The Linear (recumbent) Booster

[lawsonuw]

Well the my motor still works. It's now star wound with about a 360Kv. Fine sandpaper got rid of most of the insulation on the wires. A long dip in a micro-solderpot cleaned off what was left and made a nice joint. (Micro-solderpot, aka a hole drilled part way through an old soldering iron) Only had to re-bundle one set of strands. One of the strands was sorted into the wrong bundle shorting two of the phases.

I'll try to get it back together this weekend.

My power system

Linear Booster Power system (Small).jpg (51.16 KiB) Viewed 2018 times

On the left sticking out of the box is my key/precharge switch. Doesn't work, because the ~300 ohms of resistance it has is too much resistance when the controller beeps the motor. After that the controller refuses to start up. I'll put some smaller resistors in it to get it working. The black mass below the controller is my extra input filtering. It's ~2000uF of capacitors and a 30v TVS diode. Should keep the controller happy even with a 3-4 foot supply wire.

The last Picture is what I'm planning this to look like when it's done.

Marty

[lawsonuw]

Well... It works. It can't be putting out much more than 100w now, but right into the chain 100w is a nice boost. Changing the motor to star winding reduced power enough that the gearbox survives. Also dropping to 12:1 reduced gear stress. With the 12:1 tranny and ~360Kv motor the pedals can spin faster than I'd like. A new problem did crop up when I made this switch. With the new gearing and motor wind the glue joint between the motor shaft and pinion broke. I guess my glue doesn't stick to brass that well. I soldered it up with some silver bearing solder I had left over. So far it's stayed stuck.

If I was going to do this again I'd just stick with the motor the gearbox comes with and get the right gear ratio. This 64:1 gearbox would be about right. Similarly I'd just use any old 30-40amp airplane ESC. I'd also get THIS Staton inc freewheel adaptor. To fit the freewheel adaptor I'd glue a McMaster# 6381K464 bronze bushing to the gearmotor's shaft with Loctite 609 or equivalent. Next I'd make dimples or groves in the outside of the bushing where the set screws from the freewheel adapter would go. Finally I'd attach the adapter with the set screws and a LOT of blue threadlocker.

I'm already thinking of version two. In preparation I'll be making an electrical model of my system so I can calculate my maximum currents with a given motor and loading. So that means making a Milli-Ohm meter. (just a calibrated one amp current source, the plans popped up here a few months ago) Current plan is to take a 200kv outrunner and do a Delta to Star switch for about 120Kv. Then run a single stage belt reduction of 7.5:1 with an overload clutch to drive the freewheel.

Marty

P.S. It's a LOT quieter now. I think all the noise earlier was my gearbox screaming in pain.

[MitchJi]

P.S. It's a LOT quieter now. I think all the noise earlier was my gearbox screaming in pain.

Hi Marty,

Nice job!

How is it lubricated? You might want to contact John about this gearlube:
viewtopic.php?f=12&t=5325&p=93927#p93927

I also got some special gear lube that is supposed to make gears very quiet. The guy that blends it actually told me to get a dB meter and test the numbers!

[lawsonuw]

The BaneBots gearmotors are lubricated with what looks like a cheap lithium grease. I've since upgraded to some "extreme pressure Molly-Lithium grease" I have laying around. The Molybdenum disulfide content of this grease improves efficiency and should make the gears last longer.

Gear noise isn't a big issue anymore, and listening to the gearbox gives me an idea of how much torque the motor is putting out.

Marty

[lawsonuw]

About time I made some more updates.

First, my pre-charge switch now works. I lowered the resistance to about 30ohms and the controller beeps twice like it should.

Next I made a mili-ohm meter extension. (First picture below) It's all parts that can be found at the local Radio Shack. The 1.25 ohm 'resistor' is seven ten ohm resistors in parallel with a twenty-five ohm potentiometer in parallel. I run it off of the 3.3v output of the converted ATX power supply powering one of my chargers. The regulator drifts as it heats up, I really should add a head sink some time.

Third, I thought I'd post a picture of the gears I'm sending 100-150 watts through. It's the second picture below. They're puny and brass, yes that's a quarter to the left.

Fourth, I dropped some cash at the LBS and got a little more gear reduction for the motor. (see the last attached picture) Stuff should last a bit longer now.

I've also figured out how to install the set of Power Grips I bought a while ago. My feet are REALLY wide so the directions were no help. I've removed one of the screws holding the pedals together and screwed the outboard strap mount into it.

Last, everything seems to be holding up now. I can give the motor full throttle with about a 60rpm cadence (and my old gearing) and nothing breaks. I only use about three minutes of power on my whole commute. The motor should last a LONG time.

Marty

P.S. The duct tape is here to stay on this version.

