 100 amp fuse between the ESC and the motor to keep from cooking the ESC? Maybe like these:

https://www.vtewarehouse.com/content/electromech/fuse/html/maxi/maxifuse.php

Replacing a $0.70 fuse is a lot cheaper and wouldn't leave one stranded.

[dontsendbubbamail]

ESC is a bit light on. The 100 is a better choice but out of stock. (The 20 I recently bought might have contributed ) That being said, I am hammering an 85 at the moment and doing all I can to smoke it. Bloody thing wont break I have hit the high temp limit on it a few times though.

Noobie question: If the 85A is a bit light, couldn't you just use a 80 or 100 amp fuse between the ESC and the motor to keep from cooking the ESC? Maybe like these:

https://www.vtewarehouse.com/content/electromech/fuse/html/maxi/maxifuse.php

Replacing a $0.70 fuse is a lot cheaper and wouldn't leave one stranded.

Maybe you mean between the battery and the esc. I not sure what would happen if you tried to fuse the three phase wires between the esc and the motor. Most likely nothing good. The folks at Castle Creation told me not to put a fuse between the esc and battery. They said that blowing the fuse could also blow the esc.

Bubba

[Hillhater]

The folks at Castle Creation told me not to put a fuse between the esc and battery. They said that blowing the fuse could also blow the esc.

Bubba

Hmm ?, i would like to understand the explanation for that statement,
I know several battery packs have fuse protection built into them as standard !

[amberwolf]

If the ESC were capable of regen, and the motor was in regen mode and not freewheeling, I could understand that if the fuse blew the voltage would have no load anymore and could spike very high, potentially damaging the controller.

The only other situation I can think of is similar, in that if the controller does not have enough capacitance to suppress the switching spikes on the input side without the battery connected, then bad stuff could happen for the same reason.

THe former is not likely, as I doubt these controllers do regen, but using a freewheel would eliminate the problem anyway.

THe latter is fixable simply by adding more low-esr capacitance.

[daveo]

Thanks for the replies. Apparently the answer is a big NO on the fuse idea. Figures. It was too simple.

[spinningmagnets]

Making progress, waiting for more parts and time to fiddle with them...



There's two things I will change on the next one. The motor is so close to the back edge, I had to install one of the motor-plate bolts with the head on the inside. I will move the roller closer to the center so the motor mount bolts (4 of them eventually) will all allow the motor removal with an allen wrench from the outside, and it won't disturb the roller.

Also, I was forced to use a chamfered flush-head bolt under the roller. After the roller is moved closer to the center, I can use the same 5/16"-18 bolts as the rest of the drive (all using a 1/4" allen wrench, the single flush-head bolt uses a smaller wrench). With the current arrangement, the roller can also be changed out without disturbing the motor.



Finished the motor-shaft slotter (similar to the scrap-wood roller-shaft slotting machine, but with two discs). I re-read Grinhills second build, and he has successfully used this style of joint after remembering that older cars drove the distributor with a similar joint. I left 'just enough' slop to allow for heat expansion between the parts. The shaft shortening and cutting the two flats went easier than expected, thats a relief...

The motor mount-plate is 1/4" thick. I tried to use 20mm long screws, but they bottomed out against the skirt bearing. I needed at least 13mm to get full thread engagement, and I was able to find some chamfered flush-mount M4-0.7 screws that are 16mm long. I am somewhat annoyed the 4 motor screws use a 2.5mm wrench, but thats just the way it goes...

FD_016.JPG (109.54 KiB) Viewed 9090 times

[drifter]

Looking good Spinner, I will watch for further progress.

[Hillhater]

Making progress, waiting for more parts and time to fiddle with them...
...The motor mount-plate is 1/4" thick. I tried to use 20mm long screws, but they bottomed out against the skirt bearing. I needed at least 13mm to get full thread engagement, and I was able to find some chamfered flush-mount M4-0.7 screws that are 16mm long. I am somewhat annoyed the 4 motor screws use a 2.5mm wrench, but thats just the way it goes...

Why didnt you just use the alloy X motor mount that comes with the motor ?? .. complete with csk screws ?

[spinningmagnets]

Xmount.jpg (22.75 KiB) Viewed 8574 times

Using the stock X-mount would have been faster and easier...but where's the fun in that? Several other builds have implied that whatever the motors stator-face is bolted to will act as a heat-sink, so I thought I'd start out with an aluminum plate that was bigger than needed. I am hopeful that these parts, and the application that they will be used for will not have any heat issues. That being said, I'd rather start out too cool, and work my way towards warm.

I also had some ideas that I might explore later, concerning a fan-cooled cover to make the motor quieter. I don't know if that will work at all. Or, if it does work 'a little', whether it will be worth the effort. Thats one of the reasons for the larger plate, as a base for a cover (I can always make the plate smaller later).

edit: just remembered, another reason I attached the motor to such a wide baseplate is that I am still in the R&D phase, and if I needed to attach another motor with a different bolt pattern, I wanted to be able to swap in another motor base plate, rather than drill more holes in the side of the drive.

I also HATE the tiny phillips screws that come with it, I have already stripped one the cross-slots that the screwdriver turns, without much effort (and I was being careful). After a close examination, I now believe they are made from an unusually stiff form of butter that is covered with aluminum paint. The allen-socket screws feel much better, and it was only $1.60 for 4 of them. If I bought a box, they'd be 1/3rd that price.

On a side-note, I no longer have any reservations about the 5/16"-18 bolts I'm using. I tried to purposely strip out the threads in one hole of a 1/4" aluminum plate. Two side-by-side attached plates that had been drilled and tapped at the same time (for alignment), one hole to remain threaded.

