Mongoose DX project

oofnik

100 W
Joined
May 1, 2008
Messages
264
Location
Israel
Hi everyone,
I've been watching the giant RC motor thread with great interest over time. First off I must say that I'm all for making a forum division for this kind of stuff! Excellent idea.

Anyway I had the idea of using RC brushless motors for my e-bike project quite a while ago when the idea was still in initial brainstorming stage. Research pointed me eventually to Matt's E-cumbent and I was floored. Ever since then I've been slowly conjuring up my own design. I switched between gear drives, belt drives, different motors, different controllers, mounting methods, batteries... etc. But I've stuck with the design I have in my head right now for at least a few weeks, and I'm just about at the point where my funds are sufficient to get my hands dirty. :twisted:

I'm planning to convert my Mongoose DX bike to a commuter e-bike. I have access to a small shop here at my university so I hope fabrication will not be too difficult. Here is the crude spec list in my head:

Electrical
-TowerPro 5330-10T brushless motor (here)
-Castle Creations HV85 ESC
-LifeBatt LiFePO4 10AH 12s1p (actually 6s1p x 2 so I can charge as a 6-cell pack with my RC charger)

Mechanical
-2 stage belt drive reduction (16/72) x (16/36) = 10.1:1 ratio (not for sure on the sizes), parts from sdp-si.com
-BMX freewheel on drive output shaft to sprocket mounted on rear disc brake hub (sprocket from this guy)

I think that covers the basics. I chose the TowerPro motor because of its price, even though it seems to be well tested and checks out OK. I was originally going to go for one of the HexTronik or Turnigy motors (63-74) but it was very difficult to find information on them. Also, the protrusion of the shaft from both ends on the TowerPro will make mounting much easier I think. I'm planning on putting a small fan on the shaft to keep the motor cool like Matt did.

I hear Castle Creations is second to none in quality. Nobody else makes a high quality HV RC controller for that price either. As for the batteries... well, my university is competing in the Formula Hybrid competition again this year (check it out!) and we're planning to use a giant array of these batteries in our design. When we make our order in the coming months, I'll tack mine on as well for the volume discount. :)

Some of my concerns right now include how I will fabricate and mount the housing. I am thinking I'll end up just cutting two aluminum plates with slotted holes. The shafts and pulleys will mount to backing plates with bearings that will be adjustable to keep belts tensioned properly. I need to sit down and draw this out in Inventor when I have the time. My other concerns include weatherproofing and keeping the batteries functional during Chicago winters (less than -10C). But I will have time to work that out.

Let me know if you guys have any suggestions or comments. I would love to hear some input. I plan on keeping a detailed record of the build once it gets under way. Thank you all for running so far with this idea, I have learned a lot from here!

P.S. Maybe I could get my old thread I started back in May moved to here?
 
oofnik said:
Hi everyone,
I've been watching the giant RC motor thread with great interest over time. First off I must say that I'm all for making a forum division for this kind of stuff!

I'm planning to convert my Mongoose DX bike to a commuter e-bike. I have access to a small shop here at my university so I hope fabrication will not be too difficult. Here is the crude spec list in my head:

Mechanical
-2 stage belt drive reduction (16/72) x (16/36) = 10.1:1 ratio (not for sure on the sizes), parts from sdp-si.com

Some of my concerns right now include how I will fabricate and mount the housing. I am thinking I'll end up just cutting two aluminum plates with slotted holes. The shafts and pulleys will mount to backing plates with bearings that will be adjustable to keep belts tensioned properly. I need to sit down and draw this out in Inventor when I have the time.

Let me know if you guys have any suggestions or comments. I would love to hear some input. I plan on keeping a detailed record of the build once it gets under way. Thank you all for running so far with this idea, I have learned a lot from here!

Hi,

Matt is leaning toward selling a different design but he was planning on selling these for about $150, which is parts cost plus about $50:
http://endless-sphere.com/forums/viewtopic.php?f=1&t=6163
Pictured with more up to date information here:
http://endless-sphere.com/forums/viewtopic.php?f=1&t=6163&st=0&sk=t&sd=a&start=60#p102742

All that would be required to make it a two stage is to add a second shaft which would require increasing the length about an inch or two. Maybe he or his machinist (on his more advanced drive thread he posted that his machinist has time now) would make you one with a second shaft for a two stage.

He originally used a two stage belt but changed to first stage belt and second stage chain. He said the second stage chain is quiet and I assume more compact and more flexible.
 
Yes, I've been following Matt's build closely and must have read his build page at least 10 times over. :shock: He really does quite beautiful work.
The drive box is actually kind of tempting. I like the idea. It would save me a whole lot of machining time, which I don't even know I'm capable of at this point. Access to machinery at my school is generally pretty restricted, and my experience is not-so-much. On top of that, racking up an order on McMaster for Al plates, bearings, shafts etc. suggests a similar price even without all the machining time factored in.