[lawsonuw]

John got me thinking when he suggested using Dewalt drill motors and gearboxes for e-bikes. Normal drills are rather long and a challenge to find space for. So I started looking for angle drills. Harbor Freight has a few cheap ones, and they ever post .pdf manuals online with exploded parts views!

This and This looked interesting. What's the exploded drawing look like? ( Last two pages ) BINGO! we have a winner. I went an bought an example to dissect asap. (he local HF even had the orange one on sale )

First picture shows the guts. Gearbox is 19.25:1 with metal gears and a round lip on the case near the output for easy mounting. 3/8 - 24 right hand tap for the chuck with a left handed locking screw. The long input shaft is 7mm at one end and 6mm at the other, shouldn't be too hard to turn it into a 5-6mm shaft and slip an RC outrunner onto it.

Second picture shows the series wound motor the drill came with. Spins at about 29,000rpm and rated for 3.2 amps at 120 volts ac. (a 382 watt input motor) The iron for the field is round with some flats. It would be really easy to make a replacement field with some neodymium magnets. This would dramatically lower the max rpm of the motor (hopefully to around 6000rpm) and let it work at 80-90 volts. The armature has about a 5.5 ohm resistance. Rewinding it for a lower voltage would be difficult as the armature was dipped in resin.

So, should I mod the existing motor, or should I graft a RC motor onto this bad boy?

Marty

[johnrobholmes]

It would be easy enough to couple an RC motor into it. What size is the shaft going to the gearbox?

[lawsonuw]

...The long input shaft is 7mm at one end and 6mm at the other...

EDIT: whoops I guess it's not obvious WHICH end is 7mm and which is 6mm. (this stuff is also listed on the postit note in the first picture) The 7mm end is the one going into the gearbox. The 6mm end is ground into a double-D shape and uses a couple to connect to the motor. Note: This shaft is supported by only one ball bearing at the gearbox, the motor and coupling provides the other bearing for this shaft.

In between the ends the "shaft going to the gearbox" is like 6.5mm and NOT concentric with the ends. The small end with the double D for the motor is center drilled. (on center even from the looks of it) I'll check for a center drill on the end with the gear, and the shaft's hardness.

Marty

[lawsonuw]

The input shaft of the HF drill's gearbox is un-hardened. It has a spiral bevel gear cut on it's end with another center drill. I also re-greased the gears while it was open. I'm currently leaning towards a RC system more. Motors are available with a 6mm shaft and 400-450kv which will work with the battery/controller I have now. (an 80 volt battery for the stock motor is just expensive any way you slice it)

In other news my current BaneBots transmission is getting quieter. The gearbox now makes about as much noise as the motor. (and it sounds less like a coffee grinder)

Likely next up on my list of mods is a wind/rain shield, and a current limiter. Progress will be slow as I'm rather busy this semester.

Marty

[lawsonuw]

Well two new developments on this beast. First I figured out how to mount the angle drill! I'm replacing the top five case screws with 1.75 inch #8 sheet metal screws. An aluminum plate will be fabricated with hole positions matching the case, and inch long spacers are added between the case and mounting plate. (see my first picture) I've also drilled the mounting holes all the way through on the other case half. This will allow me to easily mark the locations of the mounting bolts on the future mounting plate.

Next I got busy in the machine shop. Made an adaptor to fit a 13T BMX freewheel onto the drill. 30mm M1 thread on the BMX, and a 3/8-24 UNF thread for the drill's chuck mounting. (the thread I actually made is a 26tpi and a little under size, darn lathe predates the use of metric threads in the USA ) Also made a shiny new field iron for the drill's motor. Even if I don't use the drill's stock motor I've got to see if this works like I'm expecting. (Magnet the field iron was designed for will use 10 )

Last Picture is what the guts of this drill's gear box look like. The gears aren't all that large and the output bearing setup is a questionable for my use, but for less than $30 I can live with it. While it was open I replaced the unknown red grease with my leftover super gear grease.

Marty

[johnrobholmes]

Nice! That looks rather beefy.

[lawsonuw]

Got the magnets for my new field in yesterday. Started gluing them in. The five silly magnets for each pole didn't like sitting next to each other! I had to glue them down in two batches. First the center and two outside magnets. (using a spare magnet to get the spacing right) Only after the glue had set for those magnets could I glue down the last two. Also, I'm glad my magnets were small. I'd have several shattered magnets by now otherwise. And even these small magnets are a pain to get apart.