I wanted the plate on the shank end of the bolt to have a smooth hole, but still snug to keep alignment. I had to keep side-milling the threads I had cut by using a drill bit (rather than order a 19/64" drill from McMaster-Carr). There were almost no threads at all, before I was finally able to strip the hole smooth.

[Hillhater]

Hey SM, you need to get off the fence and decide to go either all metric or stay "old school" ..
You are building a real "multi-national" / multi- dimensional drive there..not good in the long term ...just pick one team and support it full time !

I used to build everything with 5/16 UNF cap screws as i had boxes of every length imaginable together with tap sets & matching "nutserts" (threaded inserts for fixing to plate /sheet or blind tubing etc.
I still have a few , but i try to stick with metric stuff now as its more readily available.
5, 6, and 8 mm is all you need on this scale of build .. and of course a full range of drill sizes..inc number and letter scales.

[spinningmagnets]

Yes, my friend, I should go all metric soon. The local store has a good selection of the SAE black-oxide socket-head bolts, plus drillbits and thread-taps. If I could just stop changing the design every two days!...too many new good ideas showing up. With bicycle parts, metric is definitely the way to go.

edit: noticed today that my bikes have a lot of bolts using 5mm and 6mm hex-sockets (handlebars, seat-post, brakes, etc), so I am likely to use those in future builds.

I just hate to order a whole box of a certain bolt, then realize that a design change demands a different size. I'm just using whats quickly available until all the bugs are worked out.

With the motor mounting screws, I only meant that, I wish the wrench for them was 3mm instead of 2.5mm...much like trying to buy a 19/64" wrench. BTW, I definitely like the nutserts as an option. Fast too, drill a smooth hole in aluminum, press-in a steel-threaded nutsert. Ridges on the nutsert OD grab the hole, and once you insert a bolt and tighten it down, its good to go!

edit: I tried two nutserts on the small holes I had to drill for the longer bolts that compress the two clamps. The seat-post mount uses a 5mm wrench to install and adjust the drive, so I will use bolts on the two rotation-clamps that also use a 5mm wrench, and nutserts!
http://www.fastenal.com/web/products.ex?N=0&Ntk=Search+All&Ntt=nutsert&Ntx=mode+matchallpartial&Nty=1&searchBox=1

edit 11/01/10: A package of M6 nutserts just arrived in a padded envelope, 4 working days from their US warehouse to Kansas. The smooth part of the tube-OD is 0.388", but oddly, the ribbed portion seems to be about the same diameter? It struck me as strange for a moment, as they have recommended a 10mm drill (0.394") to make the insertion hole, and I assumed the ribs would be an interference-fit.

Went back to the web-catalogue and found that these are intended to have the neck collapse and by doing that, it expands outward into the hole (using a special tool). For that reason, the flanged and ribbed half of its length has no threads.

I collapsed one just now with just a bolt, and when this is done outside of a hole, the collapsing neck forms a thicker flange lip (0.118"). The insertion length is reduced to 0.266". The engineering drawing seems to indicate that when inserted into thin sheet-metal, the created flange sandwiches the sheet-metal between the two flanges. Or, when collapsed inside a hole thats made in thicker metal, the wall collapse forms a sideways-fold that expands outward and also inward.

I still plan to use them, but the closest drill size seems to be 10mm (0.394") so it looks like they will need a thin smear of JB / Devcon

The attachment FD_021.JPG is no longer available

[Bluefang]

Hey mate, those steel inserts look great. Do they need to be pressed in or can they just be hammered in? Also with your bearings, did you press them in or drill the hole big enough that they just slipped in?

[spinningmagnets]

Nutserts are installed into a smooth hole made by a drill-bit, then they tap in halfway with little effort, and tightening a bolt into them fully seats them (edit: depends on the diameter of the hole, I have found its hard to get a perfect fit). They can even be flush-head if you spot-face (make a smooth shallow recess) on the the insert side with a bigger drill bit by shaving off just a hairs thickness in depth. If you do get a single hole that was drilled a little too loose for the nutsert to grab properly, pull it back out and a smear of JB/Devcon will set the loose one for good.

These M6-1.0 nutserts call for a hole made with a 10mm drill, but that does not provide a tight fit. I have been using a smear of JB-Weld/Devcon with good success. As an option, you can put a bolt inside the nutsert to prevent collapse while you clamp the nutsert OD in a steel vice (in several places around the OD), and the rough vice jaws faces 'knurl' the OD of the nutsert just enough that they then have to be forced into a 10mm hole.

The nutserts that are made to insert into wood holes have fairly big spikes, but the steel ones that are made to insert into aluminum holes get by fine by just using some ribs (oops, see paragraph above). The nutserts cost more than common nuts (50 cents each?), but if you're only doing 5 or 6 holes and you don't have the $5 tap to cut threads (or, you don't want to risk breaking a tap, especially on a small hole), its not a bad option to consider.

With the two roller bearings, I got lucky and found a long tube in the pile ( 1.000" OD / 0.748" ID) at the metals supply place that had an ID that was 0.002" smaller than the OD of the bearings. The precise tubing dimensions vary a little from one piece to another.

I put the aluminum tube in boiling water, and the ID expanded 0.010", and at that point I could pick it up with long pliers and tap the tube onto the bearings with a hammer. You must move quickly, as once they touch, the temps equalize rapidly. Its called an interference fit. Too loose or too tight, and it won't work well.

viewtopic.php?f=28&t=21365&start=15#p313906

If I can only find tubing (in the future) that has the same ID dimension as the Bearing shell (0.750" ID -0.750" OD) I will have to glue them in with JB or Devcon to prevent slippage.