I've been trying to figure exactly how I want the drive system to work. Nothing is really set in stone yet except for the definite use of R/C parts. It's very hard to find reliable information on the maximum torque / speed ratings on the different types of belts out there.

I see that Matt switched over from XL to HTD pitch, and ended up even shredding a 15mm HTD belt on the primary reduction. Holy crap. I'm not planning on putting quite that much power into my build (probably peak at 3kw absolute max). From the power curves I've seen for belt loads I believe a 9mm HTD belt on the primary side and a 15mm HTD on the secondary should be sufficient.

Ehh, it's late... need to stop researching..

P.S. Oh yeah, Miles, it will be 2 belt reductions and then final chain reduction, so overall reduction will be around 18:1 or so.
I need to measure the max sprocket size clearance I have on my bike.
 
oofnik said:
It's very hard to find reliable information on the maximum torque / speed ratings on the different types of belts out there.

You can use the Gates software as a guide:

http://www.gates.com/dfproweb/index.cfm

http://www.gates.com/designflex/index.cfm?location_id=809

You can also download/access the design manuals.

For us, the torque is limited by the number of teeth engaged on the driver (smallest) pulley, for a given belt width and tension.
 
oofnik said:
Hi everyone,
Electrical
-TowerPro 5330-10T brushless motor (here)
-Castle Creations HV85 ESC
-LifeBatt LiFePO4 10AH 12s1p (actually 6s1p x 2 so I can charge as a 6-cell pack with my RC charger)

Hi,

Depending on the best price you can get on the HV85 an HV110 from John Holmes might be a better choice:

If they are still available John is selling the HV110 for $184.49.

http://endless-sphere.com/forums/viewtopic.php?f=9&t=7068
johnrobholmes said:
on Tue Oct 21, 2008
I have four HV110 ESCs right now. Go to HolmesHobbies.com and enter "ESC" into the coupon box in the shopping cart page to receive 18% off anything in my "motor speed controller" category. This includes all Castle and Tekin products I currently have. Figured some of you gearheads would like to pick up one on the cheap.

http://holmeshobbies.com/home.php?cat=2
 
I just got one of those TowerPro 10t motors and measured the constants..

Kv = 220

Io = 1.33A

Rm= .068 ohm

The Io seems to vary quite a bit, I'm thinking the bearings might be dragging intermittently. The value I gave is the lowest one I recorded.
 
Hey Fumesucker, thanks for posting that!
Values seem to be quite close to the ones posted on the website. That's good.
How is the overall construction quality of the motor? Maybe you could replace the bearings with higher quality ones?
Want to share some details about your project? :D
 
My avatar is my project, a BikeE recumbent. I'm doing a power unit to pull the bottom chain run and then a freewheel on the crank so I can use all my gears (my bike has a seven speed derailleur and a three speed hub, all the gears are in the rear).

I'm probably going to send the 10T back, I've found another model I think I'd prefer.

http://www.hobbycity.com/hobbycity/store/uh_viewItem.asp?idProduct=7709&Product_Name=TGY_AerodriveXp_SK_Series_50-65_270Kv_/_2100W
 
I figured I should post some pictures of my bike as it is currently... Nothing special, but it gets me around. Soon to be getting me around much faster. :mrgreen:

Anyway I have a few questions about my electrical system. I ended up going with the new Headway 38120S cells (12), currently in the mail. I'm planning on building two separate 6-cell packs with balance taps, no BMS, but the controller has a voltage cutoff. On the bike, the packs will be wired in series. During charging, they will be charged in parallel with a high power 6-cell R/C balance charger. Does anyone see any problems with this idea?

Also, I want to run a 5V bus to power various things like lights, the servo tester, etc. I am planning on running a switching regulator off one of the 6-cell packs, since the input on the regulator doesn't allow over 24v. Will I have balancing problems if this regulator draws < 200ma?

Let me know what you guys think. Thanks!
 
I see no problem with your charging/ discharging method, but I would suggest no more than 80% discharging for the batteries. It is the same method I use. It would be wise to set your voltage cutoff to be a bit higher than normal, say 2.4v per cell.

As for the 5v bus, it won't pull too much power. Your overall runtime will dictate how bad the bus power pack gets out of whack. If the packs are not within .2v of each other you could always charge the one bus pack separately.
 
Thanks, John. The Castle controllers have nothing in between 36v and 24v for cutoff, so I've been contemplating building a simple low voltage indicator circuit. I don't know how careful I need to be with these cells, but I don't want to risk it. The circuit I came up with won't actually cut power, but it will annoy me enough to disconnect the batteries, which works just as well :p. Plus I can see if I have any weak cells that trip regularly before the rest this way with a simple LED panel on my handlebars.

Anyway I just sent off another $300 to China for my motor, charger, and all sorts of various bits and pieces.
This is definitely my most expensive project by a long shot. It's also going to be the coolest by several orders of magnitude. :mrgreen:
 
I would like to know how this goes. I am thinking of doing the exact same thing but just dont quite understand the gearing.
 
oofnik said:
What is it exactly that you're confused about? Maybe I could explain better.