Also, I've got the motor running on 12v right now. The output of the transmission clocks in at 76.4 rpm. So about 1471 rpm at the motor. 1471rpm/12v gives me a ballpark Kv of 123 rpm/v. Factoring in the .6A idle current and 4.5 ohm terminal resistance a better estimate of the Kv is 158. Good news is with my current gearing I'll only need about 35-41v to match the motor's free speed to a 100-110rpm cadence. (34T on the crank 13T on the gearbox) Bad news is that due to the internal resistance of the motor and brushes, I'll be hard pressed to get more than 100w out of the motor at this voltage. (each brush has almost 1 ohm each!)

Back to plan A, I'll be putting in an order soon for parts.

Marty

Edit: Just hooked it up to the 30 cell NiCd on my Trek. (currently at 38v) and it spins at about 200 rpm and draws about 0.7 amps of idle current. Yes, this motor now cogs a LOT.

[etard]

You know the 36 volt drills put out around 600 in/lb torque, that seems to be enough for our application. has anybody thought about using a dewalt or other drill to power thier bike? I live in a dry climate, so I was thinking about using one with plug n play batteries in a friction type drive. I could use the trigger as my throttle, and just use all the internals of the drill. Or would the gearing of the drill make it more suited for a Curry type of drive? Does anybody see any problems with this logic?

[lawsonuw]

No problem with your logic Etard. Cordless drills make pretty good gear motors. The battle bot community often uses drill gears and motors for drive. Drills often spin 500rpm or less in low gear, which with one chain stage can be geared to the pedals, or directly to the rear wheel. They do make a bit of noise though.

Anywho, Just pulled the trigger on a bunch of parts. For this machine I ordered This cheap motor. It's a bit slower than I'd like, but a new freewheel or a few more cells will fix that easily. I also ordered 2x P# A 7C55MPSS1906 and 1x P# S40PX0MHG6M-130 from http://www.sdp-si.com. I'll replace the motor's shaft with these and set it in the same spot as the original armature. I'll also have to build up the OD of the front bearing to 22mm to make it fit, and provide a torque arm for the motor.

Marty

P.S. I also bought a bunch of parts for my 1lb battle bot Micro Vice. Check out an old version of it over HERE on YouTube.

[michaelplogue]

You know the 36 volt drills put out around 600 in/lb torque, that seems to be enough for our application. has anybody thought about using a dewalt or other drill to power thier bike? I live in a dry climate, so I was thinking about using one with plug n play batteries in a friction type drive. I could use the trigger as my throttle, and just use all the internals of the drill. Or would the gearing of the drill make it more suited for a Curry type of drive? Does anybody see any problems with this logic?

A bit pricy, but........

http://www.dpxsystems.com/ProductDetail.aspx?ID=199

http://www.dpxsystems.com/PDFs/DPX%20Web%20PDF%201.pdf

[lawsonuw]

Finally got my motor. (PayPal took it's bloody time...) I'll have some more pictures up tomorrow.

Marty

[lawsonuw]

The two pictures show how I'm mounting the new motor in the drill. Still needed is a torque arm for the front of the motor, an exit hole for the motor wires, and to remove some plastic that's rubbing on the motor.

I also took a bunch of measurements from my motor.

Code: Select all

Rm: 0.047 [ohm] measured between any two phase wires
Advance     Kv [rpm/v]    Io [amp]     Vbat [volt]
4-10        348           1.05         12.8
10-20       366           1.16         12.6
20-30       376           1.30         12.6

Some flux is leaking out of the motor can, connectors were not included, the motor is well balanced this time, 12 stator teeth 14 poles, and no louder than my first motor.

Looking at my gearing and Kv my current plan it to go up to 12 cells NiCd with some sub-C 2.4Ah cells. My system model says I'll be able to get 300 watts of power to the pedals at about 70rpm cadence, and about 4 minutes of run time.

The mount to the Linear's frame is all finished. I'll get pictures of that when the motor is in the drill case and ready to install on my bike.

Marty

[paultrafalgar]

Not wishing to dampen your enthusiasm, but I could have sworn that I read on here that Right Angle (worm/crown) gears are very inefficient?

[TylerDurden]

If the bike thing doesn't work out, there's always large animal dental work.

[lawsonuw]

Well I got the motor in the drill case and the whole mess re-installed. Switched back to driving into the slack side of the chain. I think it looks cleaner. The frame clamps and mount work great.

I've taken it for a ~10 mile shake down run. The 13T freewheel I've got is driving my nuts with its loud ass clicking. Anyone know how to shut it up? This new system pulls better than my last motor, but it's spinning a bit too slow. Maximum assist is at a cadence of 50-60rpm. When it's on this system is much quieter than the last motor. Gearbox still sounds like a coffee grinder, and it generates a subtle high pitch hissing noise. I can hardly hear the motor at all though. Oh yea, the new battery drains the battery quicker too.