There are a couple things that I dont get. First is the servotester. Does the controller plug into motor, battery, and servotester? Then the servotester runs to a 5k throttle? Second is the gearing. Can you tell me exactly what type of gears you are going to use to let you achieve high speeds with this very tiny motor? How many gears are you going to use?
 
Yes, you're right about the wiring. Basically, RC motor controllers get their throttle input from a pulsed wave input that depends on the duty cycle (ratio of time spent at 5v to 0v). This type of signal is called PWM (pulse width modulation). Same as servos. They made this interface so that it would be compatible with the already existing servo signal interface. Typically in RC cars and planes, the remote control will encode the throttle data into a radio signal and send it to the receiver over the air, which will then decode the pulses and send them to the motor controller. But in a bike application, you need to make some kind of device that emulates the output of the radio receiver to 'trick' the motor controller into thinking it's getting the right signal. There are schematics available online to build such a device, or you can just buy a servotester ready made which will do exactly the same thing.

I'm flip flopping on the gearing again. The idea to reduce my design to a two-stage reduction to decrease complexity is becoming more and more appealing. But essentially, permanent magnet electric motors (brushed or brushless) typically produce more or less the same amount of torque at any RPM. Output power is the product of torque and RPM, so you can imagine the peak power to vary roughly linearly with respect to the RPM. Therefore it is impossible to utilize the motor's full power output if it is not geared down, because nobody is going to run their bike wheels at 6000 RPM :shock:. What gearing down does is divide the RPM and multiply the torque by the same factor, keeping the power output constant (well, almost). The result is that more power can be effectively utilized.

I am planning for a gear ratio around 15:1 to 16:1. My current plan is to achieve this in two stages: one belt-drive and one chain-drive. The motor will turn a 16-tooth pulley, which will transmit power to a 72-tooth pulley. So that's a 72/16 or 4.5:1 reduction. Next, the big pulley will spin on the same shaft as a BMX freewheel, which is essentially a sprocket that turns freely one way and locks up in the other. This sprocket will have 14 teeth, and it will transmit power to the rear wheel sprocket which will be somewhere around 50 teeth. So that's 4.5 * (50/14) = ~ 16:1 total reduction. The motor will spin 16.1 times faster than the rear wheel while the rear wheel will experience 16.1 times as much torque as the motor produces (assuming 100% efficiency; actual numbers may be closer to 88% - 95%). Hope that helps you understand what I'm trying to do a little better!
 
oofnik said:
Yes, you're right about the wiring. Basically, RC motor controllers get their throttle input from a pulsed wave input that depends on the duty cycle (ratio of time spent at 5v to 0v). This type of signal is called PWM (pulse width modulation). Same as servos. They made this interface so that it would be compatible with the already existing servo signal interface. Typically in RC cars and planes, the remote control will encode the throttle data into a radio signal and send it to the receiver over the air, which will then decode the pulses and send them to the motor controller. But in a bike application, you need to make some kind of device that emulates the output of the radio receiver to 'trick' the motor controller into thinking it's getting the right signal. There are schematics available online to build such a device, or you can just buy a servotester ready made which will do exactly the same thing.
This helped a lot. What type of servotester are you going to get (or are you going to make one)?
 
Got this in the mail along with some more stuff.
Construction seems solid. Bearings appear to be OK. I'm going to go test it out now with my servo tester and HV85.

And it's got a flat machined on to the shaft already too :D
 

Attachments

  • IMG_4414_800.JPG
    IMG_4414_800.JPG
    49.8 KB · Views: 8,498
Well I've been up for the last several hours cutting copper strips out of a big sheet for my battery. I've got the cells assembled more or less how they're going to go in the bike. I'm going to have them secured to something (fiberboard? plexi?) on both the top and bottom, add balance taps, and heat shrink the entire package. The cells are each separated by little double stick foam pads to allow airflow and reduce shock.

I shot a short clip of the motor test. My oscilloscope is hooked up to one of the motor phases to watch the back EMF plus the PWM output of the controller. I think it's kind of neat. I upgraded the firmware from the older version and the upgraded startup sequence works much better. These little controllers are simply amazing.

[youtube]FuLfNrDlZaI[/youtube]

Right at about 0:40 when I say "I hate that!" I'm referring to the giant spark that occurs when the batteries are connected to the ESC. What do you guys do to prevent this from happening? I feel like that could wear out connectors quite prematurely. :?
 
oofnik said:
Right at about 0:40 when I say "I hate that!" I'm referring to the giant spark that occurs when the batteries are connected to the ESC. What do you guys do to prevent this from happening? I feel like that could wear out connectors quite prematurely. :?

http://endless-sphere.com/forums/viewtopic.php?f=2&t=7994&p=120883#p120883

Great video. Im looking forward to how you are going to mount it to the bike and get it running.
 
Back
Top