I'll be making a couple of 1-2 cell booster batteries to hook inline with my current pack. I want to confirm the best battery voltage BEFORE I buy a new battery pack. I'll also be upgrading the 5ft 12-14awg battery wire to 10awg to reduce system resistance.

@Paultrafalgar: The gears in the drill are spiral bevel gears, and they are at least as efficient as straight tooth bevel gears. Specifically since the axis of the input shaft and output shafts intersect, the gears have an approximately rolling contact and waste very little energy.

@TylerDurden: ... Last time I checked dental drills spun a itty bitty bit faster than 219rpm. Though I have to say the blaze orange case of the drill adds a whole new level of ghetto to this bike

First picture is the whole system ready to install in the bike.

Second pic is a closeup of the system installed on the bike.

Edit: the converted drill weighs 3lb 10oz (1643grams) I estimate that the whole system weighs about 6lb.

[TylerDurden]

Heavier grease in the freewheel should help suppress the noise.

IIRC, Shimano FWs are fairly quiet.

[lawsonuw]

Well, 14.4v really livened up the performance of this assist motor. Unfortunately 14.4v was also too much for the second stage planet gears from the drill.

I just ran some numbers and it looks like I could use my current motor and planned 14.4v to friction drive the rear tire of my Linear with a similar assist force as the WE hub in my Trek produces. (though only weighting 3lb + battery instead of 15lb+ battery) I'd use a light spring for the initial force on the roller and use the roller's reaction forces to generate the bulk of the normal force on the roller. Preferably the roller would be pulled off the wheel automatically when the motor was off.

Bhaa! I think it's time I stopped messing around with the cheap stuff and just do this right. Just had a hint of a snow flurry, so this project goes on the back burner till my winter bike is up and running. (a Marin Alpine trial with a Cruzbike kit and studded tires, waiting on the kit now)

Marty

[Papa]

The 13T freewheel I've got is driving my nuts with its loud ass clicking. Anyone know how to shut it up?

It's the polygonal action in combination with the roller/sprocket impact during initial meshing that is responsible for most of the noise (assuming it's lubed, tensioned and aligned correctly). Basically, after one of the chain rollers engages and is seated, further sprocket rotation accelerates the following roller's engagement velocity, causing it to 'slap' the sprocket's roller saddle.The larger the sprocket diameter, the lower the chain engagement velocity (and the lower the noise). Given the pitch of the typical bicycle chain, cogs smaller than about 15 teeth should be avoided.

[lawsonuw]

The 13T freewheel I've got is driving my nuts with its loud ass clicking. Anyone know how to shut it up?

It's the polygonal action in combination with the roller/sprocket impact during initial meshing that is responsible for most of the noise (assuming it's lubed, tensioned and aligned correctly). Basically, after one of the chain rollers engages and is seated, further sprocket rotation accelerates the following roller's engagement velocity, causing it to 'slap' the sprocket's roller saddle.The larger the sprocket diameter, the lower the chain engagement velocity (and the lower the noise). Given the pitch of the typical bicycle chain, cogs smaller than about 15 teeth should be avoided.

um, no. The 13T freewheel shuts right up if the motor is running. (have to be pedaling) It is the ratchet paws hitting the 'gear' they engage making all the racket. Heavier grease, making the paw springs weaker, shortening the paw springs, or a different brand are the solutions I'm currently considering. As it is the freewheel clicks are nearly as loud as the gears/motor.

Slept on my gear problems a bit. The broken gear in the drill looks like it is brass. I could replace those with stainless steel or hardened steel and keep them from breaking again. I also think I need to add a second bearing to the output shaft. It's quite likely that only one of the three planets is taking all the load while the other two planets sit idle due to the output shaft shifting under power. An extra bearing should fix this.

Marty

[Papa]

Marty,
Even tho I lean towards synchronous belts, I think your creation is brilliant. Thank you for making it available in such vivid, reproducible detail. I'm reminded of John Tetz's addon: http://www.recumbents.com/wisil/tetz/e-assistmetric The trend towards tire smoking HP in bike conversions leaves me disinterested. Just seems counter intuitive to walk away from a HPV with clean, unscratched pedals - just mho. I have two, fresh 14.4 3ah Li-ion packs (w/matching charger), but am still fretting over motor and drive selection. Hard-headedly, I'm trying make the jump with only one reduction.

Re: chains and sprockets

After rereading your noise complaint, the word 'clicking' finally jump off the page, (ooops!) - ok, I screwed up. However, the noise dynamics of chains and cogs, imo, is worthy of mention for those who experience it, and don't know why - it is very real. After looking at your install again, the minimal chain wrap around the 13t cog would seem to explain why it doesn't appear to be an issue on your bike. Maybe a fiberglass cover would help curb the clicking - or perhaps move the motor assy aft, and under the seat?